Tuesday 15 June 2021
Decoding - The U.S. labor market 'puzzle'
Both beginning and end of cycle characteristics
The April and May job reports disappointed, as job creations have been well below expectations, as the unemployment rate declined only very slowly and as the participation rate did not rebound. This would suggest that the US labor market is in a configuration typical of the early stages of the economic cycle. However, other indicators are consistent with an end-of-cycle configuration, which makes the analysis of the labor market more complicated than usual.
Job creations have disappointed in April and May
Assuming that the statistics are not biased (we will see at the end of the text that some questions may arise), job creations disappointed in April and May (respectively 278,000 and 559,000 non-farm jobs), coming out clearly below expectations (nearly 1 million in April and 650,000 in May). The unemployment rate fell only slightly, to 5.8% of the labor force in May (and even 8% if we take into account the 3.5 million people who had left the labor force since February 2020). Intuitively, the fact that there are so many job seekers is reminiscent of the post-recession phases.
One of the striking points about the US labor market is that the number of people in the labor force remains 3.5 million lower than in February 2020. One-third of these are above 55, which can be explained by early retirements (permanent withdrawal from the labor market) or by the fear of covid, this age category being more threatened than the others. Besides this age group, the participation rate rebounded the least for the 25-34 age group, probably because of childcare problems. In any case, it is really essential to look at the participation rate age group by age group because the participation rate calculated by the BLS makes less and less sense .
The participation rate calculated by the BLS no longer makes sense
Each month, one of the most commented variables in the jobs report is the participation rate, which is the ratio of the labor force to the total population. The "catch" is that for the total population, the BLS considers the population aged 16 and over, with no upper limit, but population ageing means that the working-age population (15 to 64 years in Eurostat statistics or the OECD) has peaked in recent quarters and is starting to decline even as the population aged 16 and over continues to grow rapidly.
Thus, the participation rate published by the BLS, which has stagnated for months, had also not recovered over the 2015-2019 period, while conditions in the labor market had improved markedly. On the contrary, a participation rate calculated on the basis of the working-age population was well consistent with such an improvement over the period and has risen gradually over the past year.
Finally, even if job creations were lower than expected, it is important to note that their pace has nevertheless been robust since the reopening in February (strong relaxation of health restrictions, in particular in the Republican States), with an average of 539,000 per month for the February to May period. Staying at this pace, employment should return to its February 2020 level by summer 2022.
But other indicators have some characteristics of the end-of-cycle periods
This may seem paradoxical in a period of high unemployment, but in recent weeks, press articles have mostly reported… on the hiring difficulties experienced by companies. In reality, the job supply observable today is consistent with an end-of-cycle period and this is understandable: in a certain number of sectors, the pandemic was managed quite smoothly with the introduction of teleworking or health protocols. Jerome Powell himself had underlined this "adaptation" during the FOMC press conference in January. In May, 48% of SMEs indicated that they had unfilled job openings, the highest proportion since the creation of this survey in 1975, whereas it is generally very low at the start of the cycle...
Read the full document below
Chief Thematic Macro Strategist