We are in the midst of a real business cycle. So this should be a feather in the cap for real business cycle theory, right? Actually, it’s looking more like the death knell of RBC theory. That’s because when we finally have an honest to God real business cycle, it looks utterly unlike anything we’ve ever seen before.
For example, consider the unemployment rate, which increased from 3.5% in February to 14.7% in April. The entire 2020 recession lasted for only two months, far less than any previous recession in US history. But the weirdness doesn’t stop there. In the next 4 months the unemployment rate fell by 6.3 percentage points, down to 8.4%.
For comparison, during the recovery from the 2008-09 recession it took an entire decade for the unemployment rate to fall 6.5 points, from a peak of 10% in 2009 to 3.5% in 2019. That means that during the recovery from the 2020 recession, unemployment fell roughly 30 times as fast as during the recovery from the 2009 recession. Read that again. I didn’t say 30% faster, I said 30 times faster.
And before anyone says this is no surprise, let me assure you that when Lars Christensen predicted that unemployment would fall below 6% by November, almost everyone thought he was being wildly optimistic. At the time, the most recent data showed 14.7% unemployment, and many people were throwing out figures like 20% unemployment for the summer months. (Christensen’s May 11 post is excellent, well worth reading.)
I suspect we may end up falling a bit short of Lars’s optimistic forecast (although it’s still very possible he will be correct.) But what is even more amazing is that this striking fall in the unemployment rate occurred against very strong and unexpected headwinds. As of June, Covid–19 deaths in the US were falling very rapidly (down 75% from April), following the previous pattern seen in Europe. Many people anticipated that there would eventually be a second wave, but it was expected to occur when the weather got colder. Instead, in the hottest part of summer the US got hit by a huge second wave, while deaths in Europe and Canada continued at a very slow pace. Many states that had been re-opening their economy reversed course, and started restricting certain business sectors. Here in California, significant parts of the economy were shut down once again.
And yet, despite that backdrop of a severe and unexpected second wave of Covid-19 cases in the US, and renewed shutdowns in many southern states, we saw unemployment fall 30 times faster than during the recovery from the 2009 recession. Imagine our recovery if the pandemic here had followed the European/Canadian pattern! Real recessions look absolutely nothing like normal US business cycles. They are radically different phenomena with radically different causes.
[In fairness, the total employment data is somewhat less impressive than the unemployment rate data, but even total employment has increased at an explosive and unprecedented rate in recent months.]
When market monetarists like me argued that the problem in 2009 was too little NGDP, i.e. tight money, we got lots of pushback on two grounds. One group argued that the real problem was real. The financial crisis was a real shock, and recoveries from financial crises tend to be slow. These pundits were not well informed on US economic history. The US has had lots of recessions associated with financial crises, and economic growth was typically quite rapid after the crisis ended. And when you asked people why a financial crisis would cause RGDP to fall, the explanations tended to center around consumer loss of wealth and a lack of access to credit. But why would a loss of wealth make people want to work less? Why would less access to credit make people want to work less? Ultimately, the answer was that the loss of wealth and access to credit reduced consumer spending and investment spending.
So the problem wasn’t too little NGDP, it was too little consumption and investment spending? Okay . . .
On the left, economists were more sympathetic to the view that a shortfall of NGDP was the problem, but suggested that there was nothing more the Fed could do. (As if the Fed had run out of paper and green ink.) Actually, we now know there were lots more things the Fed could have done:
1. Don’t pay interest on bank reserves.
2. Do much, much, much more QE, buying unconventional assets if necessary to hit the target.
3. Switch to price level targeting, which would raise inflation expectations and lower long-term real interest rates. This policy would have meant the Fed would not have tapered in 2014, or raised interest rates in 2015. Indeed they never would have stopped QE1 or QE2.
The Fed’s recent move toward average inflation targeting is a tacit admission that they blew it in the 2010s with an inappropriately tight monetary policy that caused NGDP to grow too slowly, delaying recovery from the recession.
I’m not sure the recovery in the unemployment rate during the 2010s could have been 30 times faster, but it surely could have been 4 times faster. Indeed in the 18 months after December 1982, unemployment fell by 3.6 percentage points, which is 4 times faster than the 0.9 percentage point reduction in unemployment in the 18 months after the October 2009 peak. Money was clearly much too tight in 2009.
The market monetarist view of the Great Recession is looking increasingly persuasive, and if unemployment falls to 6% by November then Lars Christensen will be crowned king of the market monetarists.
PS. Because Lars made such an extreme contrarian prediction, let’s be generous and assume he meant the actual unemployment rate in November (announced in early December) would be below 6%, not the announced unemployment rate in November, which refers to October.
READER COMMENTS
Danny
Sep 9 2020 at 8:08pm
So what is your view now about how the initial and ongoing supply side disruptions will affect AD? A lot of small and even some large businesses are failing. That in and of itself does not necessarily indicate there will be insufficient AD but presumably the Fed will need to step on the gas. You have said recently that money is still at least slightly too tight. Do you think we will have a full or near full recovery once there is a vaccine or do you think the last unwinding of the unemployment rate will be slow because Fed policy will be too tight.
