Zatera Ul

Doug at Bogus Gold pointed out an article on how (start here: On the Limitation of Predictive Models, Even When They Don’t Suck) bad predictive modeling by mathematicians and statisticians is partly to blame for the economic bubble that is now bursting. He should add physicists in there too–it was fashionable for a while for Ph.D. physicists to go into financial analysis and forecasting. Physicists model complex systems all the time–why not the stock market?

Doug rightly points out the limitations of modeling. It’s not just “garbage in, garbage out”; you can get garbage results even with good input data. In modeling, you inevitably have to leave many factors out of the model, to keep the model down to a runnable size, and you can include other ones only by approximation. This applies even to modeling single atoms (anything larger than hydrogen). I used to spend months on a single atom to get results that agreed reasonably with what was already known about it. I can only imagine that modeling the stock market would be even worse–there are irrational and semi-rational actions of humans to account for, and also the rules of the market have changed over time.

Part of the reason physicists were wandering over into the financial sector is that there is just an oversupply of Ph.D’s, thanks to subsidies from the government. Vox Day points to an older article by a physics professor about the job prospects for physicists.

There are a few advantages to getting a Ph.D in physics rather than a medical or law degree. First, with government research funding and university teaching assistant opportunities, many (perhaps most) physics grad students can get through grad school debt free. Racking up student loan debt into the six figure range would give me a heart attack. Also, generally physics students have much better working hours than medical students. Plus you don’t have to work on things that could up and die on you. Physics Ph.Ds generally go on to do interesting things, it’s just that the path for most of them isn’t nearly as clear-cut as it is for people in other fields.