The study was conducted by business school professors Brad Barber of UC Davis, Reuven Lehavy and Brett Trueman of UC Berkeley, and Maureen McNichols of Stanford University. They found that the most highly recommended stocks in 2000 returned 31.2 percent less than the market, on average, while the least favorably recommended stocks gained almost 49 percent more than the market.

"This result was particularly surprising given our prior research, which documented that the most highly rated stocks did outperform the least favorably recommended ones during the years 1986-96," said Lehavy. The results for 1997-99 were similar to those for 1986-96, but somewhat weaker.

The poor performance of the analysts' recommendations was not restricted to just a few months of the year 2000, or to just periods in which the market was declining. According to Trueman, "These poor results held for nine of the 12 months of the year. In fact, the worst month for the analysts was December, when the market was holding fairly steady." Furthermore, the researchers' results held for both tech- and non-tech stocks.

Added Barber, "Since the poor returns we document occurred throughout the year, it is unlikely that they can be attributed to the imposition of Regulation FD (Fair Disclosure) in October, 2000, which prohibits firms from revealing material non-public information to analysts." He went on to note that even if Regulation FD reduced the value of analysts' recommendations overall, there is no reason that the buy recommendations should do worse than the sell recommendations.

Concluded McNichols, "While we can't say that the poor 2000 showing is necessarily a result of increased analyst involvement in investment banking, our findings should certainly add to the current debate over the usefulness of analysts' stock recommendations to investors."