- by New Deal democrat
May data started out as usual with the ISM manufacturing index. Plus, as a bonus, official government data is finally back on schedule, only 7 months after the end of the shutdown! By which I mean to say, April construction spending was also released, on time.
To cut to the chase, the news about manufacturing was both good and bad, while that on construction was ‘meh.’
Let’s start with manufacturing first. Late last year, I began to notice that the headline regional Fed and ISM numbers were trending “less bad,” and then finally outright positive. That trend continued this month, as the headline ISM manufacturing number (blue in the graph below) increased 1.3 to 54.0 (recall that any number above 50.0 indicates expansion). The more leading new orders subindex (gray) also rose, by 2.7, to 56.8 suggesting the AI data center related Boom will continue. The three month averages, which smooth out a little volatility, rose 2.2 to 54.8 and 0.2 to 54.7, respectively:
As I have said a number of times recently, I am convinced that all of the activity surrounding AI data center construction and operation, and the affluent consumer spending secondary to the (narrow) stock market Boom associated with it are the only things keeping the US economy from being in recession at present.
That was the good news. As with last month, here’s the bad news. First, the contraction in goods producing employment continued, although it was “less bad” at 48.6 vs. 46.4 in March. The three month average is 48.3, also “less bad” than for all of last year:
This, by the way, is at variance with the official employment report as to goods-producing employment, which has been generally increasing since last October, and specificially, manufacturing employment, which has been increasing since last December. The most likely way for the two numbers to be consistent, since the ISM report is a diffusion index, is if the gains in employment are narrowly focused, but stronger than a more diffuse weakening.
But the worst news is that there continue to be widespread increases in prices paid. This did decline in April, by -2.5, but the decline was to “only” 82.1. The three month average is 81.7. The graph I show below goes back five year to show that price increases are as widespread now as they were during the worst of the post-pandemic inflation:
As I indicated last month, this is a very sharp inflationary pulse, which is going to pass right through into consumer prices for goods.
Now let’s turn to the second report, for construction spending.
in the past I have used construction to help track the long leading sector of housing; and in the wake of the Inflation Reduction Act, plus “Liberation Day,” it has also been useful to track manufacturing. But now, via tracking construction of water supply and power, it is also a useful proxy for construction of AI data centers.
In April on a nominal basis total construction spending rose 0.4%, but that was neutered by a -0.4% reduction in March. On a YoY basis, it was up only 0.7%. Residential construction spending rose 0.8%, and up 1.7% YoY. On a longer basis going back several years spending has been generally flat:
The problem with this nominally positive news is that the prices of construction materials rose 1.3% in April alone, and are up 6.7% YoY, meaning that in real terms both headline and residential construction spending was negative; in fact, the worst in nearly three years:
On the other hand, even nominally manufacturing construction continued its slide, down -1.2% for the month and -18.5% YoY (so much for tariffs bringing manufacturing back onshore!):
Finally, as indicated above, spending on power and water supply construction appear to be the best proxies for AI data center related spending. In April, power supply spending rose 0.6%, and it is up 6.8% YoY; while spending on water supply construction declined -0.5%, and is up 4.8% YoY - but is down 5% from its peak last October. Note that in the graph below I norm both to 100 as of the start of the pandemic, to show that power generation construction has increased nearly 40%, and water supply construction spending at its peak had almost doubled:
Although I won’t show the graph, adjusted by inflation in construction materials, only power supply construction spending is higher YoY.
Let’s put this all together. In the long leading sector of housing, spending is higher, but appears really to be related to the cost of materials. In the short leading sector of manufacturing, business is increasing, but spending on new manufacturing plants has been plummeting. The one sector of construction that appears to be truly increasing is power supply construction for AI data centers. And the strong inflationary pulse is continuing.
