ISM manufacturing In February positive, but with a major helping of inflation

 

 - by New Deal democrat

Most official data series remain about one month behind their normal schedule, despite the government shutdown having ended over three months ago. Thus the ISM manufacturing and services reports, as well as the regional Fed manufacturing and services reports, remain the most timely overall barometers of the economy.

In the past few months the regional Fed manufacturing reports have painted a picture of a sector recovering from the impact of tariffs - which continued to be the case in the February reports. This morning the ISM manufacturing report once again confirmed those trends. But also added one renewed source of concern.

To begin with, the headline number declined -0.2 from 52.6 to 52.4. These have been the two highest reading since August 2022:

The three month average, which I use for forecasting purposes, rose to 51.0, the first time this average has shown expansion since January of last year (recall that 50 is the dividing line between expansion and contraction).

The more forward looking new orders component, which in January exploded from 47.4 to 57.1, settled back only slightly in February to 55.8, still a strong reading:

The three month average is now 53.3, indicating a respectable increase in production in the next few months.

Even employment, which has remained a problem child in this series, continued to trend “less bad,” rising from 48.1 to 48.8:

The three month average did remain contractionary, at 47.2.

But the new cause for concern is the sudden spike in prices. These had already been a concern, in view of the tariff situation. In the latter part of last year, these readings had declined into the 50’s, indicating pricing pressure, but not nearly so much as last spring. This month the prices paid reading blew out from 59.0 to 70.5: 

This was the highest monthly reading since June 2022, suggesting that substantial inflationary pressures are building — not a good thing especially when so much of the economy appears to have been stagnating.

As I have continued to note each month, for the economy as a whole the weighted index of manufacturing (25%) and non-manufacturing (75%) indexes is more important. With the exception of one month last year, the services indexes remained in expansion. So I won’t bother with calculating the economically weighted averages today: with both sectors showing expansion, the ISM indexes indicate the economy is likely in expansion now, and the new orders component is positive for the next several months.

But as I concluded last month, and is even more true after today’s report, the caveat remains the important stagflationary pressures which have been showing up in almost all the recent data.

Weekly Indicators for February 23 - 27 at Seeking Alpha

 

 - by New Deal democrat

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

Last week commodity prices, and in particular oil, rose sharply again - probably due to what has been unfolding in the Middle East. And the US$ is still losing value. Meanwhile withholding tax payments have faltered somewhat again. Good times!

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 for my efforts.

Construction spending mainly flat nominally and declining in real terms, except for AI data center-related Boom

 

 - by New Deal democrat

Before I get to the main object of this post, there was a little kerfluffle in the financial markets this morning about a “hot” producer price index number. Specifically, PPI for final demand increased 0.5% in January. Commodity prices also increased 0.3%. The YoY% changes were 2.8%, although raw commodity prices were only up 1.6%. The below graph also shows CPI, which officially was up 2.4% YoY (although a proper accounting for shelter would probably add at least 0.2% to that number):

The bottom line is that inflation is not cooperating with those who would like to see further Fed rate cuts, since it is likely making little if any progress towards the Fed’s goal of 2.0%. Stagflation, anyone?

But let’s turn to an important manifestation of costs and spending in the real world, by way of this morning’s one- and two-months stale report for construction spending in November and December. 

For the month of December, nominally total construction spending (blue, left scale in the graph below) rose 0.3%, although it remains below its recent peak in September, and its post-pandemic peak of May 2024, and is down -0.4% YoY. The long leading sector of residential construction spending (red, right scale) rose a more significant 1.5%, but it also remains significantly below its peak of May 2024, although as I pointed out one month ago, there are certainly signs of “green shoots,” i.e., a bottoming process:

The picture is somewhat different when we account for the costs of construction materials. These rose 1.8% in November and another 0.9% in December to near its all-time highs:

On a YoY basis, they are also up 6.6%.

As a result, in real terms both total and residential construction spending fell, by -1.5% and -0.4% respectively:

The overall trend remains slightly negative for total construction spending, although residential construction spending has been trending sideways. Hence, small signs of “green shoots” by way of forming a bottom.

One big negative is manufacturing construction spending, which declined -2.5% in December, and is down over -15% from its August 2024 peak (recall that the big Boom was a direct result of the Inflation Reduction Act)

So much for tariffs bringing back manufacturing!

What has been powering the positive contributions to total construction spending, as with so much of the economy, is AI data center related spending. Power generation spending rose 0.8% for the month of December and is up 5.8% YoY. Another aspect is the water needed to cool these power plants, which declined -2.5% in December, but is up 8.1% YoY, and made an all-time highs in October:

So once again, we have an economy that is barely holding its head above water, and indeed in real terms the construction sector is declining, with just about the only subsector holding it afloat being AI data center spending, and its stock market-associated wealth effect consumer spending.

