Based mostly on ADP-Stanford Digital Financial system Lab collection for October.
Determine 1: Non-public nonfarm payroll employment (daring black), nowcast based mostly on error correction mannequin utilizing ADP-Stanford Digital Financial system Lab collection (blue), +/- 1 normal error strains (grey), Bloomberg consensus as of 10/31 (purple sq.), all in 000’s, s.a. Supply: BLS, ADP by way of FRED, and creator’s calculations.Â
The ECM used is estimated over 2021M07-2024M09, and has an adjusted R2 of 0.67, SER of 0.002. Utilizing a easy regression of first log distinction BLS on contemporaneous ADP collection yields a nearly equivalent nowcast.
Notice that it is a mechanical nowcast, based mostly solely on previous BLS knowledge and ADP knowledge. There is no such thing as a accounting for strikes and hurricanes. Goldman Sachs (Walker/Rindels) notes:
ADP’s measure of employment progress has traditionally been much less delicate to pure disasters and Homebase’s knowledge seemingly misses the influence of the present strikes as a result of its pattern is primarily made up of owner-operator-managed meals and beverage, retail, and repair companies.
Â
Â