September 29, 2010
When do recessions end?
Warren Buffett thinks the U.S. is still in a recession, declaring in a CNBC interview last week:
I think we're in a recession until real per capita GDP gets back up to where it was before. That is not the way the National Bureau of Economic Research measures it. But I will tell you that to any-- on any common-sense definition, the average American is below where he was before, or his family, in terms of real income, GDP.
I don't presume to be able to tell Warren Buffett what investment strategies work best. But I can provide some clarity on how economists use the term "recession," and hopefully shed some light on the issue that Buffett and others have raised.
At least since the work of Wesley Mitchell nearly a century ago, economists have characterized particular historical episodes in terms of the dates at which economic activity reached a peak before entering a period of decline. The low point on the way down is designated as a trough, and the episode between the peak and the trough is described as an economic recession. Here is a graph of Buffett's preferred measure, real GDP per person. It is plotted on a logarithmic scale, so that a move up or down of 10 units on the vertical axis corresponds approximately to a 10% change in real GDP per capita. The NBER dates for business cycle peaks and troughs are indicated with vertical lines on the graph.
According to the traditional chronology, the recession ends when the economy starts growing again, not when it has grown so much that indicators such as real GDP per person are back to making new all-time highs. As a matter of arithmetic, the latter standard would always give you a later date than the true low point for economic activity, and a substantially later date in the case of a downturn as deep and recovery as sluggish as the one we've just experienced.
Economists have used the term "recession" to have this particular meaning-- the episode between a peak and trough-- for 150 years or so of data. Are Warren Buffett and the common-sense Americans on whose behalf he claims to speak within their rights to suggest that we have been using the term inappropriately? I would say that they're very much justified in having an opinion on this, because the events that economists label as recessions are universally understood to affect everyone's lives. So it's quite appropriate that everyone feels entitled to talk about what the word "recession" should mean on the basis of personal and direct experience. People know they're still not back to where they were, and the data confirm that. But when economists say that the recession is over, there never was a claim that things were back to normal.
So in part it's simply a semantic issue of how a particular term is defined. Economists use "recession" to refer to a narrowly defined event, whereas many people appropriately want to be able to talk instead about the consequences that linger as a painful aftermath of the event.
But I think there's more than just a semantic issue involved here. The reason that economists say that the recession, as we use the term, has been over for more than a year is that conditions have been steadily improving rather than getting worse. True, conditions haven't improved as much as we'd hoped and expected, and they haven't improved enough to put us back where we had been before the recession. But nonetheless, conditions have been improving rather than getting worse for the last 15 months.
And that is an important statement of fact that can easily get missed in the popular discussion of these issues.
Posted by James Hamilton at September 29, 2010 07:15 AMdigg this | reddit
Thanks for this post, and your previous one about the NBER's methodology. On considering this, I realize that lay people like me use a "level based" concept of recession. Loosely, it may be phrased as "things are far worse than we've seen they could be". More formally, I think non-economist think of a recession as a time when the economy is still X% below its peak, regardless of whether it is edging upward from its trough.
So, it is a level-based concept versus a change-based concept.
Of course, I suppose that people can get used to a "new normal", if that were to take hold. So, a better metric for a level-based concept may be along the lines of: "things are X% below the most recent (last 5 or 10 years) peak.
Posted by: Realist Theorist at September 29, 2010 07:37 AM
JDH: nice post. Goes well with an iced latte and Mark Thoma's post here.
Posted by: Robert Bell at September 29, 2010 08:43 AM
Here is what is disturbing, outside of the specific definitions. I saw another log GDP graph recently, and yours shows it, too. The point of plotting the log of anything is to get a straight line.
But, starting around 1980, there is a slope change. Connecting peaks since then leads to a steadily declining slope.
You can put a straight line through log GDP back to 1800 and not see anything like this.
The U.S economy is slowly dying - Since the Reagan presidency.
Posted by: jazzbumpa at September 29, 2010 08:46 AM
JZB, really? since 1800. When did the U.S start using GDP anyways? about 1950. Before that, GNP. I know, you can approximate what it could have been, but now we're talking about approximations of approximations. Also, I wonder how reliable measuring methodologies have been over time. Maybe talk about that before making economics into politics. But then again, that is the flaw of economics. We economists need to treat it more as a science, so that it can get the respect that it deserves.
