The SEC’s weird newswire investigation

January 31, 2013

A couple of weeks ago, the WSJ’s Brody Mullins had a big story about the fact that the SEC was investigating a political-intelligence consultancy named Marwood. Marwood doesn’t seem to have done anything wrong, but the very fact that it was being investigated was, at least as far as the WSJ was concerned, front-page news.

This week, Mullins has done it again, this time with an SEC investigation of firms which provide financial data. Once again, there’s nothing in the story to suggest that any of these firms, which include Bloomberg, Dow Jones, and Thomson Reuters, have broken any insider-trading rules. And yet here’s a juicy front-page story all the same, based entirely on the fact that there was an investigation at all, regardless of whether the investigation actually discovered anything untoward.

I’d love to know the story behind these stories. It seems pretty obvious that they’re being leaked by the SEC, in a way that seeks to embarrass the subjects of the probes as much as possible. Marwood and Bloomberg and Thomson Reuters might have done nothing wrong at all, but if the WSJ determines that there’s front-page news here, then its readers are surely expected to conclude something about smoke and fire.

There’s a clear implication in the latest story, for instance, that the data companies in question (which include the WSJ’s corporate parent) did do something wrong, and that they’re just lucky the SEC can’t prove it in a court of law:

Investigators decided against filing charges because they couldn’t link the pattern to specific actions by media companies, people familiar with the probe said.

A key issue, one of the people said, was whether the government could prove in court that a time advantage for a trader of a sliver of a second—as little as a few thousandths—was enough to conduct profitable trades on confidential information.

Even so, these people added, investigators continue to have general concerns about the handling of federal economic data.

This whole thing has a decided whiff of “doesn’t the SEC have anything better to do”. For one thing, to answer the SEC’s question, it’s not at all obvious that getting information a few thousandths of a second ahead of anybody else would allow some computer somewhere to conduct a profitable trade on the information. Firstly, big economic data comes out before the stock market opens, which means that any profitable trades would have to take place either on the much less liquid out-of-hours market, or else on the bond market. Both of them are largely free of high-frequency traders.

Yes, there’s a lot of trading and jostling and positioning in the bond market in the run-up to a big data release, but I can pretty much guarantee you that all markets are in holding-their-breath mode when it comes to, say, the final couple of seconds. The traders and the algobots are short or flat or long, they’re waiting for the number, and then they’ll burst into action as soon as the number is released. If you want to trade a couple of thousandths of a second before the number is released, you’re going to be looking for a counterparty who doesn’t know what the number is but who is willing to trade anyway. It’s hard to imagine such counterparties exist.

The news agencies can blame themselves a little bit, here, because they have for many years been highly invested in the idea that if you get a certain piece of information first, even if it’s just by a fraction of a second, then you can make a huge amount of money. All of them get incredibly excited about the times when they move the market: when a story comes out, and then some financial instrument — normally a stock, but a commodity will do in a pinch — moves sharply on the news. They charge a lot of money for their real-time news feeds, and the implication is something like this:

  1. The news hits the wire.
  2. A smart trader, staring intently at his newsfeed, sees the headline cross the wire, and immediately groks the implications.
  3. The trader then puts in a monster buy/sell order, picking off a bunch of tortoises who aren’t smart or rich enough to subscribe to the wire service in question.
  4. The price moves sharply.
  5. Monster profit!

It’s a lovely story, but it’s also a fairytale: things don’t actually happen that way. In the real world, when a piece of news hits a wire, at that point it’s public. And once it’s public, the market then reflects that public information in the share price. If you’re a broker-dealer who was quoting a security at one price before the news came out, you’ll now be quoting it at a different price after the news has come out.

The key question to ask is this: how many trades happened (a) at the old price, but (b) after the news became public? Most of the time, the answer is zero, or very close to zero. News headlines often move the market, but that doesn’t mean that someone has gotten financial benefit from reading them first.

The point here is that once a headline crosses the wire, that information is, by definition, public. And if it’s public information, it can’t be insider information. There are lots of good reasons why the U.S. government and rival news agencies would be cross if one of the wires published that information a fraction of a second before the other ones did. But just because someone is cross doesn’t mean that laws have been broken, or that inside information has been traded upon. An embargo is an agreement between a news source and a journalist; it’s not something to be enforced by the SEC.

So I do wonder what the SEC thought it was doing, here, conducting what the WSJ describes as a “technically and legally complex” probe. What exactly was the SEC hoping to achieve? And why is this weird investigation, in and of itself, newsworthy as anything other than a waste of government resources?


Comments are closed.