JOLTS were reported Tuesday, and Job Openings jumped in April, putting the Beveridge Curve in territory that would have signaled full employment 10 years ago. This is great news. I think it signals that the employment strength we have been seeing in the last couple of months is not a statistical aberration, and it gives me more confidence that we will see more employment gains over the next couple of months.
One caveat, though, is that the 2000's vintage Beveridge Curve was not typical. The JOLTS data doesn't go back any farther than that, so it's hard to find versions of the relationship tying into older data. Here is one that I've found from the Federal Reserve Bank of San Francisco. The relationship in the 1970's and 1980's would target us at 8% unemployment or more with April's Job Openings level.
We all like to push our own policy preferences onto these statistical canvasses, and I'd like to do that as much as the next guy. But, the 70's and 80's had their own, minor, baby boomer generation that was hitting retirement, and I wonder if the rightward shift has to do with behavioral tendencies of an older work force that tends to be unemployed less, but for longer durations, than younger workers. Maybe a 3.1% Job Openings rate corresponds to a 6.3% unemployment rate in the current demographic context.
On the other hand, the difference between where the Beveridge Curve is now and where it was 10 years ago is roughly equal to the unusual number of very long duration unemployed, which is probably a combined product of demographics, EUI, and the recession. The shape of the distribution of unemployment durations would suggest that the Beveridge Curve has more room to shift back to the left. We are, after all, still to the right of all historical Beveridge Curve locations outside of that 15 year period in the 1970's & 1980's.
Here are graphs of the trends in Openings, Quits and Hires, and the long term levels of all the JOLTS categories (below). I would have been ecstatic if we had seen a rise in Quits along with this rise in Openings, but still, trends remain positive.
Also, here is a graph of employment flows, through May. These kind of tell a similar story to the other indicators. Flows into and out of Employment and Unemployment are roughly back at a level that would normally correspond to full recoveries. But flows into and out of the Labor Force remain high. Again, this points to a labor market mostly functioning at recovery levels, but with a large pool of marginally attached workers. I see lots of narratives about this group that are really more about the narrative authors than they are about the group itself. I suspect that all those narratives and more exist here, so if we take everyone's expectations and throw them in a blender, we'll get a decent picture of what to expect. I think there is some upward drift in the Labor Force flows over the past 20 years, having to do with the demographics I discussed above, so I suspect that unemployment will continue to retain an extra 0.2-.03% of marginal workers, even in full recovery, but most of this group will slowly re-enter the workforce.
I think this is reinforced by the Quits data. It has not shifted like the Openings data has, in relation to unemployment. I felt like it pointed to a slightly overstated unemployment rate a few months ago, but with the recent decline in the unemployment rate, the Quits/Unemployment rates seem to be in line with previous levels. This suggests that current workers see the current groups of unemployed workers as potential competitors, which, to me, argues against the narrative that the unusual group of very long term unemployed is a product of a skills mismatch. If that was the case, workers would be quitting as if unemployment was at 5%.
Whatever the factors are that are keeping unemployment over 6%, it looks to me like the nuts and bolts of this economy are ready to finish the labor recovery this year.