No, the Unemployment Rate Isn't Rising

By Matthew Yglesias, Slate

01 February 13

he headline number out of today's BLS jobs report is 157,000 new jobs. Not bad, but not great. As ever, the better information is in the revisions to past months. Here we have November up to 247,000 (from 161,000) and December up to 196,000 (from 155,000). As revised, those are both really solid months and good data estimates.

The bad news in this report, such as it is, is that the unemployment rate ticked up to 7.9 percent. Except it really didn't. You are going to see this misreported, but here's the deal. Each month the unemployment rate is calculated based on a survey of households. To do the household survey you need an estimate of the population universe you're trying to survey. And each January they redo their population estimates. These estimates are not backward-projected into the old data, meaning that from December 2012 to January 2013 you're comparing different survey universes. A gigantic shift in the unemployment rate would swamp those estimation issues, but that's not what we saw. You can't compare December's 7.8 percent to January's 7.9 percent except to say that there wasn't much change.

The better news is that the payroll survey also did a re-benchmarking and found 422,000 missing jobs from last year. The last re-benchmarking from September also found almost 400,000 new jobs. That's a recovery pattern—as the economy churns upward, initial estimates tend to undercount.