Their Forecast Wasn't Wrong -- Reality Was "Value At Risk" and the 99% Probability

The effort to measure financial risk is a good thing, in that it's an exercise in humility. This humility appears first in recognizing that you may well lose money when investing; second, it takes humility to rely on objective analysis in assessing risk, instead of trusting your instinct and emotion.

There's nothing new about measuring financial risk -- mathematicians have done theoretical studies of the problem for more than a century, particularly over the past 40 years.

Yet the unending irony comes when theoretical work spawns formulas that become the basis of financial models, in which fund managers, bank CEOs and regulators assume place an extraordinary measure of faith. How much faith, you ask? Well, "up to a 99% probability that banks would not lose more than a certain amount of money."

That's what I read last week in a Wall Street Journal column about "value at risk," or VaR. This model "assesses historical variances and covariances among different securities, informing financial institutions of the risks they're taking." And this column explained that VaR works fine until it doesn't, in this one very specific way: "VaR can't account for extreme unprecedented events," such as the rogue trader at Barings Bank, the Long Term Capital Management fiasco, and the subprime mortgage/credit crunch.

The writer went on to explain how, in recent years, thinkers like Benoit Mandelbrot and Nassim Taleb have debunked the premise of VaR and shown that the "extreme" and the "unprecedented" can and do happen far more often than the model-makers assume.

That's true, but I couldn't help but think about the fact that R.N. Elliott also understood this perfectly -- not in "recent years," but as early as 1934. And as Elliott's student, Bob Prechter has likewise said for decades than linear risk models don't work because they're based on false assumptions about "rational" behavior. Here's a quote to that effect from June 1985:

"The main reason the market 'doesn't make sense' is that the observers' premises are false. Most investors believe that they are dealing with a law of cause and effect with the market's action on the 'effect' side of the equation. The market, if it does operate in a cause and effect world, is a reading of what is on the 'cause' side. The market's behavior itself just IS. It is a manifestation of naturally rhythmic mass mood change. Price patterns reveal this basic law of human nature, and human nature hasn't changed. A student of the market, as opposed to a theorizing model-builder, will likely find himself coming ultimately to that conclusion.

"A fox only appears to be crazy if you expect it to behave like a chicken. Similarly, the market only appears 'crazy' if you expect it to behave according to the laws of physics."