While capital spending increased sharply, yet more evidence of consumer weakness

 

 - by New Deal democrat

On Friday I noted that real personal spending on goods, especially durable goods, had declined in September. If we have reached a tipping point on that metric, a recession in the near future looks much more likely, even as spending on services continues.

Late last week we also got further evidence of the bifurcation between the consumer economy and the AI-fueled production economy, in the form of durable goods orders and motor vehicle sales.

Let’s look at motor vehicle sales, updated right through November first. On a month over month basis, both light vehicle (sedans, SUVs, pickup trucks) increased, as did sales of heavy weight trucks:

That’s the good news.

The bad news is when we put this improvement in perspective by looking at the long term historical data:

Heavy truck sales carry much more, and more reliable, signal than light vehicle sales, and they always turn down sharply first. Which is exactly what they have done in the past few months. The long leading signal of housing construction turned recessionary many months ago, and now the next shoe has clearly dropped.

But the other news last week, on manufacturers’ new durable and capital goods orders, told a completely different story, as both increased to among their best readings since the pandemic:

In the case of core capital goods orders, it was the best reading since the pandemic except for one month. This is a strong uptrend that began over a year ago and really accelerated this year.

But the intersection between these two metrics is production of, and spending on, consumer durable goods. Here is headline durable goods orders (blue) vs. consumer durable goods orders (red), updated through September:

As per the above, the former was in a strong uptrend. But the latter remained flat, just as it has been for two years.

So let’s compare consumer spending on durables YoY (blue) vs. manufacturers orders for consumer durables YoY (red):

Since the latter are much noisier and more volatile than the former, I have supplied the quarterly average as well as the monthly YoY change, divided by 1.5 for scale.

In general, consumer spending on durables turns first, giving manufacturers their cue to produce more or less. This was complicated by the “China shock” beginning in 1999, where goods imports from China increased sharply, and for a generation.

Finally, here is the post-pandemic look:

Consumer spending on durables increased last autumn and winter, particularly in anticipation of T—-p’s tariffs. The YoY comparisons are still positive, but less so. As per usual, the producer response occurred afterward. 

It is especially important to reiterate than the most recent durable goods and spending data has only been released through September. The motor vehicle data, as well as other types of data such as tax withholding, have indicated a sharp slowdown since. But we’ll have to wait at least one more month to see if that has truly broadened into a contraction in consumer spending on goods.

Weekly Indicators for December 1 - 5 at Seeking Alpha

 

 - by New Deal democrat

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

There are some minor changes, but no big turns in trends. But the recent change in trend that has been further reinforced is the sharp deceleration in growth in withholding tax payments that began in October. The bulk of the evidence suggests that this is not a result of end of year tax planning, but real signal of a slowdown in labor force and payroll growth.

As usual, clicking over and reading will bring you up to date on all of the relevant economic data, and bring me a little pocket change for my efforts.

Real income rises, but real spending on goods may be turning down

 

 - by New Deal democrat

Personal income and consumption is one of the two big monthly reports on the state of the average American, in addition to the jobs report. This month it has the added virtue of being the “least stale” monthly report, as it was issued only five weeks later than scheduled. Ever since “Liberation Day” in April, I have looked for the impact of tariffs on both personal spending and manufacturers’ sales. Additionally, there has been evidence since January that income and spending might be slowly rolling over in any event. This morning’s data for September added at that concern.

Nominally income rose 0.4% and spending 0.3%. But since the PCE inflation gauge rose 0.3%, real income only increased 0.1% and real spending was flat:

[Note: with the exception of the personal saving rate, and one YoY graph, all of the data in the below graphs is normed to 100 as of just before the pandemic.]

Since real spending on services (blue, right scale) rarely turns down, even in recessions, I focus on goods (red, left scale), and on an even more granular basis on durable goods spending. In September real spending on goods (red) declined -0.4%, and on durable goods (gold) even more, by -0.6%:

Because the monthly data can be noisy, I have been particularly looking at the three month average. For durable goods, this looks like it peaked in March through May. For goods spending as a whole, the three month average only increased 0.2% for July through September, and was only up 0.5% since March through May. This is very close to rolling over into contraction.

But if the real spending side of the coin merits a yellow flag, the real income and savings side was more sanguine. 

I follow the personal savings rate because just before and going into recessions it tends to turn up as consumers get more cautious. After revisions this was unchanged at 4.7% in September:

Additionally, one of the two coincident indicators from this report which the NBER pays close attention to in dating recessions is real income less government transfers. This increased 0.1% to a new record high (blue, right scale):

On a YoY basis (red, left scale) after decelerating for almost three years, the latest data shows stabilization since May. Hence the relatively good news on the income side of the coin.

