JOLTS

I am wondering about Job Openings and Labor Turnover Survey (JOLTS) data. The reason is that I am interested in the extremely record high job vacancy rate of 6%, the moderately high unemployment rate of 5.8% and the moderately high hiring rate of 4.2% of employment in April (last month of data available). There are reports of firms having trouble finding workers, including the results of systematic surveys. Republican governors have decided to send money back to the Federal Government rather than pay the additional $300 per week unemployment benefits, because they think the over-generous benefits are causing the unemployed to turn down jobs they should take.

Others insist that the problem is childcare with unemployed parents staying home, because only about half of schools had restarted in person classes, or fear of Covid 19 keeping people home until they are vaccinated (or my favorite theory: those who are half vaccinated not wanting to combine the joys of a new job and a booster shot).

To me the question is whether there is something unusual a huge vacancy rate and a moderate hiring rate. So I went to FRED and looked at some graphs, which I feel like sharing.

It is very rare for monthly hiring to be greater than 4.2% of employment. The hiring rate has been higher only in May, June July and August 2020 and in January 2001. The enormous hiring rates last year were clearly mostly recalls of temporarily laid off workers. That is very different from matching workers to new jobs and training newly hired workers. The hint that 4.2% is some kind of ceiling is supported by the fact that it is the rate for March and April 2021 and October and November 2020 (and also way back in 2001 the pre-2020 record was 4.3%).

It makes sense that an extremely high ratio of newly hired workers to experienced workers creates problems. It is harder to explain why hiring and training limits didn’t prevent firms from posting the all time record high 6% rate of vacancies (job openings). But the key question is whether the high but not extraordinary rate of hiring is surprisingly low given the extremely high vacancy rate and high unemployment rate.

I tried to estimate a matching function hires = f(unem,Vacancy) and was surprised. I could find no evidence that there is more hiring given the job vacancy rate when there is higher unemployment (it is hard to estimate because before 2020 the unemployment rate and the vacancy rate had a correlation coefficient of -0.82 and so are extremely collinear). I also get a small regression coefficient of the log of hiring on the log of the vacancy rate of around one third. I think it is best to show this by graphing monthly hiring rate divided by the cube root of the vacancy rate

First of all this ratio has been extremely stable since 2009 — the absurdly simple matching function fits the data very well. Second, the current value 2.325 is perfectly normal. There is no anomaly. It is a general pattern that high vacancy rates just don’t imply as high hiring rates as is generally assumed (for example by me).

There was a huge anomaly in April 2020 when the economy was shutting down and there was very little hiring. There are much huger anomalies in May, June, July, and August when firms were calling back laid off workers (without having to look for new workers). But things are now normal. There is certainly no sign of lower than expected hiring due to more generous than usual unemployment insurance.