Scott Sumner
Sep 9 2020 at 10:51pm
It’s hard to know at this point, but I suspect that if there’s an effective vaccine the Fed will step up and we’ll have a fast recovery.
Derrick Miedaner
Sep 9 2020 at 8:23pm
I am continually amazed by all the brilliant people(and I mean this sincerely and not sarcastically) which continue to forecast in the middle of an extraordinary pandemic. I don’t want to call it hubris because I assume that there is some regimented methodology to their predictions, but the variety and number of variables implicated in GDP growth in these coming months is overwhelming.
Paradoxically, this thinking is what causes the fed to be reactive rather than proactive, which I don’t agree with. How does a centralized Fed take a basket full of regional economies which are experiencing disparate effects of Covid, and proactively set NGDP levels? Regional rates? Regional LIBOR rates? What possible cohesive strategy can there be in this environment?
The cat is certainly out of the bag with respect to QE, inflation, and targeted rates. I don’t know if I will see 10% inflation in my lifetime, as my parents have. The question remains then, if inflation isn’t the primary counterweight, what is? How extremely can the fed act?
Matthias Görgens
Sep 11 2020 at 9:02pm
That’s (part of) why Scott and Lars etc suggest to rely on market forecasts, instead of in-house models at eg the Fed.
Spencer B Hall
Sep 9 2020 at 10:28pm
Ben Bernanke not only destroyed the asset sides of the nonbanks’ balance sheets (turning RMBS and CMBS asset prices “upside down” or “underwater”, i.e., Bernanke was responsible for producing massive negative equity), he purged all of their short-term liabilities used for funding these impaired assets on the other side of the NBFI’s balance sheets (by paying a higher rate on interbank demand deposits than those whole-sale rates available on money market funding vehicles).
I.e., by 2011 the remuneration on IBDDs exceeded all short-term funding for clear up to two entire years out. I.e., with a policy rate that inverted the short-end segment of the retail and wholesale funding rates, shadow banking funding was either completely wiped out or the original funding spreads were made unprofitable.
It’s little wonder the housing market didn’t lead the economy out of the GFC.
https://www.corelogic.com/blog/2019/03/housing-recessions-and-recoveries.aspx
Spencer B Hall
Sep 9 2020 at 10:39pm
The FDIC’s insurance level is now @ $250,000. It was $100,000 prior to the GFC. At the onset of the GFC, the FDIC made transaction deposit insurance unlimited.
An increase in bank CDs adds nothing to GDP. Why? Because time deposit banking is an anachronism. The bank lending channel is a paradox, based on its’ balance of payments – which are all derivative deposits from a system’s viewpoint – not primary deposits.
Danielle Dimartino Booth’s in her book gets it backwards too: “Fed Up”, pg. 218
“Before the financial crisis, accounts were insured up to the first $100,000 by the FDIC. That limit kept enormous sums *in the shadow banking system* [sic].
How do you think savings products, new money substitutes, are generated? By activating savings. The only way to activate savings, put savings back to work, back into circulation, is for their owners, saver-holders, to invest/spend either directly or indirectly outside of the payment’s system.
This is the source of the pervasive error that characterizes the Keynesian economics – the Gurley-Shaw thesis.
Percentage of time (savings-investment type deposits) to transaction type deposits:
1939 ,,,,, 0.42
1949 ,,,,, 0.43
1959 ,,,,, 1.30
1969 ,,,,, 2.31
1979 ,,,,, 3.83
1989 ,,,,, 3.84
1999 ,,,,, 5.21
2009 ,,,,, 8.92 (the demand for money was immediately contractionary)
2018 ,,,,, 4.87 (declining mid-2016 with the increase in Vt)
robc
Sep 10 2020 at 4:37am
The “second wave” was a first wave in every state but LA.
There was no 2nd wave, there was slow propagation across the country.
Todd Ramsey
Sep 10 2020 at 9:18am
Even in Louisiana, the “second wave” was actually a first wave in a different part of the state.
robc
Sep 10 2020 at 3:08pm
Good point, even states are too large for aggregating data, but unfortunately most counties/cities are such small sample size the curves arent smooth and pretty.
Todd Kreider
Sep 10 2020 at 11:24am
I’m glad to scroll down to see people pointing out that not only was there no “huge second wave”, there hasn’t been any second wave yet.
The increase in Covid deaths in July was in part due to the virus hitting the south and in part because many more recorded death in July actually occurred in April and May. This was obvious to anyone tracking cases and deaths each day from late late June and early July.
After I saw unemployment drop from 15% to 13%, I predicted an Obama type situation for unemployment just before the election. Unemployment was still high in the summer of 2012 but very slowly coming down to 7.8%, an improvement enough to help him beat Romney. So I said 7.8% this October as well, which friends on the left thought was nuts. (I admitted that may be too optimistic, but I didn’t think by much.)