Posted by: Russ at September 29, 2010 09:24 AM
Did austerity get us out of the recession or did fiscal stimulus?
It is always about politics.
Posted by: Adam at September 29, 2010 09:48 AM
Posted by: Adam at September 29, 2010 10:16 AM
I understand and agree with your points, James. They are what I have to teach in my classes also.
But I think there's a valid criticism of economics & theory buried in the comments such as Buffet's. Specifically, we define "recession", but we don't define "recovery". We assume that any economy that doesn't fit the "declining" definition of recession, must therefore be in "recovery". Our theory ignores the possibility that a third (or more) situation may exist: an economy that is neither growing significantly nor declining. We have no widely used term for a side-ways economy or for a consistent GDP growth rate that is less than population or workforce growth rate. Such a situation is not "recession", but no realistic person can call it "recovery" since it will never lead back to robustness.
Posted by: Jim Luke at September 29, 2010 10:40 AM
Maybe a good word is Stagnation. Low inflation, relatively high unemployment, and low growth.
Posted by: Adam at September 29, 2010 11:17 AM
Jazzbumpa: Readers can clearly deduce you have something against Reagan. May I respectfully say your argument does not follow from the data. My hope is that you and others will learn something here.
Print off the graph and use a ruler. From the cycle peak in 1981 draw a straight line to the cycle peak in 2001. Draw another from the first cycle peak in the '50s to the '81 peak. You will prove to yourself that in the two decades after Reagan the economy grew faster than in the three decades before. This despite the price of the severe '82 recession Reagan had to pay to break the back of double-digit inflation, as well as the budgetary defense cost that bankrupted the Soviet Union and won the Cold War.
Any business cycle expert will tell you it was Reagan's free enterprise policies that grabbed hold and drove the economy upward once inflation had been broken by the '82 recession and the economy was freed of the stagflationary ball and chain of the '70s attributable to his four immediate predecessors and the various Congresses.
Or draw two lines, one from the first peak in the 50's to the peak in '81, and the second from that peak to the final peak in 2007. Their slopes (growth rates) are insignificantly different, though the complete era of Reagan and after does exhibit a smidgen less growth.
For the real story, I'd recommend you go from cycle peak to cycle peak and understand why the economy did what it did in each piecewise segment. A big chore to be sure, but when you're done you'll really know something others don't. Take the final segment -- the dismal growth from 2000 to 2007. Arguably much of the lost momentum is because of the offshoring of production to China and emerging Asia which started ramping up exponentially after the Berlin Wall went down and China soon after got access to our markets.
Of the five last presidents, Obama inherited the worst economy. Reagan inherited the second worst. George W. Bush the third worst. And his father inherited the best. It matters greatly what sort of economy these men inherited, what their policies then did to better things, and the various Congresses they had to work with. Obama's party has controlled both Houses thus far. But during Reagan's two terms, Democrats controlled the House throughout. And they certainly were not as pro-growth as he was ... by a long-shot.
Posted by: Anonymous at September 29, 2010 11:22 AM
Maybe the good folks at Measuring Worth just make stuff up. I dunno.
And there's another inflection point around 1850. Probably reflecting the industrial revolution.
Anyway, politics and economics are joined at the hip, and always have been.
One can construct a reasonable narrative around a slope change circa 1980, based on a fundamentally different approach to economic governance vis-a-vis FRD to Carter. Undoing steeply progressive income taxation, reduction in capital gains tax, deregulation, weakening of unions, the resulting great increase in wealth redistribution up, decline of the middle class, increase in depth and breadth of poverty.
The period from 1900 to WW I was almost horizontal on a log scale. Robber Baron era.
Capitalism needs consumption, and that needs a vibrant middle class that isn't tapping bubbles or borrowing to meet normal living expenses. That's how it looks to me, anyway. YMMV.
Posted by: jazzbumpa at September 29, 2010 11:47 AM
Unfortunately, the "improvement" is based upon an unsustainable level of public borrowing, just as the previous apex was based upon an unsustainable level of private sector borrowing.
The problem with directly printing credit is that it prevents citizens from buying 100% of what they produce out of net income. Printing effectively confiscates goods, and lends them back to the citizens who produced them. This creates a class of unrepayable debt, which gradually accumulates until rising defaults shake the credit system.
