Scott Sumner  

Lake Wobegon Keynesianism

PRINT
The Excellent and Admirable Al... The Silent Suffering of the No...

With the recent discussion of a possible "fiscal stimulus" coming out of Washington, I find that many people are forgetting a few basic ideas from EC101. As you may know, I do not believe that fiscal stimulus in the US has a significant impact on NGDP or RGDP. That's mostly because of fiscal offset of the demand side impact---the Fed would simply raise rates faster than it is currently contemplating, in order to keep inflation expectations close to 2%. In addition, it has little impact on aggregate supply; indeed even the sign is ambiguous. But for the moment I am going to assume that I am wrong, and that fiscal stimulus is expansionary.

The key point to understand about fiscal stimulus is that more fiscal stimulus today implies less fiscal stimulus during the next recession. That's because the expansionary impact of fiscal stimulus (in the Keynesian model) occurs because the full employment budget deficit gets larger than before, not because it is large in absolute value.

For the moment, let's assume that the full employment deficit during the next recession is $X. In that case, the larger the fiscal stimulus between now and the onset of the next recession, the smaller the increase in the deficit during the next recession, and thus the smaller the fiscal stimulus during the next recession.

But it's actually worse than this. Many pundits seem to be forgetting about the government's long run budget constraint. Even if the national debt is never paid off, it must be serviced. And that means that even in the US there are limits as to how much can be borrowed. This means that I cannot hold the future deficit constant when contemplating fiscal stimulus today (as in the previous example), I need to assume that a bigger deficit today will imply a smaller deficit during the next recession. When you read some pundits, it almost seems as if they are assuming that fiscal policy can always be more expansionary that average. But that's impossible! (Except in Lake Wobegon.)

Thus the fiscal policy being contemplated today would actually make the economy more unstable if the Keynesian model is are correct. Fortunately, the Keynesians are wrong (due to monetary offset) and hence the foolish and poorly timed fiscal stimulus that we are likely to get out of Washington will merely be a waste of hundreds of billions of dollars, nothing more.

I'd also point out that most infrastructure decisions and financing should be made at the state and local level, where funds are more likely to be spent wisely. There may be a few minor externality cases, such as I-95 though Delaware, but these can be handled on an ad hoc basis, by having a Federal agency require a minimal level of road quality in cases where interstate highways roads are mostly used by out of state residents. Ditto for mass transit. LA County just voted a small sales tax increase to fund a massive expansion of its commuter rail system. That's the level of government where these decisions should be made. People make wiser decisions spending their own money, than if they are spending someone else's money.

Alex Tabarrok
has an excellent new post pointing out one area where the federal government could help---making it easier to privatize facilities such as airports. This idea is further advanced in Europe, where infrastructure is much more likely to be privately operated, and is generally of higher quality. Hopefully Trump or his advisers will read Alex's post.

BTW, I am not suggesting that the federal government has no role here. Alex mentions some national security issues with a safe electricity grid. In addition to terrorist attacks, some worry about the impact of large solar flares. So there may be a few cases where externality issues allow for a federal role. But they are few and far between. Thus the federal government should NOT subsidize California's high-speed rail boondoggle. My fear is that Trump likes to build grand projects---I hope I am wrong.

Not a federal responsibility:

Screen Shot 2016-11-14 at 9.32.13 AM.png
PS. Newton, Massachusetts (where I live) does a poor job of maintaining its roads. But it's not due to a lack of money; the town is quite wealthy and just built a $200 million high school (3 times the normal cost). It's because voters keep voting for politicians who would rather waste money on lavish schools than fill potholes.


Comments and Sharing






COMMENTS (18 to date)
baconbacon writes:

But wait, there's more! If you buy into the Wicksellian interest rate then higher inflation will mean higher interest payments for the Treasury, which will continue to increase the deficit during the "recovery" period (as bills are rolled over. While a relatively minor problem for a country with a debt rate around its tax collection rate (~20% of GDP), the US is 4-5x its tax rate in total debt, so the longer the "recovery" the more debt will have rolled over at higher rates and eventually the higher interest payments will exceed the extra tax collection from the increase in GDP.

Market Fiscalist writes:

' That's because the expansionary impact of fiscal stimulus (in the Keynesian model) occurs because the full employment budget deficit gets larger than before, not because it is large in absolute value.'

I'm not sure I'm following this.

Take a simple Old Keynesian model where recessions are caused by a drop in I, and corrected by an increase in G. if the govt increases G outside of a recession then increased G will cause I to fall (crowding out). But then when the next recession does come along - as G is now a bigger % of (G+I) the increase in G needed to adjust for the fall in I will be smaller than if G has not increased prior to the recession.

