Yes, Wall Street Research Is Better

Despite the cost -- and the complaining -- stock analysis has clearly improved after the Spitzer settlement's forced changes

By Amey Stone

Controversy over stock research still rages on Wall Street, more than two years after New York Attorney General Eliot Spitzer's investigation into conflicts of interest between analysts and investment bankers triggered dramatic reforms in the way research is conducted.

A whole new set of rules now governs how analysts communicate both with the companies they follow and the investment bankers at their own firms. Two pages of disclosures are included at the back of every research report, showing how the stock has performed after an opinion changes and how ratings for investment-banking clients differ from other stocks covered by that firm. Analysts' pay is now tied more to performance and less to investment-banking deals, and it has declined significantly as a result.

The reforms have been costly for firms, and behind closed doors they are roundly ridiculed. More changes are coming that will continue to shape this new era of Wall Street research. For example, firms this summer will have to start handing out reports from independent outfits along with their own in-house research -- a major provision of their $1.4 billion April, 2003, settlement with Spitzer and other securities regulators.


  Already, one thing that should silence some of that closed-door sniping is clear: Wall Street research is getting better. Over the past two years it has become more objective, more original, and more accurate. "There is less regurgitated-type scribe work," says Joe Cooper, a senior analyst at First Call, which monitors and distributes analysts' reports. "Firms are doing their own proprietary research. There is more actual analysis going on."

Data from StarMine, an independent research firm that tracks analysts' stock-picking and earnings forecasting, make the case. Throughout the 1990s and until 2001, only 1% of stocks were rated sell. Now, 14% of stocks rated by the 10 biggest Wall Street firms -- the ones covered in the April, 2003, settlement -- carry the lowest rating, according to StarMine.

Furthermore, StarMine calculates that in 2002, following analysts' advice would have had a slightly negative impact on portfolios on average. But in 2003, it would have added 2.2 percentage points to returns. Not bad. (To be fair, so far in 2004, StarMine finds that analysts' advice would have had no calculable impact on returns.)


  The most important change for the better is in the quality of analysis. In the bad old days, research on the same company from different firms was often barely distinguishable. Now, written commentary in stock reports is more independent, more thought-provoking, and better represents the upside and downside potential for a stock than the bubble era's much-hyped reports. Analyst reports are also more diverse, reflecting more originality in analysis and research tools.

In numerous interviews, portfolio managers, the main audience for Wall Street research, attest to the improvement. "It's definitely going in the right direction," says Steve Goddard, portfolio manager of the New Market Fund. He sees the biggest change in the way analysts value stocks. "They are coming up with more real-world valuations vs. the pie-in-the-sky-type work you saw that was very short-term-oriented," he says.

"I see an improving trend," agrees Gary Lisenbee, president and portfolio manager at Metropolitan West Capital Management, which manages $2 billion in assets. "Clearly, there is better objectivity than there used to be." By Amey Stone


  Some of the biggest fears on Wall Street when the reforms were put in place have yet to be realized. Some industry observers thought the changes would lead to dramatic numbers of analyst retrenchments and reductions in the number of companies covered. So far, that hasn't been the case.

First Call says just as much research coverage is being generated, partly because independent firms have sprung up to do it. Now 4,158 companies are being covered, down from 4,257 in July, 2002. But that decline may reflect the absence of initial public offerings, merger activity, and companies going private more than it does research changes. Also, as of June 1, an average of 6.4 analysts covered each company, up from 6 in June, 2002.

Another fear was that all the best analysts would leave the profession due to lower pay and increased oversight. Compensation dropped from a median of $230,000 in 2001 to $155,000 in 2003, according to the analyst trade group, the CFA Institute (formerly the Association for Investment Management & Research). Although top-paid "star" analysts have left, they have been easy to replace.

"I don't buy that there has been a brain drain," says First Call's Cooper. "There may have been a temporary dip. But a few hundred thousand a year is still a very attractive compensation level" -- more than enough to attract top newcomers, he says.


  That doesn't mean further improvement can't be achieved. StarMine's analysis shows why some might conclude that bias still exists. Of the 10 top Wall Street firms, the percentage of stocks they cover with buy ratings is 40%. But of the companies they do investment banking with, 46% are rated buy on average. The percentage of sell ratings is about the same for investment-banking clients, but only 41% of them get a neutral rating, vs. 46% overall, finds StarMine.

Portfolio managers also complain that the research still reflects short-term thinking -- focusing too much on the next quarter's results -- rather than the two- to five-year perspective they would like to see. "[Analysts] get caught up in short-term issues that aren't fundamental to the long-term growth of the business," says Lisenbee. "I haven't really seen that change."

Analysts also haven't done a great job at catching the upswing in earnings (in part, probably, because they're getting less spoon-feeding from the companies they cover). Wall Street research scores a bit worse on predicting earnings since 2002, according to StarMine. In 2002, consensus analyst estimates 60 days ahead of the report were off by an average of 2.2% from the actual report. In 2004, the average percentage error had climbed to 2.8%.


  When you look at the difference between what analysts forecast for first-quarter 2004 earnings the day before the report was issued and how they actually came in, the error was even bigger. Actual earnings were an average of 7.9% better than expected -- the largest differential First Call has seen in any quarter. Cooper believes that decline in accuracy reflects an ongoing tendency for analysts to miss big changes in the underlying economy because they focus so narrowly on a specific industry. "They always tend to miss inflection points in the economy," he says. "The first quarter, they really missed it."

Matt Kelmon, portfolio manager at Kelmoore Investment, which owns the biggest names in the benchmark Standard & Poor's 500-stock index, gives Wall Street research short shrift –- both before reforms kicked in and after. For the large, highly liquid companies he follows, he doesn't believe securities analysis can add much value since the market itself constantly revalues those stocks to reflect all available information.

And even though Wall Street research is better, is it good enough that portfolio managers and individual investors are willing to pay for it? That's a question Wall Street is grappling with as today's huge financial-services firms struggle to create a new business model for research, now that it can no longer be funded from investment-banking proceeds.

These reforms are still just getting under way. But two years into the process, with lots of modifications already in place, the result so far is a welcome change for the better.

Stone is a senior writer at BusinessWeek Online and covers the markets as a Street Wise columnist

Edited by Beth Belton

    Before it's here, it's on the Bloomberg Terminal. LEARN MORE