Is there an official definition?
Journalist Clark Merrefield contacted me last week for a piece he was writing on the topic of recessions. His article appeared yesterday at The Journalist’s Resource. It’s titled “Are we in a recession? 4 things journalists should know when covering an economic downturn.”
In follow-on emails, Clark and I shared our pet peeves about sloppy journalist reporting on numbers. One of his is in #1 of his 4 tips: “Clarify the difference between quarterly changes and annual rates, especially in headlines.” That’s certainly high on my list.
While he and I agreed that the NBER’s Business Cycle Dating Committee is an important arbiter of whether a recession occurred, we disagreed, as you can see from his article, about whether its definition is the official one.
In response to his original question to me about how economists came up with the “two quarters in a row of negative GDP growth” definition, I wrote:
The answer is that I don’t know. When I taught macro, which I did as early as 1975 and as late as 2014, I would tell my students that the the “two consecutive quarters of negative growth” was economists’ “seat-of-the-pants” definition. But I don’t know where I got it. The “seat of the pants” is my term because I knew all along that it’s not the technical definition. Of course, there’s no official definition. The 8 economists with NBER are stating their opinions. They’re informed opinions but they’re opinions, nevertheless. They have no official status.
In his article referenced above, Clark writes:
It’s not a recession until the Business Cycle Dating Committee of the National Bureau of Economic Research says it is. The nonprofit economic research organization, based in Cambridge, Mass., determines when recessions start and end. Its purpose is to establish the historical record for recessions, not to rush to say whether or not the U.S. economy is currently in one. The White House calls the Business Cycle Dating Committee the “official recession scorekeeper.”
That’s too strong. Yes, it’s true that the committee “determines when recessions start and end.” But that doesn’t mean that their determination is official. Thus my statement above.
Moreover, there’s a tense problem in Clark’s statement of the issue. He writes: “It’s not a recession until the Business Cycle Dating Committee of the National Bureau of Economic Research says it is.” That cannot be true. As is well known, and as he points out, the NBER comes to its conclusion with a lag. Let’s say that it announces in December 2022 that a recession started in March 2022. To say that it’s not a recession until the NBER says it is is to say that it’s not a recession until December 2022. I think Clark meant to say “It’s not a recession unless the Business Cycle Dating Committee of the National Bureau of Economic Research says it is.” That’s a pretty different statement.
I think also that we shouldn’t get stuck in the weeds and should remember why so many economists use the “seat of the pants” definition of “two consecutive quarters of negative growth.” It’s because the NBER committee takes so long that its dating of recessions has limited usefulness. Economists want to know more quickly. Various economists have noted that, as economist Matt Rousu put it, “In the past 75 years, when the nation’s GDP has dropped in two successive quarters, it has been classified as a recession.”
Will this one be different? I don’t know. But I’ll probably know before the NBER makes its announcement.
Note: Here’s an excellent article on the issue from Phil Magness. It’s titled “Biden Borrows the Nixon Playbook on Recessions,” AIER, August 1, 2022.
And here’s Phil’s op/ed in the Wall Street Journal, “A Recession by Any Other Name,” WSJ, July 27, 2022 (July 28 print edition.) A key quote:
Economists have long defined a recession as “a period in which real GDP declines for at least two consecutive quarters,” to quote the popular economics textbook by Nobel laureates Paul Samuelson and William Nordhaus.
Update: Clark has changed the language to address my criticism and has done so with integrity, putting a note at the bottom explaining the change.
READER COMMENTS
JFA
Aug 3 2022 at 8:23am
“I think also that we shouldn’t get stuck in the weeds and should remember why so many economists use the “seat of the pants” definition of “two consecutive quarters of negative growth.” It’s because the NBER committee takes so long that its dating of recessions has limited usefulness. Economists want to know more quickly.”
I guess one could argue that saying a particular period also is of limited usefulness nothing is actually accomplished by labeling it so. I’ve never seen recession as the dependent variable or the independent variable in a regression. It’s always some actual measure. It’s only useful as shorthand when describing a particular point in time, but I wonder if the people who want the #officialdefinition of a recession to be “2 or more quarters of negative real GDP growth” whether they are achieving much economy in the language they use.
From a brief glance, it looks like the dating committee listed May 2009 as the last month of that recession, but unemployment kept going up through October. So I would say even there, there is limited utility in describing a period of time of non-negative real GDP growth and rising unemployment as “not a recession”.
I think the Biden administration is bobbing and weaving and not using the “2 quarters” definition for political reasons. The people who are saying that the administration is acting Orwellian for “changing the official definition” are also doing so for political reasons. Whatever definition you use, we should be worried about negative GDP growth.
John Hall
Aug 3 2022 at 8:30am
My recollection of my AP Economics textbook was that they defined a recession as two quarters of negative GDP, but I can’t recall if they intended that as a heuristic of not (i.e., usually as two quarters of negative GDP growth). I don’t recall them even bothering to discuss the subject in intermediate undergrad macro or my MA macro course.
Spencer Bradley Hall
Aug 3 2022 at 10:25am
The BEA should define a recession as negative real income growth.
