Man vs Machine

When Two Seconds Is an Eternity

Man vs Machine

In a previous article, we discussed the non-level playing field (to put it as charitably as possible) on which individuals who actively place their own trades compete. This message was driven home by Thomson Reuters’ announcement that it would suspend two-second early access to the semi-monthly consumer confidence index compiled by the University of Michigan, which receives about $1 million per year from Thomson Reuters for distribution rights. Naturally, Thomson Reuters wishes to make a profit on its purchase, and it does this by selling two tiers of privileged access. While the normal release time of the report is 10:00 AM EST, the company offers its regular clients the opportunity to pay a modest fee to receive the information on a conference call at 9:55. This practice is not in dispute. For an elite group of about a dozen clients, Thomson Reuters sends them the number at two seconds prior to 9:55 in exchange for a healthy payment of about $6,000 per month. In response to an investigation conducted by Eric T. Schneiderman, the attorney general of New York, Thomson Reuters has agreed to temporarily suspend this payment tier while maintaining that it is both proper and legal. It is not our place to take a stand on either question, as we are talking about privately gathered non-insider-information that may be sold to anyone at any time at whatever price the market will bear, assuming that the selling practices are fully disclosed to all purchasers and users of the information.

The question we would like to address is what happens in those two seconds between the time the elite traders obtain a market-moving piece of information and when it is given to the hoi polloi. Obviously, there is no human looking at the number and calling out a trade, as we might envision from the scene in the movie Trading Places when the orange crop data was released to the commodities trading pit. The computers are running this show, and within milliseconds, they pick off standing orders that were placed prior to the release of the information. To understand how this works, imagine the following scenario: Suppose that the SPDR S&P 500 Exchange-Traded Fund (SPY) last traded at $164.00 per share, and Joe Sixpack places a limit order to buy 1,000 shares at $163.95 per share just before the elite traders get the consumer confidence number, which we will assume is lower than expected. Joe, of course, is thinking that he will save $50.00 by not putting in a market order. Once the elite traders (or rather the machines that do their bidding) receive the disappointing consumer confidence number and estimate the new proper value of SPY to be $163.50, they will go ahead and hit Joe’s limit order (and all the others above their target price with it), and once the price hits their target, they will exit all their positions at a handsome profit. The official release of the information at 9:55 will likely afford them the opportunity to place buys to offset their original sells, as the computers owned by the regular Thomson Reuters clients put in sell orders, not realizing that the information has already been incorporated into the price. As for Joe and the hapless traders who think they are receiving actionable information at 9:55 AM, as Tony Soprano would say, “Fuhgeddaboudit!”

So where does this leave the efficient-market hypothesis (EMH)? On the one hand, we could say that when Professor Eugene Fama formulated EMH in 1965, he probably had no idea of just how efficient the market would become at incorporating information in a matter of milliseconds. On the other hand, he probably did not envision a world where a few elite organizations (the ones with the largest bankrolls and thus the most powerful computers) could consistently exploit their information advantage to profit at the expense of the smaller players. Either way, as with the movie War Games, the only winning move [for most investors] is not to play. This does not mean to refrain from investing, but rather to minimize trading activity by buying and holding low-cost index funds. To quote our prior article, “While you may not beat the market, it is likely that you will bet the majority of traders who think they can beat the market but instead get trounced by the supercomputers.”