« Assorted links | Main | The Next Crisis »
It's official.
It began (pdf) in December 2007.
Posted by Alex Tabarrok on December 1, 2008 at 04:07 PM in Economics | Permalink
Comments
Recession started in Dec, with two quarters of growth in between? WTF?, Uh, I mean, a bit counter-intuitive I believe.
Posted by: Tom at Dec 1, 2008 4:24:18 PM
It is fascinating to watch all the so-called economists line up to defend Keynes by defending FDR,
when the General Theory wasn't published until 1936, and when Keynes met FDR he regarded him
as ignorant.
This debate is truly becoming unhinged-I didn't think hagiography was economics, but I was wrong.
Posted by: superheater@live.com at Dec 1, 2008 4:24:28 PM
Recession started in Dec, with two quarters of growth in between? WTF?, Uh, I mean, a bit counter-intuitive I believe.For all intensive purposes (see unemployment statistics), 0.5% annual GDP growth is a contraction.
Posted by: Daniel Reeves at Dec 1, 2008 5:07:09 PM
It's interesting to look at which recessions have been extra long.
Look for example at the Panic of 1873. The reaction to that crisis was protectionism, which initiated the so called "Long Depression" which ended in 1896 and covers six business cycles. The first contraction is a whopping 65 month long, and the second one 38, which comes in as third in the "whopping contractions" after the great depressions 43 month downturn.
Clearly, protectionism definitely does NOT help. :) To get out of the current crisis in double time, governments should agree to open the borders, right now.
Posted by: Lennart Regebro at Dec 1, 2008 5:12:03 PM
The phrase is "to all intents and purposes," and it doesn't mean what I suspect you think it means.
Posted by: obnoxious pedant at Dec 1, 2008 5:42:44 PM
@ Obnoxious: Good call...I was thinking the exact same thing. This guy doesn't know how to use the phrase that he gets wrong. Doy?
Posted by: D at Dec 1, 2008 8:07:05 PM
It began in December 2007.
Doesn't it say that December 2007 was the peak month, and that the economy was expanding then? Wouldn't that mean it began in January 2008? Or am I reading it wrong?
Posted by: John Thacker at Dec 1, 2008 8:45:09 PM
I have what is probably a stupid question for you economists. Apparently the DOW dropped nearly 700 today, and they are blaming it on this report. What I don't understand is why an official declaration mattered so much in this regard. Investors had access to all sorts of resources that have been all but declaring this for months now. So why this 'sudden' reaction now, when it should have been plain to them the whole time?
Posted by: Leif at Dec 1, 2008 8:49:17 PM
Yes, the phrase is "for all intents and purposes." I do like the idea of "intensive purposes", though.
Posted by: David Wright at Dec 1, 2008 8:59:17 PM
"Doesn't it say that December 2007 was the peak month, and that the economy was expanding then? Wouldn't that mean it began in January 2008? Or am I reading it wrong?"
I think "peak" means that December 2007 was the inflection point.
Posted by: Jacqueline at Dec 1, 2008 9:03:07 PM
The "REASON" that the stock market averages increase or decrease are merely journalists expressing an opinion. Markets move and there is no definitive reason. If a movement occurs at the same time something else happens -- that frequently becomes the reason.
Maybe it was India or maybe it was Black Friday or the Walmart incident or maybe people had the long weekend to figure out that they want out. Or they got pushed out via fund redemptions.
Posted by: WTF at Dec 1, 2008 9:43:35 PM
If you read their explanation, I really don't think they are taking "peak" to mean "inflection point". Take the example of payroll employment - they clearly state that this series "peaked" in Dec 07, not that the *growth* of the series peaked in Dec 07.
Tom, you should read at least the first 2-3 pages of the pdf. They do explain WTF their reasoning is ;)
John Thacker - I think the idea is that the peak was reached some time during Dec 07 (say, on Dec 10). That doesn't preclude the entire economy from growing overall from Dec 1 - Dec 31; it could be that the growth from Dec 1-10 was greater than the fall from Dec 11-31. Not sure if this is really what's going on, but I hopefully you see my point?
Posted by: weichi at Dec 1, 2008 9:55:56 PM
I'm confused, too. Can someone explain? From the BEA (http://www.bea.gov/national/nipaweb/TableView.asp?SelectedTable=1&ViewSeries=NO&Java=no&Request3Place=N&3Place=N&FromView=YES&Freq=Qtr&FirstYear=2007&LastYear=2008&3Place=N&Update=Update&JavaBox=no#Mid)
Percent Change From Preceding Period in Real Gross Domestic Product
2007 (3) 4.8
2007 (4) -0.2
2008 (1) 0.9
2008 (2) 2.8
2008 (3) -0.5
Posted by: GC at Dec 1, 2008 11:37:58 PM
Pretty simple, guys. Try reading the release. GDP is an ESTIMATE. GDP and GDI are EQUAL. The estimates DIFFER. The DIFFERENCE is ERROR. GDI was estimated as NEGATIVE in 2007Q4 and 2008Q4. I repeat GDP and GDI are EQUAL. The committee judges that the GDP estimates for the period are anomalous and ambiguous. The weight of the evidence says that the economy started contracting in Dec 2007.
