Monday, 19 January 2009

The high risk nature of academic theory on risk management

One of the most refreshing things that could emerge from the current crisis in the financial markets would be for a radical rethinking of the way that finance is taught at the most prestigious business schools. Future MBA’s should be taught something more historically and philosophically oriented on the nature of risk, rather than the same mathematics that is useful in the physical sciences but is more or less useless when applied to the world of financial economics.
There has been a lot of commentary about the reasons why so many of the experts did not see the coming crisis and many well documented examples of the irresponsible and negligent failures of those should have been managing risks for the banks and the financial services industry. What is far less commented on is the fact that the underlying financial theory of risk and the probability of financial accidents arising is not just wrongly conceived, but dangerously so. None of us, including the test pilot, would entrust our lives to a new airliner which had not been robustly stress tested under the most extreme modelling conditions for aerodynamics and metal fatigue etc. and yet financial products were sold that not only were not thoroughly tested but, in assessing the likelihood of accidents or failures, the financial engineers used the wrong modelling techniques.
There is a serious conceptual problem in modelling a financial product with techniques from the physical sciences because of the sudden and dramatic discontinuities in financial time series data. Sharp gaps and other severe dislocations show that price, and economic behavior in general, cannot be adequately represented as following a trajectory which can then be analyzed in any standard statistical theory. One of the most severe consequences of this conceptual error is the fallacy, which has been more than amply demonstrated by the current financial meltdown, that the probability of large moves in asset prices can in any sense be mapped according to any kind of normal distribution, Gaussian or otherwise.
A simple example can illustrate the inappropriateness of standard statistical theory to the world of finance. There is no meaningful sense to even estimate the probability of discovering a man who is 25 standard deviations from the average height of male human beings, and yet in a well known quote by a senior Goldman Sachs executive who, when asked why one of their funds had lost more than 30% during the onset of the financial crisis, is reported to have said

“We were seeing things that were 25-standard deviation moves, several days in a row. There have been issues in some of the other quantitative spaces. But nothing like what we saw last week.”

From this quotation alone it should be clear to orthodox risk managers educated within the traditional notions of academic finance theory, that there is something profoundly ill conceived with using probability theory based upon a normal distribution, and all the Value at Risk baggage that comes with it, in order to quantify risk in portfolio construction and risk management.

What is needed is a new foundation for the understanding of how to quantify risk and the likelihood of financial collapses and contagion. There have been some promising developments in this field by writers such as Mandelbrot, Hyman Minsky and even Nassim Taleb but the field is still in its infancy. Perhaps there is no underlying logic to financial behavior and that we have to accept that financial meltdowns are as unpredictable as massive earthquakes. But the fact that many people did see our current “falling off a cliff” coming suggests that we may be in a much better position than seismologists.

No comments:

Post a Comment