Toomre Capital Markets LLC

Real-Time Capital Markets -- Analytics, Visualization, Event Processing, and Intelligence

IBM: Algorithms for Financial Markets

Toomre Capital Markets LLC ("TCM") would like to point out an interesting point of view article recently published by IBM Global Services. Entitled Algorithms for Financial Markets, this article gives some good perspective on how what the transaction environment will be like when algorithmic trading is persuasive across many market sectors.

Clearly, a significant portion of the liquid trading business will be run on machines co-located at exchange or trading center data centers. In those environments, the speed of completing the loop of receiving new market quote information, making some calculations, and then executing the trade order really matters. Cutting milli-seconds or even nano-seconds out of the process will be really important and a key to being a liquidity market leader. (Financial firms and trading centers wanting to compete in this sector should look at StreamBase Systems for their extremely fast Complex Event Processing engine.)

The more interesting thing is how is algorithmic trading going to affect less liquid securities. Yesterday the Dow Jones Industrial Average plunged more than 400 points for its biggest loss since 2001. During such market moves, liquidity typically dries up, at least temporarily. Will the algorithms for less liquid securities be "smart" enough to handle such sharp market moves? Or will the human trader be needed to handle such illiquid periods?

TCM remains skeptical of trusting algorithms during such sharp market moves. The algorithms typically are trained or checked vis-à-vis some previous trading history. There simply are too few (in a statistical sense) data points of times of extreme time or greed. Prudent institutions might well want to turn off their algorithmic trading systems until some degree of stability reappears. The interesting thing that then happens, as prudent institutions pull back from trade execution, the supposedly "liquid" securities drift to more illiquid conditions. Are the algo trading engines able to automatically adjust to the fact that there is less liquidity?

All this brings to TCM back to a central theme. Algorithmic trading presumes that there is some degree of market liquidity. What happens when there simply is no liquidity? Are hedge funds, Wall Street investment banks and other active investors properly factoring in liquidity risk? TCM would suggest not. Reader thoughts and comments are welcome.