Yellow flags from the November jobs report

 

 - by New Deal democrat

Most of the commentary about Friday’s jobs report for November was positive. By contrast, my summary - in which I averaged the two last reports to take into account the hurricane whipsaw - was much more cautious, as were the takes by a few other commentators I respect, like Ernie Tedeschi.

 

In this post I am going to delve into more detail into why I believe it is now prudent to raise the yellow caution flag about employment.

Let’s start with the totals. For many months I and many others have written that the trend in the Household Survey (red in the graph below) has been frankly recessionary - but is probably skewed to the downside by failing to take into account the surge in new jobs entrants caused by the post-pandemic immigration spike. On a YoY% basis, for the second time in three months, the Household Survey recorded a net *decrease* in jobs.

Meanwhile the Establishment Survey (blue) has been much more positive, showing jobs gains in every single month. As of November, the YoY% growth rate was 1.45%. The below long term graph subtracts 1.45% from the Establishment number and adds 0.4% to the Household number to show both current YoY levels at the zero line: 

Unsurprisingly, with the exception of one month in 1952, any time the YoY change in jobs in the Household Report has been this low, it has been because of a recession. 

What is more concerning is that, with the exception of 1952 and a number of months in the decade before the pandemic, the same has been true of gains of only 1.45% YoY in the Establishment Survey as well. Here’s a close-up of that decade:

YoY employment gains of the current magnitude or less were only measured for one month in 2013, several months near the end of 2017, and during 2019 when contemporaneously I was worried about whether a recession was in the offing.

One important consideration is population growth over this long period of time. A gain of 100,000 jobs in a month now is likely very different than a 100,000 gain back when the US population was half of what it is now.

The next graph corrects for that, subtracting the YoY% gain in the prime working at population from the YoY% in job growth. The result is the net YoY% gain over and above the prime working age population for the period:

Except during the 1970s and 1980s, when women were entering the labor force by the millions, a 0.9% YoY net gain has almost always meant a recession. Even during the 10 years before the pandemic, there were only 4 months during 2018 when the YoY gain was so low.

If that weren’t concerning enough, there is good reason to believe that job gains in the Establishment Survey are going to be revised lower for 2024. That’s because the QCEW, which is not a sample but an actual census of about 95% of all firms, and to which the jobs survey is benchmarked twice a year, has shown a great deal more slowing in the past 18 months. Here’s Prof. Menzie Chinn’s most recent update from Econbrowser:

The QCEW unfortunately is not seasonally adjusted, so the best way to compare that and the 2024 payrolls numbers is YoY. This shows a stark difference.

In June 2023, the QCEW showed a 2.5% job gain. As benchmarked, nonfarm payrolls show a 2.4% gain. But the latest QCEW report through June 2024 shows only a 0.8% YoY job gain, vs. 1.6% for payrolls through that month. If nonfarm payrolls are similarly re-benchmarked, then the *only* month going back 75 years when such a meager gain did not coincide with a recession was one month in 1952.

Further, every month I update the leading components of the jobs report, which mainly are manufacturing and components of construction jobs, as well as goods-producing jobs as a whole. And for the first time during this recovery, goods-producing jobs as a whole have stopped growing over the last two months. Here’s what they look like post-pandemic:

Since July, only 7,000 goods producing jobs have been added, or only a .03% increase. In the past 8 months, only 39,000 goods producing jobs have been added, an increase of .18%. That isn’t necessarily recessionary. As the longer-term graph below shows, there have been similar stalls in 1995, 1999, and 2016 without recessions following:

But on the other hand, outright declines in goods producing jobs have occurred for at least six months, and sometimes over a year, before about 3/4’s of all recessions going back 75 years:

Indeed, even the current 0.7% YoY gain has almost always in the past meant a recession (blue in the graphs below):

The exception is the 10 years before the pandemic:

Further, if we simply continue the trend growth for the last eight months, that would be a 0.27% job gain in goods producing jobs YoY by March 2025, which would be lower than at any point in the 10 years before the pandemic.

But as the graphs just above also show, job growth in services remains robust, at present up 1.57% YoY. While up until 2000 even that would have typically only occurred in recessions, it has been an average rate of growth since throughout the expansions as well. 

Finally, the stalling out in goods-producing jobs has been exclusively a manufacturing story. As the below graph shows, job gains in construction (dark red, right scale) and residential construction (light red, right scale, *8 for scale) continue:

As I have pointed out many times in discussing housing, residential construction jobs have almost always turned down well in advance of recessions. While housing units under construction are down -15% or so, which typically in the past has coincided with layoffs in residential construction, as of now they certainly have not.

