This study develops a new measure referred to as Active Share which is effective at identifying the best performing mutual funds. Active Share measures the extent to which a fund’s portfolio holdings differ from the portfolio’s benchmark index. Specifically, it is defined as the sum of the absolute difference between the portfolio weight of each asset in the fund and the asset’s weight in the benchmark index. Then, this sum is divided by two so that the value of Active Share is scaled between 0% and 100%.  As defined, a pure index fund would have an Active Share of 0% while a manager whose portfolio only has a 10% overlap with the benchmark index would have an Active Share of 90%.

            Quantifying the extent to which a manager actively manages a fund is very difficult. A manager can practice active management by 1) selecting individual stocks that the manager expects to outperform, 2) shifting assets across various industries or asset classes, or 3) combining both stock selection and market timing. Tracking error, which is defined as the standard deviation of the difference between a fund return and its benchmark index return, is the traditional method of measuring active management. But the authors discuss how tracking error does not accurately capture both types of active management and provide evidence that Active Share (either by itself or combined with tracking error) is a more appropriate measure.  In this study, the authors calculate Active Share for 2,650 equity funds covering the period 1990-2003. Most funds report their portfolio holdings on a quarterly basis, and thus Active Share is calculated quarterly for most of the funds in the sample.      

            For individual investors, the key result of this study is that Active Share is a strong predictor of fund performance. Funds with the highest Active Share outperform their benchmarks on a (Carhart) risk adjusted basis by 1.5% per year net of expenses. The difference in risk adjusted return between the highest and lowest Active Share quintiles is 3.3% per year. The relationship between active share and fund performance tends to be strongest for the smallest funds as measured by total net assets. The smallest funds outperform their benchmarks by 3% net of expenses. In addition, this superior performance tends to persist over time as the smallest funds continue to produce net out performance several years into the future. As a robustness check, the authors examine the relationship between tracking error and performance and do not find a significant correlation between the two variables. Thus, Active Share provides information about fund manager skill not measured by tracking error, and Active Share can help fund investors pick better performing funds.

            Overall, the Active Share measure indicates that fund managers whose stock selections deviate substantially from their benchmark do have stock selection skill and do provide value to their fund investors.