Log in

No account? Create an account


From osewalrus, a fascinating article by Paul Krugman on how the field of economics got where it is today, and why most of that field didn't see the crash of 2008 coming.

Among other things, Krugman claims that economics in the future will have to have less faith in the mathematical elegance of rational actors and efficient markets, and become more accepting of actual irrationality, as is documented by behavioral economists such as Dan Ariely (Predictably Irrational).

Alternatively (this is me, not Krugman), one can define rationality broadly to include a bunch of non-obvious costs and rewards. For example, one could argue that the adrenaline high of high-stakes gambling (not to mention the more obvious reward of commissions and bonuses) introduces an externality or a negative transaction cost that caused Wall Street traders to do more and larger trades than actually made economic sense. Classically, "too many trades" cannot possibly be a problem, since trades are always people trading away what they consider less valuable for what they consider more valuable, and thus increasing the total value in the system. For example, if coals in Newcastle are traded for ice in Greenland, both commodities actually gain value by moving to where they're scarce. And if iron bars are sold to a blacksmith, they become more valuable because they're now in the possession of someone who can do useful things with them.

There are at least two problems with this theory. First, it relies on the buyer's perception of value. If the things being bought and sold aren't actually moving, but merely changing owners to somebody else who is only going to re-sell them again, it's hard to see how any actual value (as opposed to perceived value) is being created. The collapse of the housing bubble last year didn't actually make houses less valuable as houses than before, but they lost much of the value as investments for re-sale that they had gained over the previous decade; billions of dollars of perceived value disappeared overnight.

Second, if traders have an incentive (due to commissions, bonuses, adrenaline, etc.) to make trades, regardless of whether those trades actually increase total value, it'll be hard for their customers to get accurate information on what they're buying. Financial derivatives such as multiply-repackaged pools of mortgages (which mysteriously became higher-grade every time they were repackaged) or credit default swaps seem intentionally designed to prevent buyers from knowing what they're buying; the trader's assurance that they're going steadily up may be the only available information, and may be sufficient to persuade people to buy, without being aware of the risks.


Alternatively (this is me, not Krugman), one can define rationality broadly to include a bunch of non-obvious costs and rewards.

Yes, and for the time being that's probably a good choice. But make no mistake: these are epicycles. On behalf of the whole field of behaviorism, allow me to promise you, there is no place to stop before the natural end state of "different people have different motives, and you can't really generalize across them to explain mass movements". Expand the definition of "rational" far enough -- and behaviorists do this, so we'd know -- and you can "explain" absolutely any behavior... in a way which is completely uninformative.
Yes, I suspect that by the time you've expanded the definition of rationality far enough to accurately describe what people actually do, it will be tautologous and useless.
::holds head and shakes::
owwww that hurts...make it stop!