Irrational Policy

Following up on my last post concerning Nick Rowe's application of rational expectations to public policy: it occurs to me that the way to look for evidence of policy effectiveness is to find cases where policy was, in retrospect, irrational. In other words, look for cases where policymakers were trying to achieve objectives that, in the light of later economic thought, were bad ideas in the first place. Then look at whether they succeeded in achieving those objectives.

How exactly one would go about this econometrically I have no idea, but there is at least one obvious case study, which (if you're a macroeconomist) you have probably already thought of. According to the popular story (which I'm not entirely sure is true), the boom of the late 1960's was partly a deliberate result of policy. Today's economists tend to see such booms as a bad idea, because booms (so today's theory teaches) ratchet up inflation expectations. But the concept of ratcheting up inflation expectations was virtually unknown in the 1960's. So by today's standards, the policies of the 1960's were irrational. If you believe that story, then it is at least one data point in support of the hypothesis that policy is effective.

Rational Policy

Why is it that so many economists assume that private agents are rational but policymakers are irrational?

Breaking with that rather silly custom, Nick Rowe (hat tip: Mark Thoma) applies the rational expectations concept to policymakers and reaches the conclusion that we will never be able to find evidence of policy effectiveness, even if policies are actually quite effective.

One way to think about it, perhaps, is that all we observe in the data are the effects of policy mistakes. We cannot observe policy successes because they are negative events -- non-recession, non-inflation, etc. Only if policymakers were irrational would their "successes" -- the success of perverse policies at producing perverse effects -- correspond to positive events like recessions and inflation, so that we could find correlations between policy measures and outcomes.

To be more precise, any unusual events (recessions, inflations, etc.) that we see in the data must have been unpredictable; otherwise policymakers would have predicted and avoided them. But if they were unpredictable, that means they couldn't have been correlated with any variable that is supposed to measure policy; otherwise policymakers could have observed that variable and predicted and avoided the unusual event.

I made a similar point a couple of years ago: the inability of economists to forecast recessions is actually a point in favor of the economics profession, because it means that economists who guide public policy are doing their job well and avoiding all the predictable recessions.