I want to revisit Tuesday’s article “The $1.9 trillion relief plan is not too large,” both because it quoted economist Ernie Tedeschi laying out some budget math that I don’t think is correct, and more broadly because after more conversations I think I can now describe a different — and somewhat stronger — argument against the size of the relief package than the inflation-centric one I rebutted on Tuesday.
Basically, I think the argument that the Biden relief plan will spark some huge amount of inflation is not credible, but the idea that it’s in some sense wasteful and delivers a fair amount of money that’s not strictly needed is pretty reasonable.
I don’t find this especially troubling, but others do, either for reasons of fastidiousness or based on concerns about the longer-term political consequences. But I do think it’s mathematically correct. If you shrugged off all the political constraints, you could develop legislation that’s better targeted than this.
Revisiting the output gap math
The point he and I were both making is that $1.9 trillion seems like too much stimulus if you believe in the Congressional Budget Office’s estimate of the output gap, but if you (correctly) believe the CBO is wrong about this, then $1.9 trillion seems about right. I quoted two Tedeschi tweets that made this point in terms of a monthly run rate; a CBO-based calculation says there’s an output gap of $50 billion per month, while Biden is delivering about $150 billion per month.
This chart from Tedeschi, by contrast, says that if you assume CBO is lowballing potential by about 2 percentage points, then $1.9 trillion is a good output gap estimate over a period of several years.
And that, I think, leads to the strong criticism of Biden’s COVID-19 plan — not that it’s too much money per se, but that it’s too front-loaded, kicking out a ton of money really soon while the supply side of the economy is still severely constrained by the virus rather than dribbling out smaller sums over a longer period of time to promote recovery.
One possible way that this could play out is in the form of high inflation in March and April or something. But for a whole bunch of reasons, I don’t think that’s plausible (go back to all the arguments in Tuesday’s post about market-based measures of inflation). What’s going to happen instead is that Biden’s stimulus is going to have a really low multiplier.
The Keynesian multiplier, explained
The multiplier idea is that when you pump money into a severely depressed economy, you generate more than $1,000 in GDP for every $1,000 the government spends.
You get some money, then you spend it on subscribing to my Substack, then I go buy a sandwich, and the sandwich guy goes to the store, and on and on and on. Paul Krugman’s old 1998 article on the Capitol Hill Babysitting Co-op remains the best popular explanation of how and why the multiplier works. It was a key part of arguments during the Great Recession, with progressives generally arguing that they were pushing for high-multiplier ideas that should be embraced.
So why does the multiplier matter? Well, officially, economists argue about it because they’re social scientists who want to understand the world.
But in practice:
A high multiplier means that it’s good to borrow money even if you’re worried about the outstanding stock of debt because the return on investment is really high, even if debt is in some sense bad.
Giving money to poor people (something Democrats enjoy) who will go spend it like a hot potato is generally seen as a high multiplier activity, while regressive tax cuts (something Republicans enjoy) have a low multiplier.
The Biden rescue package, however, does not seem to me like it’s a very high multiplier undertaking.
Some of that is because a bunch of this money is really for supply-side issues — it’s testing, cash for vaccination programs, and genetic surveillance. You need to evaluate that bit in public health terms, not macroeconomic terms. What economics can tell us is that controlling the pandemic is very valuable, but that’s not a fiscal policy question.
But the big thing is, these buckets of money going to state and local government (some of which are characterized as being for school reopenings, etc.) and the “survival checks” seem unlikely to be spent quickly.
Now to be clear, some state and local governments — ones that rely on a lot of sales tax revenue from commuters (New York City) or tourists (Florida, Nevada) — are in dire budget straits, and the money will have an immediate impact. But California is looking at a $15 billion budget surplus driven by the soaring stock market and the state’s progressive income tax. And by the same token, some of the people set to get an extra check are facing severe financial difficulties, but plenty of money is going to flow into the bank accounts of employed people who just add it to their “fun stuff to do once things reopen” account.
That’s a recipe for a bill that would score very poorly in terms of 2009-vintage multiplier debates. So this critique would be not that the package is “too large” in the sense of “the economy will overheat,” but in the sense of being wasteful measures that have a very low multiplier to the GDP.
