Capital Markets Update #8
Tariffs, immigration flows, bonus depreciation, tax breaks, tax cuts, industrial policy, foreign policy, etc. It all effects markets and we won’t, by any means, diminish the individual or aggregate impact of these varied and impactful shocks. But, perhaps more than anything, net new dollars invested in domestic fixed capital improvements acts as a distillation of many of the above stimuli and represents tangible growth in the US productive base, both goods and services.
US companies spent $856B on Research & Development as of Q3 2025 annualized (BEA) compared to $218B in Private Fixed Investment in Nonresidential Manufacturing Structures over the same period (FRED). More broadly, the private sector invested $887B in nonresidential structural investment in Q3 2025, annualized, compared to $1.74T invested in intangible property including software, R&D, entertainment & literacy over the same period (BEA). Both uses of funds are incredibly important to GDP growth, but it’s interesting to note the order of magnitude to which the US economy over-invests in intellectual property and therefore, by comparison, under-invests in physical properties. If you track the data series, and its unlikely you do, you will have noticed private fixed investment in manufacturing structures peaked in Q3 2024 at a $239B annualized rate and has since dropped to $218B annualized, as of Q3 2025 (FRED).
The question we’re really teasing at is can we even remotely believe the administration’s argument that the stated investment in US manufacturing is well-capitalized and will be realized through a Certificate of Occupancy (CofO) or Commercial Operation Date (COD)? More importantly, assuming some of it is true, what is the appropriate level to underwrite? According to the Whitehouse.Gov, the President has received $9.6 trillion in investment commitments from various constituencies, the majority of which look like fixed CapEx investments at face value. Spread over the remaining years of the Presidents term, that’s call it $3.2T/year in new investment. Assuming 50% of that is actually realized fixed industrial CapEx, that’s $1.6T per year in call it “non-residential structural investments” compared with the present run-rate of $887B. Which would be an absolutely unheard of 80% annual increase. We just made that 50% up, which is seemingly as good a guess as any. In reality, any fraction of the proposed haul is net positive and we hope it all comes true.
Before you change the channel, consider this. Its critical we understand how the available data accounts for investment in structures across the US economy, so that we can level set statements with reality. Fanfare and politics do us no good here. The good news is existing US macro-data generally allows us to capture projects once pen meets paper and can be officially designated as “under construction”. This is great, this is what we want. The BEA tracks industrial building construction two ways, one being a more direct survey method (AIES) the other a somewhat more indirect survey method (a PPI derivative). In the AIES survey, a bunch of companies are simply asked to quantify their expenditures on structures and equipment. Relatively straightforward: what’s the cost value of the project you’re undertaking. The Producer Price Index (PPI) based survey aggregates the value of various components of the “construction supply chain” for structures including building contracts, materials, labor, equipment utilization and so forth. The PPI survey adds these up to come up with a value for New Industrial Building Construction, for instance. That’s an actual data series. The BEA takes these two surveys and, through some weighting, derives its estimate for “Non-Residential Structures” investment for the quarter, annualizes it, and feeds the output directly into GDP. The point being, we actually have pretty good insight into ongoing projects because, provided a project has commenced, the building contract would be struck and permits pulled, the owner would report the project ongoing, the materials purchase orders are likely committed and a cost for the project holistically derived.
Looking at manufacturing discretely, Manufacturing PMI is a data series that aggregates the generic outlook for manufacturing companies. According the survey author, Institute for Supply Management, a PMI in excess of 47.5 generally indicates expansion. The present reading of 52.6 represents the 15th straight month of survey data in excess of the 47.5 threshold. Both the production and new orders sub-indices reported expansion above 50. So, something is feeding through, we just don’t know exactly to what magnitude.
What this research has really taught us is that, to some extent, two things can be true at the same time. There’s no doubt it takes every ounce of one to three years to plan, price and pre-develop a major CapEx initiative, so success must be measured in trends and momentum versus a judgement rendered at any one point in time. On the other hand, if all this stuff on Whitehouse.gov was actually going to go, we’d likely see a ton of ancillary movement across sub-sectors. Any discernable pickup in construction momentum, writ large, would logically be borne out through a steady rise in construction employment, at a minimum. Construction services employed 8.3MM people in January 2025 and employs 8.3MM people today. So, the answer very well may be stand-by for results, but trust the data. We’ll know for sure whether these investments materialize through our PMI and GDP readings; so far, they have largely not. In the end, the best data we have shows an unconstructive trend. Both nonresidential structures and manufacturing structures investment in 2025 was down from 2024 levels. Construction employment is flat, industrial capacity utilization of 76.3% is up 0.7% from January 2025 but down 2.8% from January 2019. And so forth.
We will continue to track the data and likely report back with a more thorough conclusion later in the year.