TCS Daily

Wall Street Analysts Already at Risk for Exile on Main Street

By James K. Glassman - June 14, 2001 12:00 AM

Mr. Chairman, members of the committee:

Thank you for inviting me to testify today on this important subject.

My name is James K. Glassman. I am a resident fellow at the American Enterprise Institute, concentrating in matters of economics, financial markets and technology. I am also host of the website, a cyber-think tank that focuses on matters at the intersection of technology, finance and public policy. For six years, I wrote a nationally syndicated financial column for the Washington Post. I am now chief columnist and senior consultant to Folio(fn), a financial services firm. In addition, I am a weekly financial columnist for the New York Daily News and the International Herald Tribune. I am co-author of Dow 36,000, a book on stock valuation, and author of a forthcoming investment primer titled The Secret Code of the Superior Investor. I have devoted much of my professional career both to educating small investors and to analyzing and advocating public policies in the economic sphere.

This hearing examines "whether securities analysts are providing unbiased research to investors." As a witness, I am asked to "discuss any conflicts of interest that may affect the objectivity and independence of analysts and what, if anything, needs to be done to improve the quality of information to investors."


After five years of unprecedented gains, the U.S. stock market declined sharply last year and continued to fall in 2001. Over the past 12 months (through June 11), the Standard & Poor's 500-Stock Index, the most popular benchmark, declined 13 percent. The Nasdaq Composite, dominated by technology stocks, fell 42 percent.

Many analysts were caught off-guard by the decline, which represented the first bear market in a decade. Some of the best-known Wall Street analysts, including Mary Meeker of Morgan Stanley Dean Witter and Henry Blodgett of Merrill Lynch were celebrated for making accurate recommendations of high-tech companies, but in 2000, the share prices of many of those firms, including, and Yahoo! plummeted. A recent article in Fortune magazine called Ms. Meeker "the unquestioned diva of the Internet Age" and reported she made $15 million in 1999 - a year in which the Nasdaq roughly doubled - by urging her clients to buy high-tech stocks. But now, wrote Peter Elkin of Fortune, she is "the single most powerful symbol of how Wall Street can lead investors astray."

Mr. Elkin wrote that Ms. Meeker "came to see herself not merely as an analyst but as a player - a power broker, a dealmaker, a force to be reckoned with." It is just this conflict - an erosion of the famous "Chinese wall" between the investment-banking side of a large Wall Street firm and the research side - which, in the eyes of critics, threatens the objectivity of analysts and the wealth of investors.

Meanwhile, other analysts whose stock selections turned sour have been accused of different sorts of conflicts of interest. The New York Times criticized analyst Richard Juarez of Robertson Stephens, who continually advocated purchase of iBasis, an Internet stock which dropped from $49 to $4 and which, the Times noted, Mr. Juarez was, at the time, selling out of his personal account. Laura Unger, acting chair of the Securities & Exchange Commission, recently was reported to have "warned Wall Street firms to resolve 'blatant' conflicts that surrounded the business of bringing shares to the public and then recommending them to investors."

These same concerns have led to the hearing today.

Conflicts of Interest

There is little doubt that conflicts of interest pervade the securities industry. Many of the best-known and most influential analysts work for firms that have extensive and lucrative investment-banking relationships with companies the analysts cover. A negative analysis by an analyst could embarrass investment bankers or even lose business for the firm. A positive analysis could lift the stock, make issuing more debt or equity easier, thus enriching the investment bankers. In addition, some analysts play a direct role in winning investment-banking business for a firm. A second kind of conflict involves analysts owning, buying or selling shares in companies that they cover.

But it is important to understand that conflicts of interest pervade investment banking in large part because they pervade life. A husband and wife with a three-year-old son develops a conflict of interest by deciding to have another child, but that potential conflict rarely deters them, nor should it.

