Bloomberg has an interesting article exploring how the banks are changing their risk models. Roger Ehrenberg adds some insight to the article. To me the most fundamental question that needs to be asked isn't what is the best model, but can risk even been modelled?
In all the discussion about risk and its modelling, there is always the underlying assumption that the risk can be modelled. I'm not so sure you can really model the risk. Time and again in financial markets past performance has been shown to not be a predictor of future performance (all the managed funds have that explicit on their brochures) and yet the assumption is that past risk can be used to model future risk. See the incongruence?
The market looks to be an indeterminate, chaotic system. Which to me makes the assumption of normal distribution of risk a little suspect. Being an indeterminate system, it may very well be impossible to model risk using the standard statistical methods.
Tags: Risk, Finance, Credit Crisis
