Tuesday, June 16, 2020

Why the Difference Between ADP Employment Data and BLS Employment Data and Does It Matter?

The kaleidoscope is turning.

Blake Allington emails:
Hi Robert:

Read your blog every day.

How does this FRED ADP employment data jive with the report of 2.5 million jobs gained?

ADP is still showing a decrease in total nonfarm private payroll employment.  The rate of decline has slowed a little, but it is definitely not moving in the opposite direction.

Last update was May 2020 so it’s not old data.

This is quite an impressive sight to behold.
ADP Data

Bureau of Labor Statistics Data

RW response:

First, I think it is important to recognize that short-term deviations between ADP and BLS reports are not new.

From Business Insider in 2017:
The US economy added 138,000 jobs in May.

The US economy also added 253,000 jobs in May.

Neither of those sentences has a typo. The smaller number was reported by the Bureau of Labor Statistics, and the bigger number was reported by ADP Research Institute, an arm of the HR management-software firm.

The occasional gap between the two metrics has long been a source of contention for economists. But the six-month average, which evens out lots of the monthly volatility and provides a year-to-date view, shows the most drastic diversion in years.

"It is unusual that they would be telling you such different things," Binky Chadha, the chief global strategist at Deutsche Bank, said.
Here's Seeking Alpha in 2019:
The trajectories of the BLS vs. ADP private nonfarm employment series differ, even though for much of the year, the ADP series was above that of BLS.
The divergence in changes is even more marked if one zooms in on 2019.
So other than long-term trends, it is not useful to read anything into a specific release. On top of that, the ADP and BLS use their own models which means they are guesstimating on parts of their data which is then adjusted down the road to reflect all the hard data that comes in.

Because their models are different, you can see short-term variations.

That said, I am not looking for exact pinpoint data from these releases--ever. There are all kinds of noise in the data.

I mean, do I really need to know if 130,000,000 are employed versus 130,000,001 or 131,000,000 or 129,000,000?

I know employment is way down because of the lockdowns and as the lockdowns ease, the numbers are going to improve.

When you have dramatic changes in an economy like the lockdowns, you are going to expose problems in the data. I don't care.

I am trying to understand general trends in the economy, rather than get an impossible exact number.

I like to keep in mind what the economist G.L.S Shackle wrote in Epistemics & Economics: A critique of economic doctrines (1991) P76:
The business scene and its participants can be looked on as a staging contest of rival orientations, rival ambitions, rival exploitations of the world. It is capable, for all the analyst can tell ex ante facto, of realizing some one or other of these visions in some degree, and thus of presenting an appearance of momentary or temporary orderliness during the ascendancy of one orientation and its sponsors. Or the contest may be inconclusive and sterile, and result in a period of rudderless backing and filling of the sails and of untidy, blind struggle and groping for decisive policy. It will be a kaleidic society, interspersing its moments or intervals of order assurance and beauty with sudden disintegration and a cascade into a new pattern. Such an account of the politico-economic process may at various epochs or in the course of various historical ages appear less or more suggestive and illuminating.
The kaleidoscope is now turning with regard to the economy and the economic models. During such a period, if we can sense out just one or two significant things about the economy before the pack, that can be very important. All other data can just be considered parts of the kaleidoscope continuing to settle. It's not important.

The two things I have focused on now in the EPJ Daily Alert is that, while most of the other data continue to settle down, the big things to know are that the stock market will climb (and confuse most because of the kalidescopic period we are in) and also that price inflation will climb (also confusing many).


1 comment:

  1. Excellent post. I haven't read Schackle's Epistemics and Economics:... but love the quote. We seem to be looking through the same lens at the moment.