How bad is it to be wasteful?
This is where the output gap math becomes relevant again.
If you believe the CBO’s relatively low output gap forecasts, then this really is just too big. If you’re like me and you think they’re missing the boat, the bad timing doesn’t seem so bad to me. State governments that don’t really need the extra money that they are getting will think of something to do with it over the next 2 to 3 years. Similarly, households that get checks that aren’t spent right away will spend them later. Money is money, and it will get spent. Meanwhile, the people and governments who are hard-up will get much-needed relief.
Of course, if you’re really anxious about debt, then this hand-waving “eh, it’ll work out” is not that convincing.
All I can really say to this is that with inflation-adjusted interest rates on government debt below zero, I just fundamentally think that’s wrong. It’s irresponsible not to be borrowing more money.
But then you get to opportunity costs. Granting that some kind of $1.9 trillion package was warranted, was this $1.9 trillion package warranted? I think that’s the strongest criticism Summers raised — maybe we could have cut out hundreds of billions of dollars worth of short-term low-multiplier federal spending and poured it into infrastructure projects that would get built over the next 2-3 years and have all kinds of huge long-term benefits. And I think as long as you leave it on that level of abstraction, it’s pretty compelling.
How would it work, though? After Democrats won the Senate on a promise to give a full $2,000 in direct payments, I think they basically have to deliver. And if you're going to do state and local aid, the simplest way to do it is to push money out through existing funding streams that ensure everyone gets some. Skimping on the actual virus stuff to go build broadband in northern Maine would be nuts. So then you start looking at the stuff that actually does have a strong multiplier — Unemployment Insurance, SNAP, an enhanced Child Tax Credit, etc. — which seems like a weird area to cut.
Probably the biggest opportunity cost here is in terms of not building automaticity and triggers into the package.
The federal government splits Medicaid costs with states using a formula called FMAP. Jason Furman, the top Obama administration economist, has this plan to make FMAP payments variable based on state-level unemployment rates. That’s a way of doing state and local aid that would target the funds to the states that actually need money, and would also be infrastructure to help prevent future recessions. We could make the bonus UI ramp up or down based on what’s actually happening with job openings. Lots of clever ideas about automatic stabilizers have been kicking around for years, and congressional Democrats keep refusing to pick them up for bad reasons. I wish they would pay more attention to the substantial importance of automaticity and worry less about CBO scores.
But the biggest thing I think is that on pure politics, you have to give credit where due.
Biden’s plan is very popular
The Summers criticism that makes the most sense economically is that the marginal parts of this plan are just not important enough on the merits to be worth burning down Biden’s political capital. It would be a shame for a president to make his political legacy a bunch of middle-class people getting a random, one-off $1,400 check.
But in the real world, does this critique make sense? The “political capital” metaphor always seems odd to me, because how do you know which things let you accumulate more of it versus which things let you spend it down?
I think the more straightforward thing to say is that in addition to evaluating a president’s decisions on the basis of how much good they do in the world, you want a president you like to make decisions that are popular rather than unpopular. I favor a carbon tax, for example, but there would be no point in Biden pushing through a carbon tax by a 51-50 vote on a reconciliation measure only to tempt a massive midterm backlash that leads to — among other things — its repeal.
But I don’t see any evidence in the polling that the president is costing himself anything by pushing this plan.
Polling isn’t the be-all and end-all of policymaking. But it’s hard to see exactly what the “political capital” concern here would be.
Biden’s team claims to be working on a second request to Congress for some kind of infrastructure-focused recovery plan in order to complement the rescue plan we’ve been talking about. I’ll be interested to see what that looks like. I’m a bit of an infrastructure skeptic compared to a lot of economics writers I know, since this country has a bad habit of blowing resources on dubious projects. But maybe that’s one reason I kind of like the structure of the rescue plan. By not using infrastructure spending as an emergency employment scheme, I hope we are laying the groundwork for an infrastructure plan that’s actually about infrastructure and not about “job creation.”