Podcast Center
Video Center
Webcasts
Resource Center
Events
August 02, 2004 at 08:00 PM
July 29, 2004 — A fund’s R-squared statistic can indicate whether an actively managed fund is in fact “actively managed.” Based on a screen of large-cap blend offerings, a number of funds which bill themselves as actively managed are in fact managed more like index funds. As a result, investors in these funds are paying active-management fees for what is essentially an index fund, which usually charges much lower fees.
R-squared measures how closely a fund follows the market. Index funds have an R-squared of 100, which is the highest possible, because their results are explained entirely by movements in their benchmark index. For large-cap blend funds, the benchmark is the S&P 500-stock index. In contrast, actively managed funds should have an R-squared of less than 100 because in theory they don’t stick just to the S&P 500 or their benchmark index, and, therefore, the performance of the index does not fully explain the fund’s performance. The more a fund’s composition deviates from that of the S&P 500, the lower its R-squared, and the lower the correlation with the index.
Actively managed funds seek to outperform their benchmark by deviating from its composition. After all, a fund which exactly replicates the S&P 500 obviously cannot outperform it; rather, it will have almost exactly the same return as the S&P 500. A fund can only beat a benchmark by deviating from its composition, which will in turn result in a lower R-squared statistic. However, a lower R-squared alone will not guarantee success; when deviating from the composition of the S&P 500, fund managers must still pick the right stocks. Funds with a low R-squared don’t always outperform their benchmark.
Since index funds simply mimic their benchmark, they tend to have very low expense ratios. Actively managed funds, on the other hand, must spend much more on research to find stocks which help the fund outperform the benchmark. Consequently, actively managed funds should in theory have both lower R-squared statistics and higher expense ratios.
