Big Investors Hone An Edge

Institutions' research seems to pay

Each year, equity institutional investors--mutual funds, pension funds, insurance companies, and the like--spend billions of dollars on security analysis to stay ahead of the game. Given the wide divergence in institutional performance, however, some critics have charged that such expenditures are excessive and produce little if any net benefits in trading profits.

In a new study, Scott Gibson of the University of Minnesota and Assem Safieddine of Michigan State University shed light on the issue by comparing changes in total institutional ownership of individual stocks during each quarter from 1980 to 1994 with stock returns over the same quarters.

Their finding: During the 15-year period, the 20% of stocks with the largest quarterly increase in institutional ownership (as a percentage of a company's outstanding shares) consistently posted positive returns. And the 20% of stocks registering the largest declines in institutional ownership consistently lost money. The gap between the top and bottom groups averaged 7.2% a quarter over the entire period--and a huge 12.1% from 1990 to 1994 (chart).

Of course, the fact that prices of individual stocks rose or fell in the same quarters that institutions as a group were buying or selling shares doesn't necessarily mean that the institutions' security research had paid off. Rather than reflecting new pertinent information that pushed prices in the "right" direction, it's possible that the very weight of institutional trading simply moved the market temporarily. After all, institutional ownership of stocks rose from 20% of the market in 1980 to 40% in the early 1990s.

To figure out what was happening, the two researchers compared the timing of institutional ownership shifts with the timing of earnings forecast revisions by brokerage firms. They reasoned that if the institutions were truly bringing new information to the market, then their trading should on average anticipate subsequent earnings forecast revisions by the brokerage houses.

That, in fact, is exactly what they found. Net stock purchases by institutions in one quarter tended to precede upwards earnings forecast revisions by brokerage houses in the next quarter. Indeed, the ability of such institutional trading to anticipate the direction of future earnings forecast revisions has grown stronger in recent years.

The bottom line is that the big bucks that institutions as a group shell out for research activities do seem to pay off. Compared to individual investors--who, of course, buy the stocks that institutions unload and often rely on their friendly brokers for advice--the institutions appear to be ahead of the informational curve.

Before it's here, it's on the Bloomberg Terminal.