This employment situation release January 2023 Annual baseline revisions for the business survey series and reporting population controls for the household survey series. The NFP came in at 517K, beating the Bloomberg consensus of 115K.
This is non-farm payrolls based on multiple measures.
figure 1: NFP in January 2023 incorporates baseline revisions (blue), NFP in December 2022 (tan), authors use preliminary March 2022 baseline (green) to calculate implied baseline revisions, and NFP Conceptually Adjusted Civil Employment Household Survey Series Report (red), and January 2023, Adjusted Household Series to Remove the Effect of Revised Population Controls (light red squares). The adjustments made to remove the revised population controls assume that a 1% change in adjusted employment from September 2019 to December 2022 increases overall civilian employment by 1.24%. Source: BLS (Published Jan 2023, Dec 2022), author’s calculation, Bureau of Labor Statistics.
The civilian employment series incorporates the effects of the new census population controls. I tried to remove the effect of the new population control by accounting for the change in the previous population control for the total civilian employment, which was 84,000 (officially reported 894,000). A 1.24% change in clerical employment during 2021M09-22M12 corresponds to a 1% change in clerical employment adjusted for the NFP concept. Applying this rate, civilian employment adjusted for the NFP concept is little changed. (In which case it pays to read the footnotes!). As I mentioned before, it is dangerous to place too much weight on household survey series compared to institutions because of the high variability in these series (and it is especially dangerous to look at growth rates that cover new population controls).
Notably, the growth of NFP showed an increasing trend (indicated by increasing slope on a logarithmic scale) and Employment levels, including official data for December and data containing a preliminary benchmark for March 2022 (my calculations describe here). I think the level increase from March 2022 is mainly a seasonal update.As mentioned, both of these indicate that the labor market remains tight Posted by Pawel Skrzypczyński.
The fact that measured levels of employment are attributable to seasonal factors leads one to question whether the apparent strength is due to seasonal issues. To partially address this, I plotted the 12-month log changes across several employment series.
figure 2: Year-over-year nonfarm payrolls growth with baseline revisions, seasonally adjusted (blue), unseasonally adjusted (sky blue), from employed, unseasonally adjusted (pink), and household surveys covered by the Quarterly Census of Employment and Wages Serial clerical employment is adjusted to the reported NFP concept (red), and household series are adjusted in January 2023 to remove the effect of the revised population control (light red squares). Growth rates were calculated using log differences. The adjustments made to remove the revised population controls assume that a 1% change in adjusted employment from September 2019 to December 2022 increases overall civilian employment by 1.24%. Source: BLS (published January 2023), BLS QCEW, author’s calculation, Bureau of Labor Statistics.
The seasonal adjustment issue is not a huge issue to me, as the seasonally adjusted and unseasonally adjusted 12-month (log) changes correspond to January. The higher variability of the family series is evident from the larger fluctuations; this puts the horizontal movement in context. Finally, employment covered by the QCEW fell in the second quarter (the latest available data). However, these data will be modified.
The report highlights the limitations of using the ADP series to track what happened to the BLS series on private NFPs (443K vs consensus 190K).
figure 1: Month-on-month growth rates for BLS private nonfarm payrolls (blue) and ADP private nonfarm payrolls (yellow-green), both calculated as log first differences. Sources: BLS and ADP, from FRED, and authors’ calculations.
