Monday, September 22, 2008

My Trip to Mediocristan

Since I graduated from college, I have been actively involved in commodity markets. Well, mostly. You see, for the last 2 months I have been only a casual observer of commodity markets, blogging my thoughts as an outlet for the ideas and theories I create. You see, my old employer holds a non-compete agreement on me. That is cool, though. I signed the document without coercion, and knew (mostly) what I was getting myself into.

It has been fun, though, because now (to make myself useful) I am working in a completely different job in a completely different industry. I work at a wholesale auto auction. It is fascinating stuff and a great way to spend 10 or so months. The number of cars, the speed with which the sale is performed, the small details that need to be handled to be able to successfully choreograph the movement of the cars in and out of the sale lanes on auction is amazing. And it has given me the opportunity to take a trip to Mediocristan.

When I call the auto auction Mediocristan, I do not mean that it is "mediocre" or that it is "average" in any derogatory way. There is NOTHING average or mediocre about the auction I work for. However, in the last two months I have read a lot, and one of the books I read is called "The Black Swan: The Impact of the Highly Improbable." It is by Nasim Taleb, a former trader and risk manager/quant who now discusses in his writing the concept of uncertainty - especially as it relates to the financial markets. Taleb writes about two mythical places, Mediocristan and Extrimistan. In Mediocristan, it is impossible for any one measurement to materially skew the total average of the sample.

For example, Taleb discusses human height measurements. Most everyone is going to be between 4.5 feet tall and 7 feet tall, with the average being around 6 feet tall. Outliers (dwarfs and giants) are few and far between, and the chance inclusion of one inside the sample set will not affect the sample greatly. The auto auction is much the same, as the cars all bring a positive value, and because the cars that typically sell at a wholesale auto auction are standard Ford/Chevy/Mazda type cars, and likely cost no more than $45k brand new. So all the sale prices are going to be bounded on the low side by something north of zero, and on the high side with something lower than $30k. Even if there was some Ferrari-type outlier that skewed the sale price higher, its lack of frequency would not greatly alter the average of the large number of other cars.

Extremistan is different. Extremistan is a dangerous place, where analysis of past measurements inside a sample group of occurrences are worth nothing when trying to extrapolate to a future occurrence. The financial markets are Extremistan. That is, just because the market has never lost 1,000 points in a day (for example) does not mean that it cannot happen. Quite the contrary - with the current high volatility level present in the market, we just might get there by October!

Taleb mentions that current financial modeling attempts to put a Mediocristan "face" on an Extremistan "place". Economists and Quant guys are quick to pull together histograms and bell curves, but Taleb makes the point that "past performance is no indication of future results." The funniest example he gives of assigning misplaced trust to historical events is the life of a Thanksgiving turkey. For 300 days, the turkey might write in his journal "woke up, ate breakfast." Those data points are of little value or consolation on Thanksgiving morning - since the turkey's "script gets flipped" that day!

Last week on Thursday, I was witness to my first live auction. I watched with rapt fascination to the ballet going on before my eyes. It was controlled chaos. And, on occasion, I would drift past a TV set that had CNBC on it - reporting on the gyrations and hysteria of the broader markets that day. As the gravity of the current situation in the economy began to settle over me, I thought to myself that the auction attendees that day were inhabitants of Mediocristan, and that (in markets like these) they should feel lucky to be there.

I previously referred to the Taleb book: My Previous Post

Particularly Topical: Taleb's essay on the banking implosion

Taleb's Bloomberg Interview

No comments:

Post a Comment