BOSTON -- When it comes to predicting mutual fund performance, only one criteria -- fund size -- really counts.
For the 10 years ended Sept. 30, the smallest 25% of domestic equity mutual funds outperformed their Lipper category averages 87% of the time. Small mutual funds delivered average annual returns that were almost one percentage point better over 10 years than category averages for the 10-year period ended Sept. 30.
By contrast, during the same time frame, only 48% of the largest 25% of core equity funds beat their Lipper category averages, according to new analysis by Financial Research Corp. of data on 766 core equity funds with more than $50 million under management.
That small vs. large fund differential has been widening in recent years. FRC found that since the first quarter of 1990, small funds outperformed large funds by an average of 95 basis points per year. In the shorter term, since the first quarter 1995, small funds outperformed bigger funds by an average of 142 basis points per year. Larger funds last outperformed smaller funds in 1993, according to FRC.
The small fund advantage likely is due to the extra investment flexibility and nimbleness a smaller portfolio affords, coupled with the fact that small fund managers can more easily move in and out of stock positions without affecting the price of the stock.
One piece of FRC's analysis isolated the 10 largest domestic equity funds as of Sept. 30. The average returns of that peer group (determined by fund size) were compared with those of the Standard & Poor's 500 index during the previous 11 years. On average, the peer group outperformed the S&P 500 significantly for the first six years, sometimes by as much as 600 basis points. However, as assets under management ballooned in 1995, these 10 funds began to underperform the index. During the past three years, the big fund group trailed the S&P 500 by more than 400 basis points each year.
The most common selection screen -- performance -- has absolutely no value for identifying future high performers, as both top decile and bottom decile performers showed a very strong tendency to revert to the mean in the next year, FRC found.
During the 10 years studied, top-decile performers fell on average to the 48th percentile during the next one-year period, and bottom-decile performers fell to the 52nd percentile. In fact, the tendency of outperformers to revert to the mean held true even when FRC widened the sample and looked at the top 30% and bottom 30% of performance for 12 months.
The authors of FRC study described a vicious circle for mutual funds: a track record of hot outperformance often will lead to a flood of new money, which will cause, in aggregate, performance problems for those funds in the next year.
For all time periods throughout the 1990s, FRC's research confirmed the intuitive assumption that top-performing funds outsell poorer performers by a wide margin. But high-volume net sales had a negative effect on many portfolio managers: On average, for 40 quarters of testing, the prior year's best sellers underperformed the peer group average by 0.1 standard deviations in the next year.
In fact, except for fund size, FRC's research found high recent net sales is the next best predictive factor for future performance.
In a series of case studies on individual mutual funds, FRC's report showed the $14.4 billion Oppenheimer Main Street Growth & Income Fund, the $2.2 billion Chase Vista Growth & Income Fund, the $23.7 billion Putnam New Opportunities Fund, the almost $3 billion AIM Aggressive Growth Fund, the $14 billion MFS Emerging Growth Fund and the $4.5 billion Kaufmann Fund all followed this pattern -- excellent performance when assets were small and a period when net sales soared as investors chased hot performance, followed by underperformance when fund assets swelled.
FRC analyzed three funds that defied the pattern. The now $45 billion Washington Mutual Investors and the $3.9 billion Janus Growth & Income funds have offered superior performance relative to peers, despite a huge jump in assets under management. To a lesser extent, the $88.1 billion Fidelity Magellan Fund continued to outperform Lipper category averages, but at a lower magnitude as assets increased.
FRC analyzed these nine funds because each was the largest fund in its Lipper objective, had strong performance during its first six quarters and had strong performance for four quarters just after introduction. The funds span Lipper's growth & income, growth, small-cap and midcap objectives.
FRC tested other common selection screens -- manager tenure, Morningstar Inc. ratings and portfolio turnover -- and found none had any value for predicting performance. FRC's various tests of these screens during a 10-year time period failed to produce accurate predictions of fund returns.
A fund's age did have some predictive ability, but mainly as a subset of fund size: Assets tended to grow as a function of time and strong performance.
Also worthy of more in-depth analysis were fund expenses, although FRC's research was inconclusive: Low-expense funds outperformed high-expense funds between 1994 and 1997, while high-expense funds had higher returns between 1991 and 1993. The study's authors also said a portfolio manager's total years of experience in money management (not just on a single fund) merited closer qualitative research.
Quantitatively, FRC concluded asset size and recent net sales were most reliable as fund selection screens. But the authors warned: "Mutual fund selection by quantification is a blueprint for disaster because mutual fund returns are just too unpredictable. By definition, systems-of-equations problem-solving fails in the presence of significant uncertainty. Brute computational force cannot overwhelm the forces of uncertainty. Fund selection is most certainly a divergent problem, although it may seem convergent to the uninitiated."
Qualitatively, FRC research confirmed that knowing the fund manager and the fund company and trusting gut reactions are probably the best screening factors.