Milan, I think this piece really should have included ‘in hindsight’ in the title and not buried your assertion in paragraph 25-30. The biggest issue I have with the piece is that you had a truly unique combo of bad things happening and significant non linearity in outcomes. And that is one reason why you really want to overshoot vs undershoot. Take the output gap. Did we have any idea how big the output gap would be if: a massive default of commercial mortgages was paired with massive austerity for state governments who lacked sales tax revenue from restaurants? Or individuals forced to sit on the fence? How many waves of variants would we be dealing with?
I also think this is still too defeatist a posture. The fact is we finally ended the sclerosis plaguing the American economy, we did it by generating large deficits and we now can raise interest rates back to their (very) long run norms. We can now focus on supply side issues by doing what Keynes pushed to be done on the demand side.
The thing is we knew pretty well by March that we were not going to be facing a massive default on commercial mortgages or state governments being forced into austerity — we had at worst a modest output gap to fill and at best had already filled it with the December stimulus.
But really, this is all the more reason to do automatic stabilizers!
Which March are you talking about? 2022? Even in 2021 I think it was rather touch and go…you are talking about 2 months after the ‘not so peaceful’ transfer of power?
No disagreement that automatic stabilizers are great!
In any event I would not have so much faith in a derived measure like the output gap. I think it’s ok to over shoot and almost morally required to do so when rates are at the zero bound.
Clearly, we had a big stimulus bill and then inflation happened. But post hoc is not always propter hoc. My own hypothesis is that we would be experiencing a fair amount of inflation even with a much smaller amount of stimulus, exactly because of the supply side issues. Are you aware of any analysis that attempts to break down different contributions to ongoing inflation?
Nah, I'm fine with lower unemployment in exchange for higher inflation, especially since if you look at the month-by-month, things are already returning to normal, slowly but surely. Yes, the Rescue Plan could've been arranged differently, so much of it wasn't a slush fund for Republican governors, but better to overspend the money now, instead of the reality that we're going to have no chance to spend the extra money when we "need" too down the line.
Especially when I think even if we had spent the "right" amount to make sure the right number of millions of people were unemployed to make sure nobody had to pay a dime higher for food, we would be losing the House & Senate anyway.
If the choice is 3.5% unemployment & GOP Governors get to give tax cuts and teacher raises or 5.5% unemployment and no giveaways, I'll still take the former, especially when it's pretty likely the slush fund would still exist, there'd just be no CTC at all in this new rescue bill.
1. The size of the miss matters in addition to the direction — undershooting by $500 is preferable to overshooting by $500 billion. Per CRFB, we were very close to closing the output gap after the December package, we knew ARP would overshoot by a ton.
2. This is something that Joseph Gagnon at the Peterson Institute told me that got cut for length, but one of the ways that Covid acts as a supply shock is by making some people less willing to work. That raises the natural rate of unemployment (by ~1pp per Gagnon) with the upshot being that a 3.6% unemployment rate today reflects a much tighter labor market than it did in 2018/2019. The natural rate of unemployment is adjusting back down to pre-pandemic levels but it's not clear how long that will take or how far back down it will go, but that should make us realize that we shouldn't shoot for the exact same unemployment rate than we had pre-virus to get the same tightness in the labor market.
I believe I can say without fear of contradiction that whatever policy had been passed last year would either have been too big or too small and that well after the fact we would know how big it should have been and how it should have been better structured.
Seeing that beautiful Alt-ARP layout that gives us spending on housing, CTC, climate, and raises capital gains taxes, all without tanking Biden's numbers, thus opening the door for another reconciliation bill....Ugh...my heart hurts lmao
Isn’t a big part of this the fact that covid has continued to rage contradicting expectations at the time the stimulus was passed? This helped continue pushing spending onto goods vs services and continued to cause disruption as people had to leave work for illness and many workers continued to hold out of the labor market or left entirely (older workers for instance). Then that piled into a war and now massive disruptions in China.
It isn’t at all clear that we would have gotten nearly as much inflation if COVID had effectively gone away last summer (ie no delta or omicron).
This sort of analysis always acts as if the recession was caused by a normal source of falling demand and the question really was output gaps.
But that wasn’t the actual problem. The problem was a massive pandemic had put people out of work and left businesses with reduced hours. Whether a policy would be inflationary depended heavily on how long that was expected to last and which industries would come online when.
“Markets can stay irrational longer than you can stay solvent" is generally thought to refer to the stock market, but it also can apply to politics. This inflation may be transitory, but it can last longer than the Democratic party can absorb before getting crushed at the polls.
