Public Service Announcement on the Non-farm Payroll
The large upward revisions to August payrolls released by the Bureau of Labor Statistics this morning (with jobs data for September) set the conspiracy theorists' world on fire. And they were strong, about three times the usual revision. But, as we pointed out in a report we sent to our clients earlier this week, this is a long-standing pattern. It almost always happens, whether there's an election coming up or not. Here's the relevant text from this morning's note:
August's gain was revised upward by 46,000, and July's by 40,000. Almost all the revisions, however, came from an upward revision of 101,000 to local government education in August before seasonal adjustment – a recurrent anomaly at this time of year that we wrote about in Wednesday's report. The concurrent seasonal adjustment technique distributes large changes like that backwards, so the gain was split between July and August in the adjusted numbers. Some excitable types are attributing the upward revision to political machinations, but this pattern has been around a long time. It's likely something is amiss in the BLS's collection process, and they are working on it. There shouldn't be a recurrent pattern of error like this. (Excitable types should also note that the birth/death model subtracted 9,000 jobs in September.)