One More Sign of Struggles on the Job Creation Front

Hot on the heels of the bleak employment report for June, the U.S. Bureau of Labor Statistics yesterday released another important bit of labor market information: the May Job Openings and Labor Turnover Survey, commonly known as JOLTS. Calculated Risk accentuated the positive:

“In general job openings (yellow) has been trending up—and job openings increased slightly again in May—and are up about 7% year-over-year compared to May 2010.

“Overall turnover is increasing too, but remains low. Quits increased again and have been trending up—and quits are now up about 10% year-over-year (usually a sign of more confidence in the labor market).”

The hires and quits angle was noted at Modeled Behavior as well, but at Zero Hedge Tyler Durden isn’t in the mood for even muted optimism:

“There was nothing to smile about in today’s May JOLTS release from the BLS.”

Durden focuses first on the job openings piece of JOLTS:

“Those expecting a pick up in job openings (traditionally the key requirement for an [sic] sustained increase in NFP) will have to wait some more, after the May number came at 3.0 million, the same as April.”

Longtime readers of macroblog know that the job openings data are favorites of ours, which we like to view through the prism of the Beveridge curve. This curve plots the relationship between job openings and unemployment. In past posts—here and here, for example—I have noted that:

  • The Beveridge curve appears to have shifted out over time, meaning that the amount of unemployment relative to the number of job openings has increased over the past several years relative to the patterns of the decade or so prior to the beginning of this recovery.
  • This shift in the Beveridge curve is usually interpreted as representing a change in the efficiency with which workers looking for jobs are matched with employers looking to fill jobs (though the source of the inefficiency is not completely obvious).
  • There is a debate about whether the recent shift in the Beveridge curve is a normal cyclical feature of recoveries—in which case the pace at which the unemployed are placed in open jobs ought to pick up over time—or a symptom of deeper structural problems that are likely to persist for, forgive me, an extended period of time.


For most of this year, it did appear that the Beveridge curve was moving back in the direction of its 2000–9. That progress was interrupted in May:


I don’t think that resolves the cyclical-versus-structural debate, but it certainly is not good news. And at this point it is awfully hard to believe that things are going to look better in the June version of JOLTS.

This post originally appeared at macroblog and is reproduced here with permission.