Two studies, two methodologies, two opposing points of view.

What is the secret to picking an excellent mutual fund? An endless stream of working papers, dissertations, and published studies have chased after the elusive answer. Most of these efforts, however, have ultimately proven disappointing. But the authors of two dueling studies, which tackled the question by exploring the consequences of fund style drift, hope to change this.

In the first study, Keith C. Brown, a finance professor at the University of Texas in Austin, and W.V. Harlow, a Fidelity executive, suggest that selecting style-consistent stock funds will often significantly boost an investor's returns. Mutual funds that remained faithful to their stated investment style, the authors discovered, enjoyed better returns than those that disregarded their style mandate. Consequently, chances are greater that a large-cap growth fund that remains within its style box will perform better than a large-cap growth fund that wanders into, say, the mid-cap growth or large-cap value territory.

"Do more style-consistent funds outperform less style-consistent funds?" Brown asks. "The answer is unambiguously yes. There is value in finding a more style-consistent manager."

Not so fast, says Russ Wermers, an associate professor of finance at the University of Maryland at College Park. His preliminary working paper suggests something different. "The relation between style consistency and stock-picking talent is, at best, tenuous," Wermers says. Further, he found evidence that the best stock-picking managers often indulge in a "significant amount" of style wandering.

What makes this pair of papers notable is that academic literature has paid little attention to the consequences and causes of style drift. The researchers' efforts, however, would obviously have been more welcomed if they shared the same bottom line. So what gives? The researchers suggest that one reason for this disconnect may be traced back to the distinct ways that style consistency was measured. Wermers used portfolio-based style analysis, while Brown and Harlow stuck with returns-based style analysis.

Each technique has enthusiastic supporters, which explains why debates on which method is superior have sometimes become emotional. Returns-based style analysis relies upon historical returns to provide an estimate of the average style position of a mutual fund or separate account during a given period. This type of software doesn't look at actual portfolio positions. In contrast, portfolio-based software, which is what Morningstar Principia employs, analyzes the actual securities within a given portfolio to generate style information. You can learn more about both types of analysis by reading The Handbook of Equity Style Management, Third Edition, by T. Daniel Coggin and Frank J. Fabozzi (John Wiley, 2003). Coggin suggests that the heated bickering about which method is better has largely abated and that most experts now believe that both techniques deserve respect and often are best used together.

Curious about the contradictions in the two studies, Wermers is continuing his research in an attempt to determine whether the use of different style analysis techniques is the culprit behind the disparate results. He will also explore whether some of the performance advantage that style-inconsistent funds appeared to enjoy resulted from using net returns that don't include fund costs. He expects to publish his findings later this year on his Web site at www.rhsmith.umd.edu/Finance/rwermers/. You can read Brown's study at his Web site at www.mccombs.utexas.edu/~brownk/.

In the meantime, here's some background on each study's findings:

Brown and Harlow reached their conclusions after scrutinizing the performance stats of the 3,177 stock funds that were in Morningstar's database from 1991 through 2000. The funds, which were grouped by style, fell into one of the nine main style categories: small-cap value, growth and blend; mid-cap value, growth and blend; and large-cap value, growth and blend.

The style-consistency premium was notable across the entire time period. Within certain categories, the performance gap was stunning. Stylistically pure small-cap growth and large-cap blend funds, for example, generated yearly median returns of 14.21% and 20.04% compared with 12.78% and 16.69% for their style-sloppy counterparts.

What you might find amazing is that the style-consistent funds enjoyed a performance advantage even after the researchers controlled for portfolio turnover and expenses. This is noteworthy because low fund expense ratios and modest portfolio turnover have been two of the only truly reliable indicators that could be used to predict superior performance in the future.

Arguably, the most famous fund study documenting that cheap is best was written by Mark Carhart, who is co-head of the Quantitative Strategies Group at Goldman Sachs Asset Management. In this landmark study, "On Persistence in Mutual Funds," Carhart documented that there is a direct negative correlation between a fund's returns and its expenses. In addition, Carhart concluded that another predictor of future performance is mediocrity. Dreadful funds tend to remain that way. In contrast, however, hot funds usually don't burn bright for long. A fund that generated high returns during the previous 12 months often has one year of momentum left before it struggles.

Style-consistent funds are more likely to fit into the lower-cost category; obviously, with a style-pure fund, there's not as much costly trading going on. The conclusion, then, that style consistency by itself could give a fund a competitive edge surprised John Rekenthaler, president of Morningstar Associates, a subsidiary of Morningstar. "This is a surprising result," Rekenthaler says. "When you look at a mutual fund and control for expenses and turnover rate, style consistency really shouldn't matter."

Why should style consistency alone be such a big deal? Brown speculates that the managers who remain faithful to their little square in the style box are less likely to mess up their asset allocation and stock picks than those who try to time their style decisions.

This argument makes perfect sense to Arnie Wood, president and CEO of Martingale Asset Management in Boston. "In more cases than not," he observes, "when managers make changes in a portfolio and they leave their particular style, they tend to make more mistakes as they drift. And when they drift, they tend to fall into the same trap as individual investors: They run to whatever has done best recently."