Real Disposable Personal Income (A067RL1Q156SBEA) | FRED | St. Louis Fed (stlouisfed.org)
2020-07-01 -16.6
2020-10-01 -8.3
2021-01-01 54.7
2021-04-01 -29.1
2021-07-01 -4.1
2021-10-01 -4.5
2022-01-01 -7.8
2022-04-01 -0.5
Spencer Bradley Hall
Aug 3 2022 at 11:10am
Or Real personal income excluding current transfer receipts (W875RX1)
Real personal income excluding current transfer receipts (W875RX1) | FRED | St. Louis Fed (stlouisfed.org)
2021-10-01 14440.2
2021-11-01 14463.1
2021-12-01 14465.3
2022-01-01 14441.0
2022-02-01 14484.3
2022-03-01 14442.9
2022-04-01 14490.2
2022-05-01 14512.0
2022-06-01 14471.9
John hare
Aug 3 2022 at 11:49am
As far as I am concerned, a recession is when I don’t have a backlog of construction jobs lined up. The rest of the country may or may not match.
Vivian Darkbloom
Aug 3 2022 at 12:24pm
I think Henderson has written a pretty good summary of the issue.
As for ““It’s not a recession unless the Business Cycle Dating Committee of the National Bureau of Economic Research says it is”, I can certainly see why those in power (politicians, Treasury and the Fed) like this definition. It helps deflect current responsibility. By the time the NBER gets around to it, the recession is usually history. Nevertheless, I would prefer this version which, I think, better expresses the lagging feature: “It’s not a recession unless the Business Cycle Dating Committee of the National Bureau of Economic Resarch confirms it.”
Other than the political and psychological ramifications, are there any other practical consequences of whether something is called, “officially” or not, a “recession”? I can’t think of any.
Scott Sumner
Aug 3 2022 at 1:56pm
It’s simply false to suggest that economists have defined recessions as two consecutive negative quarters for RGDP. That’s always been a useful rule of thumb, with the NBER’s dating recognized as superior. I’d say at least 99% of economists recognize 2001 as a recession year. It’s recognized as a recession by the NBER (and probably by the general public), but didn’t have two consecutive negative quarters.
As for the first two quarters of 2022, it was one of the most overheated booms in US history. The job market was tighter than I’ve ever seen. To call it a recession is to drain the term of any coherent meaning. On the other hand, it’s quite plausible that we’ll be in recession later this year.
David Henderson
Aug 3 2022 at 7:21pm
You write:
That’s false. It’s quite easy to find macro textbook authors defining it that way.
Here, for example, is Economics, 10th edition (2003) by James D. Gwartney, Richard L. Stroup, Russell S. Sobel, and David A. Macpherson:
Here are Karl E. Case, Ray C. Fair, and Sharon M. Oster, Principles of Economics, 9th edition (2009):
Lauren
Aug 3 2022 at 9:51pm
So, Scott, David, just to move this question to a next step:
When did economics textbooks start to use a shorthand definition of two quarters of decline in GDP to define recession? What was the first textbook, or who was the first textbook author, to only say that a recession is defined by two quarters of decline in GDP?
I have many textbooks on my shelves from the 1960s, 1970s, 1980s, 1990s, and 2000s. They all primarily and clearly describe the NBER’s role in determining the breakpoints of recessions as being based on a plethora of criteria.
But, somewhere along the line, something switched; and some textbooks did start using only the shorthand proxy of two quarters’ worth of decline in GDP as an actual definition without even mentioning the NBER.
When did this happen in the textbooks? Which was the first textbook or textbook author that made that switch? That’s the interesting question.
When did teachers–college teachers–start teaching that shorthand definition of two quarters’ worth of decline in GDP? Did college teachers start teaching that definition before or after it first appeared in the textbooks?
David Henderson
Aug 3 2022 at 10:41pm
Good questions, Lauren. They’re similar to the question that Clark Merrefield asked me. I don’t know the answers.
Robert Lee Coffey
Aug 4 2022 at 9:59am
This doesn’t answer your question, but it seems that many other countries use the “two quarters” rule, what with not having the NBER around.
So which came first?
Recessions existed before NBER, so how were they defined then?
Scott Sumner
Aug 4 2022 at 1:27pm
There is no official definition in the sense of “government” definition. But the NBER definition is generally accepted by economists. I don’t know of a single case (since WWII) where NBER got it wrong. We either clearly had a recession, or clearly did not. The US hasn’t had any borderline recessions since 1946. People are making this too complicated.
Prior to WWII, other terms were used (depression, slump, etc), and economists looked at variables like commodity prices, steel output and railroad shipments.
Scott Sumner
Aug 4 2022 at 1:23pm
Textbooks make mistakes. I don’t agree with the premise that the economics profession has ever defined recessions as two straight negative quarters. It’s always been just a rule of thumb. Notice that G&S mention rising unemployment. That’s always been a necessary condition. It’s why 2001 was a recession despite not having two straight quarters of falling GDP.
As for Case and Fair, the term “conventionally” is very vague. Is that a rule of thumb?
The NBER definition is too complicated for students, so textbook writers take shortcuts.
