Payroll processing giant ADP reported on Wednesday that non-farm payrolls increased by 114,000 positions in July, more than economists expected. How much should this change market expectations for the Bureau of Labor Statistics (BLS) numbers that come out on Friday?
I briefly compared changes in payrolls (job additions or losses) from the ADP and BLS reports on July 10. To recap, the correlation between the two reports changes over time, and we are currently in a period where the correlation between them is approximately 0.74.
With Wednesday’s release, ADP included revisions to earlier data. The changes were sufficiently small so that the end result did not change; the correlation of the jobs reports since the Great Recession ended through June 2011 stands at 0.74.
In sum during this period, the ADP figures have been rosier than the BLS report, overstating positions added by 130,000. On average, the ADP report has overstated monthly BLS job changes by 6,000 positions.
Let’s play the averages for a moment to gauge what we can expect on Friday. Economists expect the BLS report to show that the economy added 85,000 positions in July. If we look at how much the ADP report has overstated the BLS report, on average, in recent months, this suggests that we could expect the actual number to come in near 108,000 positions, for a 27% increase over current expectations.
What if we simply look at how ADP compared with its expectations? Economists expected the ADP figure to come in at 100,000; the actual is 114,000. If we increase the current BLS estimate by 14% (the difference in the ADP report), we arrive at just about 97,000.
That gives us a “reasonable” range of what to expect come Friday: 97,000 on the downside to 108,000 on the upside.
There is, however, one other key factor to keep in mind: The ADP figures showed that job growth slowed, from 145,000 to 114,000. But, at present, economists expect the BLS report to show that the economy created 85,000 jobs last month, up from 18,000 in June – a slight acceleration in job growth.
Does it matter that the ADP figures show a slowdown in job growth when expectations for the BLS report show an increase? In a nutshell, not really.
Growing Quickly vs. Growing Slowly
Focusing on the period since the last recession officially ended, I examined the pace of monthly changes in payroll figures. Ignoring magnitude, I used 1s to indicate periods where payroll conditions improved (including fewer losses), and 0s to indicate labor market deterioration (including a deceleration in the pace of hiring). Then I took the correlation between the two reports. I found that the two measures track monthly increases and decreases with a correlation of about 0.25.
That stated, we should not be surprised if the ADP figures show a deceleration in the pace of hiring, while the BLS report shows an uptick.