Or, more to the point, consider journalists: Surveys show that most journalists lean to the left of the political spectrum. For example, a study by the Roper Center of 139 Washington bureau chiefs and congressional correspondents found that in the 1992 presidential election, 89 percent said they voted for Bill Clinton, 7 percent for George Bush. The same study found that 59 percent characterized the 1994 Contract with America as "an election year campaign ploy" while 3 percent said it was a "serious reform proposal." Kenneth Walsh, then of U.S. News & World Report, surveyed White House correspondents from 1980 to 1992 for his book, Feeding the Beast, and found that, by a margin of 37 to 5, they had voted for Democratic presidential candidates over Republican. Yet every journalist to whom I have ever spoken claims that his professionalism overrides these conflicting political leanings. Does it? The answer is that we can judge for ourselves by reading the articles they write or the TV segments in which they appear. Some surmount the conflict; some do it. But the fact that all journalists produce something for public consumption means that individuals can see the product and judge for themselves. By the way, no one would deny that Members of Congress have conflicts of interest too. They must balance allegiances to family, donors, party, constituents, and principle. Setting precise rules on how this balance must be achieved would be fruitless and counterproductive. And, in the end, Members are judged by their actual production, their votes, and service to their district.

But back to stock analysts: their situation is similar to that of journalists except that their judgments are clearer and more easily assessed by the public. The essential problem with a conflict of interest of any sort is that it leads to poor judgment. In the case of journalists who lean left or right, it would mean a political or ideological bias that might subtly color reporting and might be difficult to discern by the public. In the case of stock analysts, it could mean that a company with poor fundamentals and poor prospects is given a positive recommendation. In this case, however, the analyst's judgment could be assessed quickly by the public. An analyst who consistently gave bad advice would be rejected as not useful either to investors or, ultimately, to the firm that employed her.

In other words, recommendations tainted by conflict of interest are decisions made in the full glare of publicity. An analyst cannot hide for long.

For this reason, conflicts of interest do not greatly trouble me. An analyst who recommends bad stocks in an effort to sell investment banking services will be an analyst whose track record - closely watched by journalists and professional tracking services - will soon lose him his job. Still, there is no forgiving an analyst who hypocritically and corruptly sells shares in a stock he is recommending. One such episode and, I believe, the analyst is finished. It is important that the full light of publicity shine on such activities.


While I believe that the perils of conflicts of interest are overrated and overstated, I do favor voluntary and extensive disclosure by analysts of their personal holdings and any other affiliations that might color their decisions. But, again, it is important not to exaggerate the benefits of disclosure. What, for example, should an investor make of the disclosure that an analyst owns 1,000 shares of a stock she recommends? That the stock may be more deficient than if the analyst did not own the stock because the analyst has an additional incentive of personal gain? Or, is the case the opposite: That the stock may be a particularly good one since the analyst does own it and thus has her own money on the line? I am not really sure that disclosure is all that helpful - except in the case of an analyst who sells stocks he recommends. In my own financial writing, however, I disclose any of the personal holdings I mention in a column, allowing readers to make their own judgments from these facts but knowing I could just be confusing them.

Performance of Analysts

The recent critique of analysts comes down to this: Biased by conflicts of interest, analysts recommend companies that do not deserve "buy" ratings. The disaster of 2000 is the evidence, and well-paid analysts who make mistakes are fair game. There is no doubt, as I will show, that last year was a terrible one for analysts, but journalists and politicians - in fact, all of us - often rush to judgment based on events that happened yesterday, without looking closely at history. Anecdotes, especially recent ones, are powerful, but they prove nothing. The essential question is this: How good have the recommendations of analysts been over time? If analysts have performed well, then the evidence would be strong that they have, in the aggregate, surmounted any conflicts of interest that may have colored their judgments.

Our good fortune is that just this question has been examined at length in a study published in the April 2001 issue of The Journal of Finance, a highly regarded publication for scholars. In the article, "Can Investors Profit From the Prophets? Security Analyst Recommendations and Stock Returns," the authors, all economists at California universities, present evidence that would almost certainly surprise many critics. They found that the consensus recommendations of analysts between 1986 and 1996 were prescient and profitable. This research reinforces earlier studies that have found that professional securities forecasters acted "rationally" - that is, with proper judgment.

