Spitzer's next crusade He's taken on Wall Street, insurance brokers and the mutual fund industry. Now for the big time. (Full story)

NEW YORK (FORTUNE) - A recently unearthed "highly confidential" Citigroup memo openly discussed the "pressures" keeping research analysts from providing investors with honest research.

In the 2002 memo, John Hoffman, then global research chief for Citi's Salomon Smith Barney division, advised Salomon Smith Barney CEO Michael Carpenter of the internal view that "implementation and enforcement of clearer and more accurate ratings is in conflict with certain paramount goals of our firm"—namely, maximizing underwriting fees. (The memo was made public in ongoing litigation filed by a Florida law firm, Babbitt Johnson Osborne & LeClainche.)

"When I read it my blood began to boil again," says New York attorney general Eliot Spitzer. "It drove home how crass that whole system was."

As a participant in the 2002 global Wall Street settlement resulting from Spitzer's investigation, Citi has since reformed its research practices.

To: Michael A Carpenter

From: John Hoffman

Re: Revision of the Ratings System

Date: March 22, 2002

HIGHLY CONFIDENTIAL

In an extensive discussion of stock ratings and a revised rating system by the Equity Research Directors last week, several general points were made which deserve your consideration and probably discussion at the Operating Committee level.

First, the equity research directors agreed that the clearest, most understandable system for our sales force and investors would be a simple Buy/Hold/Sell categorization. However, that simple system is unlikely to produce the balance in ratings sought by the regulators and the investing public. Pressures on the analysts by corporate management, by investment bankers and by selected institutional accounts cause the analyst to steer away from negative ratings.

Most of the research directors felt that they could overcome or at least lean against this positive bias with a more structured and more forceful review process. While none would expect total balance, they would expect the number of negative ratings to rise to a reasonable level (Jeff Waters pointed out that if one were to retranslate the current 1/2/3 ratings from Buy/Outperform/Neutral to Buy/Hold/Sell as they are interpreted by many of the users, the distribution is close to one-third each).

However, the research directors feel strongly that implementation and enforcement of clearer and more accurate ratings standards is in conflict with certain paramount goals of the firm. Research analysts have been told repeatedly that the primary goal of the firm is to get our equity underwriting market share ranking up into the top three. This message is reinforced in the investment banking weekly meetings and planning offsites to which research management and analysts are invited. The success of the Platinum Account program and the Equity Task Force initiatives are very dependant on the corporates' positive view of equity research and there is no mystery that this positive view hinges on a favorable stock rating. In contrast, seldom does research management hear from top management about the damage that a confusing or biased rating system has done to our sizable retail business in the form of depreciated assets and lost accounts. And seldom in recent years does research management participate in any retail business sessions.

The scope of this conflict is obvious when we compare the Investment Banking initiatives with our research coverage. The Investment Bank has designated 400 corporations as Platinum Accounts and another 1,700 as "Core" accounts. Each the corporations so designated is expected to produce $20+ million or $5+ million respectively in fees to Citigroup. Equity Research covers about 3000 companies, and since these also tend to be the leading corporations in the world, the overlap with these ID priority account lists is substantial. To have even 20% of these stocks with a negative rating (probably the minimum to satisfy the demands for balance) means that a lot of Platinum and Core accounts will carry some kind of negative equity rating.

The equity research directors question the Investment Bank's ability to accept stricter rating standards at the expense of revenues. We discussed extensively whether we could do this with a "revenue neutral" impact on Investment Banking and the answer was clearly "no". In fact, it is more likely that the new ratings standards if enforced could competitively disadvantage our Investment Bank given that historically it has been more dependant on equity research support than its leading competitors.

Given this last point, there was concern about franchise risk if we should attempt to lead the industry in setting new rating standards. These doubts were increased with the release this week of the details of the Morgan Stanley system, which at first glance looked like it might answer some of these conflicts and received a lot of publicity, but which in practice is anything but clear and understandable.

Before we institute a new, more disciplined and more balanced rating system, we need complete support from you and from all of the division heads of the firm. Are we all agreed that we want to be a leader on the topic of more accurate and more balanced ratings, recognizing that this could increase conflict with our major corporate clients, could put our bankers at a disadvantage to those who moves slowly or not at all, and will almost certainly dampen the Investment Banking momentum towards a leadership market share in our industry?

The design of a new rating system is partially dependant on your answer to this final question. I think you and I should discuss this issue privately before we get any of the business heads group involved. You have the only hard copy of this memo.