US jobs preview

Via Calculated Risk:

Important Notes:
1. The BLS reference week includes the 12th of the month. Massive COVID-19 layoffs started after the reference week (although there was a pickup in layoffs during the reference week).
2. Watch for Special Notes in the release. There could be some important announcements on how the BLS will be handling unemployment numbers and seasonal adjustments.
3. The 2020 Decennial Census was expected to increase hiring in March. This is unclear now – some of the hiring will be delayed.

On Friday at 8:30 AM ET, the BLS will release the employment report for March. The consensus is for a decrease of 100,000 non-farm payroll jobs, and for the unemployment rate to increase to 3.9%.

Last month, the BLS reported 273,000 jobs added in February (266,000 ex-Census).

Here is a summary of recent data:

• The ADP employment report showed a decrease of 27,000 private sector payroll jobs in March. This was above consensus expectations of 154,000 private sector payroll jobs lost. The ADP report hasn’t been very useful in predicting the BLS report for any one month, but in general, this suggests employment growth somewhat above expectations.

• The ISM manufacturing employment index decreased in March to 43.8%. A historical correlation between the ISM manufacturing employment index and the BLS employment report for manufacturing, suggests that private sector BLS manufacturing payroll decreased around 55,000 in March. The ADP report indicated manufacturing jobs increased 6,000 in March.

The ISM non-manufacturing employment index has not been released yet.

• Initial weekly unemployment claims averaged 2.6 million in March, way up from 213,000 in February. For the BLS reference week (includes the 12th of the month), initial claims were at 282,000, up from 211,000 during the reference week the previous month.

This suggests more layoffs (during the reference week) in March than in February.

• The final March University of Michigan consumer sentiment index decreased to 89.1 from the February reading of 101.0. Sentiment is frequently coincident with changes in the labor market, but there are other factors too like gasoline prices and politics. The decline this month was related to the pandemic.

• The BofA job tracker decreased in March to 103,000, down from 144,000 in February, suggesting fewer jobs added in March.  However, according to the BofA, this data is “stale”.

• Weather: The weather was favorable in both January and February.   It is likely some expected hiring for March was pulled forward to the previous two months, suggesting some payback in the March report.  So there was some chance that the March report would have been weaker than many expected without COVID-19.

• Conclusion: If we look back at 2005, when weekly claims for the reference week jumped following hurricane Katrina (similar to what happened in March), the economy lost 35 thousand jobs.    It is possible that job losses will be that small in the March report – it is also possible that losses could be well over 100K.   One thing is clear, job losses in the April report will be off the chart.

David Llewellyn-Smith

Comments

  1. This will depend crucially on the layoff timing and the survey week. So the surge in claims may not show up fully, at least straight away. Also, companies in the Establishment survey May record a worker as working but they may themselves report in the Household survey as not.

    So the Employment Report is likely to be a dogs breakfast. The key point is simple; unemployment has absolutely exploded in a manner unseen at any point in history.

  2. Yep, pretty much what happened. The Establishment survey showed 700k jobs lost – well below initial claims – hasn’t captured the latest results due to survey week – and the household survey was much worse down 1.4m sending unemployment up 0.9%m/m. This is just the beginning. The next report is gonna be horrible.