Many news and industry reports proclaimed 2015 to be a challenging year for active managers. For U.S. large-cap stock funds in particular, actively managed funds on average have tended to underperform their benchmarks, and 2015 was no exception. But what if those overall averages were not really representative of the active funds investors tend to favor?
In 2015, Fidelity published a research article addressing that very question about actively managed mutual funds.1 That research found that historically, using two objective criteria as filters could have narrowed the wide range of U.S. large-cap stock funds down to a smaller group that, on average, outperformed the industry as a whole and outperformed their benchmarks. In that article, analysis of historical returns for all U.S. large-cap funds from 1992 to 2014 showed that the average actively managed fund underperformed its benchmark by more than the average passive index fund did.
However, by using two objective filters to select only funds with the lowest fees from the five largest mutual fund families by assets under management (representing funds with greater potential resources), the average active fund out performed. Although averages are no guarantee of the performance of any individual fund, these results suggest that investors could benefit from using these simple filters to help select a set of appropriate funds for further consideration. (Note: These filters were chosen by Fidelity; had other filters or filter parameters been used, results would have been different.)
The analysis focused on U.S. large-cap equity funds partly due to high investor interest in that category, and partly due to a commonly held belief that a high degree of market efficiency in this category makes it more challenging for active stock pickers to outperform a broad market benchmark. (For analysis of two other market categories, see Exhibit A chart at end in the section: Outside of U.S. large-cap, average active funds outperform.)
2015 results support the conclusion
New data for 2015 further supported the conclusions of this research (see Exhibit 1 chart below). Looking just at the one-year results, the average actively managed fund did worse than the long-term average, earning 137 basis points less than its benchmark, after fees.2 In contrast, the average active filtered fund did better than the long-term average: the average low- fee fund was only 38 basis points below the benchmark, while the average fund from the top five fund families earned 54 basis points above the benchmark. With both filters combined, the average active mutual fund earned 70 basis points above the benchmark for 2015 alone, compared with a long-term average of 18. In other words, 2015 was a challenging year for actively managed U.S. large-cap equity funds in general, but the average low-fee fund from the largest fund families did much better, outperforming its benchmark by 70 basis points.
Over the longer term, with both filters in place for both types of funds, the average active fund outperformed its bench- mark (see Exhibit 1 chart below) while the average passive index fund still lagged its benchmark slightly (as one would expect from passive funds, which seek to match benchmark performance before fees). These results suggest that industry-average figures for both active and passive funds may not fairly represent the investment performance achieved by many investors.
Understanding the filters
The fee and fund-family size filters were chosen to be objective, straightforward, and intuitive. The fee filter selects funds in the lowest 25% of reported expense ratios for each fund type (active or passive).3 Funds with lower total expense ratios are able to deliver more of their gross returns to investors after fees. Because fees are clearly disclosed, investors can use this information to help them select funds. For 2015, the average filter cutoff for the lowest-fee funds was 79 basis points for actively managed funds, and 11 basis points for passive funds.The size filter focuses on assets under management (AUM), considering AUM to be a reasonable proxy for scale. For active funds, the filter selected funds from the mutual fund families with the most assets in active U.S. large-cap equity funds, because larger fund companies could use size to their advantage by committing more resources to research and trading, and the benefits of those resources can be shared across all the companies’ active U.S. large-cap funds. For passive index funds, the filter selected the top 10% of funds by size, in order to confer a similar selectivity and potential advantage (see endnotes for more information).
In the mutual fund industry, differences in fund family size can be quite large. At the end of 2015, the median amount of actively managed U.S. large-cap assets for all fund families was around $243 million, while the median for the top five fund families was more than $180 billion—more than 740 times larger.4 The largest five fund families held approximately 49% of the industry’s assets, while the smallest 50% of fund families (173 of 345) held less than 0.5% of AUM.As a result, any average analysis of the entire industry will include a high proportion of active fund families that may lack comparable resources to compete.
Better average performance is consistent with filters
Although past performance is no guarantee of future results, these filters have been remarkably consistent in identifying sets of funds with above-average relative performance over time. For rolling three-year returns, the average actively managed fund selected by both filters beat the industry average a full 98% of the time. Exhibit 2 shows how consistent this outperformance was, and by how much. In addition, a statistical test indicates one can be 99% certain that the historical long-term outperformance of the filtered average fund relative to the industry is significant.5
Implications for investors
Although these filters are not the only way to search for better-performing actively managed U.S. large-cap stock funds, many investors may find it useful to know that these simple objective criteria succeeded in identifying a subset of actively managed funds that has performed better than the general averages and outperformed their benchmarks on average (while a comparably selected subset of passive index funds still underperformed).
