Skip to main content

The ADP National Employment Report for May is worth a look, because the numbers are a bit shocking, at least compared to consensus estimates.

Compared to BLS, sometimes ADP is much higher and sometimes much lower. Misses, assuming the BLS is accurate (not necessarily a good assumption), appear to be random.


Image placeholder title

ADP and the BLS both heavily revise numbers over time, so the numbers eventually converge.

Trend Change or Outlier?

Image placeholder title


  • The average for the year is 188.
  • The 3-month average is 152.
  • The 2-month average is 149.
Scroll to Continue


If this is an outlier, then any of the averages are a better indicator than judging from the one-month total.

Consensus Estimates

  • The Econoday consensus estimate for ADP was 175,000 jobs.
  • The Econoday consensus estimate for the BLS is 180,000 jobs.

Jobs vs Employment

Last month, the BLS said Jobs Rose by +263,000 but Employment Declined by 103,000.

In the past year, the difference between jobs and employment is a whopping 1,191,000. That's a discrepancy of 99,250 every month, in favor of jobs.

In that regard, 27,000 no longer seems shocking.

I suspect the BLS and ADP are double-counting part-time jobs. Neither weeds out duplicate social security numbers.

We hear from the BLS on Friday.

Mike "Mish" Shedlock