For decades, investors and academics alike have sought the magic bullet, the one data point that would predict the future performance of mutual funds. Hundreds of studies looked at a variety of variables including past performance, expense ratio and turnover. But each factor proved to have limited value as a forecaster. A new paper may suggest the best predictor yet: trading costs. Funds with low trading costs outperform high-cost competitors by a wide margin, says the study, titled “Scale Effects in Mutual Fund Performance.” Morningstar assisted the researchers — Roger Edelen of Boston College, Richard Evans of University of Virginia and Gregory Kadlec of Virginia Polytechnic Institute — and the Chicago fund tracker is now working to develop a screen based on the research findings. “There is a lot of value in the trading cost data, and we are looking at ways to make the numbers easily available for investors,” says Russel Kinnel, fund research director of Morningstar.

While expense ratios cover management fees and other perennial costs, trading costs include brokerage commissions and the one-time bills associated with executing transactions. Researchers have long suspected that trading costs were significant. In a speech before Congress in 2003, John Bogle, founder of Vanguard Group, estimated annual trading costs at 0.7 percent of assets. But it has always been difficult to measure the exact amount of the costs or to find the number anywhere in fund documents. In the past couple of years, new technology and rules on disclosure have made it easier to calculate a figure for trading costs.

In the latest study, researchers looked at 1,700 domestic equity funds and found that the average fund had annual trading costs equaling 1.4 percent of assets, compared to average expense ratios of 1.2 percent. “Trading costs can be more significant than the expense ratio,” says Edelen, one of the study's authors. “Some funds have costs that are well above average, and that can weigh down returns substantially.”

What's In A Trade?

Trading costs include several elements. The first is brokerage commissions. In the past, finding total brokerage costs was difficult. But in response to the market-timing scandals, the SEC has begun requiring funds to disclose more data, including brokerage costs. “To get a clear idea of what a fund really costs, you need to check the brokerage commissions,” says Ric Edelman, head of Edelman Financial Services, a registered investment advisor in Virginia that is affiliated with broker/dealer SMH Capital.

Each fund must list its brokerage costs in a document called the Statement of Additional Information. In some cases, you must request the material from the company, but major fund families such as Fidelity Investments and Vanguard Group now post the document on their websites. Big funds that enjoy economies of scale may have relatively low brokerage costs, while small funds that trade often can have steep expenses. For example, Fidelity Contrafund, which has total assets of $72 billion, posts brokerage expenses equal to 0.09 percent of assets. In contrast, Fidelity Select Telecommunications, with assets of $690 million, has brokerage costs of 0.41 percent.

While brokerage commissions can be important, the bulk of trading costs typically comes from other factors; the biggest may be market impact. This occurs when a trader buys or sells. If the trader is bidding on a big order, he will tend to put upward pressure on the stock price. Say the investor wants to buy 100,000 shares at $20 per share. He may get the first 10,000 at $20, but the price will begin to rise in response to the demand he initiates. Another major cost factor is the bid/ask spread.

To calculate total trading costs, the researchers looked at actual trading data. Fund companies don't reveal every trade they make, but once a month (or quarter) they disclose their portfolios to Morningstar. Say a portfolio has 100,000 shares of IBM one month, but lists none the next. The researchers looked at data from the New York Stock Exchange to see what bid/ask spreads were for IBM. They also looked at exchange data to learn what impact trades of 100,000 shares had on the market price. Using this data, they estimated how much the fund paid to sell the shares.

Survey Says

Examining tens of thousands of trades, the researchers reached a variety of noteworthy conclusions. First, trading costs come directly out of total returns. A comparison of the one-fifth of funds with the lowest trading costs, versus the one-fifth with the highest trading costs, showed the former achieved total returns that were 1.5 percent higher. Second, for most managers, trading is not worthwhile. On average, each $1 in trading costs reduces total fund assets by 40 cents. So the typical trade results in some capital gains — but not enough to cover trading costs. Of course, there were some funds where trading added value. But many of these funds succeeded at adding value through trading because their trading costs were so low. “The evidence suggests that some managers have the skill to increase returns by trading,” says Greg Kadlec, a study author and professor of finance at Virginia Tech.

The funds with low trading costs kept these down in a variety of ways. Some tended to trade in small lots, which are cheaper to buy and sell. Some winning managers simply had low turnover: the fewer the trades, the lower the costs. In addition, some of the most efficient funds traded primarily large-cap stocks. When someone sells a big, highly liquid stock, the price may barely move. But when a trader seeks to sell a little-known stock, the seller may have to accept a low price in order to complete the trade. “If you trade Exxon, it may cost you 10 basis points, but if you trade a tiny stock, it could cost 150 basis points,” says researcher Roger Edelen.

The researchers pinpointed an additional problem area: fund flows (money moving into or out of the portfolio). About 30 percent of all trades can be traced to flows, says Alan Seigerman, chief operating officer of ReFlow Management, a company that helps mutual funds manage their flows. For example, as shareholders exit a fund, the manager must sell shares in order to raise cash. Such trades tend to be expensive, and produce poor total returns. “When a manager is forced to make trades because of fund flows, he may not be acting on his best investment ideas,” says Seigerman.

How can advisors use the new findings? Before buying a fund, make an effort to determine the total ownership cost. Start by adding the expense ratio and brokerage commissions. This should give you about 60 percent of costs. Then look at factors such as turnover and fund flows. These figures should provide a rough guide to total trading costs. To get the lowest costs, consider a large-cap fund with few flows, low expense ratio and little turnover. Some large-blend candidates include: American Funds Fundamental, Davis New York Venture and FMI Large.


Large Blend Funds That Hold Down Trading Costs.

Fund Ticker 5-Year Return 5-Year Rank Category % Turnover Expense Ratio
American Funds Fundemental A ANCFS 14.5% 3% 21% 0.61
Davis New York Ventur A NYVTX 13.5 7 6 0.88
FMI Large FMIHX 14.4 4 29 1.11
Janus Contrarian JSVAX 22.0 1 39 0.95
Neuberger Berman Partners NPRTX 14.7 3 33 0.82
Source: Morningstar Data through 6/30/07.