There is a number in this industry that everybody quotes and nobody measured. It says materials are 40 to 60 percent of a construction project. It appears in procurement software marketing, in estimating blog posts, and in sales decks aimed at electrical and mechanical contractors. It has no source attached to it.
The real number has been measured. Twice. By the federal government, under mandatory response, with the raw files posted for free download and no API key required.
For electrical contractors, materials were 31.43% of the value of construction work in 2022. For plumbing, heating and air-conditioning contractors, 32.49%.
What the Economic Census actually counted
The Economic Census asks construction establishments for two figures that matter here. The Census field names and the Census wording, exactly:
RCPCWRK, “Value of construction work ($1,000)”CSTMPRT, “Cost of materials, components, packaging and/or supplies used, minerals received, or purchased machinery installed ($1,000)”
Divide the second by the first and you have the material share of what the trade bills. Here are the 2022 values as published, in thousands of dollars, for the United States.
| NAICS | Census industry label | Value of construction work | Cost of materials | Materials share |
|---|---|---|---|---|
| 238210 | Electrical contractors and other wiring installation contractors | $244,852,717 | $76,952,362 | 31.43% |
| 238220 | Plumbing, heating, and air-conditioning contractors | $288,331,268 | $93,671,679 | 32.49% |
| 2382 | Building equipment contractors | $574,726,418 | $180,230,142 | 31.36% |
| 238 | Specialty trade contractors | $1,229,618,398 | $389,473,962 | 31.67% |
| 23 | Construction | $2,852,781,023 | $845,113,785 | 29.62% |
Two things about that table are worth pausing on. The trades cluster tightly, between 31% and 33%, and the whole construction sector sits just below them at 29.62%. Nothing in this data supports a number starting with a four, let alone a five or six.
The part that makes it hard to argue with
A single census year invites an obvious objection: 2022 was a strange year for material prices, so the ratio is an artifact of the spike.
So run the same two fields against the previous census cycle. The 2017 Economic Census publishes
EC1723BASIC with identical field definitions.
| NAICS | Industry | 2017 share | 2022 share | Change |
|---|---|---|---|---|
| 238210 | Electrical contractors | 31.31% | 31.43% | +0.12 pp |
| 238220 | Plumbing, heating, and air-conditioning contractors | 32.71% | 32.49% | -0.22 pp |
| 2382 | Building equipment contractors | 31.26% | 31.36% | +0.10 pp |
| 238 | Specialty trade contractors | 31.09% | 31.67% | +0.58 pp |
| 23 | Construction | 30.29% | 29.62% | -0.67 pp |
The 2017 underlying values, in thousands: electrical contractors, work $170,466,191 and materials $53,376,720; plumbing/HVAC, work $204,227,942 and materials $66,809,886.
Between those two reference years, copper, steel pipe, plastics pipe, conduit and HVAC equipment all went through the sharpest price movement in decades. Some categories nearly doubled. The material share of contractor billings moved by a quarter of a percentage point.
That is the finding. Not that the ratio is roughly a third, but that it is roughly a third under two completely different price regimes, measured by two separate censuses conducted five years apart, and you can reproduce both from files that cost nothing.
The stability tells you something structural. When material prices rise, contractors eventually reprice the work, and the ratio reverts. Materials hold an anchored share of a job rather than a floating one, which is exactly what makes an error in them measurable.
Where the rest of the money goes
The same census rows give the other large cost blocks, using the same denominator.
| Cost block | Electrical (238210) | Plumbing/HVAC (238220) |
|---|---|---|
| Materials | 31.43% | 32.49% |
| Annual payroll | 28.23% | 27.87% |
| Work subcontracted out | 5.75% | 7.35% |
| Everything else, including profit | 34.59% | 32.29% |
Look at the electrical column and try to fit a 50% material share into it. Materials and payroll alone would come to 78% of revenue, before a truck, a warehouse, an insurance premium, a bond, an office or a dollar of profit. The 40 to 60 percent claim is not just unsourced. It does not survive contact with the rest of the income statement.
Why a census beats a vendor blog
The most-repeated version of the 40 to 60 percent claim comes from a procurement software vendor’s blog. It carries no citation, no sample, no year, and no definition of the denominator. The same page also asserts that comparing eight to ten suppliers saves 5 to 10 percent on materials, “which data consistently shows”, while naming no data.
The distinction is not that one source is a big institution and the other is small. It is methodological, and it is worth stating in full:
- The Economic Census is a mandatory survey with statutory response requirements, not a convenience sample of whoever answered.
- It measures reported revenue and reported cost, taken from the books of establishments that actually did the work, not from an estimator’s judgment of a typical job.
- Its field definitions are published, so you know precisely what was counted as a material and what was not.
- The raw files are free, so a skeptical contractor can recompute the ratio in about ten lines of code and confirm or contradict it.
- The publisher has no product to sell that the ratio makes look better.
A vendor blog fails all five. And note the trap for anyone tempted to use both numbers: cite the census and the 40 to 60 percent claim on the same page and you have refuted yourself in public.
Three other circulating numbers that fail traceback
The 40 to 60 percent claim is not an isolated case. Three widely repeated construction statistics break when you walk them back to a primary document.
The KPMG budget figure is the schedule figure. A procurement vendor’s April 2026 post states that “only 25 percent of construction projects come within 10 percent of their original budget”, attributing it to KPMG’s Global Construction Survey. KPMG’s own 2015 report and press release give 31 percent for budget and 25 percent for schedule. The vendor applied the schedule number to budget. Separately, that KPMG survey rests on face-to-face interviews conducted in late 2014 with 109 senior leaders on the project owner side. It is not a contractor materials study, and anyone citing it in 2026 is citing twelve-year-old data.
