The BLS Job Openings and Labor Turnover Survey (JOLTS) can be used as a predictor of future jobs growth, and the predictive elements show that the year-over-year growth rate of unadjusted private non-farm job openings declined relative to last month – wiping out last month’s gain.
Analyst Opinion of JOLTS Data
The BLS stated:
On the last business day of September, the job openings level fell to 7.0 million (-284,000), after reaching a revised series high of 7.3 million in August. The job openings rate was 4.5 percent in September. The number of job openings edged down for total private (-188,000) and fell in government (-96,000). Job openings increased in health care and social assistance (+71,000). The number of job openings decreased in many industries, with the largest decreases in professional and business services (-118,000), finance and insurance (-82,000), and state and local government, excluding education (-67,000). Job openings decreased in the South region.
The unadjusted data analysis shows rate of growth is about average seen since 2010.
Market expectations from Econoday was 7.100 M to 7.125 M (consensus 7.100 M) versus actual of 7.0 million.
The graphs below uses year-over-year growth of JOLTS Job Openings – both the level of openings and rate of openings.
Last month’s graphs
This Month’s Graphs
There is a skills mismatch in the U.S. workforce – as job openings are higher than number of unemployed people.
The JOLTS Unadjusted Private hires rate (percent of hires compared to size of workforce) and the separations rate (percent of separations compared to size of workforce – separations are the workforce which quit or was laid off).
Unadjusted Hires (blue line) and Unadjusted Separation Levels (red line) – Non-Farm Private
Please note that Econintersect has not been able use the hire rate or the separation rate (or a combination thereof) to help in understanding future jobs growth. A Philly Fed study agrees with Econintersect’s assessment. JOLTS is issued a month later than the jobs data – and correlates against one-month-old data.
