Some commentators have focused on the sharp decline in civilian employment and the ISM services employment diffusion index, arguing that this indicates more weakness than agency surveys indicate.
In particular, the fact that the BLS's birth-death model captures a large share of total growth is concerning (see Forbes,For example). In fact, nonfarm payrolls data have been revised downwards over the past three months.
figure 1: Non-agricultural employment situation in different periods. Source: BLS via FRED.
The list of numbers in the birth-death model is here.
The different behavior of the NFP versus the BLS agency survey (blue line) illustrates the argument for overall labor market weakness, in which case I have included preliminary baseline revisions to the NFP level. In contrast, civilian employment in the household survey (green line) and the diffusion index component in the ISM PMI services index (tan line, left scale, below 0.5 indicates contraction) both plunged sharply.
figure 2: ISM PMI Services Employment Diffusion Index (tan, left scale, contraction below 0.5), preliminary NFP benchmark (blue, right scale), private NFP from ADP (red, right scale), and civilians from BLS Household Survey employment (green, right scale) scale), all recorded in log 2023M03=0. Sources: ISM (from TradingEconomics.com), BLS and author's calculations, ADP (from FRED), BLS (from FRED) and author's calculations.
Therefore, including the PMI indicator results in a weaker labor market than other indicators. On the other hand, the ADP indicator that continued to rise in December was not based on surveys but actual cut data. The fact that it rose in tandem with the Bureau of Labor Statistics Establishment Survey further supports the view that job growth continues.
Note further that the Philadelphia Fed's early benchmark, which contains quarterly census data on employment and wages Shows continued growth throughout June (in line with non-farm payrolls data) (Admittedly, this was 6 months ago).
