The pre-labor report labor reports
Let’s investigate these pre-BLS reports.
There are many job reports out there, but here are some of the biggies that come out right before the official government BLS employment situation (not a coincidence).
- Monster Employment Index (not seasonally adjusted) – an index geared toward online recruiting trends.
- Challenger job cuts (not seasonally adjusted) – follows announced layoffs at large companies.
- The ADP report (seasonally adjusted) – estimates private nonfarm payroll using their payroll data (very different from the BLS survey methodology).
- WANTED Technologies report (no information on seasonal adjustments on the website, but it must be, right?) – estimates total nonfarm payroll (private plus public) using “hiring demand data” stemming from online job boards.
Well, they are all generally trending in the positive direction – i.e., the “it’s getting worse less quickly” story. This is consistent with the BLS report. The trend is about all that they can match, the level value seems more like a hit or miss to me. (Note: except for the Challenger cuts, all of this data has been revised). Correction: Just heard from Charles Thibault over at WANTED – he says: “We do not ‘revise’ forecasts ex-post like ADP”. Will update if that changes. RW: Maybe there’s something to this one.
But I am a skeptical as to the exact value added from these reports. Although the ADP report does not explicitly claim this, I imagine that they do like the idea that markets generally use their number as an indicator of the upcoming BLS report on Friday.
WANTED Technologies is explicit in their claim to forecast the BLS report more consistently than does the “consensus”. And furthermore, they present a root mean square error of their forecast (a measure of how close the forecast comes to the actual data) as something below that of the BLS (see Table at the bottom of their methodology page). I am not quite clear on how they calculate the error, since they revise their data as does the ADP (again, correction: according to WANTED, they do not revise their forecast ex post, and the MSE comparison may be quite meaningful).
This gets me back to my first point, what is the value added of these pre-labor labor reports? I always thought that the various reports should focus on what are their comparative advantages. For example, the ADP payroll figures likely have information on wages that cannot be ascertained from the BLS’ survey approach. Or WANTED uses online job applications – perhaps it can provide an earlier indication of labor prospects than can the survey as a whole.
We will see what happens tomorrow.