"Kee" Points with Jim Kee, Ph.D.

Is the market rigged? That’s the take of Michael Lewis’ latest book, “Flash Boys,” which I read over the weekend (you should expect that from someone with the title of Chief Economist!). The gist of the story has to do with high frequency trading (HFT), and of the ability of firms with high speed access to information to front- run orders by locating servers closer (milliseconds) to exchanges, or by buying access to trade orders. To front run, a broker needs to find out about an order to buy a large block of stock, perhaps placed by a mutual fund or pension fund, before the rest of the market and before the order is filled. The broker then buys shares for itself, fills the order - which if large will drive up the share price – and then sell its own shares at the higher price. Front running is illegal but hard to detect, particularly with the complexities of computerized trading systems.

 

Just to be clear, from our perspective at South Texas Money Management, any issue that possibly interferes with fair trade execution deserves serious regulatory scrutiny. High frequency trading is under scrutiny and should be. But it is unlikely that any of our trades would be large enough to be impacted by these front running efforts, which are geared towards big orders at the largest institutions. Now, back to the book.

 

The book’s main protagonist is Brad Katsuyama, former Royal Bank of Canada trader (and founder of IEX exchange) who discovered that firms were engaged in various versions of front running. When discussing the phenomenon in a television interview last night, Katsuyama stated that “people will always try to use technology to game the system.” And when asked how he was able to uncover it when regulators with far greater resources did not, he replied that it was “because they have to rely on industry players.”

 

That idea - that a regulatory agency can be “captured” by the firms being regulated in order to serve their – the firms’ – interests instead of the public interest is an old one in economics. George Stigler won the Nobel prize partially for papers on the topic of regulatory capture over 40 years ago. So is the idea that people will always try to find a way to front run. Both are implied by Lewis’ book. In fact, when I saw Michael Lewis being interviewed on television, my first thought was to find (it took about a minute) economist Susan Woodward’s paper on the Securities and Exchange Commission (SEC), regulatory capture, and front running. Woodward is the former Chief Economist of the U.S. Securities and Exchange Commission and extraordinarily bright. Here’s a quote from her paper, “Regulatory Capture at the U.S. Securities and Exchange Commission,” delivered at a Milken Institute conference over 16 years ago(!). It is worth reading carefully:

  

"revealing one’s identity and trading plans to one’s broker can be costly. Large orders to buy or sell often move price. A broker who gets a large order to buy could first buy some stock himself, then buy for this customer’s order, watching the rise in price, and then finally sell the stock he bought for himself at a profit. This is called frontrunning. It is illegal. But it is very difficult to detect, and it is unlikely that institutional investors would have such a strong preference for anonymous trading venues if it did not happen with some frequency.”

  

The point is that anyone familiar with regulatory capture and the incentives to front-run wouldn’t be too surprised to find that they exist. That doesn’t mean that a certain amount of outrage isn’t warranted, and Lewis’ book is filled with outrage. The book is somewhat one-sided, which is a writer’s prerogative. It is asserted that “the little guy” is hurt by HFT actions, though little evidence is given that would show they are unable to buy stock at posted prices. And the tremendous boon to individual investors from the dramatic fall in trading costs caused by automated trading systems is downplayed. Finally, the role of providing liquidity is quickly dismissed by Lewis, though even Woodward acknowledges that “the securities markets of the Unites States have been dazzling in their effectiveness at raising capital and supporting the creation of value.”

  

But the evolution of high speed trading algorithms and how to regulate them is a serious issue, as the May 6, 2010 “flash crash” attests (the Dow lost over 1,000 points, then recovered in minutes). Regulatory change often derives a sense of urgency from outside pressure generated by influential individuals like Lewis and Katsuyama, and in that sense they are performing a valuable service. Academic researchers have already been working on solutions to the problems caused by high frequency trading.

  

A good example is the work of University of Chicago Professor Eric Budish, who back in 2010 became interested in the same firm, Spread Networks, that Lewis describes in the opening chapters of Flash Boys [Aside: Spread Networks completed a fiber-optic connection between downtown Chicago (home of the Chicago Mercantile Exchange or CME Group’s electronic exchange), and data (trading) centers in New Jersey that was hundreds of miles shorter than existing routes and milliseconds faster]. Budish proposes slowing down trading by replacing continuous-time orders with discrete (“batch”) auctions. It’s pretty technical, but the gist of it is that current, continuous markets process messages serially (i.e. “one-at-a-time”) which creates a tiny advantage to speed (learning orders early). The first in line wins (Chicago Booth). Budish asserts that frequent batch auctions would, to make a long story short, reduce the value of tiny speed advantages by “batching” or collecting similar trades for a brief blind auction. Discrete auctions are not new in economics, but they are timely. These would be “double blind auctions” in which potential buyers submit their bids and potential sellers simultaneously submit their asking prices (none of which are displayed, that’s called a “sealed bid”) to an auctioneer that then chooses a price that clears the market. Of course applied to trading exchanges it would all be automated, occurring at the “blink of an eye” (American Economic Review Papers and Proceedings). That’s just one solution; others will follow.