Like the vast majority of economists, I was delighted to see Richard Thaler get the Nobel. Anyone with a bit of sense – a group that, believe it or not, includes many economists – knows that people aren’t perfectly rational. But the assumption of hyperrationality still plays far too large a role in the field. And Thaler didn’t just document deviations from rationality, he showed that there are consistent, usable patterns in those deviations.
The question, however, is how much difference this should make to the practice of economics. And here you have a division between two camps. One says that imperfect rationality changes everything; the other that the assumption of rationality is still the best game out there, or at least sets a baseline from which departures must be justified at length.
Which camp is right? My thought: it depends on the field, for reasons not entirely clear to me. Let me talk about two fields I know reasonably well: macroeconomics, which I think I know pretty well, and finance, where I am much less well-informed in general but am pretty familiar with at least some international areas. What strikes me is that vaguely Thalerish reasoning is hugely important in one, in the other not so much.
Let me state two propositions derived from the proposition that people are perfectly rational:
1.Rational investors will build all available information into asset prices, so movements in these prices will be driven only by unanticipated events – that is, they’ll follow a random walk, with no patterns
you can exploit to make money
2.Rational wage- and price-setters will take all available information into account when setting labor and goods prices, implying that demand shocks will have real effects only
if they’re unanticipated – in particular, that monetary policy “works” only if it’s a surprise, and can’t play a stabilizing role
Now, (1) is basically efficient markets theory, which we know is wrong in detail – there are lots of anomalies. In international finance, for example, there is the well-known uncovered interest parity puzzle: differences in national interest rates should be unbiased predictors of future changes in exchange rates, but in fact turn out to have no predictive power at all. And anyone who believed that rationality of investors precluded the possibility of massive, obvious mispricing – say, of subprime-backed securities – has not had a happy decade.
Yet the broader proposition that asset price movements are unpredictable, that patterns are subtle, unstable, and hard to make money off of, seems to be right. On the whole, it seems to me that considering the implications of rational behavior has done more good than harm to the field of finance.
What about (2)? That’s where Robert Lucas came in: trying to rationalize the observed facts of business cycles with perfectly rational behavior in the face of imperfect information. This approach had a huge effect on the practice of macroeconomics, at least academic macroeconomics. But at this point we can safely say that it took the whole field down a rabbit hole. Wage- and price-setting does not reflect the best available information about future monetary policies; if it did, we’d be seeing wage contracts moving rapidly around as Kevin Warsh’s prospects on Predictit fluctuate. Everything we know suggests that there is a lot of nominal downward rigidity and a lot of money illusion in general.
And assertions that this might be true in practice, but can’t be true in theory, and must therefore be assumed away both in research and in policy have been hugely destructive.
So, rationality is a lie. But in some parts of economics it seems to be a bit of a noble lie, useful as a guide for thinking as long as you keep your tongue firmly in your cheek. In other parts, however, it’s just a disaster.
As I said, I can think of some reasons why. In financial markets, smart investors can, within limits, arbitrage against the irrationality of others. There’s no equivalent in labor and goods markets (or in consumer behavior!). But in general, the uneven applicability of behavioral thinking is surely – of course – a subject for future research.