Low Cost, Past Performance Unreliable Indicators
May 10, 2004
When fund firms tell investors they should not use past performance as an indicator for future results, they should believe it, said Andrew Clark, a senior analyst at Lipper.
When trying to choose an index-beating fund, a recent Lipper study finds that for both open-end equity funds and taxable bond funds, there is generally no difference between using low expenses as a screen versus choosing based on strong three-year returns.
Investors have often picked funds based on impressive past returns or current low expense levels. But, each of these two criteria, when used alone, are massively flawed, according to Lipper.
However, a deeper examination of expenses and their correlation to loads and share classes can help improve investor odds of selecting an index-beating fund. The report states investors should not rely on a single indicator such as expenses or returns to choose a fund. Rather, they should use a group of tools or ranking mechanisms to narrow the field of possible candidates.
In one of the more surprising findings in the study, Clark, author of the report, contends that load funds with pricier expense ratios tend to be associated with higher returns. "In general, more expensive load funds tend to be the place to look for benchmark-beating funds," Clark said.
For no-load and institutional funds, conventional wisdom holds true, as lower expenses and higher returns are more likely to go together. The study, How Well Do Expenses and Net Returns Predict Future Performance?, concludes that shareholders who bear additional expenses in load funds are usually also rewarded by higher returns. Meanwhile, those who absorb higher expenses while investing in no-load and institutional fund shares are not compensated accordingly.
Lipper examined 31 open-end equity fund classifications and 18 taxable bond fund categories from 1992 to 2003. There were eight equity and nine bond share classes involved in the study.
When choosing low-expense funds, in only three of the eight equity classes do the odds of actually receiving index-beating returns improve. For bond funds, it's only three of nine. No-load funds tend to dominate the picks where low expenses and index-beating returns are more likely to coincide.
In the remaining categories, either dominated by load funds or containing a "significant percentage" of them, filtering for low expenses doesn't necessarily improve the odds of nabbing an index-beating fund. In fact, with load shares, higher management fees (but not the highest) have a greater correlation to better returns. Clark said that this indicates index-beating fund managers tend to charge higher fees, "appropriating some of the benefits associated with their improved performance," which is consistent with the concept of an efficient marketplace.
As for past performance as a basis for choosing a fund, Clark said to beware. "Relying on past performance can at least be dubious," he said. "Performance persistence, in general, doesn't exist. Some people find it will last a few months to maybe a year. But once you get out beyond a year, it pretty much vanishes."
12b-1 Dead Weight
As for 12b-1 fees, they also are found to have a varying effect on a load versus a no-load fund. The report states that 12b-1 fees are primarily a dead-weight cost for no-load shares and that investors accept the cost because they know very little about it and even less about how to figure out its economic impact.
The same dead-weight argument holds true for load shares, but it is "mitigated to some extent by the possibility that 12b-1 fees help counteract redemptions with new sales in order to avoid selling into market weakness." Additionally, brokers and planners are counseling investors not to sell and thereby incur a service fee after periods of poor performance in the shares with sales loads.
"I'm not advocating that 12b-1 fees are a good or bad thing here, but the funds that were able to beat their index also tended to have more 12b-1 fees," Clark said. "That's a rather interesting phenomenon."
Copyright 2004 Thomson Media Inc. All Rights Reserved.