Thomson Reuters (TRI) (my employer, but I’m not speaking on their behalf here) pays the University of Michigan a seven-figure sum every year. In return, it gets the distribution rights for the university’s closely-followed bi-monthly consumer confidence survey. It’s an arm’s-length commercial transaction: free enterprise in action. But it’s also controversial.
Up until now, Thomson Reuters has orchestrated the distribution of the data using a three-tiered system. The top level of high-frequency traders pay a reported $6,000 per month to get the information at 9:54:58am. The next level of Thomson Reuters subscribers, paying substantially less, get the data two seconds later, from a conference call which takes place at 9:55am. Finally, at 10:00am, the report is made freely available on the web.
As of this Friday, however, the first tier will be eradicated. You can still pay to get the data five minutes early, but you won’t be able to get the data five minutes and two seconds early. Thomson Reuters didn’t want to make this move — it has said repeatedly, and accurately, that it’s entirely kosher for news and information companies to distribute non-governmental data to fee-paying subscribers. And as Henry Blodget says, Thomson Reuters is hardly alone in this: there’s a massive marketplace out there of information providers charging hedge funds enormous sums for timely access to various datasets.
So what’s going on here? The first thing to note is that the only official investigation, here, is being led by Eric Schneiderman, the New York attorney general. His office is important, because it means he can wield the Martin Act; the SEC can’t do that, and the investigation is not, legally, centered on Reg FD or any kind of insider-trading statute. The Martin Act is an incredibly broadly-written statute dating back to 1921, and allows New York’s attorney general to prosecute just about anybody on Wall Street for just about anything, so long as the behavior in question can be construed to be “contrary to the plain rules of common honesty.” (Quoting the appeals court decision in People v. Federated Radio Corporation, 1926.)
The second thing to note is the way that Schneiderman’s investigation is being linked to high-frequency trading: after all, the ability to make money in the space of two seconds is pretty much by definition limited to high-frequency traders. The hedge funds who paid for that two-second head start never bothered to filter the information through a human being: they just got their computers to receive the information directly, and set them loose on the market. Again, that’s entirely legal, although there is something rather dystopian about it as well.
Of course if your main target is high-frequency traders, then going after the pre-release of a consumer confidence index is silly: it’s like attacking the use of guns by banning the sale of paper targets. But it makes sense to me that Schneiderman is concentrating just on the two-second window, rather than the five-minute window: while charging human beings for information is a normal and established part of the U.S. economy, there’s something qualitatively different about HFT algobots.
What’s more, there’s one big difference between the University of Michigan consumer confidence data, on the one hand, and the valuable proprietary information which is bought and sold every day, on the other: the UMich data is valuable precisely because it is made public, and the act of making it public moves the markets. (Sometimes.) The people who buy early access to the data don’t particularly care about the data itself, and wouldn’t pay nearly as much money for it if it wasn’t about to be made public. What they care about is the market reaction to the data: they want to be able to position themselves, in the market, so as to be able to profit from that reaction.
The UMich consumer confidence data in no way qualifies as “inside information” under the SEC definition of the term. But it is market-moving information. And the distinction between the two seems to be slowly being erased. As I wrote last year, there seems to be a general feeling, somewhere in the air, that if a medium-sized group of financial professionals get simultaneous access to potentially market-moving information (the people who subscribe to the 9:55am conference call, for instance, or the people who dial in to Lehman Brothers squawk-box calls), then that’s fine. But if an even smaller group of financial professionals gets the information even earlier, then a line has been crossed.
None of this makes much logical sense. But you don’t need a tightly-argued jurisprudential philosophy to bring a Martin Act prosecution; all you need is to convince a court that something is not fair. Call it principle-based regulation, if you must. Under the letter of America’s insider-trading laws, no one did anything wrong here. But put the law to one side; look at the way that the public reacted to the news of what Thomson Reuters was doing. To my eyes, it looked like a diluted version of the reaction to the news of what Henry Blodget was doing at Merrill Lynch, when Eliot Spitzer went on the warpath against him. Grizzled and cynical Wall Street types said, basically, “everybody knew this was happening, what’s the big deal” — but the broader public and news media was genuinely shocked, and that shock was all that Spitzer needed to extract a big settlement.
So I think that Thomson Reuters is smart to suspend the top tier of access to the UMich data, at least until Schneiderman finishes his investigation. The two-second window is entirely justifiable on legal grounds — but when you start talking about algobots paying large sums to get a eye-blink head start on market-moving information, potentially making millions of dollars while doing so, then an aggressive New York AG has all he needs to start making lots of hay. It’s probably best to sacrifice the two-second window, at least temporarily, lest Schneiderman start drawing a bead on the five-minute window, next. No one wants him to start getting ideas about the multi-billion-dollar market in financial information more broadly.