Can you find any textbook that says 2001 was not a recession?
Richard A.
Aug 3 2022 at 3:00pm
GDP should equal GDI. Real GDI came in at +1.8% in the first quarter of 2022. The GDI figure suggests that we are not in a recession.
Take a look at the post, The Ratio of GDI to GDP, at Ecobrowser.
Phyecon
Aug 3 2022 at 10:24pm
The reference to scorekeeper is apt, but in a more nuanced way. In baseball there exist events that are officially defined and matter for the real game, e.g. ball, strike, run, out. There are other events that are “officially” defined, but do not matter for the real game, e.g. hit, error, quality start, winning/losing pitcher. A Recession is much more like the latter. There can be an accepted simple/objective definition such as two quarters of negative RGDP change, but really it does not matter really. Just as wins/losses for a starting pitcher can be misleading so too can the label recession.
In other contexts journalists used the term “jobless recovery” and I have suggested calling our current state a “jobfull recession”
Thomas Lee Hutcheson
Aug 3 2022 at 10:43pm
Why would anyone want an “official” definition of “recession?” What decision should be made differently at time x if at time x a “recession” is declared, or even less if it is declared that a recession began at x-n?
Mark Z
Aug 4 2022 at 12:39am
I’m inclined to repudiate any definition that outsources itself to a committee. That’s not really how definitions work. The ‘two quarters of declining RGDP’ may be a bad definition by some standard, but the alternative should be something like some specified sustained increase in unemployment over some specified period of time. That way whether a definition applies to an event isn’t dependent on the arbitrariness of who happens to sit on the relevant committee at the moment.
Michael
Aug 4 2022 at 7:13am
The two quarter “rule” likely came from the observation that that rule has tracked past recessions well. The problem is, two negative quarters isn’t the essence of what makes a recession (at least to me), so in weird economic times it might fail. Like all rules of thumb, it’s not perfect.
I think you have the right idea about what a better definition or rule of thumb would be. An increase in unemployment beyond a specific threshold within a specific amount of time. That’s the approach adopted by former Fed economist Claudia Sahm, and I think it is preferable because it gets to the heart of what a recession is: a self-perpatuating spike in unemployment.
Vivian Darkbloom
Aug 4 2022 at 9:23am
Robert Barro has an interesting article at Project Syndicate in which he notes that there have been no “false positive” indications of “recession” under the two-consecutive-quarter rule when compared with subsequent opinions of the NBER:
https://www.project-syndicate.org/commentary/two-consecutive-quarters-negative-us-growth-predict-recession-since-1948-by-robert-j-barro-2022-07
Scott Sumner
Aug 4 2022 at 1:31pm
Yes, it’s a good rule of thumb, albeit not useful in early 2022. But it’s not the definition.
Michael
Aug 5 2022 at 7:05am
I’m amazed at how many people who should know better not understanding the difference between a “rule of thumb” and a precise definition.
vince
Aug 4 2022 at 10:26am
Wikipedia seems to identify the origin of the two-quarter rule. It quotes a NYT article written in 1974 by Julius Shiskin, Commissioner of Labor Statistics. He offered a “rough translation of the bureau’s qualitative definition of a recession into a quantitative one that almost anyone can use.”
He actually provided several alternatives, the easiest being two consecutive quarters of real GDP decline. He called them useful guidelines, especially “for those who are not experts in business cycle analysis.”
Here’s a related question: What’s a depression?
Bob Bell
Aug 4 2022 at 3:17pm
Many good comments, and an excellent, informative article. From a layman’s perspective (that means I don’t know anything) it seems the issue boils down to the timing. The NBER “pronouncement” is probably the most accurate but is such a lagging indicator that the recession could be over before you knew you were in one. For people, especially politicians, who believe the government can “fine tune” the economy that’s not much use. It’s like having a driver assist system in your car that announces that you’ve driven over a cliff instead of announcing that you’re about to drive over a cliff. And as many commenters have pointed out, while the 2 quarter rule of thumb is not official (of course, neither is NBER) it’s been pretty darn accurate historically.
Scott Sumner
Aug 5 2022 at 5:07pm
“it’s been pretty darn accurate historically.”
Yep, but not this time. It’s almost comically inaccurate this time. Imagine if the Fed thought we were in a recession and tried to “stimulate” the economy. Thank God they know better.
Michael
Aug 5 2022 at 8:38am
The economy added 528,000 new jobs in July, crushing the expected add of ~250,000.
Unemplyment fell to 3.5%, just 0.1% off of a 50-year low. Basically back at prependemic levels.
Slight upward revision to the June jobs number, too.
This is not anyones version of a healthy economy (inflation), but I don;t think there is any sensible definition of recession that includes the unemployment rate being at/near a 50-year low.
diz
Aug 9 2022 at 1:30pm
Without a formal definition this discussion will seem to inevitably devolve into semantics and opinion mongering.
I find myself growing more unimpressed with the opinions of experts over time. I can have my own opinion of which movie is the “Best Picture” and which college football player is the best. Why should I care what the Academy or the Heisman selection committee says? And don’t get me started about those scientists with their “Doomsday Clock”.
Comments are closed.