Do economists get their information from principles textbooks?
This isn’t a rhetorical question, I’d really like to know. Here’s what we do know:
1. Standard macro theory say that the fiscal multiplier (from the demand-side) is zero whenever a central bank targets inflation. Even Paul Krugman is careful to point out that the argument for fiscal stimulus only makes sense when at the zero bound.
2. Most textbooks tell students that modern central banks target inflation.
3. Most textbooks tell their students that fiscal stimulus is effective, and don’t put in any qualifiers about whether the central bank is targeting inflation.
4. Most economists who talk about fiscal stimulus in the press seem to assume that economic theory predicts it’s effective whether you are at the zero bound or not. So that’s my question; where do economists get this idea? I know where I got the idea, I was taught the crude Keynesian fiscal stimulus model in college. But I later unlearned it. Is it possible that most economists learned the model just as I did, and never unlearned it?
Here’s an article that Saturos sent me, but there are 1000 others just like it, so I’m not trying to single out Laura Tyson and Owen Zidar—they are clearly in the majority.
BTW, notice that if you eyeball the graph showing demand-side tax cuts work, the downward sloping line seems to almost completely disappear if you remove a single observation (2008)?
Tags:
21. October 2012 at 06:39
Maybe it’s just me but articles like that one always feel like politics steering economics, instead of the other way around.
21. October 2012 at 06:45
Actually, I wasn’t even complaining about that. I was raising doubts about the test of supply-side effectiveness:
Why should hours worked (I’m assuming that’s what “employment growth” means) be the only margin of adjustment? What about quality of effort, productivity factors of both workers and firms? What about human capital? And wasn’t the whole point of the cuts supposed to be about the propensity to invest? OK, so the study is focussed on the short run only. It’s plausible that there won’t be a hugely elastic response from high income earners immediately. But I think it’s clear that the Republican plan is the old supply-side one, which theoretically was never supposed to be about short-term boosts.
What is your opinion on this Saez paper? http://elsa.berkeley.edu/~saez/piketty-saez-stantcheva12thirdelasticity_nber_v2.pdf
They don’t find the cross-counrty evidence supportive of high elasticity.
21. October 2012 at 06:48
My point being: if lower MTRs on managers make them work harder, their firms become more efficient, and you won’t necessarily catch that effect locally.
Also, if a tax cut is going to be offset by a later tax increase, shouldn’t Ricardian foresight withhold the increase in labor-supply? That’s my main doubt about Romey: in the long run I think the US will repay its debt not default. And that is inevitably going to be done with higher rates. So lowering taxes rather than tackling spending is just kicking the can down the road, and so Romneys cuts shouldn’t be that stimulative so long as spending stays high. (I’m pretty sure Ricardian Equivalence isn’t completely wrong.)
21. October 2012 at 07:01
But yes, I’ll concede that Romney is lying about his main job-creation plan. He won’t create 12 million jobs in 4 years by lowering taxes on the profits of small businesses (doesn’t change marginal incentive to hire in the short run).
21. October 2012 at 07:20
[…] Scott Sumner questions the validity of the majority on fiscal stimulus: Most economists who talk about fiscal stimulus in the press seem to assume that economic theory predicts it’s effective whether you are at the zero bound or not. So that’s my question; where do economists get this idea? I know where I got the idea, I was taught the crude Keynesian fiscal stimulus model in college. But I later unlearned it. Is it possible that most economists learned the model just as I did, and never unlearned it? […]
21. October 2012 at 07:25
Contrast the unreliable and conflicting results from fiscal stimulus and the very straightforward implication of NGDP growth:
http://thefaintofheart.wordpress.com/2012/10/21/the-berkeley-view-fiscal-stimulus-rules/
21. October 2012 at 07:37
It’s a conflation of “instrument” and “stance”.
Arguments that fiscal policy is effective are based on an assumption of no change in the monetary policy *instrument* (which, of course, means that the stance will be changing).
Scott’s position is (or at least appears to be) that the correct thing to do is to assume no change in monetary policy *stance*, which would be achieved through a central bank that was endogenous to the system adjusting the policy instrument.
