Wednesday, November 3, 2010

In Which I Don't Understand What Smart People Are Saying

. Wednesday, November 3, 2010

I guess I'm an idiot, as I can't understand simple things. Paul Krugman says:

One clear result of the midterms is that we won’t have anything like a further round of stimulus. And this, in turn, means that the narrative all the Very Serious People will tell is that fiscal policy was tried, it failed, and that’s that.

To support this he presents two points of data, and only two. They are:

1. Germany's economic decline has been worse than the U.S.'s.

2. Germany's government spending has been higher than the U.S.'s.

From this, he concludes:

3. U.S. government should spend more in order to boost growth. (Or, alternatively, the U.S. didn't really try fiscal stimulus at all.)

As far as I can see, that conclusion is a non sequitur given the data he's presented. He updates the original post to say "Just to be clear, I’m not saying that the Germans were big Keynesians; the point is that neither of us were". Fine. But if there is a correlation between government spending and economic growth during this downturn, that correlation is very clearly negative given the data he's given us. How can we conclude from this that the election narrative that "fiscal policy was tried, it failed" is wrong? The data that he's given supports that very conclusion.

This is not a sophisticated analysis, and there are all kinds of relevant variables that aren't included. But Krugman doesn't say which of those might be mitigating factors. He doesn't qualify the data he presents. It's a really strange conclusion for him to reach. As I've noted before, Germany really messes with the standard Keynesian analysis. Tyler Cowen built off of that post here.

Brad DeLong writes about this post, but doesn't square the circle either. Like Krugman, DeLong is much smarter than me, so I guess I'm missing something obvious. I wish they'd point out what it is.

Also note that this crude analysis supports this recent study on the fiscal multiplier, since Germany has a fixed exchange rate with many of its trading partners, while the U.S. does not.

What am I missing?


Latinamericanist said...

What you don't (seem to) understand is that Krugman, in this particular post, addresses one specific argument, namely
You often hear the US experience contrasted with Germany: America, we’re told, went for Keynesian policies, while Germany chose austerity, and Germany did better.

And what he shows is that both parts of this statement are wrong. The evidence he presents is neither for nor against fiscal stimulus - instead, for whatever reason he feels it's important to shoot down this one particular argument.
I don't know how common the argument is in this form, and who the people you "often hear" are - if Krugman is talking about Cowen then yes, he's missing the point. But he might be talking less about academic economists and more about his commenters and policy types. I wish he were more explicit, but that's a different question.

Kindred Winecoff said...

I see what you're saying, but in the context of his other writings this just doesn't make sense. What he wants to say is that *no one* did Keynesian stimulus, so we can't criticize Keynesianism based on the record of those countries. But what he actually shows that is the country that did more Keynesian stimulus had worse outcomes. This does not support the argument (from context) that more Keynes = better.

Which is why I'm not sure exactly what he's trying to show. It doesn't make any sense in the context of his other writings. I would expect this type of argument from an Austrian, say, not from Krugman.

Is there an implicit argument about thresholds? If so, I've missed it, and I read everything Krugman writes. So I really don't know where he's going with this.

Latinamericanist said...

you're making a ceteris paribus assumption that Krugman doesn't make.

Krugman - in this and the German Miracle post -
addresses a very specific argument about Germany, namely that it is an example of austerity working great. And I think he shows convincingly that it's not (lets put aside whether that's a strawman or not) - both because it's not working so great and because it's not an example of austerity.

You, then, take the step of "correlating" these two and turning them into "the country that did more Keynesian stimulus had worse outcomes". But obviously there are lots of other factors - not least the depth of the initial shock that matter here. Krugman doesn't have to bring them up, because he doesn't make a ceteris paribus argument. But if you want to turn this into correlation-type argument, obviously you need to address them.

So I still think you're reading way too much into a very simple "fact-checking" type of post by Krugman. (who is, btw., I think quite wrong on many things Germany, especially his policy advice for the German government).

Kindred Winecoff said...

I guess the difference is that I'm considering this in the context of Krugman's other writings on fiscal stimulus, in which he has strongly argued that we need more of it.

I don't think this is a simple fact-checking post, because he argues that neither Germany nor the U.S. has pursued Keynesian stimulus. The data he presents shows that gov't consumption in Germany has gone up 7 percentage points in two years. He says this is not Keynesian stimulus. I wonder what would be. 10 points? 15?

If there is a negative correlation at 7 points, how do we know that the sign will reverse at 15?

The German case doesn't support his view. I really wonder why he keeps bringing it up.

dan davies said...

There isn't a negative correlation. Look at the charts and match up the time axes. Germany has "done worse" because it had a larger initial fall. This was basically because it's a very export-oriented economy, and because the German policymakers' initial reaction was to go all Austerian (because German policymakers always want to try to free ride on other peoples' surplus via the exchange rate).

The larger initial fall obviously can't be caused by the increase in government consumption because it preceded it. But if you look at Germany vs US real GDP post 2009 Q4, it's clear that Germany's done better.

Kindred Winecoff said...

Dan -

I did. Not sure what you're seeing that I'm not, but Germany's govt consumption increased pretty linearly across the period, and was greater than the U.S. for all but two quarters (Q3-Q4 2008). During that time Germany fell harder and faster than the U.S. After that time, Germany's govt consumption was still higher than the U.S., and they recovered more slowly until Q1 2010. Germany's recovery picked up after that... which is precisely when govt spending tapered off a bit.

I agree that there are mitigating factors here, and I wrote that. But it is not at all clear from these graphs that more govt spending would have helped Germany.

In Which I Don't Understand What Smart People Are Saying
There was an error in this gadget




Add to Technorati Favorites