VaR
I have recently been involved in a number of debates about the efficiency/usefulness of VaR (Value at Risk). Much like Macs, you either love it or hate it. In this post I have examined both the usefulness and the effectiveness of the VaR calculation by looking at the pros and cons of its use as a risk management tool.
In financial math VaR is a popular measure used to gauge the risk of loss on a specific portfolio. The VaR itself is a single number that represents the minimum loss that a portfolio can suffer in a period of time, for a pre-calculated percentage of worst cases with the caveat that the absolute weights of the portfolio are not changed during that period.
Proponents of VaR believe that the process behind the calculation is as important as the result because it forces organizations to confront their exposure to risk and set up proper risk management functions to attempt to mitigate the revealed risk. They acknowledge that the benefit of VaR is not in the single number but in everything that goes into calculating the threshold.
A second advantage would be that VaR calculates the minimum loss that will occur one day in a hundred (assuming 99th percentile) enabling the risk manager to allow the company to take all the risk it wants within the VaR limit, as statistically the firm will remain safe. However, outside the limits indicated, nothing can be accounted for and the risk manger is on his own. Proponents of VaR believe that that though it cannot account for everything, the fact that VaR provides the decisive line between what is acceptable risk and what is not proves it to be a useful tool for risk management.
The problem with VaR begins with its definition, which fails to account for the fact that the actual loss could be far deeper than the minimum. Since these losses are buried deep within the ‘tail’ of the portfolio they go largely unobserved.
Another problem with VaR is that it seems to factor in correlations between risk factors that come up during the time the data is gathered (normally VaR takes into account data from the past few years). However, these correlations tend to breakdown in times of turbulent markets. Furthermore, banks are only required to reveal a number but not the method with which it was calculated. A bank can say its VaR is down twenty percent but that does not necessarily mean that the risk is down twenty percent, as the decrease could just be a result of assumptions made by the analyst. The problem once again is that companies and governments both stress too much on the number that VaR shows but do not scrutinize the assumptions driving that number, resulting in the public getting a false sense of security.
The biggest complaint with using VaR has to be the argument for using normal distributions to create scenarios. The normal distribution is a symmetric bell shaped curve that is focused on the center and decreases on each side, and hence has less of a tendency to produce extreme values. Working with the assumption of a normal distribution (in regard to the real life market) seems to be the most unrealistic choice. Simply put, normal distribution fails to capture what is going on in the lower tails of the distribution. Even if we were to accept the assumption that actual portfolio returns would be normally distributed, the market would not be representative of this due two major factors: Firstly, VaR leads to a bunching of risk that contradicts the diversification principle. Secondly, VaR does not take into account the increased complexity that is added by aggressive trading and the use of derivative instruments.
In conclusion, although VaR does have some benefits, it cannot be solely relied on as a measure of risk as it is not designed to deal with extreme shocks or “black swan” events. Rather, by using volatility and correlations calculated from past market movements, VaR estimates a singular risk number that in no way utilizes or takes into account market expectations or movements. As a result I believe that VaR is ineffective as single risk management tool (especially) in the turbulent markets we are faced with today.
Note: General Motors has filed for bankruptcy protection, Chrysler is on its way out of bankruptcy after filing less than two weeks ago…