Larry Swedroe, director of research at BAM Advisor Services in suburban St. Louis, observes that the Brown-Harlow study indirectly makes a strong case for index funds, which are inherently stylistically pure. "We would suspect that index funds would compare favorably since passive funds would logically have the greatest persistence of style, which should equate to the best performance," he says.

Swedroe, who has written a trio of pro-indexing books, says it is unfortunate that the research didn't directly compare actively managed funds, which don't stray from their style mission, with index funds.

There weren't enough index funds in the study period to generate any meaningful conclusions, Brown explains. The finance professor, however, warns that no one should assume that the researchers believe that indexing is superior. "The message isn't that you shouldn't hire active managers and just index," he insists. "We are finding people who are producing reliable alpha, but they are the ones who stick closer to their mandate."

In his study, Wermers looked at all domestic equity mutual funds, which had at least 50% of their assets invested here, that existed anytime between January 1975 and December 1994. The 2,892 funds were culled from the database maintained by the Center for Research in Security Prices (CRSP) at the University of Chicago and Thomson Financial's institutional ownership files. The styles of the stocks held within the funds were characterized by size, book-to-market value and momentum.

One of the most intriguing findings was the great success of the study's biggest style busters. The top 5% and 10% of funds, ranked from top to bottom by their active style drift, respectively earned 18% and 17.3% average annual net returns. In comparison, the average net return for all funds in the study was 15.8%. Wermers suspects that part of this superior performance resulted from the fact that these style-inconsistent managers were more likely to hold portfolios of small stocks, which enjoyed greater returns during the study's time period.

But the performance advantage for style-sloppier funds continued to hold, to a point, further down the food chain. The top quintile of style-active funds, as well as the second quintile, each generated 16.1% annual net returns. The performance, however, bottomed out in the third quintile at 14.4%, before the numbers started improving. The lowest quintile, which contained funds that rarely deviated from their style mandate, generated a return of 15.6%. And the 5% of funds least likely to stray from their style mandate generated a return of 16.2%.

Wermers also looked at stock-picking skills of this universe of managers. Once again, he said the managers who weren't shy about deviating from their style produced substantially higher returns. The top quintile of style-active managers oversaw portfolios that beat their style benchmarks by 1.8% a year, while the median quintile enjoyed no extra performance. Even more dramatic, the top 10% of style-inconsistent managers beat their style benchmarks by more than 3%.

Certain types of funds were more prone to style-creep than others. "Our study finds the growth-oriented funds have higher levels of style drift than income-oriented funds, and small funds have higher levels than large funds," Wermers says. "Managers who have better career stock-picking track records and higher levels of career portfolio turnover tend to engage in trades that cause more active style drift. Further analysis shows that these managers deliver superior future portfolio performance, which indicates that they are not simply overconfident."

Steve Hardy, president and founder of Zephyr Associates Inc., a major source of returns-based style analysis software, is familiar with the contradictory papers, but declines to choose a favorite. In an interview, however, Hardy expresses misgivings about forcing a manager to fit into a certain style niche. "The more you constrain someone, the less likely they can do their job well," he suggests. At the same time he scoffs at the futility of so many managers who trample over style considerations as they chase hot stocks. It's only the truly skilled managers like Bill Miller, Peter Lynch and George Soros, he says, who can pick stocks within very loose parameters and not crash and burn with the ambulance chasers.

What Hardy seems to suggest is that the success of a style-consistent fund could lie in whether or not a fund manager toes the style line willingly. "The issue, then, is whether managers who are style-consistent are that way because they are constrained in some way," he says.

Hardy says he suspects that a manager who, for instance, only invests in small-cap value stocks because that's what he knows and that's what he likes, and that's what he thinks about every day, should do well. But the manager who invests exclusively in small-cap value because he could lose his job doing otherwise is more likely to fail.

If you ultimately conclude that style loyalists do hold an edge, pinpointing these funds won't necessarily be easy. The most recent quarterly analysis of funds by Standard & Poor's suggests that most equity funds, over lengthy time periods, are style creepers. During the three-year period ending in June 2003, for instance, S&P calculated that only 54.1% of domestic funds stuck to their stated style. During the past five years, that statistic drops to 37.4%. "The message of the story," says Srikant Dash, S&P's index officer, "is really that investors and financial advisors who are advising them should keep track of style changes."

According to S&P's three- and five-year statistics, large-cap value and growth funds were the most likely to remain within their style boxes. During the five-year period, 63.4% of large-cap growth and 63.5% of large-cap value funds stuck to their discipline. In contrast, mid-cap and small-cap blend funds were singled out as the biggest style violators. During the five-year period, a mere 13.2% and 11.2% of these mid-cap and small-cap funds remained style-consistent. Of course, it should be noted that S&P's numbers were generated using returns-based style analysis. What would have happened to the numbers using portfolio-based style analysis? It's anybody's guess.