If citizens were issued a rebate check when printing took place, prices would go up, and there would be no net change in debt. If the Fed simply gave newly printed money to the treasury, the treasury would spend it with no net change of debt. In effect, printing would be converted into a highly regressive tax.
Printing credit leads to an ever increasing level of unrepayable debt, until defaults shake the credit system. This is how every major country on earth can simultaneously wind up with too much debt to repay, even net exporters.
Think of the magic money elves that double the money in everyone's wallet overnight. Prices double, but citizens still buy the same goods as they did before with their cash. If instead, the magic money elves put a credit card in everyone's wallet, prices would still double. Citizens could buy half of the goods they did before out of cash, and would have to put the rest on the magic credit card. Now the citizens owe the elves half of what they produced, and can't repay it.
Posted by: Ekim at September 29, 2010 12:19 PM
This is really a problem of economists. If you want to reserve a definition for a term that others will easily confuse, you must also create a terminology for others to use for what is important to them. Due to this deficiency, I suggest economists abandon the term to common usage and create a new one and refrain from using it in ordinary speech.
Posted by: Lord at September 29, 2010 12:28 PM
A recession means the economy is going backwards. The timing of the recession end just says when the economy stopped going backwards. But Buffett's point is that saying that the recession is over doesn't match with the real decline in living standards that continues after the recession ends.
Is there an economics term or national committee dedicated to examining when the economy reaches the point when the economy reaches a state that was better than any time before? The NBER dates the peaks before the recession but doesn't identify the recovery dates to those pre-recession peaks, does it?
Posted by: uber_snotling at September 29, 2010 12:45 PM
Nice picture, but it doesn't go back far enough. The last like experience to the current malaise (household balance sheets way out of balance) was GD1. This graph is just another version of the econometric forecasts based on data only from the post-war experience - a stastical non-starter IMHO, more misleading than informative.
As for Buffet's point, it may go deeper than merely saying we have not reached the old peak. He may be saying we will go down again before reaching the old peak. That is, he may be trying to say that the malaise has not been cured - we are still sick and not even true recovery mode yet. I, for one, would agree with that. More specifically, I would agree with saying that if a second dip comes before reaching the old peak that we are still in the same recession, not a new one. Just seems a better use fo the language.
Posted by: don at September 29, 2010 12:51 PM
I guess, Mr Buffet was not talking as an economist but with politics on his mind.
From what I have read about psychology I remember that people are not much happier, or they were before this crisis, than they were in e.g. 1950s when real GDP per capita was several times smaller. And that people tend to be happier when things are improving rather than staying still, no matter what the starting point in your bank account is.
From this point of view the classical and non-fuzzy peak-trough economists' definition of recession would sound more appropriate than the back-to-where-we-were definition.
The catch is that the recession was not long enough to distort people's memories on how they were doing before it started and they tend to compare with those times.
But if you imagine a young person who entered his/her adult life after the onset of the recession, he would likely see things as improving and rather shiny.
Posted by: John at September 29, 2010 01:16 PM
Further adjust the data by looking at private nominal GDP (total GDP less total gov't spending, including transfers) per capita; one cannot possibly argue on this basis that the economy is "recovering".
Then, to further understand why there is no private sector "recovery" occurring, and there cannot be one, subtract private health care spending from private nominal GDP to see what has occurred to private per-capita GDP after expenditures for private health care spending.
I dare say that most readers here will be not a little bit surprised, if not shocked, to find out what has happened to private GDP per capita after the growth of health care spending since the 1980s.
And many e-CON-omists argue that we need to continue to grow health ("disease") care spending at 7-8%/yr., a bankrupt notion that will result in the same for the country.
Finally, to the picture add the 25- to 40-year decline in per-capita US oil production, and it is no surprise why the US has deindustrialized, decapitalized, financialized, and feminized the economy over the period.
In effect, the US has been getting older, sicker, and poorer in terms of energy, incomes for the bottom 80-90% of households, production, and capital accumulation per capita for 20-25 years or more.
No Keynesian, supply-side, socialist, imperial, or "free trade" policies can pull us out of this with Peak Oil, population overshoot, and increasing risk of a breakdown of trade and diplomatic relations with China hereafter.
