Active share makes a point of the benefits of being different

In 2006, Dr. Martijn Cremers and Dr. Antti Petajisto, then of the Yale School of Management, introduced the concept of active share.¹ They noted that active managers strive to outperform their benchmarks in three ways: through stock selection, through factor bets (e.g., overweighting a specific industry), or through a combination of both.

Because many managers favor one approach over the other, investors lacked the means to quantify active management across funds. The traditional approach of using tracking error alone was insufficient because a diversified stock picker could have a low tracking error, erroneously implying a low level of activity.

As an alternative, the researchers developed active share, which measures the fraction of a portfolio that differs from its benchmark. Active share is calculated by summing the absolute value of the differences between the security weights in a portfolio and the corresponding weights of the fund's benchmark, and then dividing by two; the result is measured on a scale from 0% to 100%. A portfolio with no holdings in common with its benchmark would have 100% active share, while a portfolio that is identical to its benchmark would have 0% active share.

Drs. Cremers and Petajisto argued that active share might be useful in two distinct ways. First, it can describe the potential for a fund to outperform, since logically a fund must differ from its benchmark in order to outperform it. Second, when combined with tracking error, active share can help to more accurately discern different types of management styles.

The researchers applied their active share analysis to a total of 2,647 U.S. equity mutual funds over a 23-year period (1980 to 2003). For each year and each fund, the stock holdings were reported on three separate dates, resulting in a total of 48,354 observations. They drew several conclusions from their work:

• The active share of an individual fund is extremely persistent over time.
• Funds with low active share (20%–60%) tend to underperform their benchmarks over time (these funds were deemed closet indexers).
• Funds with the highest quintile of active share (>80%) consistently outperformed their benchmarks over time and after fees.
• Funds with smaller asset bases tend to have higher active shares.
• Tracking error by itself is not related to fund returns, suggesting not all dimensions of active management are rewarded in the market.

Arguably the most significant finding of their research is that the most active stock pickers have enough skill to outperform their benchmarks even after fees and transaction costs. However, Drs. Cremers and Petajisto also drew conclusions about the universe of actively managed funds, suggesting that funds may be broken down into four different categories, defined by active share and tracking error.

Academic research has shown that active share can be a useful statistic in determining the potential for a manager to outperform, particularly when paired with tracking error. However, as with any other financial metric, active share provides only a partial picture of a fund and can be skewed by the universe in which that fund invests. Investors also need to consider a fund's objective, its role in a portfolio, the degree of risk, and its fees before determining whether it makes sense to include that fund in a diversified portfolio.

1 “How Active Is Your Fund Manager? A New Measure That Predicts Performance,” M. Cremers and A. Petajisto, 2009.