VaR

I recently got asked by a headhunter to consider a job as a project/implementation manager for a company that specializes in VaR, or "Value at Risk." That's high finance talk for software that risk departments in investment banks and other financial institutions use to evaluate the extent of the damage they could bring upon themselves, considering their current strategies. You load up all of your positions and it tells you how risky your portfolio is, and perhaps suggests ways of hedging your bets with others so that your investments might be safe and sound.

During the interview I asked a bit about what went into it. At first I was thinking, wow, this might be my opportunity to find out first hand about strange attractors and self-organizing systems, but the interviewer, an MBA and PhD in mathematics (because you have to be in order to understand this stuff) assured me that VaR uses normalized calculations, just very complex ones. I lost about 95% of my interest at that moment, but I was still curious, so we went on.

He was most confident that in all but very rare cases the calcs could pinpoint and closely predict riskiness, and even alleviate most of the risk due to chaos (the unexpected), by helping to stabilize everything that was easily assimilated into knowable and predictable patterns. This was a few weeks before the financial markets imploded.

I'm sure that Lehman Brothers, Bear Stearns, Wachovia, Fanny and Freddy, for instance, were all using some sort of VaR system to help stave off their own tribulations.

But when I think again about Laura Ward's "Whistle Me Higher" (see previous post) and all of the figures that may have been taken as indicators that may have suggested financial peril (i.e. gambling casinos, radioactivity, pandemonium, superstition, the sweeping up of toxic money by people in breakout suits), I am left wondering: who's got a better system? True enough, like all art, Laura's work is a bit of a Rorschach, something like the Delphic Oracle pouring out of the mouth of a Python Priestess high on geologic gasses, later interpreted by a throng of specialized priests.

There are many oracular parallels throughout the history of the world, everything from the reading of tea leaves to the bumps on one's head. The Tao De Ching, for instance, grew out of a practice people in ancient China had of turning over tortoise shells in order to seek a correspondence between static visual patterns and those being lived. Nowadays we have complex visualizations of numerical data, (i.e. graphical display) which attempt to convey information in a way a body can understand, alongside explanations that, after all, mathematics is built upon embodied metaphors, bringing us oddly back to where we had begun, with a little added perspiration (and perhaps precision).

Add to that the concept of expertise that Malcolm Gladwell puts across in Blink, or the theory that our neurologies enable us to make complex statistical-based decisions on an unconscious level, and you begin to gain a kind of insight into the possibility that the somewhat mystified and mistrusted operations of art-making may not be, in many ways, too different from those more reliable domains, such as the sciences.

This does not say that one should listen only to Laura, and ignore the complex, though normalized, mathematical formulas. But we might attempt to think, and experience, for ourselves and learn to weigh both into our judgements, in whatever proportion seems fit for the context. In other words, use the statistics the machine can give you, but also those that we give and receive as a system of biological machines. True, there is much that is quite primitive in there, but tempered by our other facilities (i.e. those that have built the complex mathematical models to begin with), we might find we have a much bigger net to catch things in.

It comes down to being able to pay attention to, and integrate, more varied data. And also to develop a kind of optimized paranoia for detecting patterns in emergence.