Wednesday, July 17, 2013

Hammer, Architects of Electronic Trading

Most of us rarely think about all the technology that’s required to keep markets humming—at least not until there’s something like the flash crash. We take it for granted that brokerage firms will process our trades in a timely fashion (ever more quickly) and that our account balances will always be accurate. Of course, a lot of talent, effort, and money go into building and maintaining this infrastructure.

Stephanie Hammer talked to key technology experts and in sixteen interview-style chapters takes the reader behind the scenes—and into the future. Architects of Electronic Trading: Technology Leaders Who Are Shaping Today’s Financial Markets (Wiley, 2013) is written with the layman in mind, which suggests that Hammer probably did a lot of skillful editing.

As I read this book I found two themes especially intriguing. First, how long a technological edge lasts. (I’m always interested in the fleeting nature of edges.) And second, what kinds of changes we can expect in the not too distant future. Let me share some of the often conflicting answers the interviewees provided.

So, how long does an IT edge last these days?

“An edge becomes stale extremely quickly if it is horsepower related. It is a question of months maybe.” (p. 19)

“Not as long as it used to. Advances in technology are leveling the playing field. …Software and ‘wetware’ also provide an edge, and the good news is that you can keep a software edge longer because it is harder to commoditize.” (p. 40)

“[A] pure, network speed advantage may last between three and six months.” (p. 55)

“It’s hard to say. Months. Maybe a year or two.” (p. 85)

“Putting an expiration date on an edge is tough because the duration of an edge varies greatly according to what it is. Coding (non-algorithmic) is almost commoditized. Hardware and networking are ascendant.” (p. 150)

Looking into the future, the interviewees pondered what the next “quantum leap” in trading technology would be. They suggested a range of possibilities, among them: quantum computing; the ability of wireless transmissions to cross the Atlantic and, soon thereafter, the Pacific Oceans; and complex event processing (CEP) platforms. Other interviewees pointed to technologies currently in use that will undoubtedly become even more sophisticated, such as FPGAs (field programmable gate arrays), GPUs (graphical processing units), and cloud computing. By the way, for those aspiring to “affordable desktop supercomputing,” dollar for dollar GPUs “offer 10 times the processing power of traditional central processing unit (CPU) technology.” (p. 33)

Naturally, interviewees addressed issues of high frequency trading and data centers. But technology is also used to help control risk, with an emphasis on the word “help”; technology is not a panacea. And then we have the “big data” that “firms can utilize for pattern recognition, to discover new linkages between datasets and by extension, potentially valuable correlations among products, markets, and even events.” A caveat here: “the reality of the current situation is that firms are having varying degrees of success in terms of the value that they are deriving from data.” (p. 146)

Architects of Electronic Trading offers a cogent account of the various components that make up this complex field. I am certainly better informed for having read it.

No comments:

Post a Comment