"Let's say you walk by someone's office and they're sitting there doing nothing. You're going to assume that they're not doing everything they're supposed to do," says Evans. "It may be that the optimal strategy for a mutual fund is not to trade anything; maybe things are going well as they are. But if people see me not trading and think I'm somehow shirking responsibility, then I might be incentivized to trade just to look like I'm an active trader."
Other studies have shown that fund managers will even make value-reducing trades just to, in effect, look busy. That's less likely to happen, says common sense, if the manager has skin in the game.
Manager ownership also has consequences for expenses, though they aren't as clear-cut. Morningstar data show that among core stock funds, expenses don't vary that much by management ownership levels within fund categories. In fact, equity funds with no manager stakes had lower expense ratios (relative to their category peers) than those with invested managers. Among bond funds, though, the lowest expense ratios within categories are those with the biggest manager stakes (more than $1 million).
To be sure, correlation isn't necessarily causation. But given all the variables that influence investment outcomes—a realm we can fairly label "luck"—investing is supposed to be all about process. If the process is sound and you stick with it, goes the theory, it should perform in the long run whatever the short-term variables encountered on the way.
When you see managers with little or none of their own money in the game, it's only fair to wonder about the process. Morningstar, writing in 2008, noted that funds it designated "picks" had about seven times the absolute level of manager investment than those it called "pans." (You can learn what manager owns what by looking at the fund's Statement of Additional Information, usually a supplement to the prospectus.)
It could be just coincidental that manager ownership corresponds with fund performance the way it does. But, for most investors, that's probably not a prudent assumption.