I don't want to impose on TOD by quoting my whole piece from my own site about this, but I blogged about this the other day.

In short, the kinds of predictions they're talking about are so completely at the mercy of things beyond the scope of the model that the prices they arrive at are useless.

The worst scenario they came up with, the $262/barrel one, was related to the fall of the House of Saud.  Trying to map that kind of geopolitical event to a precise price level is beyond silly.  Anyone here could easily cook up a sets circumstances under which that same central event results in a price substantially higher or lower than $262.

But golly gee, it sure makes for some nifty headlines, doesn't it?  I'm beginning to think that this kind of prediction and the resulting breathless articles are nothing more than the econometric and media equivalents of the movie SAW--a cheap thrill that no one uses as a basis for real action, and that makes you feel good that it hasn't happened (at least not yet).

Careful. "Saw" is a great movie.
Yep, I liked SAW a lot, also.  But my point is that it didn't change how people live their lives--they had the thrill and the entertainment, but then went back to business as usual.
Why the umbrage?  They're talking about scenarios that involve massive shut-offs of oil overnight.  If Saudi (or Venuezuela or Nigeria) just quit we probably would be talking about over $200/bbl.  Their cost vs supply regression probably has a lot of error but we all know that it would be bad.
I get upset over that kind of predictions (and the inevitable breathless media coverage) because they're meaningless.  As I said above, you simply can't predict the prices caused by those hugely disruptive events, and tossing around specific numbers is wildly misleading.  If one or more of those things were to happen, the actual price level could be much higher or not nearly as high as they're predicting, thanks to factors their model can't take into account.

When I was in economics graduate school, I had a professor who liked to say some times you can make up a totally convincing, precise argument based on econometrics that's laughably wrong.

I'm an economist by training and I have great respect for people who work with statistics and economic modeling; whether they're explaining relationships between factors in the past or trying to predict the future, they provide a very useful service.  But I also think that one of the main things we all should keep in mind is the limits of such analysis, and how incredibly easy it is to step over the line from reasonable conclusion into utter gibberish without even realizing it.

[quote]and how incredibly easy it is to step over the line from reasonable conclusion into utter gibberish without even realizing it.[/quote]

Goes with the job, doesn't it?

Sorry, thousand apologies, but I just had to say that. grin

I would say $262 is a somewhat low. My rough rule of thumb is a short term spike of 50% increase for each 2% shortfall in supply. Saudi produce about 10% of world oil at present, that's 5 multiplicative 50% increases, giving approximately $500 from a starting price of $68. That should fall back to between $300 and $350 fairly quickly (within a week or two) but thereafter any price decline would depend on how soon the shortfall is reduced by restoration of supply or demand destruction.