Posted by: benamery21 at Dec 1, 2008 11:56:49 PM
BTW, real GDP needs to grow at least as fast as population for per capita GDP to stay constant. This means in a country like the U.S. with significant population growth, we are actually going backwards if we're standing still. It also means folks who want to change their talking points to say everything was hunky-dory until Dec 2007; in fact, dag-nab-it, that was the PEAK, so it must have been as GOOD AS IT GETS, or EVEN GOODER; should hang their heads in shame and slink away.
Posted by: benamery21 at Dec 2, 2008 12:02:17 AM
WTF, thats what I thought to begin with. But if you look at the timeline you see the index dip sharply right after the news was announced. It could have been mostly a coincidence, but to me there seems to be at least some correlation there.
Posted by: Leif at Dec 2, 2008 12:59:18 AM
I think the point is you either have a (relatively) objective standard or you have a committee clusterfudge.
The economy has been growing below trend for a while. The IMF has a 3% standard for recession, which is dumb, but if they stick to it then at least it's an objective standard.
If you don't stick to an objective standard, then you also have to guess at what the committee thinks the label means and they have to guess at what we are going to think it means.
Posted by: Andrew at Dec 2, 2008 5:20:57 AM
This seemed odd to me, so I thought it'd be interesting to check and compare
how NBER handled the recession in 2000. Here's what I found:
Adding it up, it looks like we have yet another example of a politicized and
profoundly dishonest group of academics.
Posted by: Mark Amerman at Dec 2, 2008 6:20:43 AM
"GDI was estimated as NEGATIVE in 2007Q4 and 2008Q4. I repeat GDP and GDI are EQUAL."
Yeah, we know. Look at q2 numbers. 2.8% is a big jump in the middle of a recession. I repeat... blah blah...
Posted by: Tom at Dec 2, 2008 8:55:25 AM
Mark,
The guy at your links, just like a number of people here, is for some reason choosing to pay attention to only one indicator - GDP, which apparently peaked in Jun 08. But if you RTFPR, you'll see that they look at other indicators as well, in particular payroll employment, which peaked in Dec 07.
The first post you link to actually puts the lie to the idea that NBER is politicized, though the author seems not to realize it! He reproduces an NBER table that shows when various data series peak relative to the start of the recession (as called by NBER). It shows that since 1980, the NBER has been *very* consistent about how they determine when a recession starts: they always pick a month within 2-3 months of the peak in payroll employment (three times out of 5 recessions choosing the peak month itself). In fact, with the exception of 1973, this holds back to 1960!
So the picture is clear: the NBER looks primarily at payroll employment, and other factors are secondary.
Now maybe you want to argue that this is a bad methodology, and GDP or whatever is a better indicator than employment. Fine. But it's very clear what the NBER is doing, and it looks to me like they are being very consistent about it.
Posted by: weichi at Dec 2, 2008 11:14:01 AM
weichi,
Possibly I'm mistaken. But I don't think so.
If I go to wikipedia and look up "recession", here's the first sentence:
In economics, the term recession is generally used to describe a situation
in which a country's GDP, or gross domestic product, shrinks for at least 2 consecutive
quarters.
But then I could have written that myself as that's exactly what I thought the word
meant.
Possibly payroll employment is a key indicator, more important than GDP, and we really
need to put most of our attention on its rises and falls but in that case why not simply say
that large organizational employment has fallen?
Is that so obscure? Is that so hard to understand?
Why is it necessary to gut the original meaning?
Also, and a little off-topic, I can't help wondering if we should be putting all this emphasis
on payroll employment even if the subject was being discussed straightforwardly. Correct me
if I'm wrong, but isn't the payroll survey predominately government workers?
Pretty easy to manipulate that, isn't it? If a government wants to avoid "recession"
just hire more civil servants. Simple. Likewise any attempt to reduce the size
of the government workforce -- not that I'm aware this has ever happened in the real world -- is
by definition a recession.
Finally I hope you caught this, from BizzyBlog's post:
At one point, there was some degree of economic consensus that the recession began in July of 2000
and ended in March of 2001 (2007 note: info at link was changed after publication of this post, and
references to what was just noted are no longer there). By the time the "non-partisan" (uh-huh) NBER
declared its end (third paragraph at link), the time frame had changed from March 2001 to November 2001.
How convenient.
I'll reiterate. Up until 2006 or possibly 2007, NBER had asserted that a recession began
in July of 2000 and ended in March of 2001. Then sometime in 2006 or 2007 someone at NBER wiped
this assertion off their website and (without explanation) replaced it with the assertion that
the recession in question began in March of 2001 and ended in November of 2001.
