Every month, three data streams arrive on the desks of people who study the construction and manufacturing economy. Most readers look at each one separately. Looked at together, however, they tell a story that none of them tells alone — and that story has significant implications for manufacturers whose products depend on the construction pipeline.
Here is what the data is saying, and why it matters more than the headlines suggest.
The Manufacturing Economy Looks Healthy. Look Again.
The S&P Global US Manufacturing PMI hit 52.4 in March 2026 — its seventh consecutive month above the expansion threshold of 50, with new orders posting their strongest rise since October 2025. The ISM Manufacturing PMI confirmed a second straight month of expansion in February at the same reading. Business confidence is reaching a 13-month high.
These are the numbers that get quoted in earnings calls and trade press. They are real. But, they are also incomplete.
Inside those same reports are these realities:
- employment remained in contraction
- input and output prices surged to their highest levels since June 2022, driven by steel, aluminum, and tariff-driven cost increases
- supplier delivery times lengthened to levels not seen since October 2022.
And critically, the ISM report noted that a meaningful portion of January’s expansion — the month that broke a 12-month contraction streak — appeared to reflect pre-tariff stockpiling rather than genuine end-user demand. Stockpile-driven PMI readings are well-understood as false signals. When the stockpile is built, orders drop.
So the manufacturing economy is expanding, but on fragile foundations: rising costs, contracting employment, stretched supply chains, and at least some demand that is borrowed from the future.
The Housing Economy Is Weaker Than Its Forecasters Expected
John Burns Research and Consulting — one of the more rigorous and accountable forecasting firms in the housing sector — published their 2025 forecast vs. actual scorecard recently, and the pattern in the misses is instructive. Every single category came in softer than forecast: job growth at +0.5% against a +1.1% forecast, resale sales flat against a +3% forecast, resale price appreciation at +0.6% against +2.3%, single-family rent growth at +1.9% against +3.0%, apartment effective rent growth at +0.2% against +1.5%.
That is not random forecasting error. Every miss lands on the optimistic side. The firm acknowledges the culprits honestly: immigration declined faster than projected, household formation slowed, and macro uncertainty eroded buyer confidence at precisely the moment the spring selling season needed it most. Their 2026 baseline was already cautious when they set it. Since then, the Iran conflict has pushed oil prices higher, mortgage rates have spiked, and their builder conversations are pointing to a muted spring. They are actively revising single-family starts forecasts downward in real time.
Remember Covid? Any economic forecast during that tumultuous time was revised hourly. In fact, the housing market during COVID is the definitive case study in forecast futility. In January 2020, the U.S. was posting its strongest housing starts since December 2006. By the second quarter, residential fixed investment had collapsed at a 35.6% annualized rate. Every major forecaster called for a prolonged downturn. They were wrong within months. Despite dramatic increases in uncertainty about health, the macroeconomy, and social circumstances — conditions that would logically predict a sharp housing downturn — house prices increased roughly 10% in real terms in 2020, then rose another 15% in the year to July 2021. The average American home gained nearly $100,000 in value between mid-2020 and mid-2022, roughly three-quarters of which exceeded anything the pre-pandemic trendline would have predicted.
The forecasters didn’t just miss the direction — they missed the magnitude by an order of magnitude, in both directions, within the same 24-month window. The shift from a pessimistic outlook of a sharp correction to a more optimistic one took years to fully resolve, and the question of why the bleak forecasts missed the mark so badly is still being studied.
The lesson isn’t that forecasters are incompetent. It’s that housing is a system with too many interacting variables — rates, migration, remote work, immigration, consumer sentiment, fiscal policy — for any model to hold its shape when one or more inputs changes suddenly and simultaneously. COVID changed nearly all of them at once. The Iran conflict, tariff uncertainty, and AI-driven employment anxiety are not COVID. But they are the kind of compounding, simultaneous shocks that make any single-point forecast — including JBREC’s — a starting position rather than a destination.
For manufacturers whose products flow through the residential construction and repair-and-remodel channels, the JBREC data confirms what Home Depot and Lowe’s earnings calls have been saying for two consecutive years: the demand environment is genuinely soft, and the structural conditions for recovery — rate relief, consumer confidence, household formation growth — are not yet in place.
The Construction Data Reveals a Market in Two Pieces
ConstructConnect’s™ February 2026 snapshot and Spring Forecast add the project-level granularity that turns the macro picture into something actionable. Total construction starts were up 4.4% year-over-year in February and 20.1% year-to-date. Both numbers sound constructive. Both numbers are substantially misleading.
Strip out private offices — where ConstructConnect™ categorizes data centers — and the story changes entirely. That single category was up 1,364% year-to-date through February. It accounted for roughly 25% of all nonresidential building spending in 2025, up from 6% in 2022. Remove it from the nonresidential growth calculation and the 22.8% NRB growth rate falls to 11.5%. The top project in February was a single Meta data center in Lebanon, Indiana valued at $10 billion — more than half of all February residential construction starts combined.
