CONTACT: Stanford University News Service (415) 723-2558
The securities research analysts who advise your broker on the best stock market picks may not be trying to mislead you deliberately, but beware of their rosy attitudes. Maureen McNichols and Patricia O'Brien have done some research of their own, turning up evidence that equity analysts' earnings forecasts are persistently overoptimistic.
McNichols and O'Brien, associate professors of accounting at Stanford University's Graduate School of Business and London Business School, respectively, cast a broad net: The study included 523 analysts covering 3,774 different companies at 129 different brokerage firms listed in a Standard & Poor's database. McNichols and O'Brien found that as a group, analysts are less likely to report bad news than good because they self-select what they cover. That is, they add coverage of stocks when their information is favorable and drop coverage when their information is unfavorable. As a result, the bad news is rarely reflected in their last forecast report or recommendation about the issue.
The researchers also found that return on equity was greater for stocks the analysts had just added to their coverage than for stocks with previous coverage. Not surprisingly, return on equity was lower for stocks that are dropped, probably because the stock was abandoned after an analyst became disillusioned with its performance. Return on equity was lower still for stocks not covered at all by sample analysts.
McNichols suggests several possible explanations for these trends. First, if the stock analyst identifies hot stocks that benefit the brokerage firm's clients, the analyst enhances his or her reputation for picking winners and providing timely information. Second, a good story is easier to sell than a bad one. When the brokerage mails out its research reports, good news is relevant to a broader audience and will get a higher hit rate of revenue-producing trades executed by the firm. Any investor is potentially interested in trading a hot stock. Only a smaller group of investors already holding the stock would be interested in selling it. Third, analysts prefer to maintain good relationships with corporate management because management is a potential client for investment bank services and a big source of information.
McNichols and O'Brien discovered that analysts also drag their feet when they deliver bad news. They found the median number of days between upgrade ratings of a stock was 98. By contrast, the number of days between downgrades was 127. One possible explanation is that when analysts get preliminary negative information, they sit on it a while to see if the next spurt of information about the company will be positive. Others may detect initial negative signals, such as product quality problems, but not wishing to jeopardize their relations with the company's senior management, will wait until someone else puts out the bad news--and then hop on the bandwagon.
Of course, says McNichols, equity analysts have a strong incentive to maintain their credibility. Without it, they cannot survive for long. Analysts must sometimes advise clients of impending doom. One critical constituency is made up of fund managers who invest millions of dollars in money pools. McNichols speculates that analysts may advise fund managers by phone of upcoming corporate trends that may be negative, allowing the fund managers to rearrange their portfolios and look smart before the bad news hits the street in official earnings announcements. In some cases, analysts may simply drop such a stock from their coverage.
"Self-Selection and Analyst Coverage," Maureen McNichols and Patricia O'Brien, GSB Research Paper #1412, December 1996.
By Barbara Buell