With respect to the "stimulus" in 2020-2021, I think it is fair to say that there wasn't a demand-driven output gap that just needed more money being spent. But there were plenty of opportunities to spend money on things that would have reduced the supply issues. For example, there is really no good reason that rapid tests could not have been freely available by the fall of 2020. With a little incentive payment, I'm sure we could have gotten tests that turn a color when negative and stay that for a day so you can carry it around. Workers could have tested every day before shift. Diners could have shown them to get into restaurants. Yes, this would have been expensive and a major logistical effort, but the RoI would have been enormous.
More broadly, though, stimulus discussions nearly always male aggregation errors. Yes, if there are idle resources due to a lack of demand, you can have a free lunch by paying them to do something, because that doesn't take resources away from any other economic activity. But that doesn't mean you can spend money on just anything. Too frequently these programs end up spending on resources that aren't actually idle. The best example is probably road construction, where even during the financial crisis contractors were already very busy. Throwing money at them didn't do anything to alleviate the demand shorfall.
Much better in all these discussions would be focusing on eliminating supply bottlenecks (some of which would actually reduce spending) and supporting incomes at the bottom where a lack of savings may result in consumption of capital (e.g., selling a car that could be used to get to a better job, not taking a training course) and hurt the economy's long run potential.
This is not relevant to the contents of the article, but in the accompanying photo Chuck Schumer is wearing his mask all wrong, like he's wearing it as an inverse-MAGA instead of as a tool to reduce his risk of getting Covid.
I actually do not think the spending was too big. although I would much prefer if the transfers had been off set by taxes. Politically, however, the problem was putting what I think are low priority items in the BBB. A CTC is almost a pure transfer with only the slight deadweight loss of less effort (if any) by the high income taxpayers who would pay for it. Child care credit, on the other hand does skew child care from informal to formal channels. And the continued failure to create a generous automatic VAT-financed unemployment insurance system is a mystery and a shame.
Others are noting the lack of discussion of monetary policy, but I want to point out something specific. Nominal rates in 2008-2009 were low, but the Fed made a serious category error on thinking about interest on reserves (IOR). In an attempt to keep its balance sheet reasonably sized, it started paying IOR. The result was that, regardless of nominal interest rates, the IOR set a floor on the rate banks were willing to lend at: why lend money at 1% when you earn 2% keeping it in your Fed account? So rates looked much lower than they really were.
My understand was that it was a risk mitigation issue after 2008 to get firms to hold more credible reserves. It doesn't usually make sense to loan out at 1% return unless there is no risk, and the only place that is very true is with the fed.
Trump's fault. But he was pro-stimulus the whole of 2020 and basically right on the economy. Pelosi was trying to wrangle too much money the entire year and playing politics. It's politics and she played the game well, but she wasn't advocating for what the country needed.
1. We can never truly know ahead of time how much stimulus would be “just right” to preserve the economic status quo.
2. Economic cycles create healthy pressure on companies and states, “creative destruction” much like natural forest fires. The real problem is the impact on individuals.
Therefore my takeaway from living through both 2008 and 2020 is that in the future I’d rather see the government do massive direct cash assistance to individuals and choose to risk overshoot rather than undershoot, but don’t do things like bail out states (2020) or banks (2008). We have tools to gently or even aggressively correct excess demand if needed. Underemployment is harder to fix, takes longer to fix, and does more harm.
I agree that getting it just right > overshoot > undershoot but the magnitude of the miss matters — you'd rather undershoot by $500 than overshoot by $500 billion. And per CFRB, we knew that ARP was going to overshoot big time when it was being written.
Milan are you trying to tell me the "money printer go brrrrrr" memes were actually bad policy??? Who could have predicted this.
Milan, I think this piece really should have included ‘in hindsight’ in the title and not buried your assertion in paragraph 25-30. The biggest issue I have with the piece is that you had a truly unique combo of bad things happening and significant non linearity in outcomes. And that is one reason why you really want to overshoot vs undershoot. Take the output gap. Did we have any idea how big the output gap would be if: a massive default of commercial mortgages was paired with massive austerity for state governments who lacked sales tax revenue from restaurants? Or individuals forced to sit on the fence? How many waves of variants would we be dealing with?
I also think this is still too defeatist a posture. The fact is we finally ended the sclerosis plaguing the American economy, we did it by generating large deficits and we now can raise interest rates back to their (very) long run norms. We can now focus on supply side issues by doing what Keynes pushed to be done on the demand side.
The thing is we knew pretty well by March that we were not going to be facing a massive default on commercial mortgages or state governments being forced into austerity — we had at worst a modest output gap to fill and at best had already filled it with the December stimulus.
