Did reliance on VaR contribute to the current economic crisis?

I am sure many people read the Risk Management cover story written by Joe Nocera featured in this week’s New York Times Magazine. The article provides some background on Value-at-Risk, and discusses whether proper use of it could have helped prevent the current financial crisis or whether reliance on it helped cause the crisis. There have been many discussions on the article already including Barry Ritholtz’s blog.

Many people interviewed expressed greatly differing views, some of whom had incredibly strong feelings on the subject. Nassim Taleb and David Einhorn had similar comments and believe that VaR is more or less a useless fraud. Taleb continues to refuse to even talk about the subject. Einhorn says “VaR is like an airbag that works all the time except when you have a car accident.” I think it’s important to understand that VaR is just one tool used in risk management. You shouldn’t rely on an airbag as your sole means of protection in case of a crash. Just as you should still wear your seat belt and have anti-lock brakes and other safety features in your car, you should also use other tools aside from VaR to manage portfolio risk. You must also use the tools correctly.

Nocera includes the caveat “assuming a normal market” in his definition of VaR on the first page. I believe that caveat is part of the problem and contributes to why many frown on and even attack VaR. I have seen numerous firms simply delta adjust their option positions when calculating VaR. This leads to highly inaccurate measures since that approximation breaks down completely when you need it most – in the 1% tail.

A good risk system and VaR measure must use Monte Carlo simulations with appropriate option pricing models to calculate VaR. There is no simple closed-form solution. It must also take account of the fat-tailed nature of the short-term returns of equities and other primary assets. Other points to consider are to make sure to use appropriate risk models for the time period being measured, perform stress testing on your risk system, compare risk to your P&L, and look at trends in all of the above. Looking at a single VaR number on its own is not an appropriate risk management practice.

Do you agree with Taleb or do you think VaR has a place in risk management if used properly? Share your thoughts below.

Get updates to this blog delivered to your inbox.