Missing the May Jobs Story

The coverage of the May employment numbers put far too bright a spin on the data in the report. For example the NYT article was headlined "Several Signs the Economy is Reviving." The Post told readers that "Job Growth Strengthens the Economy."

Admittedly this was a tough report to read -- 157,000 jobs isn't bad -- but reporters should be doing their homework. There is good reason to believe that the establishment survey (the one showing 157,000 new jobs) is now overstating employment because its imputuation for job growth in new firms not captured in the survey. As I pointed out in my jobs byte, the Labor Department is imputing more jobs in new firms this year than in 2006, when the economy had been growing very rapidly. The possibility of a substantial overstatement of job growth in the establishment survey was brought home by the release of Business Employment Dynamics (BED)data for the 3rd quarter of 2006. These data are a near census of employers, since they are based on unemployment insurance filings. The BED showed private sector net job creation of just 19,000 in the 3rd quarter. This compares to almost 500,000 jobs reported in the establishment data. (Thanks to Nouriel Roubini for keeping an eye on this one.)

There are other important items suggesting weakness in this report. The average hourly wage grew at just a 3.2 percent annual rate over the last three months. It had been growing at nearly a 4.0 percent rate last fall. It is unusual to see such a rapid deceleration of wage growth in a strong job market.

Finally, the household survey is certainly showing real signs of weakness. The employment to population ratio has fallen from 63.3 percent in March to just 63.0 percent in the April and May data. This corresponds to 600,000 fewer people having jobs. That doesn't happen in a strong labor market.

The long and short is that this report showed serious evidence of economic weakness, which is why some of us bears are now thinking we may have been right all along.

--Dean Baker