Scott notes how fast unemployment fell this year but doesn’t say how it shot up in an unprecedented way. He points out how slowly unemployment fell from a high of 10% in 2009 but doesn’t include that it took 18 months for it to climb from a low of 5% in April 2008 to a high of 10% in 2009.
Scott Sumner
Sep 10 2020 at 4:19pm
Yes, but at a national level there was a second wave.
Todd Kreider
Sep 10 2020 at 7:41pm
If you insist, but epidemiologists won’t say that. If there is a second wave, it will start in a few weeks.
Scott Sumner
Sep 11 2020 at 11:55am
It’s silly to argue about terminology. The fatality rate rose sharply after June, and that’s what matters for the purpose of this post, that’s what matters for the economy. The public couldn’t care less about how epidemiologists define terms.
Michael Sandifer
Sep 10 2020 at 6:12am
Yes, I was on Twitter when Christensen made his prediction and I agreed with him at the time, based on my model, but thought that unemployment might not get to 6% until 2021. I thought the fear regarding the virus might not allow the unemployment rate to fall quite so quickly, but we’ll see. Most commenters I read disagreed with Christensen, but the market monetarists and fellow travelers didn’t see his prediction as absurd, though perhaps bold.
Thomas Hutcheson
Sep 10 2020 at 8:10am
This is probably true but it would be hard to distinguish that view from the belief, say based on the way the Fed had allowed the TIPS to fall even further away from its supposed target, that there was little more the Fed WOULD do. The rather underwhelming reaction of the TIPS to the Fed’s announcement shows that the markets also doubt the Fed’s resolve treating it more like the promise of a dieter who plans to start next week.
It would be a shame if the pandemic market the end of RBC theorizing when what is needed is rather theorizing about how a real supply shock can trigger and interact with a demand shock and optimal policy response to it.
More generally, it seems that theorizing ought to move in the direction of how to respond both to supply and to demand shocks that are concentrated in some sectors more than others. This could have implications, I suspect, not only for monetary policy (which assets should it buy even if NGDP targeting remains valid as the higher level target) but for fiscal policy (beyond just creating automatic stabilizers like unemployment insurance and following the NPV rule) as well.
Matthias Görgens
Sep 11 2020 at 9:07pm
Why should it matter what assets the central bank buys?
bill
Sep 10 2020 at 9:06am
I’m glad you highlighted this prediction. I recall thinking it was too optimistic at the time. Very glad to be wrong!
What do you make of the ongoing very high rate of new UI claims? Is it a good sign in the sense that if the unemployment rate is falling in the face of such high job loss that we must also be creating new jobs at a very high rate? I wonder which sectors are absorbing these workers?
Scott Sumner
Sep 10 2020 at 4:21pm
I have no idea what to make of the new claims data; we are really in uncharted territory.
bill
Sep 10 2020 at 7:26pm
Thanks!
TMC
Sep 10 2020 at 1:01pm
The tone is also different than 2009, more anti-regulatory rather than regulatory. Remember the 1099s every business was going to be required for every purchase over $600? A lot of people were sitting on cash waiting to see what was going to come down the regulatory pike, whereas today, there is a lot more comfort (outside of retail).
Scott Sumner
Sep 10 2020 at 4:22pm
I see no evidence that there is less regulation than in 2009–maybe different types of regulation.
Bob Murphy
Sep 11 2020 at 12:30pm
I’m not a proponent of RBC itself–as an Austrian, my own theory is a monetary-disturbance-causes-real-effects guy–but this statement from Scott is a non sequitur:
“We are in the midst of a real business cycle. So this should be a feather in the cap for real business cycle theory, right? Actually, it’s looking more like the death knell of RBC theory. That’s because when we finally have an honest to God real business cycle, it looks utterly unlike anything we’ve ever seen before.”
The reason this particular RBC was so quick, is that the “real” disturbance–namely, government lockdowns and/or fear of coronavirus–was quick in its onset and exit.
But if an RBC theorist explains some earlier crisis, e.g. the housing boom/bust, by pointing out that workers sucked into building homes in Nevada and California need time to move and find other jobs, then the quickness of the 2020 recession is hardly an argument against them.
Going the other way, suppose an asteroid destroys Earth. That would be a “real recession” where GDP goes to zero, forever. We wouldn’t then say (from our vantage point on the space station before our air ran out), “Oh, because this definitive RBC cycle has a bust that’s lasted 12 years and counting, it means the 2020 recession must not have been an RBC one after all.”
Scott Sumner
Sep 12 2020 at 2:34pm
Bob, You said:
“But if an RBC theorist explains some earlier crisis, e.g. the housing boom/bust, by pointing out that workers sucked into building homes in Nevada and California need time to move and find other jobs, then the quickness of the 2020 recession is hardly an argument against them.”
But that’s not what happened in 2007. Housing construction fell by more than 50% between January 2006 and April 2008, and yet the unemployment rate stayed low. It was the decline of in NGDP during late 2008 that caused the severe recession.
The stronger “real” argument is that the banking crisis of late 2008 was a real shock that greatly worsened the recession. But then why was the recovery so slow?
Comments are closed.