(Okay, Scott would no doubt cringe at the use of the phrase “policy instrument” as it sounds suspiciously like “interest rate”, but you get the point …)
I would criticise (my characterisation of) Scott’s view in two respects:
First, that this should only produce fiscal multipliers that are zero in (ex ante) expectation, conditional on the central bank’s information. In principle, if the CB is operating with a well-specified model of the economy, this should result in ex post (i.e. realised) multipliers that are indistinguishable from white noise. In practice, the (supply side) structure of the economy is constantly changing and the CB is only learning about this over time just like everyone else. As such, one would expect non-zero fiscal multipliers to persist for arbitrarily long periods of time in any given sample. Which brings me on to my second point …
Second, fiscal policy does not only affect (aggregate) demand. It affects the micro-structure of supply, too. A central bank that (a) needed to empirically disentangle what was demand and what was supply; and (b) was lexicographically focussing primarily on inflation would therefore still produce non-zero multipliers.
21. October 2012 at 08:03
Saturos, let me put out a differing view on the top 2-3% of SMBs that matter.
If you go to any given small town in America, say under 1M (larger towns too, but small matters in my analysis)….
You’ll find that many SMBs are created by partnerships, you’ll see two three founders that in many cases have other businesses they are in as well.
I grew up in Canton, Ohio. A giant test market for the nation, so I saw both home born companies (non-franchise) and a steady influx of franchisees, that were expanding outside their home base of somewhere else in the US. Bankers would tell them to go test the franchise in Canton.
Note: the more normal your market is, the crazier the infomercials you see on TV. 1 in 12 infomercial products work, so in big expensive markets that skew one way or another, you only see the winners. But in hyper-normal test markets, you see the other 11 A LOT. This gives you a very different take on the level of choice that the rest of the country has. The same works with Franchisees.
Anyway, many SMBs start out a partnerships, and many guys who already have a successful trucking or plumbing or home building business, who play poker together will get together and buy the Hooters franchise for Canton.
Profits on this Hooters are treated as income IF the $ leaves the corporation.
So the guys with the Hooters are being HEAVILY incentivized to open ANOTHER Hooters – in Canton, maybe Akron, but that’s a 30 minute drive, and they got other things they do.
Meanwhile one of the owners, he’s got a son who’s starting a landscape business – he wants to use robot lawn mowers.
If the Hooters owner could simply take his Hooters profits and put them UNTAXED into his son’s robot landscaping service.
THEN and only then is the agent able to make the correct decision without giant inefficient tax incentives pushing him to OVER HOOTERS Canton, and under robot landscape Canton.
This is a very common situation. Go talk to an accountant in a small town who’s clients are all both good friends, and own lots of stuff all over town.
They will RAVE at the idea of treating SMB income as untaxed capital gains as long as it moves from one SMB to another.
This is the kind of stuff I’m 100% SURE Romney understands and Obama has no clue about. I honestly don’t think most Econ eggheads have a clue, because they don’t see SMB firms in the correct way.
We do the same thing with Rule 1031 for taxes on sales of building profits. If you buy another building in 90 days you pay no taxes.
It isn’t enough to “reduce corp taxes” – we need to FAVOR SMB owners, and we need to understand their actual structures and incentives…. so that the hegemony, the smart guys in each small town, are MOST ABLE to best control their resources investing around the town.
21. October 2012 at 08:15
Morgan, absolutely taxes on capital should be eliminated – but you won’t get strong near term effects. It changes the incentive to invest, not the incentive to bear short-run variable costs.
How are “smart guys in each small town” hegemonic if they need “FAVOR”s from government?
21. October 2012 at 08:16
Saturos, I’m very dubious of the Saez claims. There’s a reason why almost every country in the world has lower tax rates on capital income than labor income.
The left has no plausible hypothesis as to why Europeans worked as hard as Americans in the 1960s, but now work much shorter hours. The right does have a plausible hypothesis. You can’t beat something with nothing.
21. October 2012 at 08:21
John,
1. Of course we are focusing on expected multipliers, that’s the only sensible way to approach the issue. And the central bank moves last.
2. I specifically mentioned demand-side fiscal stimulus, that is, stimulus that doesn’t plausibly increase AS.
So I don’t accept either of your criticisms.
21. October 2012 at 08:26
The left has no plausible hypothesis as to why Europeans worked as hard as Americans in the 1960s, but now work much shorter hours
And yet, they are not that much poorer…
I think you had a good argument in an old post – that in the long run states got the same revenues after sharply lowering MTRs. The Laffer curve is real, in the long run.
