Economist Online
SUBSCRIBE:

Monday, December 24, 2007

Recession or not?

With recent econ stats all looking (surprisingly) good, why people still fuss about recssion? Here included are some heated debates on the topic: 3 videos from CNBC's Kudlow and one quant piece from Jim Hamilton.

http://www.cnbc.com/id/15840232?video=612803140
http://www.cnbc.com/id/15840232?video=613298026
http://www.cnbc.com/id/15840232?video=613298033



The bears must wait another quarter

by James Hamilton

Currently available data on consumer spending make it very unlikely that we'll see negative real GDP growth for the fourth quarter.

The BEA reported on Friday that seasonally adjusted real consumption spending grew at a 6.5% annual rate between October and November. As Calculated Risk has noted in the past, this takes much of the guesswork out of predicting what the fourth-quarter numbers will turn out to be. The quarterly consumption component of real GDP is just the average of these monthly numbers, so we now have two of the three months that will go into 2007:Q4 consumption spending.

Following CR, blue bars in the graph below represent the growth rate of the average real consumption during the first two months of a quarter relative to the average real consumption over the first two months of the previous quarter; (my numbers differ slightly from CR's in that I've measured growth rates from changes in the natural logarithm, which correspond to continuous compounding and make the algebra a little simpler). Green bars in the graph report the actual growth rate for the quarterly consumption component of real GDP. The October and November 2007 data imply an estimate of the growth rate of real consumption spending of 3.2% during the fourth quarter of 2007.


pce_dec_07.gif

The two-month estimate has an R2 of 80% in predicting the actual quarterly numbers, and a root mean squared error of 80 basis points. The latter can be used to construct a 95% confidence interval for the 2007:Q4 forecast. The 3.2% estimate for 2007:Q4 implies a confidence interval between 1.6% and 4.8%. That Normal confidence interval can also be bootstrapped, by the way-- only 4 out of the 69 errors since 1990:Q3 exceed 160 basis points in absolute value.

With 2007:Q4 real consumption growth in the ballpark of 3%, it's extremely unlikely that we'll see the negative real GDP growth for 2007:Q4 that some analysts had been predicting before the strong November retail sales figures came in.




--
**************************************
Paul D. Deng
Department of Economics
Brandeis University
IBS, MS 032
Waltham, MA 02454
http://www.pauldeng.com/