In conclusion, there is sufficient cause for concern to raise a yellow caution flag about the trend in employment growth. But there are nowhere near sufficient reasons to hoist a red warning flag.

Weekly Indicators for December 2 - 6 at Seeking Alpha

 

 - by New Deal democrat

My “Weekly Indicators” post is up at Seeking Alpha.

This week whipsawed the data that was heavily influenced by Thanksgiving week. 

The tone of the short leading and coincident data remains positive. The negativity of much of the long leading data is becoming more problematic.

As usual, clicking over and reading will bring you up to the virtual moment as to the state of the economy, and reward me with a penny or two towards lunch for calculating and organizing it for you.

November jobs report: the expected monthly rebound masks deeper declining trends

 

 - by New Deal democrat



To understand this month’s jobs report, let’s start with last month’s, where I wrote that “there were some signs of real weakness in this report that do not appear to be hurricane-related. But Hurricane Milton, as well as the strike, had an impact, so take this report with a gigantic helping of salt.”

So everyone, including me, expected a big rebound this month, and we got one. As I’ll get into below, though, it is especially important to average the two months together to get a better idea of the trend.

Below is my in depth synopsis.

HEADLINES:

• 227,000 jobs added. Private sector jobs increased 194,000. Government jobs increased by 33,000. the two month average was an increase of +131,500.
• The pattern of downward revisions to the last months reversed this month.. September was revised upward by +32,000, and October by +24,000, for a net increase of +56,000.
• The alternate, and more volatile measure in the household report, showed a decrease of -355,000 jobs. On a YoY basis, this series has *declined* by -725,000 jobs, which remains consistent with recession, as it has for months. This is the second time in three months this measure has shown a YoY decline.
• The U3 unemployment rate rose 0.1% to 4.2%. Since the three month average is 4.167% vs. a low of 3.7% for the three month average in the past 12 months, or an increase of over 0.4%, this means the “Sahm rule” is back in effect.
• The U6 underemployment rate also rose 0.1% to 7.8%, 1.4% above its low of December 2022.
• Further out on the spectrum, those who are not in the labor force but want a job now declined -180,000 to 5.486 million, vs. its post-pandemic low of 4.925 million in early 2023.

Leading employment indicators of a slowdown or recession

These are leading sectors for the economy overall, and help us gauge how much the post-pandemic employment boom is shading towards a downturn. This month they were again mixed, but tilted towards negative:

• the average manufacturing workweek, one of the 10 components of the Index of Leading Indicators, rose 0.1 hour to 40.7 hours. This remains down -0.8 hours from its February 2022 peak of 41.5 hours, but on the other hand is only -0.1 hour below its 18 month high.

• Manufacturing jobs rose 22,000. But this only reversed half of the -44,000 strike-related decline last month, so the two month average is negative.
• Within that sector, motor vehicle manufacturing jobs declined -400. The two month average is -3,200. 
• Truck driving increased 2,900. The two month average is +950.
• Construction jobs increased another 10,000. The two month average is +9,000.
• Residential construction jobs, which are even more leading, rose by 1,400 to another new post-pandemic high.
• Goods producing jobs as a whole rose 34,000, but because they declined -42,000 last month, the two month average is -4,000. This is especially important, because these typically decline before any recession occurs. As I wrote last month, “in the absence of special factors this would be a serious red flag for oncoming recession.” Thus the net two month decline is worth at least a yellow flag.

• Temporary jobs, which have generally been declining since late 2022, rose by 16,000, although the two month average is -850. These are down over -550,000 since their peak in March 2022. This appears to be not just cyclical, but a secular change in trend.
• the number of people unemployed for 5 weeks or fewer rose 97,000 to 2,209,000. The two month average is an increase of +32,500.

Wages of non-managerial workers

• Average Hourly Earnings for Production and Nonsupervisory Personnel increased $.09, or +0.3%, to $30.57, for a YoY gain of +3.9%. Their post pandemic peak of 7.0% in March 2022. This is equal to their recent low in July. Nevertheless, and importantly, this continues to be significantly higher than the 2.6% YoY inflation rate as of last month.

Aggregate hours and wages: 

• The index of aggregate hours worked for non-managerial workers rose 0.1%, vs. last month’s revised unchanged level. This measure remains up 1.4% YoY, which is higher than its trend for the past 12+ months.