The McKinsey overrun range is not in McKinsey. The same post says “the average cost overrun sits between 28 and 33 percent, according to McKinsey Global Institute research.” That range does not appear in Reinventing Construction, and no McKinsey publication containing it could be located. Treat it as unverified. The related habit of restating McKinsey’s $1.6 trillion figure as an amount the industry “wastes” is also wrong on its face: the report describes it as an upside opportunity from closing a productivity gap, not a loss.
The material waste statistics dead-end at Wikipedia. A material management software vendor publishes a statistics page for MEP contractors. On that page, “industry budgets assume 2.5-5% of materials will be wasted”, “actual waste rates run 10-15%”, and “construction accounts for 23% of total U.S. solid waste” are each attributed to Wikipedia. A fourth, “up to 30% of materials delivered to job sites end up as waste”, is attributed to a university idea-pitch event page. A fifth is the KPMG 2015 budget number, laundered through two other vendors’ blogs and re-dated to 2025.
The pattern is consistent enough to be a working rule. If a construction statistic about materials has no primary document behind it, it is usually because there is no primary document behind it.
The arithmetic that makes 32% commercially interesting
A one-third material share is not an academic curiosity. It sets the leverage on every pricing error that reaches your invoices.
Start with the census ratio and take a shop billing $5,000,000 a year, on the electrical figure of 31.43%.
- Materials at 31.43% of $5,000,000 is $1,571,500.
- A 5% overcharge across that spend is 0.05 x $1,571,500, which is $78,575.
- As a share of revenue that is 0.05 x 31.43%, which is 1.57%.
Now the part that requires an assumption, and it should be labeled as one. Net margin varies enormously across specialty trade shops by market, mix and year, and no verified industry-wide contractor net margin figure is cited in this study. So rather than assert one, here is the same overcharge scored against a range of net margins. Each cell is the share of net profit consumed, computed as (0.3143 x overcharge rate) / net margin.
| Overcharge on materials | Cost as % of revenue | At 3% net margin | At 5% net margin | At 8% net margin |
|---|---|---|---|---|
| 1% | 0.31% | 10.5% of profit | 6.3% of profit | 3.9% of profit |
| 2% | 0.63% | 21.0% of profit | 12.6% of profit | 7.9% of profit |
| 5% | 1.57% | 52.4% of profit | 31.4% of profit | 19.6% of profit |
On the $5,000,000 shop, a 3% net margin is $150,000 of profit, and the $78,575 overcharge is 52.4% of it. At an 8% net margin the same overcharge is $78,575 against $400,000, or 19.6%.
Read those columns as sensitivity, not as prophecy. The point that survives every column is that material pricing sits on a base roughly ten times the size of a typical net margin, so a small percentage error in materials lands as a large percentage error in profit. The plumbing and HVAC figure of 32.49% makes the leverage marginally worse, not better.
Two reasons the invoice price is not the real price
If materials carry that much leverage, the obvious follow-up is whether the number printed on a supply house invoice is reliably the agreed number. Two audited disclosures suggest the channel’s own pricing mechanics are more layered than the invoice implies.
Rebates are real but small, and they are not on your invoice. WESCO International is the only company in electrical distribution that publishes a rebate ratio. Its FY2025 10-K states that “Supplier volume rebate income as a percentage of net sales was 1.4% in 2025, 1.3% in 2024 and 1.4% in 2023”, with receivables under the supplier rebate program of $264.0 million at December 31, 2025. Across ten consecutive 10-Ks the figure runs from 1.1% to 1.6%. That is a useful calibration in both directions: rebate economics in this channel are measured in low single digits of sales, not the double digits sometimes implied, and they settle behind the invoice rather than on it.
Suppliers sometimes invoice at list and settle later. Watsco’s FY2025 10-K discloses that it has “arrangements with certain suppliers that invoice us based on a list price rather than on our net buy cost for specific inventory”, with the supplier providing the difference through what Watsco calls a pricing claim advance. At December 31, 2025, $98,512 thousand was recorded as a reduction of inventories for pricing claim advances, of which $71,882 thousand was provided by Carrier and its affiliates.
That is an audited, quantified, manufacturer-named illustration of a simple structural fact: in this channel, the price on the document and the price in the deal are different objects, reconciled after the fact. Both disclosures describe distributors buying from manufacturers, not contractors buying from distributors, so do not read them as a claim about your counter ticket. Read them as evidence that list-price invoicing followed by later adjustment is normal practice in the supply chain immediately above you.
Limits
What the Economic Census measures, and what it does not:
- The denominator is revenue, not job cost. These percentages are materials as a share of value of construction work, which is the industry’s billings. That is not the same as “percent of job cost”, and it should never be silently relabeled as such.
- The data is from 2022 and is the most recent available. The Economic Census runs on a five-year cycle. The 2022 cycle was published in November 2024; the next reference year is 2027. These are not current-year figures and nothing here should be read as describing 2026 conditions.
- Plumbing and HVAC are not separated. NAICS 238220 combines plumbing, heating and air-conditioning contractors into one industry. The census does not split them at this level, and any separate plumbing figure or separate HVAC figure claimed from this source would be fabricated.
- Employer establishments only. Firms with no paid employees are outside the frame.
- These are industry aggregates. They describe the sum of every reporting establishment in a NAICS code, and the variance across individual shops is real and large. A service-heavy shop and a new-construction shop will not look alike, and neither will look exactly like 31.43%. Nothing here predicts any single contractor’s material share.
- The worked example above is an example. The 3%, 5% and 8% net margins are stated assumptions chosen to bracket a plausible range, not measured values, and no contractor net margin source is cited in this study.
What does travel well is the structure. Materials are close to a third of billings for these trades, that share held steady through the worst material price shock in decades, and anything that moves the material number by a few percent moves the profit number by a lot.