Every December a contractor reads a forecast and makes a decision about next year's capacity. This year the forecast says starts will fall 5.4%, and a lot of people are going to make that decision badly.
Not because the number is wrong. Because it is an average across categories that are moving in opposite directions, denominated in dollars whose value is itself changing by close to 9%.
One question, four answers
"Is 2027 a down year" has four defensible answers depending on what you build and how you count.
Nearly the entire projected decline sits in the bottom bar. Most contractors are looking at one of the top two.
The nominal-versus-real distinction is doing a lot of work here and it is worth being explicit about. A nominal decline of 4.6% in non-residential starts alongside a real increase of about 1% is not a contradiction — it means the dollar value of work started falls while the physical quantity of work rises, because each dollar buys less. With materials inflation expected to peak above 9% around the turn of the year, that gap is wide enough to flip the sign on the number you are planning against.
For a contractor, the real figure is usually the more relevant one. You are not bidding dollars. You are bidding crews, weeks, and equipment against a quantity of work that has to get built.
The pipeline argument
The strongest reason not to read a starts decline as a demand collapse is the planning data that sits ahead of it.
The Dodge Momentum Index tracks nonresidential projects entering the planning stage, and it leads actual construction spending by roughly a year. It finished 2025 up 37% against 2024, and grew 7% in December. Projects that entered planning through 2025 and 2026 still have to be built, and a large share of them will be built in 2027 and 2028 regardless of what happens to new starts in the interim.
This is also why the sources appear to disagree with each other. ENR titled its 2027 forecast coverage "Full Steam Ahead" while reporting a 5.4% starts decline in the same piece. Both are defensible: the flow of new work slows, the stock of planned work is unusually large. Dodge's own economists expect non-residential starts excluding manufacturing and transportation to accelerate in 2027, on planning activity in data centers, healthcare, and recreation.
The mix shift nobody is pricing
Buried in the office forecast is the line most likely to matter to a mid-sized GC or specialty contractor: renovation is projected to overtake new construction in 2027 within office.
That is a different business, not a smaller one. Renovation work carries different risk, different sequencing, and different margins than new build. It runs in occupied buildings, on compressed schedules, with more unknowns behind the walls and more change orders as a result. The estimating discipline that works on a greenfield build is the wrong instrument for it, and change orders that never get billed are where renovation margin quietly goes.
A contractor whose 2027 plan is "do less of what we did in 2026" may be planning for the wrong year entirely. The more accurate framing for a lot of firms is "do a different mix of roughly the same volume, with more of it in occupied space."
What to actually do with this
1. Find your line in the forecast before you change anything. Residential remodeler, mechanical sub on healthcare, civil contractor on utilities — these are three completely different 2027s. The headline is an average of all of them and describes none of them.
2. Do not cut trained crew against a forecast. The 2026 evidence is that labor stayed tight through a soft patch rather than loosening with it, with data center work absorbing crews that would otherwise have freed up. Releasing skilled people into that market to meet a projection is expensive to reverse, and you will be bidding against your own former employees' new crews.
3. Separate nominal from real in your own numbers too. If your revenue grows 6% while materials inflate 9%, you built less work for more money and your capacity utilization fell. Track volume alongside dollars or you will misread your own year the same way the headline misreads the industry's.
4. Re-qualify your pipeline against the categories that are actually growing. Data centers, healthcare, recreation, and renovation. If your business development is still pointed at categories entering a real-terms decline, the forecast is a reason to redirect effort rather than to shrink.
5. Hold the escalation discipline. A down year in starts tightens bidding, and tight bidding is exactly when escalation clauses and short quote-validity windows get conceded to win work. That concession has a price, and it does not get cheaper because the market got harder.
How much to trust any of this
Less than the confidence of the presentation suggests, and the forecasters would say the same.
Dodge revised its 2026 growth expectations sharply upward between the April and July 2026 releases. ENR's framing of the same dataset is considerably more bullish than the raw decline implies. Materials inflation is a forecast, not a measurement, and it moves the nominal and real figures in opposite directions. Nobody publishing a 2027 number right now has a good track record on 2026 numbers published in 2025.
Which argues for a specific posture rather than paralysis: plan against the line of the forecast that matches your actual work, keep the capacity you have trained, and revisit in April when the first real 2027 data arrives. The contractors who get hurt by a forecast are rarely the ones who read it wrong. They are the ones who made an irreversible decision on a number that was always going to be revised.
6 Signal works with contractors on the systems behind the business — estimating discipline, financial control, and the AI infrastructure that keeps them running. For the cost pressure inside this forecast, read Your Fixed-Price Contract Is a Tariff Bet; for the labor side, read Nearly 3 in 10 Contractors Report Enforcement Effects; for the financial control layer, read Margin Gone Before the Job Closed.
Sources
- Total starts up 10.9% in 2026 and down 5.4% in 2027; non-residential starts down 4.6% — ENR 2027 Forecast
- Inflation-adjusted 2027 outlook (residential ~+2%, non-residential ~+1%, nonbuilding ~−10%); office including data centers −4.9%; renovation projected to overtake new construction in office; non-residential excluding manufacturing and transportation projected to accelerate on data center, healthcare and recreation planning; materials inflation expected to peak above 9% — Dodge Construction Network, 2027 market trends
- Dodge Momentum Index full-year 2025 up 37% against 2024, December growth of 7%, and its roughly one-year lead on nonresidential spending — Dodge Construction Network and Construction Dive