Sound Engineering: A control theorist crashes the market
B. Ross Barmish is currently working on a National Science Foundation grant to discover how control theory might apply to financial markets. Barmish, a professor of electrical and computer engineering at the University of Wisconsin-Madison College of Engineering, specializes in control theory—a field of applied mathematics that looks at the control and stability of systems. His current research, though, has him trying to convince financial scholars of ideas that run counter to their orthodoxy. Hear Barmish discuss his work and why he enjoys being a little inflammatory with his research.
Scott Gordon: Can a professor of Electrical and Computer Engineering revolutionize the way we look at financial markets? B. Ross Barmish of the University of Wisconsin-Madison College of Engineering certainly thinks so, and he’s hoping to stir up debate by taking his ideas right into the financial research community. Barmish has a National Science Foundation Grant to apply something called control theory to trading stocks. Here’s a rough version of how that works.
B. Ross Barmish: Let’s imagine a rather trivial scenario where you’re trying to fill up a bathtub with hot water to a desired temperature, let’s say 90 degrees. Well, if you’re a modeler, the obvious thing to do is start writing thermodynamic and fluid-dynamic equations and eventually solve those complicated equations and reach a conclusion that you should turn the hot water to such-and-such a level and the cold water to a different level, i.e. faucet control. On the other hand, one could take the point of view that a model is entirely unnecessary. Add more hot water if the bath water is too cold, and add more cold water if the bath water is too hot. Well, this analogy can be carried over into financial markets quite easily, where instead of using water, one uses money, and the question of increasing or decreasing the level of investment is analogous to increasing or decreasing the water flows.
Scott Gordon: Now, the approach also involves carefully crafted algorithms and simulations that use large amounts of real stock-market data. But Barmish is basically saying that you can dispense with empirical models of market behavior. That pits him against the orthodoxy of financial thought.
B. Ross Barmish: What I like to do, therefore, is go into that community and actually explain my work, even if it is provocative. Matter of fact, I would say my goal oftentimes is to be intentionally provocative in order to stimulate good debate. My goal is to push the envelope just a little bit so that I actually get response, and that serves as a feedback for me. After all, given that my research area is feedback control theory, let’s put it into practice.
Scott Gordon: Ultimately, Barmish does all that button-pushing in the financial community because he feels a responsibility to make sure he actually contributes something to financial thought, not just the study of control theory.
B. Ross Barmish: My belief, even though we’re doing research in the control area, is that the ultimate accountability for this work really goes to the end user, who are the people in finance—hedge funds, banks, all users, investors, of the theory.
Scott Gordon: For more information on Barmish’s work, look up his faculty page at engr.wisc.edu.