Thursday, 28 January 2010

Is behavioural economics moralising?

A slight misunderstanding here (in my opinion) from datacharmer at the Bluematter blog.
Tim Harford: "...[was too fascinated with] behavioural economics. But the financial system did not fail because of some psychological trait, but because it was riddled with damaging incentives"


In other words, let's get over evil bankers and the belief that moralizing will change the world, and let's focus on incentives.
But behavioural economics is not about moralising, nor is it about blaming individuals for making mistakes in the hope that they won't make them again.

It's about understanding why and how people behave the way they do; so that when we do design incentives, they will work.

Of course people respond to incentives. What else would they respond to? But if you think they respond consistently, transitively, stably, and (in the technical economic definition) rationally*, to incentives, you'll be sorely disappointed.

Even the most famous examples of Nudge-style policymaking: choice architecture (the salad at eye level), pension or organ donation defaults, or the fly in the urinal: all of these are incentives. Cognitive incentives, which in many cases are more influential than traditional material incentives. But more important is the combination of the two. Any material incentive, because it has to be communicated to people in order to affect their behaviour, must carry with it a cognitive incentive of some kind.

If you want to design the right incentives for the financial system, or the business you run, or your family or yourself, what could be more important than knowing exactly how people will react to them?

* Note that just because people aren't rational, doesn't mean they are irrational or, for that matter, unpredictable. More on that another time.

3 comments:

Adam Ozimek said...

Looking only at the portion of the quote you've pasted, I think it may be a sensible, if slightly sloppily worded framework.

If the crisis was due to market participants behaving rationally to the incentives before them, then the incentives are the more direct cause, since the agents behaved as expected. If the incentives were designed such that rational agents responses to them would not create a crisis, and the crisis was caused by the agents irrational response to those incentives, then the irrationality of the agents is more direct cause than the incentives.

You could argue that designers of the incentives should have taken irrationality into consideration, but I don't think behavioral economics has brought us to a place yet where the irrationality of agents is broadly predictable. In some cases, like those you give, irrational behavioral responses are more well known, and designers of incentives should predict them. But broadly speaking, and you're the expert here so correct me if I'm wrong, irrational behavior that is predictable and not very dependent upon particular, situational circumstances is the exception rather than the rule.

Leigh Caldwell said...

Thanks Adam - very clearly put.

I would agree with you that much of the existing behavioural economics research is quite specific and is essentially a study of phenomena rather than principles. This is one of my key frustrations with it - if we are calling it economics, it should have a model as well as empirical data. If it's just data, then it's just psychology.

There is a vague sense within the field that we should be doing more, and there are a few people - including myself - who are working quite closely on the cognitive models behind the phenomena. We have made some progress towards developing broad, general predictive principles; but so far they are neither mainstream nor robustly tested.

So while I still don't think behavioural economists are just moralising, I do agree with you that they mostly don't yet have the tools to provide solid solutions to crises or recessions.

The main job for the field now is to develop exactly those tools.

Tom Hickey said...

"Is behavioral economics moralising" requires an investigation into the role of facts and norms, description and prescription in economic reasoning and rhetoric. To the degree it is a science, economics is descriptive and facts are the arbiter of hypotheses and truth-claims. To the degree that economics enters policy-making it also involves value judgments and justifications for prescriptions based on economic reasoning.

The science of economics make no policy statements. Rather, it involves conditional statements about expected outcomes of different policy options. Behavioral economics and cognitive economics have contributions to make here.

Economists working in such (for now) "peripheral" in these fields should take heart from the fact that a lot of mainstream neoliberal and New Keynesian economics is based on theoretical assumptions that are either unproven, untestable, or even disproved. And prescriptions based on mainstream economics, often represented as being scientific, are regularly used in policy-making.

Political economy, on the other hand, appeals to both facts and values in approaching policy-making. Political economy in this sense is the provence of people trained in economics who are also social and political philosophers.