Are you just saying that when the next recession come along then (in the Old Keynesian model) G will need to be increased further no matter what level it had been brought to ahead of the recession?

"There may be a few minor externality cases, such as I-95 though Delaware, but these can be handled on an ad hoc basis, by having a Federal agency require a minimal level of road quality in cases where interstate highways roads are mostly used by out of state residents. "

The modern European solution to this issue is (often privately built and run) toll roads. No externality.

baconbacon writes:
But then when the next recession does come along - as G is now a bigger % of (G+I) the increase in G needed to adjust for the fall in I will be smaller than if G has not increased prior to the recession.

The size of the increase in G that is necessary is related to the size of the recession. Unless you are claiming that a larger G during an expansion phase will lead to a smaller recession Scott's point appears correct.

Kevin Erdmann writes:

I think the infrastructure thing is even worse. Localities prefer to have control over their infrastructure spending. So, the only projects that get routed through Washington are the projects that failed the local cost/benefit test. So they get repackaged as jobs programs that will bring a federal injection of money to the local economy. Federal infrastructure spending isn't just spent carelessly. It's actually targeted at failed projects.

Market Fiscalist writes:

@baconbacon

I just meant that a recession caused by a given % fall in I, would need less additional G (in absolute terms) to "fix" the recession the bigger that G is as a % of GDP.


Thaomas writes:

I'd like to see a federal role in allowing states and local governments to borrow at near federal LT bond rate for infrastructure projects that pass NPV tests and for the federal government to start using such tests. (I assume that Newton's potholes and new IRS computers would and the 200 million HS and high speed rail would not.)

Thaomas writes:

On the political economy of Fed accommodation. with Republicans in power, there will be less political pressure on the Fed to have tight money just as Republican concern about the deficit will evaporate.

baconbacon writes:
I just meant that a recession caused by a given % fall in I, would need less additional G (in absolute terms) to "fix" the recession the bigger that G is as a % of GDP.

You are functionally stipulating that the recession is smaller because G is larger. Recessions are not measured against G or I, but potential GDP, which includes both. If you assume that the next recession will be smaller due to a larger G you can get to this conclusion, but it does not appear to have any empirical or theoretical backing.

Khodge writes:

Trump is, of course, big on extravagant spending. The current Republicans, however, seem to be intent on holding the line on federal intrusion in state matters. Moreover, they are extremely unlikely to grant a boondoggle to California, given the current political battle lines.

Hazel Meade writes:

Let's have a mission to Mars.

Make America Great Again!

MikeP writes:

If you assume that the next recession will be smaller due to a larger G you can get to this conclusion, but it does not appear to have any empirical or theoretical backing.

Indeed, I came here to say exactly the opposite.

G crowding out I almost tautologically means that production people don't want to invest in is crowding out production people do want to invest in. It is hard to imagine that a recession along that course isn't deeper than otherwise.

Then again, the more people who get paychecks from the government for doing nothing people actually want, the less boom and bust there is in the real economy. So maybe the assertion that greater G leads to shallower recessions has some merit. The cost, of course, is an economy in perpetual doldrums -- a Great Stagnation, if you will.

Andrew_FL writes:

Didn't you already use this title?


@Hazel Meade-You know, you joke, but one of his closest allies is New Gingrich...

Philo writes:

"Even if the national debt is never paid off, it must be serviced. And that means that even in the US there are limits as to how much can be borrowed." This seems not to follow: the interest payments--no matter how large--on the existing debt could (in principle) always be financed by further borrowing.

john hare writes:

@Kevin ----- Federal spending actually targeted at failed projects is something I have missed before. Thanks for that sentence.

@Philo ------- If the debt were so high that the whole national budget couldn't even pay the interest, then that would be a hard limit as no responsible party would continue loaning to that nation. Obviously the cut off is well below this point, but there are hard limits somewhere along the line.

Scott Sumner writes:

Market, I am saying that if fiscal policy is more expansionary that average during expansions, it will be less expansionary than average during contractions. And less expansionary than average means (in the Keynesian model) contractionary.

Luis, I agree.

Kevin, Good point.

Andrew, Very likely, I'm getting on in years. Hopefully the post is not identical.

Philo, That would lead to an exponential growth in the debt as a share of GDP, and it would eventually blow up. You'd have default or hyperinflation.

TMC writes:

"but one of his closest allies is New Gingrich"

Well, he was the last one to balance a federal budget. Hope he does the same again.

John S writes:

If the US Debt/GDP ratio got too high, couldn't the Fed purchase and retire a good chunk of that (as the BoJ has done in Japan) to lower the debt burden to a manageable level?

(I'm asking hypothetically; I know Fed actions are supposed to be taken without regard to the current level of gov't debt.)

Comments for this entry have been closed
Return to top