21. October 2012 at 08:51
Economists seem to me too eager to not learn the underlying assumptions of each model, and what do they mean. This results in applying models when they do not make sense, because the assumptions are no longer reasonable.
Nicholas Taleb characterized this problem in an interesing way. He asked 2 fictional characters the question “A fair coin has been tossed 99 times, and it has landed heads on all 99 of them. What is the probability of the 100th toss to land tails?”
Character #1 replied 50%, since the tosses are independent.
Character #2 replied 0%, because it is more likely that the coin is rigged than a fair coin to land heads 99 times in a row.
While the analogy is a bit too crude, it does highlight the problem at hand: Character #1 does not stop and ask himself if the underlying assumptions of his model apply in this case.
The left has no plausible hypothesis as to why Europeans worked as hard as Americans in the 1960s, but now work much shorter hours. The right does have a plausible hypothesis. You can’t beat something with nothing.
Sometimes saying “I don’t know” is the right thing to do. Creationists have answers to several things scientists do not know the answer to.
21. October 2012 at 09:00
Yes, clearly we should continue to scientifically puzzle over the mystery of why people don’t spend more hours earning Euros when most of each one must be paid to the government.
I don’t think you can blame the guy for following the standard texhnique for solving math problems, where you take the assumptions explicitly stated in the question as given.
“What is 2 + 2? Answer: 3. I was lying to you about the second 2.”
21. October 2012 at 09:22
Scott wrote:
“BTW, notice that if you eyeball the graph showing demand-side tax cuts work, the downward sloping line seems to almost completely disappear if you remove a single observation (2008)?”
This would be an excellent point if you combined it with Krugman’s claim that fiscal stimulus only makes sense when at the zero bound. The ZLB was hit on December 17, 2008. There are three observations during which we were at the ZLB throughout (2009-2011) that appear to apply a rather large weight pushing the slope downward.
What amazes me is that no one noticed this until now.
21. October 2012 at 09:38
I think most principles books don’t really talk about the central bank targeting inflation, they might make an off-hand comment about it. They give the impression of a highly discretionary Fed. I know you’ve posted about this before, but maybe you can talk about what the mechanism is in which fiscal stimulus can work, assuming the Fed doesn’t offset it — it works through changes in velocity, right?
21. October 2012 at 10:07
Saturos,
I don’t think you can blame the guy for following the standard texhnique for solving math problems, where you take the assumptions explicitly stated in the question as given.
“What is 2 + 2? Answer: 3. I was lying to you about the second 2.”
I explicitly said the analogy was too crude. The problem is that in modeling you are asking yourself the question. Therefore, it makes sense to ask if the assumptions given/implied in the question are reasonable. And my claim is that lots of economic commentators are not doing that.
21. October 2012 at 10:14
It depends on what is meant by “inflation targeting”. If by “inflation targeting” one means “the central bank attempts to hit some inflation target using conventional monetary policy, but does nothing beyond that if the zero lower bound prevents it from hitting the target; and the instant it is able to hit the target it do so without any regard for the expectational consequences of such a commitment”, then the Keynesian claim that you assail here is actually right. Assuming separability of consumption from government spending in consumer preferences, then the multiplier is strictly greater than 1.
This is because in the baseline model, the only general equilibrium response that can cause consumption to decrease in response to an increase in government spending is an increase in the expected path of the real interest rate. But when the interest rate is pinned at zero, and the central bank follows purely forward-looking policy, this doesn’t happen. In fact, the real interest rate declines somewhat, since inflation increases in response to government spending. (This is where the “greater than 1” part comes from in New Keynesian models. Without any inflation response or other twists to the model like rule-of-thumb consumers, the multiplier is exactly 1. My view is that this inflation channel is very weak, so that in these models the “multiplier” is basically just 1.)
If by “inflation targeting” one means “the central bank comes up with an intertemporally optimal policy that involves commitment to low rates in the future in order to come closer to the inflation target today”, then things become a lot more complicated, and it is unclear whether the “Keynesian” logic above still holds. There is, in fact, no precise answer, because this kind of “inflation targeting” is not completely specified: to hit the inflation target today, you must choose a policy that fails to hit the inflation target in the future. For a central bank to follow this policy, it would need some kind of loss function that weighed deviations from target in the future against deviations from target today and specified the optimal balance.
