How High the Firewall? Separating Investment Banking from Research
At the March 2000 peak of the stock market, buy recommendations outnumbered sell recommendations 73 to 1. “Now, being wrong doesn’t necessarily mean that something illegal happened,” Glenn Hubbard, dean and Russell L. Carson Professor of Finance and Economics, told incoming students at a panel discussion called, “How High the Firewall? Separating Investment Banking from Research,” the year’s first special session as part of the Individual, Business and Society: Tradeoffs, Choices and Accountability curriculum. But the collapse in corporate governance scandals of companies analysts were cheerleading highlighted the research flaws.
New York Attorney General Eliot Spitzer’s investigation into why securities analysts didn’t uncover financial fraud or earnings manipulation and alert investors to problems before stock values fell resulted in a sweeping settlement with leading investment banks. The settlement required investment banking firms to sever links between research and investment banking; to end the practice of compensating analysts for equity research; to make their analysts’ recommendations public; and, for a five-year period, to contract with three or more independent research firms to provide research to the banks’ customers. The industry as a whole had to pay fines exceeding $1.4 billion.
Considering the apparent conflict of interest in financial institutions providing multiple services — such as investment banking for corporate clients and brokerage for retail clients, where the revenues for underwriting greatly exceed brokerage commissions — it might seem logical to completely separate the two functions. But many economists are concerned that the settlement firewall is too high. They argue that a similar separation between commercial and investment banking was unsuccessful and that the synergistic benefits of sharing information across multiple services in a firm outweigh the potential costs of conflict.
“I think many practitioners generally agree that it is important to have increased disclosure for investment analysts — and, for that matter, credit-rating agencies and auditors — that would reveal any interests that they have,” Hubbard said. “There should definitely be codes of conduct to control conflicts of interest. But there is some disagreement in the context of trade-offs.”
“Regulation will help,” said Trevor Harris, managing director at Morgan Stanley and codirector of the Center for Excellence in Accounting and Security Analysis at the School. “But it will also hurt quite a lot.”
The trade-off, he said, is that retail investors, who were hurt most in the scandals, will now suffer even more from tighter controls on information. “I think the way the market is going to work in the short term is that you’re going to see much less good research made available to retail clients,” Harris said.
Sallie Krawcheck ’92, CFO and head of strategy at Citigroup, Inc., expressed similar concerns, saying that an unintended consequence of the 2000 Regulation Fair Disclosure, which was created with a similarly good intention of combating selective disclosure of material nonpublic information, is that analysts and executives have become much more guarded about what they say, reducing disclosure to everyone. “I’m sure that the information that is getting out there now is getting out to everybody all at once,” Krawcheck said. “But the problem is that the regulation has retracted the stuff that we will actually tell people.”
Another problem is that the settlement does not address more systemic flaws, such as research techniques and analyst incentives. “If you put up a firewall, the analysts can change to some degree,” Krawcheck said. They no longer can talk to investment bankers without a lawyer present, for example. “But what they’re not doing is changing the way they do their research and put out research,” which is itself a conflict of interest.
Too many analysts rely on information the companies provide, rather than doing their own independent analysis, she said. As a result of lazy research habits and personal relationships with institutional clients, analysts can forget that the investor, not the company, is their client, and can become susceptible to pressure from buy-side analysts and company clients to make positive calls that keep stock prices up. The system offers incentives for analysts to make models look “pretty,” even if they don’t always hold up, Harris said.
“I’m a CFO now,” said Krawcheck, who during the 1990s was the most successful analyst on Wall Street because of her willingness to make controversial calls and who, as director of research at Sanford C. Bernstein & Co. in the late ’90s, early 2000s, issued more sell ratings than the rest of Wall Street combined. Now, she said, “I don’t like the analysts who are mean to us. It upsets me and I think they’re wrong and I don’t want to talk to them very often.” Nevertheless, she thinks it is important for analysts to remember that they have “one client and one client only,” she said. “If the company gets upset with you, then so be it — that’s not your client.”
Harris, among other economists, suggested resolving this conflict by doing away with the rating system. “What does it really mean to say that a stock is expensive or is a buy or a sell? Is it in a 3-month period, a 6-month period, a 12-month period or a 2-year period?” Harris said. “Any single word to represent where the risk-reward trade-offs are at a point in time never makes any sense. Because the reality is that we all knew — including the technology analysts at the time, at least the good ones — that the stocks were overvalued. The problem was, when was the price going to change?”
Certainly, the rating system could be improved, Krawcheck said. But she strongly disagreed with trashing it entirely. “We did this at Bernstein for a while, and our analysts lost their focus,” she said. “Instead of driving toward the goal of the business, which is to make money for clients, they ended up doing some really terrific, highly esoteric and completely useless research. If you try to take away the ratings, my experience is it ends up being research for research’s sake.”
Instead, she said, improved research depends on analysts severing relationships with the companies they cover. “The company is selling something,” she said. Once her analysts at Bernstein “stopped listening to what was being sold and did their own research,” she said, “they got more accurate.”
The Slippery Slope
Without excusing any of the scandals, Krawcheck said, “I think it’s pretty easy to see how Wall Street got itself into trouble over the past several years. If you look at the course of financial services history, what happened a few years ago was actually pretty predictable.”
Until the mid-1970s, when commissions were deregulated, analyst compensation came from fixed commissions. After deregulation, commissions fell 75 percent very quickly. Many Wall Street firms went under, and others joined forces to get through it. The firms that survived decided to have analysts and investment bankers share information, work together on deals and, ultimately, serve both the company and investor clients.
“Wall Street went down what was a slippery slope,” Krawcheck said. “This step wasn’t so terrible and the next step wasn’t so terrible and the next step wasn’t so terrible. But before you knew it, the analyst was caught in a classic conflict of interest” with investment bankers issuing stock and researchers setting its price. “The investor is on one side and the company is on the other side,” Krawcheck said. “The investor wants a lower price and the company wants a higher price. And the analyst is put in the situation of saying, ‘OK, which way do you go?’”
None of these situations were black-and-white, she said. Most were shades of gray, but all involved a corporate culture that encouraged analysts to serve two sets of clients. And, while the firewall might resolve this particular conflict of interest, Krawcheck warned that ethical conflicts pop up everywhere.
“I can assure you, if you are going to work on Wall Street, you’re going to work with some very smart people who are good at saying, ‘Well, how about if we do this? Can we do this and this?’” Krawcheck said. “The answer is always yes. You can get from one point to the other, but at some point you need to trust your gut and say, ‘No, we just went a step too far,’” even “when your boss says it’s fine and the lawyers say it’s fine.”
As for regulation, Krawcheck and Harris agreed that its success will be measured by the retail sector. “If retail investors who feel like they were harmed can come back and feel like they can use Wall Street research with any degree of confidence,” Krawcheck said, “that’s how we’ll know we’re successful.”