But really, this is all the more reason to do automatic stabilizers!
Which March are you talking about? 2022? Even in 2021 I think it was rather touch and go…you are talking about 2 months after the ‘not so peaceful’ transfer of power?
No disagreement that automatic stabilizers are great!
In any event I would not have so much faith in a derived measure like the output gap. I think it’s ok to over shoot and almost morally required to do so when rates are at the zero bound.
Clearly, we had a big stimulus bill and then inflation happened. But post hoc is not always propter hoc. My own hypothesis is that we would be experiencing a fair amount of inflation even with a much smaller amount of stimulus, exactly because of the supply side issues. Are you aware of any analysis that attempts to break down different contributions to ongoing inflation?
This article by the SF Fed says that Covid stimulus (CARES + ARP) increased the inflation rate by ~3pp: https://www.frbsf.org/economic-research/publications/economic-letter/2022/march/why-is-us-inflation-higher-than-in-other-countries/
Nah, I'm fine with lower unemployment in exchange for higher inflation, especially since if you look at the month-by-month, things are already returning to normal, slowly but surely. Yes, the Rescue Plan could've been arranged differently, so much of it wasn't a slush fund for Republican governors, but better to overspend the money now, instead of the reality that we're going to have no chance to spend the extra money when we "need" too down the line.
Especially when I think even if we had spent the "right" amount to make sure the right number of millions of people were unemployed to make sure nobody had to pay a dime higher for food, we would be losing the House & Senate anyway.
If the choice is 3.5% unemployment & GOP Governors get to give tax cuts and teacher raises or 5.5% unemployment and no giveaways, I'll still take the former, especially when it's pretty likely the slush fund would still exist, there'd just be no CTC at all in this new rescue bill.
Two things:
1. The size of the miss matters in addition to the direction — undershooting by $500 is preferable to overshooting by $500 billion. Per CRFB, we were very close to closing the output gap after the December package, we knew ARP would overshoot by a ton.
2. This is something that Joseph Gagnon at the Peterson Institute told me that got cut for length, but one of the ways that Covid acts as a supply shock is by making some people less willing to work. That raises the natural rate of unemployment (by ~1pp per Gagnon) with the upshot being that a 3.6% unemployment rate today reflects a much tighter labor market than it did in 2018/2019. The natural rate of unemployment is adjusting back down to pre-pandemic levels but it's not clear how long that will take or how far back down it will go, but that should make us realize that we shouldn't shoot for the exact same unemployment rate than we had pre-virus to get the same tightness in the labor market.
Re 1. I saw that this was asked elsewhere, but this makes me curious where you'd be indifferent to the tradeoff.
Which of the following would you prefer:
1. 500 billion overshoot or 500 dollar undershoot?
2. 500 billion overshoot or 500 thousand undershoot
3. 500 billion overshoot or 500 million undershoot
4. 500 billion overshoot or 5 billion undershoot
5. 500 billion overshoot or 50 billion undershoot
6. 500 billion overshoot or 250 billion undershoot
7. 500 billion overshoot or 500 billion undershoot
And where's your line?
I believe I can say without fear of contradiction that whatever policy had been passed last year would either have been too big or too small and that well after the fact we would know how big it should have been and how it should have been better structured.
Seeing that beautiful Alt-ARP layout that gives us spending on housing, CTC, climate, and raises capital gains taxes, all without tanking Biden's numbers, thus opening the door for another reconciliation bill....Ugh...my heart hurts lmao
Isn’t a big part of this the fact that covid has continued to rage contradicting expectations at the time the stimulus was passed? This helped continue pushing spending onto goods vs services and continued to cause disruption as people had to leave work for illness and many workers continued to hold out of the labor market or left entirely (older workers for instance). Then that piled into a war and now massive disruptions in China.
It isn’t at all clear that we would have gotten nearly as much inflation if COVID had effectively gone away last summer (ie no delta or omicron).
Yes, if Covid disappeared that would fix a big part of the supply side issue
This sort of analysis always acts as if the recession was caused by a normal source of falling demand and the question really was output gaps.
But that wasn’t the actual problem. The problem was a massive pandemic had put people out of work and left businesses with reduced hours. Whether a policy would be inflationary depended heavily on how long that was expected to last and which industries would come online when.
This sort of analysis also ignores non-fiscal aspects of public policy. (Or pretends they don’t matter much when it comes to economic growth.)
Yeah… the Fed really should try to predict and react to Federal outlays.
Shor and Summer pilled.
If this is true, shouldn’t the inflation work its way out of system?
