I finally had a chance to read the Cato Unbound symposium that Todd Z mentioned below. It’s well worth the read.
I should add to my earlier post on this that I am not, as it happens, hostile to the pursuit of behavioral economics; far from it. But I do share some of the skepticism about how it goes about the intellectual enterprise – for example, the way in which concepts in moral psychology that I would regard as unavoidably “deep,” such as trust, get operationalized, in order to fit into a certain form of testable experimental design, into concepts that are much flatter and on the surface, such as “confidence,” to quote a recent conversation with a psychologist friend in the field.
Indeed, as I’ve occasionally observed here at VC, if rationalist economics and behavioral economics are in some sense at odds and competing, or at least complementary, in another sense, they are much more deeply linked. They are linked in being “surface” theories of human nature. Neither of them embraces a deep view of human psychology or human nature – they are each minimalist and reductive, each in its own way.
There are contrasting approaches in economics that do embrace deeper views of people, more precisely, more “committed” views of human psychology. One category would be institutional economics insofar as it embraces one or another form of sociological theory of institutions. A theory of institutions that matters to economics, however, in a social sense will almost always be one that concerns legitimacy.
The other category would be theories of value based around deep commitments about human nature – moral psychology – and the defining work would be the Theory of Moral Sentiments. But all that stuff that Akerloff and Schiller and other people today want to say about animal sentiments, etc., is all essentially founded in what might as well be called the moral psychology school of economics – and yet does not want quite to speak its name. It hides behind Keynes, in a peculiar feat of intellectual history, when really it is Smith, the moral psychologist.
It’s not to say that one or another is right or wrong. They describe different kinds of things and are better seen as complementary. Confidence, in the sense that some behavioralists assert it as a conceptual proxy for trust, isn’t quite that – it is a stronger concept that confidence as its operationalization, for example. And legitimacy is far more than the artifact of an observed tendency of people to obey institutional precepts; even the term precepts is conceptually slightly different from commands or orders.
There are plenty of reasons why one would want to design experiments around minimal commitments; there are plenty of reasons why minimal commitments are not enough to explain things. The problem with surface theories of human nature is that they describe the ‘man without qualities’. The problem with deep theories of human nature is that they eventually wind up with Dr Freud or something equally non-falsifiable – the ‘man with too many qualities’.
That is at the level of intellectual inquiry. But the Cato Unbound discussion – and the end of the Ferguson piece, as well as my comment on the move that a certain application of behavioral economics makes, parallel to a move concerning rationality and deliberative democracy – is about its political and moral application. That’s a separate discussion from the intellectual qualities of behavioral economics. However, I found myself interested by a side issue, which is the final thing I want to raise here. It is that the Cato symposium regarded the moral and political issue as being one of paternalism – whether that is true or not, that was the framing moral question in the discussion.
I would have thought, however, that rather than “paternalistic” policy, the phenomenon that the Cato critics were raising was much more a question of “therapeutic” social policy. There are reasons, of course, to be concerned about them both if one is a libertarian – and even if one is not, quite possibly. But they are not precisely the same thing. Paternalistic or therapeutic? Culture of paternalism or culture of therapy? Discuss. (Update: I woke up this morning thinking that maybe the Cato Unbound discussion really is about paternalism; I’m still interested in the distinction, though. Also, responding to one of the comments, I don’t think that falsifiability or non-falsifiability is the only relevant criterion; falsifiability on examination turns out to be more complicated than usually expressed.)