Last Friday (August 6, 2021), the US Bureau of Labor Statistics (BLS) released their latest labour market data – Employment Situation Summary – July 2021 – which reported a total payroll employment rise of 943,000 and a 0.5 points decline in the official unemployment rate to 5.4 per cent. The results are strong even though the US labour market is still 5,702 thousand jobs short from where it was at the end of February 2020, which helps to explain why there are no fundamental wage pressures emerging.
Overview for July 2021:
- Payroll employment increased by 943,000.
- Total labour force survey employment rose by 1,043 thousand net (0.69 per cent).
- The seasonally adjusted labour force rose by 261 thousand (0.16 per cent).
- Official unemployment fell by 782 thousand to 8,702 thousand.
- The official unemployment rate fell by 0.5 points to 5.4 per cent.
- The participation rate rose 1 point 61.7 per cent.
- The broad labour underutilisation measure (U6) fell by 0.6 points to 9.2 per cent.
For those who are confused about the difference between the payroll (establishment) data and the household survey data you should read this blog post – US labour market is in a deplorable state – where I explain the differences in detail.
Payroll employment trends
The BLS noted that:
Total nonfarm payroll employment rose by 943,000 in July, following a similar increase in June (+938,000). Nonfarm payroll employment in July is up by 16.7 million since April 2020 but is down by 5.7 million, or 3.7 percent, from its pre-pandemic level in February 2020.
In July, employment in leisure and hospitality increased by 380,000 … employment in leisure and hospitality is down by 1.7 million, or 10.3 percent, from its level in February 2020.
In July, employment rose by 221,000 in local government education and by 40,000 in private education … Since February 2020, employment is down by 205,000 in local government education and 207,000 in private education.
Employment in professional and business services rose by 60,000 in July … … [but] … is down by 556,000 since February 2020.
Transportation and warehousing added 50,000 jobs in July … the industry has recovered 92.9 percent of the jobs lost during the February-April 2020 recession (-575,000).
The other services industry added 39,000 jobs in July … … [but] … is 236,000 lower than in February 2020.
Health care added 37,000 jobs in July … [but] … is down by 502,000 since February 2020.
Employment in manufacturing increased by 27,000 in July … [but] … is 433,000 below its February 2020 level.
Employment in information increased by 24,000 over the month … [but] … is down by 172,000 since February 2020.
Employment in financial activities rose by 22,000 over the month … [but] … is down by 48,000 since February 2020.
Employment in mining increased by 7,000 in July … [but] … is 103,000 below a peak in January 2019.
Employment in retail trade changed little in July (-6,000) … [but] … is down by 270,000.
The first graph shows the monthly change in payroll employment (in thousands, expressed as a 3-month moving average to take out the monthly noise). The gray lines are the annual averages.
The data swings are still large and dwarf the past history.
The US labour market is still 5,702 thousand jobs short from where it was at the end of February 2020 and the commentary from the BLS above tells us how this shortfall is distributed across the sectors.
The next graph shows the same data in a different way – in this case the graph shows the average net monthly change in payroll employment (actual) for the calendar years from 2005 to 2021.
The final average for 2019 was 168 thousand.
The final average for 2020 was -785 thousand.
The average for 2021 (so far) is 617 thousand.
Labour Force Survey – employment rises strongly
The data for July 2021 reveals:
1. Employment as measured by the household survey rose by 1,043 thousand net (0.69 per cent).
2. The labour force rose by rose by 261 thousand (0.16 per cent).
3. The participation rate rose 1 point 61.7 per cent.
4. As a result (in accounting terms), total measured unemployment fell by 782 thousand to 8,702 thousand and the unemployment rate fell by 0.5 points to 5.4 per cent.
The BLS note that both unemployment measures:
The unemployment rate declined by 0.5 percentage point to 5.4 percent in July, and the number of unemployed persons fell by 782,000 to 8.7 million. These measures are down considerably from their highs at the end of the February-April 2020 recession. However, they remain well above their levels prior to the coronavirus (COVID-19) pandemic (3.5 percent and 5.7 million, respectively, in February 2020). …
The number of long-term unemployed (those jobless for 27 weeks or more) decreased by 560,000 in July to 3.4 million but is 2.3 million higher than in February 2020. These long-term unemployed accounted for 39.3 percent of the total unemployed in July.