Now, my view is that the Fed is implicitly following some variant of the second kind of inflation targeting, and that this probably does make the fiscal multiplier lower than 1. So in some sense I agree with you. But I wouldn’t claim that the “>1 fiscal multiplier” view is just some ignorant relic from undergrad textbooks; it is consistent with a straightforward way of thinking about inflation targeting and Fed policy, one that was probably appropriate in the past and (depending on your exact view of the Fed’s policy regime) could still apply now.
Btw, you keep saying that the “multiplier is zero under inflation targeting”. This is not true, as far as I know, in any model unless government spending is perfectly substitutable for private consumption spending. Otherwise, the neoclassical model (which is what inflation targeting effectively mimics) gives a model above zero, perhaps well above zero depending on the relevant elasticities.
Moreover, if we do assume that government spending is perfectly substitutable for private consumption spending, then the multiplier is also zero in the core New Keynesian model regardless of the specification of monetary policy. So, if this is your view of the world, you don’t even need to bring up inflation targeting to say that Keynesians are wrong.
21. October 2012 at 10:51
Also,
Yes, clearly we should continue to scientifically puzzle over the mystery of why people don’t spend more hours earning Euros when most of each one must be paid to the government.
That is not what I said. I was objecting to the use of a bad argument to defend a hypothesis, not to the hypothesis itself.
21. October 2012 at 11:39
In the case were many fiscal sub-jurisdictions coexist under a single monetary authority targeting inflation (e.g. US states, Canadian provinces, Euro zone members, etc…), wouldn’t make sense for jurisdiction with inflation lower than the target (and insufficient AD) to implement fiscal stimulus and for those over the target to adopt fiscal austerity?
21. October 2012 at 12:34
Mark, Good point.
Matt, You are making this way too complicated. I explicitly abstracted from the zero bound problem–I realize that fiscal stimulus is more controversial when you get to the zero bound.
And I was only looking at demand-side fiscal stimulus. I agree that there are some types of fiscal stimulus (such as employer-side payroll tax cuts), which increase RGDP without any increase in NGDP. I’ve said that repeatedly. Some stimulus like infrastructure spending might boost AS. I’m focusing on demand-side effects here. I specifically discussed an article that was looking at the demand-side effect of tax cuts.
21. October 2012 at 13:05
How about this one for the economists not reading their own textbooks…. Supply side theory says that uncertainty resrains investment and growth. Temporary adjustment to tax rates, such as the “Bush tax cut” create uncertainty. Put that together with the nearly constant adjustments at the margins, and the continuous arguements between lower taxes for all and higher taxes for some. Supply-siders should be fighting to make it more difficult to change tax policy.
Supermajorities for all changes to tax policy!
21. October 2012 at 17:57
From the article:
“An increase in consumption, which still accounts for about 70 percent of G.D.P., fuels increases in demand, and that leads companies to create more jobs.”
ugh. This article makes me want to pull my hair out. The last time I checked, increases in productivity, which is the root of economic growth and prosperity, requires INVESTMENT. Where do you get the capital for sustainable investment? Saving, also known as, REDUCED CONSUMPTION!
Crude Keynesianism indeed. If this is typical mainstream macro, and I’m pretty sure that it is, then mainstream macro is a probably a net drain on society through the transmission of terrible fallacies that justify anti-growth, anti-stability decisions both in policy and personally. Sigh.
21. October 2012 at 18:58
John Papola:
You said it. That is one of my pet peeves too.
21. October 2012 at 20:26
ssumner:
Of course, I mean, that explains why you still cling to the crude Monetarist model, while other economists have unlearned it.
Monetarist theory predicts that a constant rules based inflation is effective whether there are inflation generated malinvestments or not.
22. October 2012 at 04:40
Morgan, thanks for the informative post, and especially for:
“THEN and only then is the agent able to make the correct decision without giant inefficient tax incentives pushing him to OVER HOOTERS Canton, and under robot landscape Canton.”
22. October 2012 at 10:49
There’s a lot of junk in economics textbooks that’s much worse. It starts with loanable funds(banks actually issue credit first and look for reserves later), then goes to rational expectations(which assumes that the errors are random and the movements are independent), then goes to the quantity theory of money(which falsely assumes the velocity of money is constant)–I would actually argue that the quantity theory of money is wrong in the medium term as well, possibly even in the long-run, to efficient markets(the strong-form, I have no problem with the weak form), and that’s just the stuff that I can think of at the top of my head. I’m sure there’s a lot more stuff than that.