Normally the second coincident metric looked at by the NBER, real manufacturing and trade industries sales, is also reported with a one month delay at the same time as personal income and spending, but this month that was not the case.

In summary, this was a mixed report. On the positive side, although growth has slowed, the positive trend in real income is intact, as is the neutral trend in personal saving. On the negative side, real spending on goods and in particular durable goods declined, with the latter having made at least a temporary peak back in springtime. The former must increase at least 0.1% (subject to revisions) in the next report in order for the three month average not to decline.

Jobless claims: Holiday seasonality enters in a big way

 

 - by New Deal democrat

The good news is, we are back to the normal weekly jobless claims releases. The really good news is that this week’s number, except for one week in 2022, was a new 50 year low! The bad news is that Holiday seasonality is very much in play, so take the good news with multiple grains of salt.

To give you an idea of how much seasonality, look at the decline that was seasonally “expected” vs. the actual number, per this week’s report:

“The advance number of actual initial claims under state programs, unadjusted, totaled 197,221 in the week ending November 29, a decline of 49,419 (or 20.0%) from the previous week. The seasonal factors had expected a decrease of 21,172 or -8.0% from the previous week.”

But to the numbers: seasonally adjusted initial claims declined -27,000 to 191,000 last week, and the four week moving average declined -9,500 to 214,750. With the typical one week delay, continuing claims declined -4,000 to 1,939,000:

To show you the seasonality at work, here are the last two years starting November 1 of non-seasonally adjusted claims (orange) vs. seasonally adjusted (blue):

A big decline in claims always occurs during Thanksgiving week. This year’s decline was signficantly bigger than the two prior years.

As per usual, the YoY% changes are more important for forecasting purposes. So measured, initial claims were down -15.1%, the four week average down -1.9%, and continuing claims up 3.6%:

Needless to say, this is positive. But I strongly suggest we wait for next week’s inevitable big seasonal increase, and average the numbers before popping any champagne corks.

ISM services for November generally positive and improving

 

 - by New Deal democrat

Probably the most important economic news this entire week was this morning’s ISM services report. Services are about 75% of the economy, and this report was for November, which means it is the most wide-ranging and current datapoint we have at the moment.

And the news on this front was almost all good. The headline number (blue in the graph below) improved to 52.6 from last month’s 52.4. Employment was less bad, improving to 48.9 from 48.2. Prices paid decelerated (a good thing) from 70.0 to 65.4. The only (slight) disappointment was that new orders (gray) were less positive at 52.9 vs. last month’s strong 56.2 [note: all graphs via TradingEconomics.com]:

My short term economic forecast gives 75% weight to this metric (gray) and 25% to the manufacturing survey (blue), and also averages over 3 months to cut down on noise. For the headline number, the three month economically weighted average was 51.0:

For the more leading new orders metric, the economically weighted three month average was 52.0:

Needless to say, both of these were expansionary if weakly so, but with evidence of a slightly improving near term forecast.

The retreat in the prices paid metric was particularly good news in comparison with preceding months:

But as with the manufacturing survey, the regional Fed surveys, and this morning’s ADP report, the bad news (even if “less bad”) is that employment appears to be contracting:

For the working and middle class as a whole, the question is whether payroll gains via wages more than make up for th apparent slight loss in the number of jobs. Unfortunately, for that at the moment we only have shadows on the wall.

Production weakens while private employment declines

 

 - by New Deal democrat

Although not published by the federal government itself, the Fed’s measure of industrial production relies on some federal data, and thus it was not updated during the government shutdown - which means that this morning’s update is likewise stale, being for September.

Industrial production has been much less central to the US economy since the “China shock,” but it remains important for the goods producing sector. In September, headline industrial production rose 0.1%, while manufacturing production was unchanged. The above graph normalizes both measures to April 2022. As you can see, between spring 2022 and late 2024, production generally declined before surging in the first six months of this year. Total production exceeded that level just barely in July, while manufacturing production has stalled without reaching that level:

Here is the longer term historical look since before the “China shock”:

Finally, I would be remiss without noting the poor ADP employment report for November this morning. The below graph shows industrial and manufacturing production for this year, together with the ADP employment trend and the official payrolls number, all normed to 100 as of April:

Employment has stalled since then, and production *may* have during this summer, but there have been plenty of noisy such periods before. So far consumer spending fueled by the surging stock market and the resulting “wealth effect” have more than counterbalanced that weakness.