I spoke with Robert Barro of Harvard yesterday about the stimulus bill, fiscal policy, and related issues in macroeconomics.
I wanted to speak with Professor Barro after reading his piece in the Wall Street Journal about the multiplier on government spending. The piece, which argued that the multiplier has historically been much lower than the Obama administration hopes, produced a tremendous amount of response -- from Paul Krugman, Brad DeLong, Greg Mankiw, Matt Yglesias, and Tyler Cowen (some of them several times). And that response was notable, in part, because it turned into a reflection on the "standards" of the stimulus debate itself. I was interested to hear what Barro thought about his critics this debate.
He was admirably patient with my questions:
Conor Clarke: What I am trying to do is sort of apply a barometer to
modern macroeconomics and see where the profession is, because I am
sort of confused by a lot of things.
Robert Barro: [laughs] Probably the fault of the profession.
Well, one thing I am confused by is where all of this resurgent interest and fiscal policy came from. That's very broad. But where do you think it came from? When I took macroeconomics in college there was not a lot about fiscal policy.
It came from the crisis and memories of the Great Depression and the fact that monetary policy seems to have done not a tremendous amount, and conventional stuff doesn't look like its going to work anymore. And it's about grasping at straws to try and find something else.
And I take it from the Wall Street Journal piece you wrote last week... well, the piece is just specifically about measuring multipliers, but I take it that you are fairly skeptical in general that fiscal policy will boost aggregate demand.
Right. There's a big difference between tax rate changes and things that look just like throwing money at people. Tax rate changes have actual incentive effects. And we have some experience with those actually working.
What would you say is the best empirical evidence there?
Well, you know, it worked to expand GDP for example in '63 and '64 with the Kennedy/Johnson cuts. And then Reagan twice in '81 and '83 and then in '86. And then the Bush 2003 tax-cutting program. Those all worked in the sense of promoting economic growth in a short time frame.
I'm the middle of a study where I am trying to estimate this overall, going back to 1913 -- sort of constructing some measure of the overall effect of the tax rate at the margin, at the moment. I'm just looking at that now, actually...
You're talking about the multiplier on a dollar of...
Well both things, but here I'm talking about the tax rate stuff. Get some measure of the effect of marginal tax rate that comes from the government -- federal, state, local. And then you can see what it looks like going down or going up and how the economy responds. And then, in addition to that, the government might be spending more or less money on either military stuff or not on military stuff. And we can estimate that at the same time. With the government spending stuff, the clearest evidence is in wartime. It's not that it's the most pertinent, but it's the clearest in terms of evidence because it's the dominating evidence at those times, especially during the world wars.
Do you read Paul Krugman's blog?
Just when he writes nasty individual comments that people forward.
Oh, well he wrote a series of posts saying he thought the World War II spending evidence was not good, for a variety of reasons, but I guess...
He said elsewhere that it was good and that it was what got us out of the depression. He just says whatever is convenient for his political argument. He doesn't behave like an economist. And the guy has never done any work in Keynesian macroeconomics, which I actually did. He has never even done any work on that. His work is in trade stuff. He did excellent work, but it has nothing to do with what he's writing about.
I'm not in a position to...
No, of course not.
I'm not in a position to know things like the degree to which Paul Krugman counts as a relevant expert on new Keynesian economics.
He hasn't done any work on that. Greg Mankiw has worked in that area.
Greg Mankiw is, I guess, skeptical of spending for the same reasons
that you are: he says that there's some empirical evidence -- I think
he cites the Christina Romer study from 15 years ago -- that a dollar of tax cutting has a larger impact than...
The Romer evidence is very recent actually. It's an ongoing project.
I thought it's from 1993 or something like that. Maybe that's something else.
They have a current thing that's going to be presented at Brookings at the next meeting, where they have some estimates of how the economy responds to tax changes. It's not really looking at tax rates. It's looking at tax revenue, which is not the same thing. That's mostly what Greg was referring to, which is going to be presented in a few months.
I would need to go back and check. But one question, and I think Greg Mankiw raises this question as well, is, Why does this set of evidence depart from what seems like the standard Keynesian theory that a dollar of spending would have a larger multiplier than a dollar of tax cutting?
I don't think it is really confusing at all, because when you cut taxes there are two different effects. One is that you cut tax rates, and therefore give people incentives to do things like work and produce more and pay more -- maybe, depending on what kind of taxes. And then you also maybe give people more income. This income effect is the one that's related to this Keynesian multiplier argument, where it's usually argued that government spending should have a bigger effect. So that's the income effect. But the tax-rate effect, inducing people to do things like work and produce more and invest more, is a whole separate effect, and that could easily be much bigger than the multiplier thing, than the income thing.
This might just be my confusion, but the inducement to work, is separate from the idea of boosting aggregate demand and consumption in the short run.
Oh it's exceptionally different. But the experiment is that the government is doing something by changing the tax system to lower its collections -- by, for example, a tax cut. The response of the economy to that is not going just to isolate this business of giving people money. It's also going to have these incentive effects, more than tax rebates, on economic activity. It's going to be a combination of those two things -- income effects and incentive effects. One piece looks like this sort of multiplier stuff, which is analogous to government spending -- probably because the government spending has a first-round effect where it comes in and directly affects the aggregate demand -- and then in the second round it sort of looks like a tax cut. That's why the government spending thing is bigger in textbooks: because it has this first round in addition to all these subsequent ones.