The future of economics research - Chaos theory and behavioural psychology

Hi,

http://www.timesonline.co.uk/tol/comment/columnists/article5689642.ece

In recent weeks, journalists and market pundits, like Anatole Kaletsky at The Times, have predicted that economic research is going to branch out into a number of different areas, leaving behind the 'flawed' principles that have driven much of recent economic study. Bell curves/ normal distributions or Monte Carlo VaR models, as tools for risk managers, have come in for particular criticism. 'Chaos Theory' and 'behavioural psychology' are two of the more popular topics that have been suggested as new avenues for economists to explore. I'm not going to dwell on what Chaos Theory is, or any other new theory for that matter, (search on wikipedia if you'd like a general explanation) but what I am going to address is how these theories, and the implications taken from them, aren't actually new and have been used in a handful of areas of the financial system for a number of years now.

Recently, I've heard statements from Wall St. CEO's saying that the markets are in chaos. If markets are indeed 'chaotic', and 'Chaos Theory' will be to economics what The Theory of Relativity was to Physics, what implications does that have for real world practitioners? Pension funds still need to match assets and liabilities despite this chaos. Insurance firms who have minimum capital requirements remain exposed to risk but must meet these restrictions despite the huge amount of uncertainty flooding our economic system.

The good news is that you don't need to rewind all the way back to Poincare and restart your education in Economics from scratch. You also don't have to wait for new 'chaos' research to start being published to take advantage of the findings that will come out of this new research. The reason is this: the conclusions one can draw, given the fact that markets present as chaotic, and the conclusions one can draw from the fact that markets are driven by human beings rather than bell curves, have been utilised in certain pockets of the financial system by cutting edge risk managers for a number of years. These methods are relatively new and they have recently been thoroughly tested through the ravages of the credit crunch, and, put simply, they work. The expertise and the experts are already out there if you know where to look.

One problem, however, is that these new techniques were far from ubiquitous as the credit crunch struck. In fact, they were very rare. Had they not been, many firms would have been in a much better position to cope. During the past two decades, as the complexity of financial products has exploded, many risk managers and many senior financial managers made the mistake of believing that the bell curve inspired statistical models they used to measure a firm's risk were somehow so sophisticated as to be able to predict the future. This led them to accept risk they didn't fully understand because it fit into their model.

Now, if markets present as chaotic, and you are measuring risk using a bell curve model, you should realise that, amongst other things, you are leaving yourself exposed to a sudden change in market environment. Bell curve Monte Carlo models might be sophisticated maths but they aren't magic. They also can't predict the future and in fact, they are quite poor at predicting the future. To set large amounts of strategy and risk appetite based off of bell curve Monte Carlo model output, without further risk analysis, is a big mistake if you believe markets are chaotic.

Of course, risk managers will learn. One crucial lesson they must embrace is this: any statistical risk model (and not just the popular bell curve Monte Carlo ones) MUST be seen for what it is - nothing more than help for a company to understand the economy in which it is operating, and most importantly the changing face of the risks to which it is exposed. Fortune tellers these models are not. There are occasional market environments when these models should be ignored. A risk manager's skill comes in knowing when to ignore the model to make his decision or knowing when to use his model output to shape his decision.

These risk and capital models, whether Bell curve inspired or otherwise, are one bullet in the risk manager's arsenal. The weaknesses and failings in the model must be fully understood and planned for. The key lesson is to change attitude from an unthinking and unquestioning strategy based on flawed models to the following:

1. A desire to push the boundaries of understanding with regards to the niche you hold in the economy. Statistical models of a variety of different types are needed to do this but they must be pulled apart and questioned by the risk managers who built them. Risk managers need to be some of the smartest people in their company.
2. Once your niche is understood as much as possible, you can set about formulating a strategy based off of market expectations and ALSO dynamically hedging this strategy to defend against unexpected events. And you will have many unexpected events.

Chaotic markets present excellent opportunities as well as problems, and any business strategy must be dynamic and flexible enough to change as the market environment dictates. One lesson this crisis has taught us is that the moment you think the dust has settled and you know what is around the corner is the moment you're surprised.

Happy hunting,
Andy Shaw