Just as important, even though 2015’s average actively managed fund did worse than the long-term average, the average filtered fund did much better than the average fund for the year, and beat its benchmark by 70 basis points after fees. Of course, averages never tell the whole story, and any one particular fund may do better or worse than the average, particularly over short time horizons. Prudent, informed research is always an important part of identifying funds that fit an investor’s objectives. However, we believe the results of applying these criteria continue to suggest that certain straightforward and objective filters can be a helpful starting point for investors seeking to identify above-average actively managed equity funds.
Outside of U.S. large cap, average active funds outperform
U.S. large-cap stock funds are a commonly cited example of how difficult it is for active funds to outperform. Fidelity’s research has shown that in the other largest equity fund categories (international large-cap equity and U.S. small-cap equity), active managers have had a better record of outperforming their benchmarks, even when all funds are considered (see Exhibit A on right).
1 See Leadership Series article ”U.S. Large-Cap Equity: Can Simple Filters Help Investors Find Better-Performing Actively Managed Funds?” (May 2015).
2 A basis point is 1/100th of percentage point. So, 137 basis points is 1.37%.
3 Expense ratio is the total annual fund operating expense ratio as reported in the fund’s most recent prospectus.
4 Data as of December 31, 2015. Source: Morningstar, Fidelity Investments.
5 After making an adjustment for overlapping data, the statistical significance of this outperformance was evaluated in a two-tailed test, and resulted in a t-statistic of 4.43. A two-tailed test is a method for computing the statistical significance of a parameter inferred from a data set, in terms of a test statistic. In this case, the test statistic of 4.43 indicates a 99% likelihood that the results are significant and not random.
© 2016 FMR LLC. All rights reserved.
Fund selection: Our main analysis focused on all U.S. large-cap, foreign large-cap (“international large-cap”), and U.S. small-cap equity mutual funds tracked by Morningstar between Jan. 1, 1992, and Dec. 31, 2015, including all blend, value, and growth funds within each category and including actively managed and passive index funds. We included funds that did not exist for the entire period (closed or merged funds), to reduce survivorship bias. We eliminated funds identified as passive that were labeled as “enhanced index,” and eliminated funds with tracking error greater than 1% (which are unlikely to be actual passive index strategies despite their identification in the database). For international large-cap funds, we eliminated funds benchmarked to a price index, for greater comparability. See below for benchmark indexes included and definitions.
Our analysis began with the entire set of funds with available data from Morningstar at any point over the full period: 2,013 actively managed mutual funds, and 115 passive index mutual funds. We selected the oldest share class for each fund as representative; where more than one share class was the oldest available, we chose the class labeled as “retail.”
For U.S. large-cap equity, average fund counts for each subset of selected funds are as follows: Unfiltered (full set of funds available): active 831, passive 50. Fee filter only: active 220, passive 13. Size filter only: active 79, passive 5. Both filters applied: active 46, passive 3. Total fund counts for international large-cap equity funds: active 432, passive 29; average fund counts for performance calculation: active 218, passive 11. Total fund counts for U.S. small-cap equity funds: active 704, passive 43; average fund counts for performance calculation: active 295, passive 18.
Averaging excess returns: We used Morningstar data on returns from Jan. 1, 1992, through Dec. 31, 2015. We calculated each fund’s excess returns on a one-year rolling basis, relative to each fund’s primary prospectus benchmark and net of reported expense ratio, for each month. We used an equal-weighted average to calculate overall industry one-year returns for each month. (We chose equal weighting for the averages in order to represent the average performance of the range of individual funds available to investors, rather than asset weighting, which may introduce bias into an analysis.) For filtered subsets of funds, average excess returns ascribed were the one-year forward rolling returns, calculated monthly. All filtered subsets were rebalanced monthly. If a fund closed or was merged during a one-year rolling period, its returns were recorded for the months that it was in existence, and the weighting of the remaining funds in the subset was increased proportionally for the remainder of the year.