Over at Rortybomb Mike writes: “In economics there is something called Knightian Uncertainty. In quant circles, it can be called ‘model risk.’ …”
Because Mike is reproducing an analytical mistake that is common in the financial industry, I want to discuss this. Knightian uncertainty is unmeasurable risk. That is a completely different entity from unmeasured risk. If somebody puts together a bad model, they are choosing not to measure risk that can be measured by a more careful model. Then, the problem isn’t Knightian uncertainty, it’s just poor quality risk estimation.
Mike continues: “Now that we’ve just lived through an empirical experiment in how well the best modeling can predict tail risk, I tend to look closely at [tail risk in climate change models]. And the uncertainty there has me worried.”
I think that it is very hard to defend the view that what happened to our financial system had anything to do with “how well the best modeling can predict tail risk”. The problem had a lot more to do with the expunging of negative data from the inputs. Linda Lowell over at HousingWire covered this in a piece titled “Those who bury history are doomed to repeat it”.
As far as I can see there’s very little evidence of some grand human struggle with Knightian uncertainty when it comes the financial debacle.
Update 07-01-09: Maybe Paul Wilmott agrees with me.