“Markets can stay irrational longer than you can stay solvent" is generally thought to refer to the stock market, but it also can apply to politics. This inflation may be transitory, but it can last longer than the Democratic party can absorb before getting crushed at the polls.
It was always most likely to get crushed at the polls.
American politics just see-saws between the parties, there’s not a substantial logic to it.
With respect to the "stimulus" in 2020-2021, I think it is fair to say that there wasn't a demand-driven output gap that just needed more money being spent. But there were plenty of opportunities to spend money on things that would have reduced the supply issues. For example, there is really no good reason that rapid tests could not have been freely available by the fall of 2020. With a little incentive payment, I'm sure we could have gotten tests that turn a color when negative and stay that for a day so you can carry it around. Workers could have tested every day before shift. Diners could have shown them to get into restaurants. Yes, this would have been expensive and a major logistical effort, but the RoI would have been enormous.
More broadly, though, stimulus discussions nearly always male aggregation errors. Yes, if there are idle resources due to a lack of demand, you can have a free lunch by paying them to do something, because that doesn't take resources away from any other economic activity. But that doesn't mean you can spend money on just anything. Too frequently these programs end up spending on resources that aren't actually idle. The best example is probably road construction, where even during the financial crisis contractors were already very busy. Throwing money at them didn't do anything to alleviate the demand shorfall.
Much better in all these discussions would be focusing on eliminating supply bottlenecks (some of which would actually reduce spending) and supporting incomes at the bottom where a lack of savings may result in consumption of capital (e.g., selling a car that could be used to get to a better job, not taking a training course) and hurt the economy's long run potential.
This is not relevant to the contents of the article, but in the accompanying photo Chuck Schumer is wearing his mask all wrong, like he's wearing it as an inverse-MAGA instead of as a tool to reduce his risk of getting Covid.
I think Schumer got vaccinated in December, the picture is presumably shortly after ARP passed Congress in early March
I actually do not think the spending was too big. although I would much prefer if the transfers had been off set by taxes. Politically, however, the problem was putting what I think are low priority items in the BBB. A CTC is almost a pure transfer with only the slight deadweight loss of less effort (if any) by the high income taxpayers who would pay for it. Child care credit, on the other hand does skew child care from informal to formal channels. And the continued failure to create a generous automatic VAT-financed unemployment insurance system is a mystery and a shame.
“Child care credit, on the other hand does skew child care from informal to formal channels”
Thereby making it less cost effective.
That was my point. I think a child tax credit is better than a child care tax credit or subsidy or direct provision
I don’t think it’s much of a mystery — people don’t like paying higher taxes.
Others are noting the lack of discussion of monetary policy, but I want to point out something specific. Nominal rates in 2008-2009 were low, but the Fed made a serious category error on thinking about interest on reserves (IOR). In an attempt to keep its balance sheet reasonably sized, it started paying IOR. The result was that, regardless of nominal interest rates, the IOR set a floor on the rate banks were willing to lend at: why lend money at 1% when you earn 2% keeping it in your Fed account? So rates looked much lower than they really were.
My understand was that it was a risk mitigation issue after 2008 to get firms to hold more credible reserves. It doesn't usually make sense to loan out at 1% return unless there is no risk, and the only place that is very true is with the fed.
Trump's fault. But he was pro-stimulus the whole of 2020 and basically right on the economy. Pelosi was trying to wrangle too much money the entire year and playing politics. It's politics and she played the game well, but she wasn't advocating for what the country needed.
Trump's big idea on the economy was to repeal the ACA and cut taxes for the rich so I wouldn't go so far as saying he was "basically right"
I thought his big idea was that the Fed should have an easy money policy? The other two seem like him accepting the default Republican position...
Good article.
Oversimplifying, but here’s roughly how I see it:
1. We can never truly know ahead of time how much stimulus would be “just right” to preserve the economic status quo.
2. Economic cycles create healthy pressure on companies and states, “creative destruction” much like natural forest fires. The real problem is the impact on individuals.
Therefore my takeaway from living through both 2008 and 2020 is that in the future I’d rather see the government do massive direct cash assistance to individuals and choose to risk overshoot rather than undershoot, but don’t do things like bail out states (2020) or banks (2008). We have tools to gently or even aggressively correct excess demand if needed. Underemployment is harder to fix, takes longer to fix, and does more harm.
I agree that getting it just right > overshoot > undershoot but the magnitude of the miss matters — you'd rather undershoot by $500 than overshoot by $500 billion. And per CFRB, we knew that ARP was going to overshoot big time when it was being written.
I agree if we compare actual 2013 to what you just made up about 2025, 2025 sounds way better!