Posted by: Nemesis at September 29, 2010 01:43 PM
John: "But if you imagine a young person who entered his/her adult life after the onset of the recession, he would likely see things as improving and rather shiny."
I doubt it. I think he or she would be very worried about gettig a job and repaying loans for an education gained under very different expectations about what would happen on graduation day.
Posted by: don at September 29, 2010 03:29 PM
I'm unable to locate a table of quarterly real GDP per capita data. BEA has real GDP, but not per cap, that I can find. Could you tell us where you got it?
Posted by: jazzbumpa at September 29, 2010 03:34 PM
jazzbumpa: Both real GDP and population are available from FRED. I converted monthly population to quarterly by averaging the three months of the quarter. I multiplied GDP (in billions of 2005 dollars) by 1,000,000 to get GDP in thousands of 2005 dollars, so that the ratio of GDP in thousands of dollars to population in thousands of people is GDP in dollars per person. Take natural logs and multiply by 100 to get the series plotted.
Posted by: JDH at September 29, 2010 04:44 PM
The layperson finds a technical term interesting, begins to use it, and often uses it badly. Then, common usage evolves differently from the technical usage. If technical usage changes, we become unable to read past research without checking publication dates and consulting old references on common usage.
In this context, why are we discussing real GDP per capita instead of real GDP. The official definition makes no reference to per capita. Using per capita GDP would have recessions starting earlier and ending later based on positive US population growth. Under the Buffet definition, we would be in recession if real GDP growth dropped below population growth.
Simply define "economic recovery" as the period between the end of recession and the point at which we return to the previous peak of real GDP. Buffet will still argue because it is not per capita, but "economic recovery" should be consistent with "recession".
Posted by: Rick M at September 29, 2010 04:46 PM
Rick M: I use GDP per capita because it is the measure proposed by Buffett. My point is that he reaches his conclusion not because he uses this measure, but instead because he is simply thinking about recessions in a different way.
Posted by: JDH at September 29, 2010 06:01 PM
Excellent. Thank you.
Posted by: jazzbumpa at September 29, 2010 07:00 PM
What if the expansion phase after reaching the first low point ends before the previous peak has been reached; and the next contraction phase is longer lasting and has a deeper trough than the previous one. So the expansion is just an intermediate one within a longer-term contraction event. Is the second contraction phase a new recession or still the old one? What do economists say about such a case? How do they name it? A depression?
Posted by: rootless at September 30, 2010 03:13 AM
I have no complaint with people trying to think things through, but we ought to keep in mind here that we are arguing over definitions of a label. There can be preferences, but no objectively right answer, in such cases. Any given label will hide information, as well as reveal it. NBER's methodology cannot be "wrong", but may hide more than it reveals, or reveal less than some other methodology.
I'm tempted to conclude with "who cares?", but will resist the temptation. When it comes to politics, we are going to use words like "recession", and that makes arguments over definition unavoidable. In the end, though, policy will be driven by judgment - political as well as economic - and labels are just labels, and labels don't improve judgment.
Posted by: kharris at September 30, 2010 04:41 AM
Words are extremely powerful things. That’s because the day-to-day language we use affects our subconscious mind, and the beliefs that reside within the subconscious. These subconscious beliefs, in turn, shape and determine subconscious strategies that dictate the vast part of our behavior as humans. This is not to deny the existence of consciously formed strategies and resultant actions, of course. Think iceberg as a rough analogy.
Now the idea of carefully choosing one’s words scales up from the individual to the larger society. Words publically spoken in today’s media – especially such emotionally laden words as recession – reach many individuals and affect both social consciousness and the collective social unconscious.
In his CNBC comments, Mr. Buffet says he defines current economic conditions as recession because per capita real GDP has not yet gotten back to where it was before. If you are the type of person who puts weight on what Mr. Buffet says because he is considered an expert, and you then believe we are still in a recession, that very belief will have an impact on your actions in an ongoing sense. Multiply this by the number of people across the country who are so swayed, and you get some idea of the magnitude of the effect on overall societal actions. Words are powerful and they ramify. Hence clarity and truth should be strived for.
