Today's initial GDP reading of 0.6% was the lowest since the first quarter of 2007. GDP was this low just 3 quarters ago, so it's not that out of the ordinary. And it's important to remember that the number will be revised two more times before it is final. It could go higher, but it could also go lower. Below we highlight the historical final annualized GDP numbers going back to 1947 (4Q '07 is the only one that is not a final number). We also highlight with red dots any time that GDP was less than 1%. While a drop below 1% often times means we'll end up getting 2 consecutive quarters of negative growth, it doesn't always mean it will happen. In fact, while the NBER declared a recession in 2001, we haven't had 2 consecutive quarters of actual negative GDP growth since 1991.
Subscribe to Bespoke Premium today.
It's correct that there will be two more revisions.
"Quarterly GDP reports are broken down into three announcements: advance, preliminary, and final. After the final revision, GDP is not revised again until the annual benchmark revisions each July. These revisions can be quite large and usually affect the past five years of data."
Do not forget, however, that there comprehensive revisions conducted approximately every five years and likely due in 2008.
http://bea.gov/scb/pdf/2007/07%20July/0707_ta.pdf
The BEA published new estimates (comprehensive) for 2001 only in 2003 and then the NBER has to change its definition to recession with which not everybody agrees. For example, decreasing employment is usually accompanied by increasing productivity. What is better?
Accuracy of quarterly revisions is well described in:
Reliability of the NIPA Estimates of U.S. Economic Activity
By Dennis J. Fixler and Bruce T. Grimm
http://bea.gov/scb/pdf/2005/02February/0205_NIPAs.pdf
They are not too much accurate - mean absolute revision is around 1%. Important is that current estimate (advanced, preliminary, final) contains no information on next revision.
Posted by: KIO | January 31, 2008 at 05:56 AM