The Productivity Commission has just posted on its website the proceedingsÂ of a 2007 roundtable on behavioural economics. The most provocative piece is by QUT’s Paul Frijters (who mistakenly gets a UQ designation), discussing Eldar Shafir’s opening keynote. Frijters’ discussion (which starts on p35 of the PDF) is a neat summary of the problems involved in applying behavioural economic research to policy. A few snippets:
After presenting his list of anomalies and giving us nice anecdotal research to make it indeed entirely plausible that these all exist and are important for real-life decisions by people of flesh and blood, Shafir attempts to draw some policy conclusions. Shafir’s policy conclusions are a little weak, it has to be said. His paper is full of the usual â€˜we should take into account’ and â€˜we may also need to reconsider’ and â€˜policy should be carefully crafted’, but he essentially leaves it up to the reader to work out what any item on the list means for policy advisers. Indeed, he gives no real suggestions for new policies, but rather mentions policies that were enacted in an era preceding â€˜anomaly economics’, presumably as evidence that â€˜anomaly economics’ has a point. …
Implicit in Shafir’s piece is that it is actually possible to think through the consequences of these anomalies for policies. This requires their integration into the current economic policy toolkit of Homo Economicus, and policy directed at correcting market failures. Shafir himself does not undertake to do this, but explicitly calls for it to happen.
It is not just Shafir who can not integrate them. To my knowledge, no-one in this literature has even been remotely able to incorporate a multiple of items on The List into actual modelling or into a blurred overall vision of human choice behaviour. We still cannot do what the classical economists thought too hard to attempt.
If even the leading authors in this field cannot integrate the lessons of the anomalies into mainstream policy advice (and even his list is far shorter than the full one), it should be clear that it is going to be extraordinarily difficult to do. Indeed, I think we should simply admit that it is not going to happen: we are stuck with a fairly unorganised and large set of anomalies which we are never going to be able fully to integrate into a theory that does justice to each of them. Unless we are prepared to simplify The List into a couple of much simpler rules of thumb, it is just too hard to come up with a framework that really makes intellectual sense.