I already discussed the July employment situation in a previous post but I found interesting to make a final update taking into account the release of several data, analysis and comments from Donald Trump. Before starting, it’s important to remember that the U.S. employment report presents statistics from two major surveys, the Current Population Survey (CPS; household survey) and the Current Employment Statistics survey (CES; establishment survey). The most important thing is that, according to the BLS, “For both surveys, the data for a given month relate to a particular week or pay period. In the household survey, the reference period is generally the calendar week that contains the 12th day of the month. In the establishment survey, the reference period is the pay period including the 12th, which may or may not correspond directly to the calendar week.”
Looking at high frequency indicators, the signals range from slower improvement to a deterioration. The Census Bureau’s weekly Household Pulse Survey showed that the number of employed Americans declined by about 6.7 million from mid-June through mid-July, including a 4.1 million plunge from the first to the second week of July.
In addition, real-time labour estimates produced by academics at Arizona State University and Virginia Commonwealth University showed that the proportion who are employed fell again in July.
Source: Bick A. and Blandin A. (2020), Real-Time Labor Market Estimates During the 2020 Coronavirus Outbreak∗
Meanwhile, Saint Louis Fed economists, that use real-time data from Homebase to track labour market trend, show “that since the week of June 12, the recovery in employment has slowed down and slightly reverted.” However, as you can see below, by comparing the weeks related to the survey periods, the situation improved a bit from June to July.
Sources: Current Population Survey/IPUMS, Homebase and authors’ calculations (Dvorkin A. and Bharadwaj A.)
The Labor Department data also show that employment conditions started deteriorating in mid-July with initial jobless claims snapping a 15-week streak of declines in the week ending July 18. However, latest figures showed an improvement by month-end. Applications for U.S. unemployment benefits fell by 249,000 to 1.19 million in the week ended Aug. 1 (the lowest since the pandemic started).
— Christophe Barraud🛢 (@C_Barraud) August 6, 2020
Now looking at monthly surveys, results were somehow disappointing. On the one hand, ADP reported Wednesday that private payrolls increased by just 167,000 well below the Bloomberg consensus of 1.2 million and prior figure of 4.314 million in June. In the meantime, the two ISM surveys (Manufacturing and Services) showed that the employment component remained in contraction territory in July, respectively at 44.3 and 42.1 (well below the expansion threshold of 50).
All in all, most of employment indicators imply that the current consensus for NFP is probably too high at +1.5 million (downwardly revised from +2.0 million), yet, the bright spot is likely to come from State and Local Government (Education) jobs amid positive seasonal adjustment. Bill McBride (Calculated Risk Blog) was the first to flag that point. He noted that “Every year, state and local governments let about 2 million teachers go in late Spring, and then hire them back at the end of Summer. Since this happens every year, the BLS adjusts for this seasonal pattern in the monthly employment report. However, in 2020, state and local governments let almost 1.2 million teachers go in March, April and May, Not Seasonally Adjusted (NSA)”. It means that after seasonal adjustment, we can expect a sharp bounce in this category.
— Bill McBride (@calculatedrisk) August 2, 2020
The other contrarian point came from Trump statements. President Trump said Tuesday and reiterated Wednesday that there is “a big number coming out Friday on jobs”. While presidents usually receive a preview of the employment report before its official release, it tends to be the prior evening.
To summarize, high frequency and monthly indicators suggest that employment growth was weaker in July with some deterioration seen around mid-July. Due to the survey period, this phenomenom is unlikely to be fully reflected in the report. In addition, seasonal adjustment issues related to Education will boost artificially the headline by several hundred thousand jobs in July. Trump positive comments also imply that the worst scenario (negative print) is very unlikely.