The following graph shows the monthly employment growth since January 2008, which shows the massive disruption this sickness has caused.
To put the recent period in perspective I took out the extreme observations (outliers) between March 2020 and October 2020 and repeated the graph.
This graph shows the current recovery more realistically.
The Employment-Population ratio is a good measure of the strength of the labour market because the movements are relatively unambiguous because the denominator population is not particularly sensitive to the cycle (unlike the labour force).
The following graph shows the US Employment-Population from January 1950 to July 2021.
While the ratio fluctuates a little, the April 2020 ratio fell by 8.6 points to 51.3 per cent, which is the largest monthly fall since the sample began in January 1948.
In July 2021, the ratio rose by 0.4 points to 58.4 per cent.
It is still well down on the level in February 2020 (61.1 per cent).
As a matter of history, the following graph shows employment indexes for the US (from US Bureau of Labor Statistics data) for the five NBER recessions since the mid-1970s and the current 2020-COVID crisis.
They are indexed at the employment peak in each case and we trace the data out for each episode until one month before the next peak.
So you get an idea of:
1. The amplitude (depth) of each cycle in employment terms.
2. The length of the cycle in months from peak-trough-peak.
The early 1980s recession was in two parts – a short downturn in 1981, which was followed by a second major downturn 12 months later in July 1982 which then endured.
1. Return to peak for the GFC was after 78 months.
2. The previous recessions have returned to the 100 index value after around 30 to 34 months.
3. Even at the end of the GFC cycle (146 months), total employment in the US had still only risen by 8.3 per cent (since December 2007), which is a very moderate growth path as is shown in the graph.
The COVID collapse was something else but the recent gains are evident, although the gains have been slowing relative to early in the recovery.
Unemployment and underutilisation trends
The first graph shows the official unemployment rate since January 1994.
The official unemployment rate fell by 0.5 points to 5.4 per cent in July 2021 (a significant fall).
The official unemployment rate is a narrow measure of labour wastage, which means that a strict comparison with the 1960s, for example, in terms of how tight the labour market, has to take into account broader measures of labour underutilisation.
The next graph shows the BLS measure U6, which is defined as:
Total unemployed, plus all marginally attached workers plus total employed part time for economic reasons, as a percent of all civilian labor force plus all marginally attached workers.
It is thus the broadest quantitative measure of labour underutilisation that the BLS publish.
Pre-COVID, U6 was at 6.8 per cent (December 2019).
In July 2021 the U6 measure was 9.2 per cent, a decline of 0.6 points on the previous month.
Most of this fall was due to the decline in official unemployment, given that the part-time for economic reasons cohort (the US indicator of underemployment) was “about unchanged” according to the BLS.
Ethnicity and Education
The next graph shows the evolution of unemployment rates for three cohorts based on educational attainment: (a) those with less than high school completion; (b) high school graduates; and (c) university graduates.
As usual, when there is a crisis, the least educated suffer disproportionately.
In the collapse in employment, the unemployment rates rose by:
- 14.1 points for those with less than high-school diploma.
- 13.0 points for high school, no college graduates.
- 5.9 points for those with university degrees.
The period since April 2020 has seen the unemployment rate fall by:
- 11.5 points for those with less than high-school diploma meaning the unemployment rate is now 3.8 points above the March level.
- 11 points for high school, no college graduates meaning the unemployment rate is now 4.6 points above the March level.
- 5.3 points for those with university degrees meaning the unemployment rate is now 2.3 points above the March level.
In the last month, the change in the unemployment rate has been:
- a fall of 0.7 points for those with less than high-school diploma.
- a fall of 0.7 points for high school, no college graduates.
- a fall of 0.4 points for those with university degrees.
In the US context, the trends in trends in unemployment by ethnicity are interesting.
Two questions arise:
1. How have the Black and African American and White unemployment rate fared in the post-GFC period?
2. How has the relationship between the Black and African American unemployment rate and the White unemployment rate changed since the GFC?
1. All the series move together as economic activity cycles. The data also moves around a lot on a monthly basis.
2. The Black and African American unemployment rate was 6.8 per cent in March 2020, rose to 16.7 per cent in May and is at 8.2 per cent in July 2021. In the last month, it fell by 1 point.
3. The Hispanic or Latino unemployment rate was 6 per cent in March 2020, rose to 18.9 per cent in April and is at 6.6 per cent in July 2021. In the last month, it fell by 0.8 points.
4. The White unemployment rate was 3.9 per cent in March 2020, rose to 14.1 per cent in April and fell to 4.8 per cent in July 2021. In the last month, it fell by 0.4 points.
The next graph shows the Black and African American unemployment rate to White unemployment rate (ratio) from January 2018, when the White unemployment rate was at 3.5 per cent and the Black or African American rate was at 7.5 per cent.
This graph allows us to see whether the relative position of the two cohorts has changed since the crisis.
If it is rising, then the unemployment rate of the Black and African American cohort is either rising faster than the white unemployment rate or falling more slowly (or a combination of that relativity).
While there is month-to-month variability, the data shows that, in fact, through to mid-2019, the position of Black and African Americans had improved in relative terms (to Whites), although that just reflected the fact that the White unemployment was so low that employers were forced to take on other ‘less preferred’ workers if they wanted to maintain growth.
In April 2019, the ratio was 2.1 (meaning the Black and African American unemployment rate was more than 2 times the White rate).
By April 2020, the ratio had fallen to its lowest level of 1.2, reflecting the improved relative Black and African American position.
As the pandemic hit, the ratio rose and peaked at 1.8 in October 2020.
In July 2021, the ratio was 1.71 – a fall of 0.6 points from the previous month, which indicates that the relative position of the Black and African American cohort is improving.
Special Topic: Business Employment Dynamics
The BLS publish a quarterly dataset derived from the Business Employment Dynamics survey of private sector establishments.
The latest publication (published July 28, 2021) – Business Employment Dynamics Summary – for the fourth-quarter 2020 provides some interesting insights into how much change goes on in the labour market relative to the net changes reported in the labour force survey and payroll data that I analysed above.
In the December-quarter 2020, employment grew in net terms by 2,043 thousand using the establishment survey.
However, underpinning those changes, the BED data tells us a very richer story.
There were 8,756 thousand gross job gains and 6,713 thousand gross job losses.
So, a much larger amount of job creation and job destruction was going on.
Even during the March-quarter 2020, when the pandemic started to really hit, the aggregate data reported a loss of jobs of 773 thousand overall in net terms.
But driving that net figure, we know that there were 6,963 thousand gross job gained which were more than offset by the 7,736 thousand jobs lost.
The net figure reported is the difference.
The point is that even as the economy was shedding a massive number of jobs in the March-quarter, there were thousands of gross jobs created.
The following graph tracks the Gross Job Gains and Losses from the March-quarter 1995 to the December-quarter 2020 (latest).
The difference between the lines is the net rise or fall in employment, such that if the blue line is above the orange line, then employment is rising and vice versa.
The George Bush recession in the early 2000s and the GFC is notable.
Another interesting insight is available because the data disaggregates the gross job gains into those job increases in existing firms that are expanding and new entrants who have just opened.
Similarly, the BLS disaggregate the gross job losses into those from existing firms who are contracting and firms who disappear altogether (close down).
So in the December-quarter, 81 per cent of the gross job gains came from existing firms expanding, and, 80 per cent of the gross job losses came from existing firms contracting.
Those ratios are very stable
We can also examine the employment dynamics by firm size.
The following table shows the proportion of total gross job gains and losses by firms size and in the final column, the share of existing employment by each size class.
You can see there is a lot of job creation and destruction at either end of the firm size distribution but less so in the mid-range firm sizes.
I will leave the implications of that for another day.
The July 2021 US BLS labour market data release suggests that the labour market continues its recovery although there is still a long way to go.
The US labour market is still 5,702 thousand jobs short from where it was at the end of February 2020.
That is enough for today!
(c) Copyright 2021 William Mitchell. All Rights Reserved.