Today’s U.S. employment report is not as good as it seems
Head of Macro Analysis
Summary: The U.S. August employment report published by the Bureau of Labour Statistics was the most awaited indicator this week. The initial market reaction to the decrease in the unemployment rate to 8.4% vs est 9.8% and 10.2% prior was positive but, if we dig a bit into data, we realize that the non-farm payrolls report is not as good as it seems and there are many caveats.
The decrease in the unemployment rate to 8.4% vs est 9.8% and the addition of 1371k new jobs in August vs est 1350k may come as a surprise to many. Indeed, it follows the recent release of a lot of business cycle indicators and labor market statistics indicating evidence of deceleration in economic activity in August and in September in the United States. Just in the last few days, we had three clear indications of labor market slowdown in the previous month: (1) The August ISM Services Employment subindex, which is tracked by economists to forecast the evolution of the official unemployment rate, was out at 47.9, suggestion a net contraction of hiring; (2) the increase in private employment in services and in the industry slowed down in August based on the latest ADP National Employment Report; and (3) jobless claims barely came down in August and PUA claims in (which refers to the federal Pandemic Unemployment Assistance program) spiked in the week ending August 29 at 29 million (+2m from the previous week).
Therefore, today’s BLS report might be at first a bit confusing. Actually, digging into data, it also confirms what other statistics have pointed out recently: the improvement in the labor market has stalled following the initial V-shaped rebound that occurred when the lockdown has been lifted. In order to remove noise from monthly data released by the BLS, we prefer to refer to three-month modified moving average (mma) data in these unusual circumstances. Using this method, its appears that the NFP report is not glorious at all. The three-month mma is down at 2628k from 3080k in July (see the below chart). In addition, if we subtract the hiring of new temporary Census takers (+237,800 in August to a total of 288,204), the non-farm payrolls decrease at 1133k, which is significantly lower than the official headline. Actually, employment shortfall remains a severe issue in many sectors, such as hospitality and leisure, and almost in all sectors hiring is still well below pre-COVID level.
But there is worst reinforcing the idea of long-term scarring:
- Unemployment duration is jumping fast with the share of those unemployed for over 15 weeks increasing to 60%.
- The number of people that permanently lost their job rose from 3400k from 1200k at the beginning of the year and all indicates that it will get much worse in the coming months.
The reality of the U.S. labor market is that many furloughed people are not being re-employed due to the spread of the virus and persistent economic uncertainty in many states. In this context, we increasingly think that the US government will have no other choice but to extend as long as necessary furlough scheme (especially via the PUA) in order to avoid that a massive part of workers falls into poverty, thus increasing the U.S. inequality issue. The exact same situation is already happening in Europe where Spain’s labour minister has confirmed a few days ago that the erte furlough scheme does not have expiry date. Basically, all of this is just Universal Basic Income or UBI through the backdoor, which is being slowly implemented on both side of the Atlantic.