Such large revisions raise doubts about the abilities of the Bureau of Labor Statistics' sampling procedures. Amplifying those doubts, the BLS last month indicated that it estimates the annual benchmark revisions will add some 810,000 to nonfarm payrolls as of the March 2006 benchmarking month. Against the backdrop of such large revisions, it's hard to have much confidence in any inferences to be drawn from the reported data. -- David Resler, Nomura EconomicsRevisions of a million jobs here, a million jobs there - pretty soon you're talking real employment! I don't know if the BLS is using a bad methodology or what, but when the size of the revisions is substantially bigger than the original data, the statistics are unusable. The other economists making comments based on the BLS's data must just like to see themselves in print, because comments based on useless data are, well, useless.
Friday, November 03, 2006
Somebody's Got it Right
A number of economists commented on the October jobs numbers in the Wall Street Journal. Here's my favorite: