Tuesday, March 09, 2010
Can "stimulus" work?
Ludwig von Mises argued that debt financed government spending and central bank credit creation are both capable of creating an artificial boom. Hayek concurred (at least on monetary policy). This artificial boom is the source of the business cycle, in the Austrian account.
And certainly, if by "work" we mean "increase GDP and reduce unemployment," stimulus spending might well do this; it has to have some effect, doesn't it?
Well, there's a debate of sorts between Menzie Chinn of Econbrowser and Russ Roberts of Cafe Hayek over the effects of the ARRA stimulus bill (and particularly the CBO estimates thereof). Chinn has been arguing for some time that it's been effective, while Roberts is skeptical. While there are some interesting issues involved, I'll bypass those and note a couple of oddities in Roberts' argument.
1. Many who are of an Austrian bent (I guess Roberts is Austrian) seem to be adamant that government cannot boost GDP or reduce unemployment with fiscal or monetary policy. Yet this stuff is the source of the boom in the Austrian business cycle. So why is this such an important issue for these guys?
2. Roberts appears to be arguing that the CBO estimates are not estimates at all, because they are based on theoretical models. I object. I've earlier tried arguing with Chinn that these results are not empirical measures because they are based on theoretical models, including simulations and counterfactuals. But that's quite a different point. Roberts must know that all economic propositions about the effects of some event are based on comparison with the counterfactual case where the event didn't occur. Such estimates are necessarily theory laden.
Hence my comment (on both blogs and now here): "Economics necessarily involves that which is not seen as well as that which is seen, as Bastiat points out. And the cost of an action can never be realized, as Buchanan shows us. The use of a model or theory is inescapable in economics... 'what didn't we see,' 'what would have happened,' etc.
So CBO's approach is an analysis of what stimulus actually did; such analysis necessarily requires a counterfactual, based on an underlying model of what would have happened otherwise."
3. Roberts seems to accuse the CBO of partisanship. Evidence? None offered, and I couldn't find any in reading the CBO study. They used mainstream macro, and spelled out their methods fairly clearly.
I actually share Roberts' skepticism about stimulus (and mainstream macro). But I dislike poorly reasoned critiques that add nothing to our understanding. They don't teach us anything, and simply strengthen our opponents' beliefs that we've nothing of value to offer. In a time when economic analysis is getting a terrific bashing, it needs to be done more seriously and carefully than ever.
Thanks for your thoughtful observations. I hope to post something in response at Cafe Hayek in a little bit. Just want to say here that I didn't mean to suggest that CBO was partisan and not sure where you got that impression.
I think many people, including Menzie Chinn, thought you said this as well.
You wrote "They [CBO] have done NO independent non-partisan analysis of what actually happened."
If you didn't mean to say this, I hope you'll correct the error. I look forward to your update on Cafe.
including my response to your comment.
Perhaps I'm wrong, but I think Chinn emphasized non-partisan because he has many commenters who accuse him, and CBO, of partisanship. Chinn clearly has his preferred sides, and says so, but it makes sense to emphasize CBO is apolitical.