ITHACA, N.Y. -- Can you trust your analyst to pick the best performing stocks? Not always, suggests an award-winning study by two business school professors.

When the analyst recommends investing in a newly public company whose initial public offering is underwritten by the corporate financing arm of the analyst's investment bank -- something that happens often -- the choice is likely to be biased and not the best, the study shows.

The study, "Conflict of Interest and the Credibility of Underwriter Analyst Recommendations," by Roni Michaely, associate professor of finance at Cornell University's Johnson Graduate School of Management, and Kent Womack, a former Ph.D. student at Cornell's Johnson School and now at Dartmouth College, was cited in Business Week this May 15 and won a top award at this year's Western Finance Association annual meetings.

The researchers tested two hypotheses. The first is that underwriters' analysts have superior knowledge about the potential performance of IPO stock (they don't, according to the study). The second is that underwriters' analysts may make decisions that are biased in favor of the IPOs their investment banking firms underwrite.

"There's a conflict of interest in that the analysts really wear two hats," says Michaely. "Using the first hat, they try to serve their clients -- the investing public -- and using the second hat, they help the investment banking arm of their firm to push through stocks they underwrite. They have a strong incentive to do so since a large portion of their compensation comes from that part of the business."

Michaely and Womack found that the long-run performance of IPO stocks recommended by analysts whose investment banks were lead underwriters was much poorer than the performance of stocks recommended by non-underwriting analysts. The difference was significant -- more than 50 percent when mean- and median-adjusted buy-and-hold stock returns were measured over two years following the IPO. Before the study, most researchers believed stock performances were about the same.

Interestingly, the very same analysts made better recommendations on IPOs when their investment banks were not the lead underwriter, Michaely and Womack found. "It is not the difference in analysts' ability to value firms that drives our results," they wrote. Rather, it is "a bias directly related to whether the recommender is the underwriter of the stock."

The study also showed that stock prices of IPO companies recommended by underwriting analysts dropped on average, while those recommended by non-underwriting analysts rose during the Securities and Exchange Commission "quiet period," 30 days before analysts issue recommendations on an IPO, and that lead underwriter analysts made 50 percent more "buy" recommendations on the IPO stock underwritten by their bank than did non-underwriting analysts.

In addition, the study revealed that the market reacted better to the announcement of "buy" recommendations by non-underwriters than underwriters. Stock prices of firms recommended by non-underwriting analysts had an excess return of 4.4 percent on the day of the IPO, while those recommended by underwriting analysts had an excess return of only 2.7 percent when adjusted for overall market movement.

The authors posit that underwriting analysts have incentives to issue positively biased recommendations on firms that their banks take to market and are aware of the biases or, alternatively, are unaware but believe, often inaccurately, that IPOs underwritten by their banks are simply the best.

The study sample involved 391 firms that conducted IPOs in 1990 and 1991, as reported in Investment Dealers Digest, and analysts' recommendations of companies that completed IPOs in 1990-91, as reported in First Call. To determine what might lead to investment bias, the researchers also reviewed the literature on current practices and surveyed 31 professional analysts, 26 of whom responded.

The paper was published in the Review of Financial Studies (Vol. 12, No. 4) and was selected by the editors of the journal for the runner-up award for the best papers published in that volume. The Barclay's Global Investors/Michael Brennan Prize was awarded at the Western Finance Association annual meetings this June in Sun Valley, Idaho.

For information, or a copy of the study, contact the people listed at the top of this release or visit this web page:http://www.gsm.cornell.edu/faculty/michaely/front.html.