More clarity in jobless claims: both post-pandemic seasonality and regime change at work

 

 - by New Deal democrat

As we move further into the calendar year, we are getting more clarity on what has been happening with jobless claims. As a reminder, I look at these because historically they have been a good short leading indicator for the unemployment rate and more broadly for the economy as a whole. And to cut to the chase what I see happening with claims is *both* residual post-pandemic seasonality *and* a change in regime to lower claims that started at mid-year last year.

First, to this week’s numbers: initial claims rose 4,000 to 212,000. The four week moving average rose 750 to 220,750. And with the typical one week delay, continuing claims declined -31,000 to 1.833 million:

The one week delay in continuing claims this week is likely unusually important, because this reports (like initial claims last week) for the week including President’s Day and so had one fewer day that government offices would be open.

Note, though, that the post-pandemic seasonality, in which claims make lows at the turn of the year, rise to highs at mid-year, and then decline back down through the end of the year, looks more apparent now as we see initial claims and their four week average moving higher since January.

But as usual, the YoY% change is more important for forecasting purposes; and it is here that we see continued evidence of a change in regime. On a YoY basis, initial claims were lower -7.1%, the four week average down -2.5%, and continuing claims down -0.8%:

This is the trend of generally lower claims we have seen since the end of last June. This trend is a positive for the economy over the next few months. Again, the reason for this regime change may have to do with the nearly complete cessation in new immigration, and fear and deportations driving immigrants already here not to show up for work or to file claims, leading to lower jobless claims. Since we have also seen a gain in AI data center-related employment, it could reflect a lower number of layoffs in nonresidential construction employment as well.

Finally, since we are close to the end of the month, let’s take a look at what this likely means for the unemployment rate (red in the graph below, left scale) in the next monthly jobs reports:

The downward trend in jobless claims strongly implies a declining trend in the unemployment rate over the next several months, towards 4.2% or even 4.1%. 

Per Prof. Edward Leamer’s forecasting method, a yellow flag “Recession Watch” is warranted

 

 - by New Deal democrat

In a landmark paper almost 20 years ago, UCLA Prof. Edward Leamer wrote that “Housing IS the Business Cycle.”

In that paper he concluded:

In the years before recessions … consumers contribute a total of 65% of the leading weakness. In contrast, business spending contributes only 10%…. The temporal ordering of the spending weakness is: residential investment, consumer durables, consumer nondurables consumer services….

I’ve been writing for many months that the housing sector has been recessionary. I’ve also noted that another important leading indicator from the real economy, truck sales, has also been recessionary. This month three more of the items on Leamer’s list, albeit in noisy terms, also appear to have turned recessionary.

The first of those is personal spending on durable goods, which I noted last Friday. Here is the graph I used then, normed to 100 as of one year ago:

The second is motor vehicle sales for December (blue in the graph below), which was reported as part of last Friday’s GDP release, and is the quinessential “consumer durable” on Leamer’s list. These declined sharply to an annualized rate of 14.810 million, the lowest monthly number in four years, and over -17% below the peak month last March, in which a SAAR of 17.892 units were sold:

The three month average of 15.5 million is -9% below the peak average earlier last year of 17.0 million. This is significant because, as the historical graph below indicates, typically recessions have begun once motor vehicle sales have declined roughly -10% from peak. In both the above and below graphs I also show heavy truck sales (red), which almost always decline first and more decisively, and recover later. As I’ve noted above, these have already been recessionary:

The third metric is manufacturers’ new orders for consumer goods, which were reported last Wednesday. These declined -1.4% for the month, but more importantly with the exception of last April and May were the lowest in almost two years:

They were also negative YoY, just as real personal spending on goods also turned negative YoY in that most recent report last Friday:

In other words, every single item on Leamer’s leading checklist has turned negative except for consumer spending on services (which I question because frequently services spending has continued to increase throughout recessions).

On Tuesday I noted that since September three of the four of the typical monthly data upon which the NBER has traditionally relied upon in determining if a recession has occurred - payrolls, real sales, and real income excluding government transfers - have been essentially flat:

I characterized this as “an economy barely holding its head above water.” By the terms of Prof. Leamer’s forecasting method, a yellow flag “Recession Watch” is warranted. Should the three month averages of motor vehicle sales and new orders for consumer durables turn more negative, a “Warning” would be warranted.