This is not a broadly growing construction market. It is a market with one very large, very hot engine and several cold or stalling cylinders. Total residential construction was down 22.3% year-to-date, with single-family down 25.7% — consistent with and actually slightly worse than JBREC’s full-year 2025 number. Manufacturing facility starts were flat. Hospitals were down 36.5% in February. Colleges and universities were down 49.8%. Warehouses, the post-COVID darling category, came in essentially flat. Construction employment shed 11,000 jobs in February against a historical average February gain of roughly 10,000 — a 21,000-job swing concentrated almost entirely in residential sub-trades and civil work.
What Three Sources See That One Source Misses
The manufacturing PMI tells you activity is expanding but employment is contracting and costs are rising — a setup for margin compression rather than growth. JBREC tells you the housing demand engine that drives residential construction and R&R spending is operating well below capacity, with no near-term catalyst for recovery. ConstructConnect tells you that whatever construction growth exists is concentrated in a narrow set of categories that most building products manufacturers don’t meaningfully serve.
Taken together, these three data streams are describing the same underlying condition from three different angles: a bifurcated economy in which capital is flowing heavily into AI infrastructure, data centers, and select megaprojects, while the broad middle of the construction and manufacturing market — the segment that generates specification volume for HVAC systems, plumbing products, faucets, building envelope materials, and light commercial equipment — is soft, delayed, or contracting.
The regional data confirms this geographic concentration. Indiana is up 773% year-to-date in nonresidential starts. Illinois is up 295%. North Carolina is up 209%. These are almost entirely data center and large commercial starts. Texas — still the largest absolute construction volume state — is down 9.6%. Florida is up a modest 13%. New England is down 27%. The geographic reallocation of construction dollars is moving faster than most manufacturers’ sales territory strategies reflect.
The Strategic Implication: Delay Is the New Default
The conventional specification strategy was designed for a market where projects moved on reasonably predictable timelines from design through construction. The data across all three sources suggests that market no longer describes most of the pipeline.
What the data describes instead is a split between committed capital — projects with locked funding and unstoppable momentum, primarily data centers and megaprojects — and hesitant capital, the broad middle of the market where interest rates, tariff uncertainty, and cost volatility are producing delay rather than cancellation. The adjective distinction matters enormously for manufacturers. A cancelled project exits the pipeline. A delayed project stays in it — but it is being quietly reconsidered the entire time it sits there.
Budgets are reworked during delays. Contractors reassess. Alternatives reappear. The original specification — the one your sales rep worked to establish, the one the rep counted as a win — is often no longer being actively defended by anyone in the value chain. Which means that in a delay-dominated market, winning the specification is no longer sufficient. Staying specified through the delay cycle is the actual competitive objective.
The manufacturers who understand this are beginning to treat delayed projects not as wins on hold but as live opportunities with their own risk profiles — asking which specs are sole-source versus substitutable, whether the project is fully funded or being reworked, whether competitors are already having conversations. The ones who don’t are discovering at bid time that a specification they thought was secured quietly migrated to a competitor who stayed engaged while they moved on.
The Opportunity the Headlines Are Missing
The temptation in a market dominated by data center headlines is to chase data centers. Most building products manufacturers cannot win there at scale — those projects are dominated by companies with specialized capabilities in large-format industrial construction, AI cooling infrastructure, and power systems.
The real opportunity, as the ConstructConnect data makes clear when you look past the office category distortion, is in the underfollowed segments: manufacturing facility construction running at 104% growth on a 12-month rolling basis, civil subcategories including water and sewage treatment and power infrastructure, and healthcare pockets where spending is uneven but present. These categories collectively represent more than double the dollar volume of all data center starts — $192.9 billion in 2025 nonresidential starts outside of offices, against roughly $85 billion in data center-related construction.
That is where the specification volume lives. That is also where the delay risk is concentrated, because those are the projects sitting in hesitant capital territory — real, funded in principle, moving forward eventually, but vulnerable to substitution in the meantime.
The construction market has changed its shape. The manufacturers who recognize that shift in the data — not just in the headlines — are the ones who will hold their specifications when those projects finally move.
__________________________________
List of Works Consulted
- ConstructConnect. Construction Economy Snapshot: March 2026. Cincinnati: ConstructConnect, Inc., February 28, 2026.
- ConstructConnect. Spring 2026 Construction Forecast. Cincinnati: ConstructConnect, Inc., 2026.
- Dallas Federal Reserve Bank. How the U.S. Might Outgrow Pandemic-Era Housing Unaffordability. Dallas: Federal Reserve Bank of Dallas, August 2024.
- Institute for Supply Management. ISM Manufacturing PMI Report on Business: February 2026. Tempe: ISM, March 2, 2026.
- John Burns Research and Consulting. 2025 Housing Forecast vs. Actual and 2026 Shifts. Irvine: JBREC, March 2026.
- Richmond Federal Reserve Bank. The Housing Market and the Pandemic. Richmond: Federal Reserve Bank of Richmond, Q4 2020.
- S&P Global Market Intelligence. US Manufacturing PMI: March 2026 Preliminary Flash Estimate. London: S&P Global, March 2026.
- The Brookings Institution. Quantitative Easing and Housing Inflation Post-COVID. Washington: Brookings, October 2025