I sometimes see this statement. Given the pattern of our data as of June 2022, it’s clearly relevant. Red arrows indicate local maxima.
figure 1: Civilian employment (blue, left log scale), nonfarm payrolls (tan, right log scale) as of July 24, 2022, both at 1000 years old, seasonally adjusted. Source: BLS via FRED.
This argument is based on conjectures from the BLS birth/death model (see Discussion here) used to process data from corporate surveys are most likely to go off track at inflection points, especially in The Great Depression and the Covid Pandemic.
Looking back at the past four recessions, we see the following patterns:
figure 2: Civilian employment (blue, left log scale), nonfarm payrolls (tan, right log scale) as of July 24, 2022, both at 1000 years old, seasonally adjusted. The NBER uses shades of grey to define the peak and trough dates of the recession. Source: BLS via FRED, NBER.
image 3: Civilian employment (blue, left log scale), nonfarm payrolls (tan, right log scale) as of July 24, 2022, both at 1000 years old, seasonally adjusted. The NBER uses shades of grey to define the peak and trough dates of the recession. Source: BLS via FRED, NBER.
Figure 4: Civilian employment (blue, left log scale), nonfarm payrolls (tan, right log scale) as of July 24, 2022, both at 1000 years old, seasonally adjusted. The NBER uses shades of grey to define the peak and trough dates of the recession. Source: BLS via FRED, NBER.
Figure 5: Civilian employment (blue, left log scale), nonfarm payrolls (tan, right log scale) as of July 24, 2022, both at 1000 years old, seasonally adjusted. The NBER uses shades of grey to define the peak and trough dates of the recession. Source: BLS via FRED, NBER.
Only in one instance did the civilian employment series (household survey) peak before the nonfarm payrolls series (institutional survey). Now, these are the final revised series. What does the series look like in “real time”, i.e. observers assessing the onset of a recession at the same time? The corresponding graph is shown in Figure 6-9.
Image 6: Civilian employment (blue, left log scale), nonfarm payrolls (tan, right log scale) as of June 7, 1990, both in 1000, seasonally adjusted. The NBER uses shades of grey to define the peak and trough dates of the recession. Source: BLS via ALFRED, NBER.
Figure 7: As of May 4, 2001, civilian employment (blue, left logarithmic scale), non-farm employment (tan, right logarithmic scale), both at 1000 years old, seasonally adjusted, as of May 4, 2001. The NBER uses shades of grey to define the peak and trough dates of the recession. Source: BLS via ALFRED, NBER.
Figure 8: As of February 1, 2008, civilian employment (blue, left logarithmic scale), nonfarm payrolls (tan, right logarithmic scale), both at 1,000 years old, are seasonally adjusted as of February 1, 2008. The NBER uses shades of grey to define the peak and trough dates of the recession. Source: BLS via ALFRED, NBER.
Figure 9: As of April 3, 2020, civilian employment over 16 (blue, left log scale), nonfarm payrolls (tan, right log scale), both at age 1000, seasonally adjusted. The NBER uses shades of grey to define the peak and trough dates of the recession. Source: BLS via ALFRED, NBER.
In real time, there were two times (2001, 2007) that the number of households turned a month earlier than the establishment, and the two turning points were at the same time (1990 and 2020). In the revision, the peak of the civilian series was advanced once (1990) and then delayed by two months (2001). The NFP peak is shifted back once (2001 recession).
I don’t think people will get much from the flattening of the current family series. With that in mind, here is a review of both series, as well as the family series adjusted to the NFP concept. The latter series was developed in response to criticism from conservatives that the traditionally built series that relied on the life/death model was error-prone after the 2001 recession (ie, critics argued that employment was undervalued). The study series, shown in red below, does not encounter the same problems in estimating the births and deaths of companies (but does have a relatively small sample size associated with household surveys).
Figure 10: Civilian Employment Over 16 (blue), Civilian Employment Adjusted for Nonfarm Wages Concept (BLS Series LNS16000000) (red), Nonfarm Employment (tan), Bloomberg 7/25 Consensus (tan squared), all in 1000 log scale, seasonally adjusted, as of July 24, 2022. Source: BLS calculations by FRED, BLS, Bloomberg and authors.
Civilian employment fell in June, while civilian employment adjusted for the NFP concept rose (-315, +131), while NFP rose 372. As of today, the Bloomberg consensus for July is +255.