The authors of the new study - Brad Barber of the University of California at Davis, Reuven Lehavy and Brett Trueman of Berkeley, and Maureen McNichols of Stanford - looked at a database of 360,000 pieces of advice from 269 brokerage houses and 4,340 analysts over 10 years. They gathered recommendations regarding each stock into a consensus in one of five groups - from 1 (most favorable) to 5 (least favorable). Consensus ratings, by the way, are easily available for free on the Internet, from such financial websites as CBS MarketWatch and Yahoo! Finance.

Every time an analyst initiated coverage of a stock or changed his or her rating of a stock, the consensus was recalculated by the researchers, and, if necessary, the stock moved by the researchers into a new group. Over the 10-year period, the group 1 stocks returned an annual average of 18.8 percent while the group 5 stocks returned an annual average of just 5.8 percent, with the other groups arranged in order between them. The stock-market benchmark over this period returned an average of 14.5 percent.

The researchers then applied controls for market risk, size, book-to-market ratios and price-momentum effects. They found that the highest rated stocks still outperformed the lowest by a wide margin. The group 1 stocks beat the benchmark by 4.1 percentage points, and the group 5 stocks trailed the benchmark by 4.9 percentage points.

These results are truly exceptional. Rare, for example, is the mutual fund that can beat the Standard & Poor's 500-Stock Index by four points over 10 years. In fact, the benchmark has beaten a majority of funds over the past two decades. The results of the Barber study suggest that analysts are truly able to pick winners.

Last month, the four researchers produced a follow-up to their study, examining, in the same manner, the four years from 1997 to 2000. In the first three of those years, a portfolio of the group 1 stocks generated an annual average return that was 4 points higher than the market as a whole while the least-favored stocks generated a return that was 9 points lower. But the final year, 2000, was a debacle for analysts - five standard deviations away from the previous 13 years' results. During that year, the most highly recommended stocks produced a return that was 31 percentage points below the benchmark while the least-favored stocks did best - 49 percent above the benchmark. The year 2000 appears to have been an anomaly, an outlier - though, clearly, more research is needed. Still, Dr. Barber and his associates looked at issue of conflicts by simply examining whether significantly more investment-banking activity occurred during 2000. It did not. Dr. Barber calls the year "a mystery," with a performance completely at odds with those of the previous 13 years.


Mystery or not, the year 2000 does not in itself provide enough evidence on which to base a new set of conflict-of-interest regulations. My own assessment of stock analysts is that they are, for the most part, solid and conscientious professionals who try their best to find good companies in which investors can put their savings. Are the judgments of some of them biased to the point of error by conflicts of interest? Of course. And if those analysts are wrong enough times, then clients won't trust them and will move elsewhere. Allan Sloan, writing in Newsweek, is correct when he wrote last week that analysts "became the bad guys when the bubble burst 15 months ago, and America began one of its favorite activities: searching for someone to blame."

The truth is, many analysts were right for most of the 1990s, as stock prices rose substantially, and were wrong in 2000 and the beginning of 2001, as stock prices fell sharply. Overall, however, they have done much better than the laws of chance would allow - better, it appears, that mutual fund managers and newsletter writers.

But even if their performance were poor, I would not favor this committee's writing laws to order certain kinds of disclosure or forbidding certain conflicts. Individual firms and the securities industry as a whole have strong incentives to increase disclosure and to limit conflicts in order to increase public faith in markets - and, more important, client faith in their companies.

If ever there was a case of transparent, well-monitored information on which the public can make its own judgments, it is stock-market analysts' recommendations and ratings. They are out there for all to see, to criticize, to respond to.

In the end, however, the recommendation of an analyst is only one tool in an investor's kit. Personal observation of companies in action, examination of income statements and balance sheets, news stories and even word-of-mouth all go into investors' decisions to buy and sell - as they should. Many investors, however, rely on analyst research because they do not have the time or inclination to do their own. Analysts have a professional and moral obligation to make sure that research is the most honest and thoughtful they can offer. Most of them, it appears, live up to that obligation. But scrutiny of analysts, in a forum like this one, is appropriate and beneficial. Precipitous legislative action is not.

Thank you, Mr. Chairman and members of the committee.


TCS Daily Archives