JPMorgan's (JPM) loss (May 2012) is another example of the disastrous effects of market risk underestimation. The orthodox models as referred to by Pr. Benoit Mandelbrot (see reference below) do not accurately forecast market behavior as they are mere simplifications of a much more complex reality.
The financial crisis that began in 2007 largely resulted from the inability of regulators, central banks and large market participants to fully understand the complexities that had been introduced into the market. Innovations in trading methods and complex derivative products are driven by the need for market participants to shift risk and cleanse their balance sheets in order to meet regulatory liquidity requirements. Today, we find ourselves in the midst of an undesirable mix of tight economic and financial conditions. The deadly embrace between over indebted sovereigns and over-leveraged banks has created a vicious cycle.
Deficient risk management systems based on the neoclassical school of thought propose easy but ineffective methods of forecasting risks. In the case of CDOs, these were based on Mr. Li's Gaussian copula function put forward in early 2000 which assumed that correlation was a single constant number. Rating agencies and banks grabbed the opportunity and shaped their assessments based on this function. While the formula was new and complicated for non-mathematicians, banks were happy to classify their products according to this single risk measure and as such sell their CDO products without much effort. It is impossible for professionals at the time not to have been aware of the danger of this function's limitation as correlation cannot be constant over time.
A measure of risk should take into account long term price dependence and the tendency of volatility to cluster. Wild prices and fat tails led Mandelbrot to view the financial series as a fractal series. The pursuit of certainty is vain and risky. People should accept the reality of risky markets and stop pretending otherwise...
By identifying markets' structures, what we essentially learn from Mandelbrot is that a measure of risk should take into account long term price dependence and the tendency of bad news to arrive in waves. Our focus shouldn't be on how to predict prices; but on how to foresee risks:
Opportunities are in small packages of time, large price changes tend to cluster and follow one another. If there was a large price change yesterday, then today is a risky day. (Mandelbrot, The (Mis)Behaviour of Markets - A Fractal View of Risk, Ruin and Reward, 2008)
This is a delicate statement, meaning that we may not be able to forecast market direction, but during its clustery periods we either exit the market and thereby reduce the chances of loss or trade its volatility. Each cycle, each market, each strategy and each underlying asset requires an appropriate dedicated model. Nothing can be generalized to an entire portfolio. To analyze a market's dissimilarities is difficult and tricky. Each field needs its own appropriate - in the zone - skilled person.
It is important to think in terms of affordable risks before thinking of potential gains. Forecasting might require sophisticated mathematical, economic and financial tools in order to account for variable and uncertain factors such as people's hopes and fears, natural catastrophes and new inventions. But most of all it requires a sophisticated mind to grasp a market's structure and humility to accept it.
Complex interconnectedness in the financial system leads to under-pricing of risk. This status quo will persist as long as human beings exist; it's part of how markets function.
Apart from the fact that we will all be dead in the long run (Keynes, 1936), there is another thing that we can be sure of: outperformance will be wiped out due to market dislocations. Risk measures require the modeling of possible variations in market prices according to probable and extreme circumstances.
Statistical risk measures cannot capture all of the possible risks. Standard statistical models cannot capture the emotional probabilities or intensities of investors including their greed, fear and sudden consensus reversal. The aim is not to develop entirely new and or completely infallible forecasting tools, desirable as they may be, practically it would be impossible. Markets keep shifting and in so doing they break models. Learning how to build new risk models, that describe extreme events while taking into account diverse risk factors, is essential. The implementation of any model in practice is as much an art as it is a science.
By accepting the reality of uncertain markets, financial practitioners acquire the elements to develop enhanced portfolio management tools. The mystery of the financial markets is the key to its existence, understanding this is the key to our survival. Survival in multifractal markets requires us to "Mind the Risk."