E-piphany

The ADP figures are the best indicator we have when it comes to predicting Payrolls; unfortunately it isn’t a great indicator. Moreover, Friday brings the December data, which along with January are really difficult months to seasonally adjust so “anything goes.” With those caveats, I am very interested in the fact that ADP (which measures private-sector jobs from companies who use ADP for payrolls…tending toward larger companies) is running so far behind Non-Farm these days. Over the last 12 months, as the following chart shows, Non-Farm Payrolls have outpaced ADP by what is historically a fairly large amount.

NFP has been running very far ahead of ADP

At some level that is not completely surprising, since we would expect that government (not covered by ADP) has been a prime source of jobs over the last year. Oddly, though, that’s not the case. According to Bloomberg, total government payrolls actually declined from Nov 2008 to Nov 2009, by 42,000 employees (from 22.543mm down to a mere 22.501mm local, state, and federal employees are working hard for you…one in every six working stiffs toils for the government).

Thus the lagging ADP either means that the jobs are all being added by smaller companies or, anyway, non-ADP customers…which seems unlikely, since there is no credit available for small businesses at the moment…or this is a statistical artifact, perhaps having to do with the way the series is smoothed on the ADP side or collected and reported on the Labor Department side.

This makes me concerned about the consensus for December Payrolls, which is for a flat report. It looks like economists are just projecting the NFP trend, but that would represent (after today’s figure) an 84k outperformance of ADP; which would bring the 12-month total to +647k.

If the 12-month total instead remained flat at +516k, then NFP would be printing -131k.

There may well be other systematic things going on; as I remind people repeatedly there has never been a period like this in the data so nothing you extract from the data can carry very high confidence. But I would think this month that there is a greater chance for a downside surprise in NFP than an upside surprise.