ADP said payrolls surged 220,000. Expectations were for a reading of 210,000.
Last month was revised to 209,000 from 191,000.
Here’s the breakdown:
- Goods-producing 24,000
- Service-providing 197,000
- Construction 19,000
- Manufacturing 1,000
- Trade/transportation/utilities 34,000
- Financial activities 8,000
- Professional/business services 77,000
Here are the charts:
Around this time last month, we reported that some analysts were rebelling against the weight given ADP’s ability to predict the actual jobs number from the BLS. The contention was that it is too reliant on BLS data and not on its own figures, making it sort of useless. What’s more, the analysts said, ADP wasn’t even using BLS data properly, and that it often ended up with a number closer to the previous month’s official figure.
Then ADP totally proved them wrong by basically nailing the number, which came in at 192,000.
In a note this morning, HFE’s Jim O’Sullivan says Friday’s print will prove another test, but that the critics seem to have outkicked their coverage:
…while we have never viewed the ADP series as a very reliable monthly indicator, and would prefer if the data were more clearly independent from the BLS series, we think the scathing criticism is not merited. Realistically, the ADP series will never make the employment report redundant, but it does at least provide another gauge of whether trends are suddenly changing. As with jobless claims, its message a few months ago was that the sudden weakening in payrolls in December and January was exaggerated.
For the record, the median ADP miss, without regard to sign, has been 36K for the initially reported data in the 18 months since Moody’s became the compiler. That is a clear improvement relative to the 62K median miss in the prior 18 months. However, it is slightly larger than the 31K median consensus miss in the last 18 months; the median consensus miss was 44K in the prior 18 months.
Let’s see what happens.