• The index of aggregate payrolls for non-managerial workers was rose 0.4%, and is up 5.3% YoY. This increase may be just noise, but at least for this month it reverses the slow deceleration since the end of the pandemic lockdowns. With the latest YoY consumer inflation reading of 2.6%, this remains powerful evidence that average working families have continued to see gains in “real” spending money.

Other significant data:

• Professional and business employment rose 26,000, but the two month average is a decline of -10,500. These tend to be well-paying jobs. Although the YoY comparison therefore improved this month, they are only higher YoY by 0.4% - a very low increase that has *only* happened in the past 80+ years immediately before, during, or after recessions. 
• The employment population ratio declined another -0.2% to 59.8%, after a -0.2% decline last month, vs. 61.1% in February 2020.
• The Labor Force Participation Rate declined another -0.1% to 62.5%, after a -0.1% decline last month, vs. 63.4% in February 2020. The prime 25-54 age  participation rate declined -0.3% to 83.5%, vs. 84.0% in July, which was the highest rate during the entire history of this series except for the late 1990s tech boom.

SUMMARY

On a month over month basis, this report was very positive, as with the exception of the labor force participation rate and the employment population ratio, everything rebounded - as expected.

That’s why looking at the average of the past two months is so important. And there, the news isn’t so good at all. In addition to upticks in the unemployment and underemployment rates, not only did manufacturing, motor vehicle production, professional and business jobs, and temporary help jobs decline further, but for the first time, so did goods-producing jobs as a whole. For the last four months, there has been less than a 0.1% gain, and only a 0.2% gain for the last eight months. Even since the accession of China to regular trading status, such meager gains have signaled at least weakness if not outright recession.

There certainly were bright spots, as construction, including residential construction jobs, continued to plow ahead. The downturn in trucking jobs reversed. Those who want a job now but have not applied for one also decreased. And aggregate hours worked and aggregate payrolls for nonsupervisory workers both increased. This suggests that consumer spending will continue a net positive in the next few months.

Last month I closed with “I would take 60% of this month’s decline as temporary, but 40% real.” This month’s report is confirmatory of that hypothesis, with the two month average gain being 131,500, and the three month average 173,000. In other words, the trend of deceleration in the jobs market is continuing without abatement. If this trend continues for another 12-15 months, it will be negative - in other words, signaling a recession.

Jobless claims: neutral - with an extra grain of salt

 

 - by New Deal democrat

As I cautioned last weekend in my “Weekly Indicators” update, we have entered that period of the year where Holiday seasonality means take everything with at least a little grain of salt. For example, this year Thanksgiving was almost one full week later than lat year.

With that caveat, initial jobless claims for Thanksgiving week this year increased 9,000 to 224,000. The four week moving average increased 750 to 218,250. Continuing claims, with the typical one week lag, declined -25,000 to 1.871 million:

As per usual, the YoY% changes are more important for forecasting purposes. So measured, initial claims were up 3.7%, the four week average up 0.3%, and continuing claims up 2.9%:

On the face of it, these comparisons are a little weak, since they are all higher YoY, but not nearly enough to warrant any special concern. Still, take even that statement with a little extra caution because of seasonality.

Looking at tomorrow’s unemployment rate for November, the suggestion is that absent the impact of immigration unemployment should be in the area of flat to 5% (as a percent of a percent, left scale) higher than one year ago. Since, per the gray line (right scale) which shows the actual unemployment rate, one year ago was 3.7% in November, that means trending towards an unemployment rate of 3.7%-4.0%:

This is all neutral - with a grain of extra salt.

ISM non-manufacturing shows that services continue to power the economy forward. Are they inflationary?

 

 - by New Deal democrat

Because services are roughly 3/4’s of the economy, I now pay a lot of attention to the economically weighted average of the ISM manufacturing and services indexes. Since the accession of China to normal trading status with the US, a downturn in manufacturing alone has simply not been enough to forecast recession - which has again been true in the past two years.

This morning the ISM non-manufacturing (i.e., services) index again came in positive, at 52.1, while the more leading new orders subindex came in at 53.7. Their three month weighted averages are 54.3 and 56.8, respectively.

Since the three month average for the manufacturing index is 47.4, and for the new orders component 47.9, that means the economically weighted three month averages are 52.6 for the total indexes, and 54.6 for the new orders components.

This means that the economy is nowhere near a recession for the next few months, as services continue to power it forward.

An interesting question is whether the strength in services, which as you can see above includes continued strong pricing pressure, translates into continued elevation in the non-shelter services portion of the CPI and PCE indexes. I haven’t done a comparison, but it very much looks like a significant correlation to calculate going forward.