As for this still being a recession, it is not. The recession ended in summer 2009. That was evident to knowledgeable economists at the time. Recession derives from the Latin recedere – to recede. It denotes withdrawing or shrinking, which is exactly what happens when the level of GDP falls. Economists named it well. Recovery is the stage the economy is in now. Recovery comes from the Latin recuperare – to recuperate, to regain. The recession ended and the recovery is now underway. When the doctor says the fever has broken and you are recovering, that news has a wonderful psychological effect that transmutes into a physiological one that accelerates the healing. Words are powerful, and brilliant economists in the past gave us just the right words to describe the important processes of the business cycle: recession, recovery, and expansion. Given the (asymmetrical) wave-like nature of the economic process over time, peak and trough are apt as well.
Posted by: JBH at September 30, 2010 05:30 AM
You're right that it is a semantic issue, but the confusion here is that the pure economic term deosn't suficiently capture the concept that makes sense to most people, because of the simple assumption that recession means bad and no recession means good.
Imagine a huge decline in GDP of perhaps 10%. that would viewed as terrible and be especially bad on the way down. Now imagine that the decline stops and economic "growth" resumes at an annual rate of 0.1%. No matter how pure you want your economic terms to be, everyone will still feel like the situation hasn't changed much (and it really hasn't).
Another issue is that NBER decides, so either: 1) they are using some defined formula that anyone else can use and we don't need them to declare a recesion is over; or 2) they are applying subjective assessments and there's no way to directly compare one recession to another. Which is it?
I agree with Warren about redefining the terminology, but I would go further. In terms of calling a double dip, I would count it as a double dip only GDP had returned to the previous high adjusted for the intervening population increase.
Posted by: MGKurilla at September 30, 2010 06:43 AM
If you use the real per capita GDP data at measuring worth and calculate trend growth rates there are two breaks. One prior to 1850 so that the trend growth rate from 1790 to 1850 is 0.9%.
The second trend is from 1850 to 1950 when trend growth is 1.6%.
The third trend is since 1950 when growth is 2.1%.
Basically the trend growth around 1850 stems from the birth of modern capitalism with the limited liability corporation becoming the driving institutional force behind growth.
The 1950 break represents the introduction of the modern mixed economy with large scale central government and the Fed actively attempting to manage growth.
The measuring worth data may not be the best data since the recent research showing that maybe the old data showed more volatility than was actually the case. But this should not impact the long term trends as both the more recent data and the older data depend heavily on information collected by the census every decade. So essentially the old and new data agree about the economy once every decade and just disagree about how volatile the economy was between the census based benchmarks.
Posted by: spencer at September 30, 2010 07:00 AM
Great post professor. Buffet's value based investing made him a lot of money but that doesn't mean he is a very good economist. I always laugh when the talking heads quote Buffet as if he were the economic god.
Posted by: Ricardo at September 30, 2010 07:17 AM
Is it really any surprise that it's costing more money to keep an aging obese population alive to 85?
Posted by: Anonymous at September 30, 2010 07:23 AM
Warren never said that he sees himself as an economist, James. One might think : That is why Warren is so succesful in business ..
But Warren said :
“The way the tax system has gotten tilted toward guys like me over the last 20 years is -- as opposed to the middle class, you know, in my view, is a little obscene,” he said. “I just think that -- when a country needs more income and we do . . . they should get it from the people that have it.”
Warren is a nice guy, James. He is a honest crook and likes GS shares ..
Posted by: Johannes at September 30, 2010 10:14 AM
So, what's the word for what Warren Buffett's talking about? Coz, it seems like an important concept.
Posted by: PG at September 30, 2010 11:38 AM
MGKurilla is right on point when describing the popular understanding of the word recession. In the popular, lay understanding, given an operational economic definition by Buffett, the economy is in a recession when the economy is bad, and in a boom when the economy is good.
The meaning of words is mutable. For example, the accepted medical term for a person of disabling low intelligence (e.g. "idiot" and "imbicle") is changed every decade or two because the old term inevitably become derogatory. The words "Negro" and "Colored Person" have developed new negative connotations. The gender inclusive sense of the masculine words ("he" v. "he or she," or "mankind" v. "humanity") goes out of fashion. "Cool" comes to mean not just low in temperature but something that is interesting or fascinating. "Cheap" comes to mean not just low in cost, but low in quality or miserly. The word "Execution" whose original sense is to carry out an act, for example by causing a court judgment to take effect or signing a document, has come to have the more narrow sense in the popular sense of carrying out a death penalty judgment.
