If ever there were a time-honored investment rule, it's the importance of portfolio diversification. Since no one can know what stocks, sectors, or asset classes will do well in the future, it's best to have your money in many pools. One or two might spring leaks, goes the theory, but that won't necessarily hurt the others.
One objective of diversification is to avoid excessive correlation—the tendency of certain asset classes to move in tandem. A development that hurts housing, for example, could hurt mortgage lenders as well, so it's good to own sectors that aren't overly correlated to housing. Obviously there are limits as to how much one can minimize correlation, as reflected in the aphorism that in a crisis—like the one in 2008—"all correlations go to 1." (Correlation is expressed as a coefficient, where 1 represents perfect correlation and zero represents none.)
But is the opposing position a risk as well? Is it possible to be over-diversified, to have your assets spread so thin that you forfeit meaningful gains?
Just as a little thought experiment, ask yourself what "perfect" diversification would look like: You'd own equally allocated bits of every security on the market—better yet, of every company in the economy, listed or not. That means you'd never do better than the overall economy, which historically hasn't grown as fast as the market—a subset of the economy, if you like. You might be gaining a measure of short-term safety, but at a significant long-term opportunity cost, namely the gains you could have gotten with a bit more concentration of risk.
Even if the average investor runs little risk of nth-degree diversification, there are ways in which misunderstanding diversification can hurt. For instance, the notion that the more asset classes you own—large- to small-cap stocks, bonds of all sorts, commodities, real estate, and other "alternatives"—the safer you are. Studies have shown that beyond some point, adding new asset classes doesn't help much.
"It's hard to be over-diversified in a theoretical sense," says Paul Zemsky, head of multi-asset strategies at ING Investment Management. "But from a practical perspective, you get probably 90 percent of the diversification [benefit] with the first seven or 10 asset classes, and after that it's diminishing returns."
Let's say you find 15 or 20 sector funds with no overlapping holdings. You could invest a little in all of them, but at some point you encounter a practical problem. "The individual is getting very modest diversification benefit," says Zemsky, "but he's increasing the number of funds he has to keep track of."
Indeed, if you hold too many funds—and by extension too many stocks—you end up with what Gene Goldman, director of research at El Segundo, Ca.-based Cetera Financial Group, calls a "closet index fund." And a costly one at that. "Given that most of these managers have higher expense ratios than index funds, you can in essence underperform the markets that way," says Goldman.
While diversification is generally "a winning strategy," says Vanguard founder John Bogle, "if you diversify by using a large number of funds, which in turn hold a large number of stocks, the problem is less over-diversification [than that] you're paying too much for it. You might be paying 1 or 2 percent—3 percent even if you're paying sales loads. That's a killer."
The biggest single source of diversification benefit derives from the first allocation decision most investors are likely to make: how much you'll put in stocks and how much you'll put in bonds. Indeed, says Goldman, "When we build asset-allocation portfolios, we don't want to interfere with the stocks to bonds to cash allocation, because then you start changing the client's overall investment objective."
So you can establish your stock/bond preference and limit yourself to a moderate number of sectors or funds. The next question is what number of holdings is optimal for a fund? For equity funds, INGs Zemsky suggests, 60 to 80 stocks represent "a moderate-risk" active manager.