September 26, 2003

#123: Betting the Ranch (interim report)

P. Krugman

This is an interim Squad report while Paul Krugman is away on book tour promoting his latest recycle-job, "The Great Unraveling."

All Squad readers should circle October 3rd on their calendars. This is the day the Bureau of Labor Statistics makes its annual benchmark revisions in the payroll survey of U.S. business establishments. As mundane as this topic may sound it should have Krugman and the Democrats awake at night with worry. We've often pointed out that PK and the Dems have few arrows left in their anti-Bush quivers over economic policy. They are now down to just one arrow--JOBS. For example, in the recent debate between Democratic presidential candidates the words "jobs" or "jobless recovery" were used more frequently than any others except "and", "I" or "the." But the upcoming October payroll revisions may kick the "jobs prop" out from under them.

Here's some background. Whenever Krugman and the Democrats say that we have lost 1.1 million jobs in the U.S. since the recession ended in November 2001 and cite this as evidence of a jobless recovery, they are relying on the Labor Department's survey of payrolls at business establishments, a sample of 400,000 firms. But Krugman's problem, which he may have to confront after October 3rd, is that this survey is probably the most revised data series produced by entire the Labor Department. There are many reasons for this but the main problem is that the establishment survey in real-time does a poor job of surveying small firms and it barely counts jobs created in new, start-up firms. Jobs in these areas are particularly important at turning points in the economy and, not surprisingly, that is when the establishment survey contains the largest revisions as more complete information becomes available.

For example, during the last "jobless recovery" in 1992 the establishment real-time survey showed a loss of 55,000 jobs initially, but was later revised to show a gain of 1.4 million jobs. That's a swing of nearly one and a half million jobs! Thus Krugman and the Democrats currently are betting the ranch on a real-time data series that is KNOWN to be unreliable.

Will the Oct 3rd revisions follow the 1992 precedent and show that job loss was much less than initially reported and that jobs are now increasing? No one knows for sure. But Krugman has good reason to be worried and we can afford to dream a bit. For example, what would he say if the jobs data were in fact revised upward and then, after that, the 3rd quarter GDP report, also announced in early October, showed continued strong growth? We're hoping we get a chance to find out.

[The Truth Squad is a group of economists who have long marveled at the writings of Paul Krugman. The Squad Reports are synopses of their discussions. ]

Posted by John Weidner at September 26, 2003 3:40 PM
Weblog by John Weidner