Friday, 10 April 2009
Eric Falkenstein has an interesting post which highlights some of the problems in the study of behavioural economics (in this case, behavioural finance).
I have to agree with his premise, though my conclusions are a bit different.
I've just reread some of Nudge and am partway through Animal Spirits. Both books cheerlead for the behavioural cause - though in each case, one author - Thaler and Shiller respectively - seem to be much more closely associated with it than the other - Sunstein and Akerlof.
However both books exemplify the problem that Falkenstein identifies. The field is full of effects without explanations.
You can easily list a whole string of cognitive biases which can be easily demonstrated - I show the effect directly to audiences in presentations, by running a price anchoring experiment. But the behaviourists rarely seem to propose a good underlying model of how these effects arise.
Having just completed another book, The Making of an Economics, Redux, I can see why this might be. There is a surprising split between theorists and empiricists (experimental economics or data analysts), and most of the behavioural work has come from the experimental side. Indeed most theorists working on behavioural economics are busy trying to disprove it - come up with an apparently irrational effect, and they'll show how it can in fact be explained within their rational models.
Nothing wrong with that in principle - the key distinction of the scientific method is that hypotheses must be falsifiable, and we move forward by gradually disproving the wrong bits and leaving the best parts in. But in most sciences, the theorists propose something new and the experimenters try to disprove it. In behavioural economics it is the other way around!
Mainstream economists, understandably, want to have a solid theoretical foundation before trying to incorporate behavioural economics into the fields they study. This is perceived as part of how "real" economics is done. And, as far as I can tell, the behaviourists haven't stepped up yet.
Perhaps it's a cultural divide that has kept the theorists out of the field, or perhaps there really is an underlying flaw in the behavioural ideas, which the experiments are skirting around. While I believe that the behavioural results are real, I would not like to be defending a PhD thesis on them just yet. It's imperative that we find out whether these results can actually form the basis of a new economic model.
I'm going to start to attack this myself by proposing a theoretical decision-making model which might explain certain anomalies in consumer response to prices. That will be in the next posting.
Update: orgtheory roughly shares my concerns.