VaR for Beginners

One of my last entries discussed the VaR principle. In class, we have been discussing Value at Risk in more detail. We have used the normal distribution to calculate estimates of Value at Risk. I was exploring the financial crisis links on the class website and decided to learn more about VaR. The discussion I found is VaR for beginners:

http://baselinescenario.com/2009/01/04/risk-management-var/

Another article that discusses VaR in the New York times is much more dense. The beginners article helps clarify the NY Times:

http://www.nytimes.com/2009/01/04/magazine/04risk-t.html

Value at risk uses a normal distribution to predict the confidence level for a company. This confidence level reveals how much money the company needs to keep on hand to cover their losses. At the 99% confidence level, VaR tells the company the maximum amount they will lose 99% of the time. VaR calculated using a normal distribution ignores fat tails that could arise in a distribution. These fat tails are high risks that may occur seldomly, but still could cause huge losses to the company. In my previous blog, I discussed that this was the problem with using only one risk calculation to determine the risk of an investment. The beginner article discusses other problem found with VaR. Using the normal calculation just to make prediction easier causes mathematical error. Not all events occur in patterns like a distribution either. The world changes. Just like how we showed the CAPM assumptions are unrealistic, it is unrealistic to assume this perfect world normal distribution. Just as the CAPM assumptions cause error, the VaR normal distribution assumption also causes error.