December 01, 2008
The Business Cycle Dating Committee of the National Bureau of Economic Research announced today that the eleventh U.S. postwar recession began in December of 2007.
As has often been the case historically, the announcement itself is a bit anticlimactic, in that pretty much everybody had already reached the same conclusion. I was interested to see that the Business Cycle Dating Committee explained the apparent non-recessionary behavior of real GDP in 2008:Q1-Q2 just as we did last September in terms of the statistical discrepancy between GDP and GDI. From the NBER:
The committee believes that the two most reliable comprehensive estimates of aggregate domestic production are normally the quarterly estimate of real Gross Domestic Product and the quarterly estimate of real Gross Domestic Income, both produced by the Bureau of Economic Analysis. In concept, the two should be the same, because sales of products generate income for producers and workers equal to the value of the sales. However, because the measurement on the product and income sides proceeds somewhat independently, the two actual measures differ by a statistical discrepancy. The product-side estimates fell slightly in 2007Q4, rose slightly in 2008Q1, rose again in 2008Q2, and fell slightly in 2008Q3. The income-side estimates reached their peak in 2007Q3, fell slightly in 2007Q4 and 2008Q1, rose slightly in 2008Q2 to a level below its peak in 2007Q3, and fell again in 2008Q3. Thus, the currently available estimates of quarterly aggregate real domestic production do not speak clearly about the date of a peak in activity.
Mechanical algorithms based on the monthly values of non-farm payroll employment, the index of industrial production, real personal income excluding transfer payments, and real manufacturing and trade sales are updated independently (with slightly different specifications) by Professor Jeremy Piger of the University of Oregon and Professor Marcelle Chauvet of the University of California, Riverside. Their separate indexes are both now indicating a 99% probability of recession.
The Econbrowser recession indicator index, which is based solely on the GDP numbers, has not yet triggered a recession call. I am expecting it to do so when the 2008:Q4 GDP numbers are released at the end of January.
As on so many other matters, Calculated Risk was way ahead of the curve. CR began in March of 2008 to add shading to his graphs to indicate that a recession began in December 2007.
Posted by James Hamilton at December 1, 2008 10:57 AMdigg this | reddit
Maybe your official recession index has not called it, but I remember seeing you in San Francisco in April at which time you said to me that you thought we had been in a recession since about the end of 2007. So, congratulations.
Posted by: Barkley Rosser at December 1, 2008 11:37 AM
Professor, with this behind us, now, it is time to work on a 'Econbrower Depression Indicator Index.'
Because, that is what we are in (year one, now underway, of five).
Posted by: jg at December 1, 2008 12:21 PM
What wonderful news! Now Bernanke can finally begin to stimulate the economy!
Greenspan started pushing up the interest rate in June of 2004 and the slow descent began in September 2007. What a masterful job. They engineered the recession within 4 months. Wow, are they brilliant! I just love command economies. Don't you!
And to think his successor is even better as he works toward engineering both a recession and massive inflation at the same time. Not stagflation but recesflation. These guys are amazing. They can work tricks we never dreamed of.
I'm not cynical. Hey, my 401-K hasn't fallen 40% in the past year.
Posted by: DickF at December 1, 2008 01:41 PM
An off topic question but I'll ask anyways: Given that the Fed has agreed to buy MBS I assume that this will be done via their magic checking account. But is it potentially self-sterilizing in that now the Fed will be receiving payment (in)directly from mortgage payors and thus can cancel any cash they receive?
Posted by: Rob at December 1, 2008 02:15 PM
I was the first! I said that the recession began in Dec 2007, on Feb 5, 2008.
So even earlier than Calculated Risk, or Laxman Achuthan.
Posted by: GK at December 1, 2008 02:34 PM
The Econbrowser recession indicator index won't identify a recession until 2 months after the NBER labels it. Its also several months after numerous internet sources labeled it, and some of them use systematic techniques, which they also publish. And, its a year after the recession started.
Could you explain to us the value of the Econbrowser recession indicator index, as in, how would one of us make a better economic decision by using it or the data it produces.
Posted by: Mike Laird at December 1, 2008 02:57 PM
I think when the recession started is not a pressing question now. The pressing question now is whether this will be just a "recession" or an actual depression.
Posted by: Hal at December 1, 2008 11:24 PM
Mike Laird: Do you have any use for the NBER announcements? If not, then certainly our index would be of no use to you, either.
The historical simulations reported in my 2005 paper with Marcelle Chauvet indicate that for the previous 8 business cycle turning points, the call from the GDP-based index would have come between 11 months earlier and 3 months later than the NBER announcement. This time it looks like it's going to be 2 months later.
In addition to sometimes giving a substantially earlier call, our algorithm offers an independent and entirely objective way of generating these statements, whose potential benefit I mentioned here:
outsiders might wonder (perhaps without justification) whether the dates of [the NBER] announcements are entirely independent of political considerations. For example, there might be some benefit to the presidential incumbent of delaying a declaration that a recession had started or accelerating a declaration that a recession had ended. For these reasons, it is worth exploring whether one could perform a similar function using purely objective summaries of the data.
As I think I have also mentioned to you on one of the previous occasions on which we have discussed the merits of the index, I recall talking to a reporter in January of 2003. At that time the NBER had not yet made its declaration that the recession ended in November of 2001. When I told the reporter that in my opinion the recession was definitely over, he looked at me as if I were from Mars and must know absolutely nothing about the economy. So I certainly think there would have been a benefit at that time of having an index like this in the public domain.
Granted our anticipated January announcement will be even more anticlimactic than yesterday's NBER statement. But I suspect there will come a future occasion, perhaps as soon as announcing the end of the recession, when the separate Econbrowser index does have some news value. At that time I have a hunch your criticism will change from "why tell us something we already know" to "obviously your algorithm is wrong".
Posted by: JDH at December 2, 2008 06:31 AM
Looking back at old posts, your recession indicator index was:
The threshold for calling a recession is 66%, so wouldn't this be read as saying that we were not in a recession in any of these three quarters? Or does new data coming in cause a revision of old probabilities?
If the index turns out to be in disagreement with a consensus call that the recession started in late 2007:Q4 then of course it would not be the only analysis which was turned upside down by the unprecedented events of the past year. I imagine that a number of results are going to have to be reconsidered from scratch in the aftermath of current events.
Posted by: Hal at December 2, 2008 11:26 AM
Hal: The index itself is never revised, because, by definition, it is an assessment of where the economy stood at time t-1 given information available at time t. When that number gets above 67%, the economy is declared to have been in recession at t-1.
Having made that declaration, there is then a second question as to when the recession began. That determination will be based on looking at assessments of where the economy was at t-j given information available at time t. This by definition is revised-- the assessment of 2008:Q1 given 2008:Q2 data is different from the assessment of 2008:Q1 given 2008:Q3 data.
My current guess, based on looking at some "what if" scenarios, is that the index will shoot well past 67% when the 2008:Q4 data are available, and that at that time we will declare that the recession started in 2007:Q4. That's based on my telling the algorithm a number I'm guessing for 2008:Q4 GDP.
See the write-up here.
Posted by: JDH at December 2, 2008 11:40 AM