Now that you’ve seen what ACA Insight has to offer, don’t be without it. Subscribe now!

The weekly news source for investment management legal and compliance professionals

Current subscribers - please log in to the website in the upper right-hand corner

News October 25, 2004 Issue

MFS’s Pozen: No Unbundling, Please

Instead of requiring advisers to "guesstimate" the value of research, the SEC should require advisers to disclose average commissions per share.

So said MFS Investment Management chairman Robert Pozen, speaking at last week’s SIA Mutual Fund Reform Update conference. Pozen said his views represented a majority of people on the NASD’s soft dollar task force, but added that a "significant minority" held a different viewpoint.

Pozen trotted through a quick, but entertaining, history of the soft dollar issue. He called the "lawful and appropriate assistance" standard in the 1986 release "the most Pandora’s-box-opening thing that the Commission has done in years." He recalled telling SEC staffers at the time: "You’ve opened the door for every soft dollar practice in the world."

Then came the SEC’s 1998 soft dollar study, which found that soft dollars were being used for things that had nothing to do with Section 28(e)’s original intent. "Did the Commission take enforcement action?" asked Pozen. "Did the Commission put out a clarifying release? No." The SEC, he said, "has been remiss," jokingly adding that New York attorney general Eliot Spitzer surely will be bringing an enforcement case against the SEC for having done nothing about 28(e), despite its "obvious awareness, with scienter" of the problems.

He continued: "Now, having done nothing for all this time," the SEC last year started bringing revenue sharing cases, which he said were "really pretty good in terms of revisionist history, in terms of rewriting the law." The cases are "pretty brutal," he said. "The Commission has been so lenient on this subject and then all of a sudden has gotten very tough." He said that he doubted that many people would attempt to litigate the cases. Pozen also pointed out the irony of the sequence of events, noting that the SEC adopted the prohibition on directing fund brokerage for distribution about a year after it started the enforcement cases. "Usually, you go the other way around, but apparently not in this case," he said. In his view, the SEC should have banned fund brokerage for distribution prospectively, and not brought the cases.

Turning to the outstanding issues: Pozen agreed that soft dollars should be spent only on things that the investment manager legitimately can say is helping him do a better job in the investment process. However, he pointed to a complication in the SEC’s "intellectual content" standard for research, predicting that it will "turn out to be a real constraint" — the fact that Section 28(e) applies not only to research, but to brokerage-related services, as well. Delivery systems that might not count as research, he said, "may be legitimate adjuncts to the trading process." He gave the example of fixed protocols "and the pipes that go along with that." Those, he said, are very important to the execution process. While trading platforms may not be "research," he said, "they may be legitimate, enhancing parts [of the brokerage process] and therefore deserve protection."

Pozen then turned to the third-party / proprietary research debate, and the related topic of unbundling. What the problem really comes down to, he said, is that "in principal, we all want to treat Street research and third-party research the same." In practice, however, managers know exactly how much they are paying only for third-party research. "You can pretty much tell what you are getting and how much you are getting." Do managers know how much they are paying for Street research? "Clearly, no," said Pozen. "I’ve spent a lot of time on this problem," he said, calling it a very difficult issue. "Any attempt to value the research component in the brokerage commission we get from the Street would be in the area of guesstimation," he said.

To illustrate his point, he gave the example of services provided by Goldman Sachs. MFS, he said, might be Goldman’s 1001st client. Goldman could value the research provided to MFS using its incremental cost (which would "probably be zero") or average cost (which is probably "very high"). While Pozen said he didn’t know whether Goldman should use incremental or average cost. "I do know this: Goldman’s value of the research that they’re providing is quite different from than MFS’s value." He also pointed out that MFS might value Goldman’s research differently than would other money managers. He also noted that Goldman’s assistance might include very nonquantifiable services, such as Goldman analysts training new MFS analysts. "How do you put a dollar value on that?" he asked. "I can guesstimate as well as anybody else but the problem is that it is an extremely difficult valuation."

And that, he said, is the crux of the matter. "If we had the same sort of relative certainty" about the price of Wall Street research as for third-party research, "we would all decide to disclose what we are paying."

As an aside, Pozen noted that the third-party research people have "clearly won Chairman Donaldson’s heart." He said he was "quite sympathetic" to their concern that putting a dollar value on their research, but not on proprietary research, will put them at a competitive disadvantage.

Pozen clearly opposed the notion of requiring advisers to unbundle the cost of research. But "what would be worse," is if advisers were required to unbundle and didn’t get valuations from brokers, he said. While he noted that advisers "may agree or disagree" with a broker’s valuation, "at least it gives us a framework to say that."

Valuation "is going to turn out to be really the stickiest issue in the widget," he said. "We can spend a lot of time debating on what percentage of the brokerage commission should be allocated to which function." Pozen said that he has personally come to view capital commitment, which "arguably isn’t a soft dollar service," as perhaps the most important part of the brokerage commission. "I’d be willing to pay 6 cents a share for somebody who steps up and buys the last 200,000 of my remaining shares" of a thinly-traded stock that is being unloaded very quickly, he said. And unbundled commission disclosure would mean that advisers would have to specify that "22.4 percent went to research, 16 percent went to capital commitments for brokerage," he said. "We can all make up the numbers, I just don’t think that they have great degree of certainty."

Because valuation "is a very, very tricky area," Pozen said that he is advocating the use of average commissions per share. "That we actually know," he said, "If we get average commissions per share, we know what electronic trades cost. I think everybody can then make a pretty good ballpark notion without trying to make believe like we know the exact value of things." If people are paying 5 or 6 cents versus 4 cents a share, "that would tell them something about what sort of other services they are getting, other than execution." Managers wouldn’t have to segregate it or unbundle it, but "you would have a pretty good idea that they were either bargaining hard for you or they’re not taking as many ancillary services."

To work, he said, managers would have to state what percentage of their trades were done on a principal basis and what percent were done on an agency basis "because otherwise people would figure out some way to stuff everything into the principal basket." He also said that international commissions would have to be broken out, since in most countries commissions are calculated using basis points, not cents per share. "To talk about them in the same breath is quite confusing."