It doesn't matter the underlying methodology is, there's no way to square this with "consistency".
Posted by: Mark Amerman at Dec 2, 2008 1:43:44 PM
Mark,
First, please read the press release Alex linked to. They *do* explain their thinking, and provide answers to the many objections I'm seeing. Perhaps their answers are wrong, but at least address their actual arguments! For example, here is the second paragraph:
"A recession is a significant decline in economic activity spread across the economy, lasting more
than a few months, normally visible in production, employment, real income, and other
indicators. A recession begins when the economy reaches a peak of activity and ends when the
economy reaches its trough. Between trough and peak, the economy is in an expansion."
Later on:
Q: Is the expansion of real GDP (as measured using the product-side estimates) in the first
quarter of 2008 consistent with the identification of a recession starting in December 2007?
A: The committee considers a range of indicators of economic activity, and many of them
suggest declining activity in the first quarter of the current calendar year. These include payroll
employment and the income-side estimates of domestic production.
Q: In December 2007, was there a clear peak in economic activity or was there a flat period
around that time?
A: The committee found that economic activity measured by production was close to flat from
roughly September 2007 to roughly June 2008, while activity measured by employment reached
a clear peak in December 2007. The committee judged that the weight of the evidence suggested
that the peak occurred in December 2007."
Second, I don't know what the best definition of the term "recession" is. Maybe they are wrong, and the two-quarter definition is right! But that doesnt' mean that they are "politicised".
Finally, you are misunderstanding BizzyBlog's claim about the 2000/2001 recession. He claims that someone (in his bad link) at some point said the recession started in Jul 2000. I'm sure someone did, but it wasnt't the NBER! They have *always* said the recession started in Mar 01. In fact, in the linked document, the NBER state they have not changed any recession dates since 1975. I doubt they are lying about this.
However, in 2004 they apparently *were* considering whether to move the start date back. From this article in the NYT: (http://www.nytimes.com/2004/01/22/business/22WIRE-RECES.html?ex=1228366800&en=8affe769b8f7bb99&ei=5070)
"The NBER last changed a decision on dating a recession back in 1975. But the group's job this time was complicated by a number of factors, including a revision of the statistics making up the gross domestic product.
That revision, released last month, showed that the GDP, the country's total output of goods and services, shrank at an annual rate of 0.5 percent in the July-September quarter of 2000. Before the revisions, the GDP had remained positive through all of 2000 and turned negative for the first three quarters of 2001."
For whatever reason, the NBER never did change the start date.
Also, note that BizzyBlog was writing in 2006. He must have been looking at *revised* GDP numbers, which were different than the GDP numbers NBER was looking at in 2001 and 2003.
Posted by: weichi at Dec 2, 2008 3:29:55 PM
One more thing:
"Correct me if I'm wrong, but isn't the payroll survey predominately government workers?",
You are wrong. Just google "payroll employment survey", the first link will tell you what it is.
OK, another final thing. The president of the NBER? Martin Feldstein, author of 30+ op-ed pieces in the WSJ, including Sep 2008's "John McCain Has a Tax Plan To Create Jobs". Not exactly an anti-free-market leftist, which I think is the implication of the "politicized" charge.
Posted by: weichi at Dec 2, 2008 3:49:50 PM
weichi,
I wonder if it would be possible for the NBER to define objectively just
what it's criteria are. You are, if I understand you correctly, and NBER's
statement would seem to support your interpretation, asserting that the main component
of NBER's definition is the second derivative of the GDP function. As long as the
GDP is not growing but growing at an increasing rate, we are, according to this
definition, in an expansion, and as soon as the rate of increase slows down
we're in a recession.
With this definition not only is it possible that the economy could be growing in
a recession but it's even quite likely.
That violates my personal understanding of what the concept means. To me recession
means an economic loss of ground, and reading the comments of others, even supporters
of NBER, that's what it means to most people.
Also I know from reading other economists, say for example european economists,
that they also have the same, apparently erroneous, conception. For example, France,
would not assert it's in the middle of a recession if it's rate of GDP growth fell
from 2% to 1.5%. Yet if I understand the NBER they would assert that this would
be a good part of what they mean by "recession".
And also, there's the clear implication that it's not only possible but even likely
that the U.S. could be in an economic expansion at the same time that it's economy
is actually shrinking.
Once again this violates my intuition and my sense of what the word has meant
in the past, and what it means elsewhere.
Posted by: Mark Amerman at Dec 2, 2008 4:40:24 PM
weichi,
Thank you for pointing out my misunderstanding of the nature of the "payroll employment survey."
Somehow I had thought it was something different. Is there, or was there, an employment
survey focused only large institutions? Whatever.
Posted by: Mark Amerman at Dec 2, 2008 4:58:18 PM
