Recession Risk - Is it High?

Manufacturing Labor Market

Growth in hours worked gives us another layer of information on the labor market besides just growth in payrolls. It combines the length of the workweek with growth in payrolls. In the August BLS report, aggregate hours worked in the manufacturing sector improved 0.5% monthly. That was the largest increase since June 2018. Improvement was partially due to the easy comp of -0.7% monthly growth in July. It was the weakest reading since November 2011.

Improvement wasn’t all because of the comp though as yearly growth improved from -0.3% to 0.1% despite a tougher comp. Growth in non-supervisory aggregate weekly hours worked wasn’t as strong which signals the pay growth went to managers. Low paying workers didn’t get as much of the gains. Specifically, monthly growth was 0.2% and yearly growth was -0.8%. This weak result is correlated with the contraction in manufacturing production in July and the weakness in overall manufacturing job growth.

Finally, we look at look at overtime hours worked. For all private sector manufacturing workers, average weekly overtime hours worked fell from 3.3 hours to 3.2 hours. That’s the lowest total since April 2017. For production and non-supervisory workers, it stayed at 4.2 hours which is below last year’s peak of 4.7 hours. On a non-seasonally adjusted basis, it increased from 4.3 hours to 4.1. As you can see from the bottom chart below, non-seasonally adjusted yearly growth was -8.5%.  

Recession Indicators

In the graphic above, the chart on the right shows macro, valuation, liquidity, and sentiment percentiles. The higher the percentile, the higher the recession risk. As you can see, the current percentile is about 90%. It’s way higher than in April 1995 and April 1985. I don’t agree with the usage of the categories which is why I don’t see a recession coming soon. However, I must show the bearish thesis to be fair.

Firstly, I disagree with the duration of the cycle being used as a recession indicator. Just because a cycle is long doesn’t mean total growth was high. Total growth isn’t impressive this cycle because average yearly growth has been the lowest in many decades. Wage growth is high, but there isn’t high inflation, so I discount this category. 

That’s the same reason I discount labor market conditions. I don’t follow CAPE closely because it has been elevated for the past few decades. In this cycle, it was boosted by the extremely low 2009 earnings. Regardless of whether you think the Fed got too hawkish last year. It has been supporting markets this year. That discounts the monetary conditions category.

As you can see from the chart below, manufacturing new orders have been declining just like yearly growth in manufacturing overtime hours. The manufacturing new orders index of 47.2 is closing in on the recession threshold. Obviously, the counter argument here is August non-manufacturing new orders were strong as the index increased 6.2 points to 60.3.

Less People Don’t Want A Job Now

Consistent with the increase in the prime age labor force participation rate which matched its cycle high in August at 82.6%, the chart below shows the 12 month average of the percentage of the prime age population that isn’t in the labor force that doesn’t want a job now fell to 16%. It’s amazing how high this reached a few years ago. There was about a 2% increase in those not wanting a job from the end of the recession to 2016. 

Based on the past 2 cycles, it seems like the percentage increases during the beginning of the expansion and peaks during the middle of it. Given how relatively full the labor market is now, it’s surprising this many people don’t want a job. It might be pushed higher by the percentage of millennials still continuing their education at that age.

Let’s look at aggregate weekly hours worked for the entire private sector. Growth was relatively weak here too. Yearly growth improved from 1.1% to 1.2%. A 1.1% reading in July was the lowest yearly growth rate since August 2016. That 2016 growth rate was the weakest reading of the cycle. Fullness of the labor market is suppressing job growth. To be clear, the weakness in demand growth is also playing a role. It’s a combination of both.

Results were worse for production and non-supervisory workers. Yearly growth improved from 0.7% to 0.8%. July’s reading was the lowest of the cycle. Growth in both manufacturing and the overall labor market has been worse for non-supervisory workers. On the other hand, these workers had higher wage growth. Overall, hourly wage growth was 3.2%. These workers had 3.5% growth which was the highest since December 2018 (the highest growth of the expansion).

Amazon To Hire 30,000 Workers

In the August Job Cut report, there were only 25,000 hires announced. That will change in the September reading because Amazon just announced it is hiring 30,000 full and part time workers and will hold a career day on September 17th. The firm will pay workers at least $15 per hour and give them benefits. It also includes on the job training which makes sense in a tight labor market. 

These jobs will be in corporate offices, tech, and customer fulfillment. I’m guessing most of the jobs will be at fulfillment centers. These are very tough jobs which should be replaced by robots over the next 10-20 years. Even if these aren’t the best jobs, it’s still good for all workers because it limits the supply of people looking for work which helps boost wage growth.


The labor market is mostly full and demand growth is weakening, so we are seeing some weakness in many labor market indicators. Good news is wage growth is very strong, so the consumer should have a strong holiday season absent the tariffs. Tariffs will be the wild card in this scenario. 

Currently, the trade war is also causing marginal weakness in the labor market. The Job Cut report stated the trade war caused 10,488 cuts. That number should increase in September with the latest round of tariffs going into effect. 

Spread the love

Comments are closed.