The term "recession" is not immune that kind of linguistic shift.
One word that could be appropriated for Warren Buffett's concept, although this would itself be a mutation of the meaning of the word, would be "Downturn." A "Downturn" could be defined to mean a period when per capita GDP is below a previous peak, while an "Upturn" could be defined to mean a period when per capita GDP is above all previous peaks.
Note that one can never be in an upturn, by this definition, unless one is also in a period that is not a recession, and that one is always in a downturn, by this definition, when one is in a recession. But, by this definition, one can be in a downturn while simultaneously being in a "recovery" as economists use the term. (Indeed, presumably, one is no longer in a "recovery" when one has exceeded all previous GDP peaks - "recovery" itself is a word that sometimes means "improving from a setback" and sometimes means "successfully having been completely restored from a setback").
Our problem right now is that the recession was particularly deep, yet the recovery has not been particularly rapid. This is hardly a surprise given the recession. It was the longest since the Great Depression as economists define recessions and the length of a recovery is correlated strongly with the length of a recession.
Posted by: ohwilleke at October 1, 2010 09:51 AM
I think slump would be a good term for it. I would suggest panic for the decline but that may be more precipitous than what ordinarily occurs.
Posted by: Lord at October 1, 2010 10:40 AM
"Recovery is the stage the economy is in now. Recovery comes from the Latin recuperare – to recuperate, to regain. The recession ended and the recovery is now underway. When the doctor says the fever has broken and you are recovering, that news has a wonderful psychological effect that transmutes into a physiological one that accelerates the healing."
The analogy you use is telling, as well ("the fever has broken"). Unfortunately, I think the patient is just in temporary remission, leading to some foolish equity bets right now, as well as giving policy makers a false sense of well-being. My guess - we better accept the Fed's inability to help in this malaise, and recognize the need to stem the current account leakage, before it is too late.
Posted by: don at October 1, 2010 01:29 PM
A mea culpa and a couple of thank you's.
Somehow I thought this was Menzie's post, but I see it's James. Sorry for the error, and thanks, James, for the links to get the data.
Anon @ 9/29, 11:22 - I strive diligently to be data driven. That's why I wanted access to the raw data. And connecting peaks is exactly how I want to go at it. Just eye-balling what I see here, the 60's were the best, by a lot. The idea of what a Prez inherits cuts both ways - the things that follow busts are booms - so Reagan's results - boosted by recession recovery and his own fiscal irresponsibility - are less impressive than you suggest.
The big point is that with the passage of time, the peak to peak slopes are clearly in decline.
Posted by: jazzbumpa at October 1, 2010 03:27 PM
"The National Institute interprets the term “recession” to mean a period when output is
falling or receding, while “depression” is a period when output is depressed below its
previous peak. Thus, unless output turns down again, the recession is over, while the
period of depression is likely to continue for some time. We do not expect output to pass
its peak in early 2008 until 2012."
Real economists properly define their terms. Real economists can also identify when a person wants to use the word depression on national television, but cannot, due to psychological reasoning.
Posted by: Alexander at October 1, 2010 05:09 PM
Economies have been growing more or less continuously since before the industrial revolution. The NBER's terms simply ignore this fact -- they assume that the "business cycle" has no long-term trend. If there's no long term trend, then it does make sense to define recession and expansion the way they do. But to build a terminological taxonomy that ignores such an overwhelming factor as long-term growth is simply wrongheaded and stupid.
Warren Buffet is neither wrongheaded nor stupid, and neither are most people who have lived longer than the duration of a single cycle. Politicians are often both, and economists may not be stupid, but their "publish or perish" intellectual economy apparently allows them to be wrongheaded quite successfully.
Posted by: Dean at October 2, 2010 12:05 PM
I think people need to think in terms of three terms. Recession, which is the time that we are headed down into the valley, recovery, the time when we are going back up the next hill, but have not reached the same altitude as the previous peak, and expansion, which is the period when we are setting new records for things like real GDP per capital and industrial production. Today it is pretty clear we are in a recovery, but a long way from being in an expansion.
Posted by: Dirk at October 4, 2010 02:09 PM
All I can really say is that I'll be so incredibly glad when America starts to talk about something OTHER than the fact that we're in a recession.
Posted by: Josh at October 4, 2010 08:01 PM