The paradox remains: despite below-consensus additions of 99K in private ADP-Stanford NFP (well below consensus of 144K). The cumulative change of ADP is above the cumulative change of CES, while the first benchmark is below.
ADP survey increase in private NFP from 2023M03 is 3.2mn to August, More than the 2.8mn reported in the current official CES series until July.
In contrast, the first benchmark shows 819 thousand less in March 2024 than the current official series. Given the increase in independent NFP from March 2024, the increase in employment is 2.0mn (not 2.8mn until July).
Figure 1: Cumulative growth in private NFP as of March 2023 according to CES (blue), according to ADP-Stanford Digital Economy Lab (tan), according to the first benchmark (red square), all in 000’s, sa Source: BLS, ADP by FRED, BLS, and author statistics.
The ADP-Stanford Digital Economy Lab series is not survey-based, but actual check-cutting activities. Preliminary Benchmark depends on QCEW (unemployment insurance). While QCEW is a census, that census data is updated. Additionally, some employment will not be included in the census, as some employment is not reported by firms to the tax authorities. Therefore, Goldman Sachs thinks that the effective downward revision is about 300K, instead of 818K, for the NFP as a whole (there is no different figure reported by GS for the private NFP).
Source link