Stock analysts still rarely utter that four-letter word - "sell" - but investors who know how to listen for other clues can stay ahead of the pack.
As part of a $1.4 billion settlement nearly four years ago, regulators demanded that the biggest U.S. brokerages insulate research analysts from the pressures of winning investment banking business, a move that consumer advocates hoped would free analysts to more candidly warn investors when a stock was in trouble.
But fewer than one in 10 reports, or 7.1 percent, carry a "sell" signal today, down from a peak of 10.4 percent in 2003, according to industry tracker StarMine Corp.
Little wonder, some observers say, that investors don't appear to be making much use of the increased research analysis called for in the global settlement reached in late 2002 and finalized in 2003 among the Securities and Exchange Commission, several state attorneys general and 10 of the biggest U.S. brokerage firms.
In fact, recommendations emanating from those 10 firms were more optimistic after the settlement than before, concludes a new book, Financial Gatekeepers: Can They Protect Investors?
"The global settlement has done little, if anything, to change the recommendations made by the 10 settlement firms or their long-term investment value for investors," wrote Leslie Boni, an assistant professor at the University of New Mexico's business school, who contributed a chapter in the book based on her analysis of the firms.
Specifically, Boni found that, as a group, the 10 firms had reduced research coverage between 2000 and 2004 and had issued fewer of the lowest ratings on stocks. Meanwhile, the number of positive and neutral ratings remained constant.
In addition, the overall volume of analysis today isn't adequate, contends John Coffee Jr., a Columbia University securities law professor and author of another new book, Gatekeepers: The Professions and Corporate Governance.
Coffee asserts that conflicts of interest at brokerage houses have been reduced but not eliminated, and that research coverage has been reduced significantly since the global settlement, particularly for the smallest capitalization stocks.
For their part, brokerage firms vehemently disagree with critics, saying their research committees now scrutinize coverage and that they have changed analysts' pay incentives to reward accuracy.
"Stock research is far superior to what it was" a few years ago, in large part because analyst pay is now tied directly to performance, said Matthew Carpenter, U.S. director of Citigroup Investment Research in New York.
Citigroup was one of the 10 firms involved in the settlement, along with Bear Stearns, Credit Suisse, Goldman Sachs, JPMorgan, Lehman Brothers, Merrill Lynch, Morgan Stanley, US Bancorp Piper Jaffray and UBS Warburg.
Michael Dritz, an independent equity analyst who provides research reports to UBS clients under the settlement arrangement, also believes there are fewer conflicts today and fewer cases of companies "freezing out" analysts who issue unflattering reports.
So why are there still so few "sell" recommendations?
Blame selection bias, he said.
"Most analysts initiating coverage on stocks are looking for companies with strong prospects, not losers," Dritz said. To be sure, he said, analysts could be far quicker to alert investors when their opinion turns negative.
But, research firms say, rather than looking at the ratio of buys to sells, investors can measure an analyst's or a firm's track record by benchmarking their recommendations to movements in stock price.
Good analysts, and firms to watch, are the ones whose "buy" recommendations produce higher stock-price returns than do their "sell" recommendations, experts said. The "buys," in turn, should outperform their industry benchmarks.
In 2002, the stocks of companies that had aggregate analyst "sell" ratings actually went up more than the "buys," according to independent research firm StarMine Corp.
The following year the momentum swung, and stocks with positive ratings outperformed by 2.2 percentage points. Last year, the positive ratings outperformed by 1.33 percentage points.
"Analysts actually subtracted value in 2002. Portfolios were slightly worse, on average, than throwing darts," said David Lichtblau, vice president at StarMine. "Fortunately, there was a big turnaround in 2003."
Wall Street says it has succeeded in devising new pay incentives for research departments that reward accurate calls, not participation in helping firms' investment bankers win underwriting business.
Until then, investors can look for other signals to obtain useful information out of stock research, experts said.
Firms like StarMine (www.starmine.com) and Investars (www.investars.com) publish free data on the performance of stock ratings by brokerage houses and by independent research firms.