How to fix Wall Street's analyst mess.

The Harvey Pitt/William Webster debacle has relegated Wall Street's other big story—New York Attorney General Eliot Spitzer's stock research reform plan—to the back page.

Until last week, it seemed that Spitzer, SEC officials, and big Wall Street firms had hammered out a global settlement to the issue of stock analyst independence. Ever since the allegations that Salomon Smith Barney analyst Jack Grubman colored his stock ratings to help win banking business for his firm, Spitzer has been angling to separate Wall Street's research and banking businesses. Under the draft agreement, large investment banks—even those whose reputations hadn't been tarnished like Merrill Lynch, Citigroup, and Credit Suisse First Boston—would pony up $1 billion over five years to establish a new industrywide stock research panel. This panel would, in turn, hire independent companies to generate analyst reports for small investors, which the big brokers would distribute alongside any proprietary research they still wished to conduct.

The deal seemed too good to be true—the round $1 billion price tag, the grand compromise that would eliminate all future claims, and the guarantee of ironclad independence. In fact, it seemed a cheap way (yes, divided among big firms over several years, $1 billion can seem cheap) for Wall Street's big dogs to put a thorny problem behind them.

But even as top Wall Street executives seemed to embrace the idea publicly, they worked behind the scenes to kill it. Now, according to the Wall Street Journal, instead of creating a new panel, Spitzer's settlement may just call for Wall Street firms to hire new executives to guard analysts from interference by investment-banker colleagues and ensure that brokerage firms also distribute research reports from other sources.

The new settlement is a nice idea. Citigroup, for example, has already taken significant steps in this direction, bringing in a respected outsider to run a newly constituted research arm. But, like the $1 billion panel, this in-house solution doesn't address the most fundamental problem of research: The people who consume it, by and large, don't pay for it. Under both Spitzer proposals, they still won't. Making research truly independent and responsive to customers would require a far more significant change in the structure of the business and, as importantly, in the mind-set of individual investors. And it would probably mean investors would pay for more than they're accustomed to in order to trade stocks.

The poor economics of research on Wall Street are indicated by the fact that at a time when bolstering research might seem a competitive advantage, many big Wall Street firms are sharply reducing their research staffs. Well-regarded analysts have been pushed onto the pavement because research decoupled from investment banking simply doesn't generate revenues.

There are lots of reasons for the lousy economics of research. The changes and reforms that have benefited investors during the past 25 years—the ending of fixed commissions on stock trades in the 1970s, discount brokerages in the 1980s, online trading and the advent of decimalization in trading in the 1990s—have been bad news for Wall Street firms. They created relentless pressure for firms to lower transaction costs. When you charge $50 instead of $300 for a trade, it becomes impossible to pay analysts out of revenues generated by trading commissions.

The '90s brought a temporary cure in the form of Jack Grubman and CNBC. Favorable coverage by a friendly analyst became part of the suite of services that investment bankers offered corporate clients. And the rise of celebrity analysts duped gullible investors, who believed that they were getting reliable information from their television. The stage was set for abuses by Grubman and others.

The same force that impelled ever-lower transaction costs led individual investors to believe that they should not have to pay for stock research. Fifteen years ago—before CNBC, CNNfn, Yahoo!, and CBSMarketwatch.com—the only way to get analysis about stocks was to pay for it, either indirectly by conducting business with a full-service brokerage firm or directly by subscribing to the Wall Street Journal or an investment newsletter. In the bad old days, you even had to pay to get access to SEC filings.

Now, of course, a plethora of good investment-related information is available gratis on the Internet and the airwaves—the SEC's Edgar system and Yahoo! Finance, to name just two sources. Given this, and the fact that the markets were rising generally, it didn't make much sense for the growing army of do-it-yourselfers to pay for research in the 1990s.