Original research · Cost
The Roofing Inflation Index
Twenty-six years of federal price data, split into the two things a roof is made of. The answer is not the one the industry usually gives.
Written by HyreRoof Research Primary-source research and fact checking
The finding
Read this before the charts
Every index on this page measures a producer price or an employer payroll figure. A producer price index tracks what a manufacturer receives for what it ships. It is not a retail price, not a delivered price, not a quoted price, and above all not the price of a roof on a house.
An installed asphalt roof is materials plus labour plus tear-off and disposal, plus scaffolding or staging, plus general liability and workers’ compensation cover, plus permit and inspection fees, plus warranty reserve, plus the contractor’s overhead and margin. No series published by any federal agency measures that bundle for a single-family home. Anyone who tells you "roof prices rose 255% since 2000" has taken a shingle-plant index and put a homeowner’s label on it.
What this page can honestly tell you is how the two largest input costs behaved relative to each other and relative to general inflation, and that comparison is both defensible and, it turns out, surprising.
Why the usual explanation does not survive the data
The standard account of roofing cost inflation is a labour story. Skilled trades are scarce, the workforce is ageing, crews are hard to keep, and so the price of a roof climbs. It is a plausible story and it is repeated everywhere, usually without a number attached.
The federal statistical system publishes both halves of that story separately, and has done for decades. The Bureau of Labor Statistics runs a producer price index for the industry that makes asphalt shingles, and a monthly payroll survey that reports average hourly earnings for roofing contractors specifically. Putting the two on the same axis takes an afternoon. It very rarely gets done, and when it is done the labour story does not hold up.
HyreRoof analysis: over the twenty-five years to 2025, roofing labour costs rose at 2.99% a year compounded, against 2.53% a year for consumer prices generally. Roofing wages beat inflation, but only just: the kind of margin you would expect from any occupation that is not in decline.
Asphalt roofing materials rose at 5.20% a year, more than double the inflation rate, for a quarter of a century. That is the anomaly, and it is on the materials side.
The five instruments, and the bases nobody quotes
| Series ID | What it is | Index base | First observation | What it actually measures |
|---|---|---|---|---|
WPU1361 | Prepared asphalt and tar roofing and siding products (PPI commodity) | 1982 = 100 | January 1926 | What manufacturers received for shingles, roll roofing and cap sheets. |
PCU324122324122 | Asphalt shingle and coating materials manufacturing (NAICS 324122) (PPI industry) | June 1984 = 100 | December 1975 | The output price of the whole shingle-and-coatings industry, not one product. |
PCU23816X23816X | Roofing contractors, nonresidential building work (NAICS 23816X) (PPI industry) | December 2007 = 100 | December 2007 | What roofing contractors received for installed nonresidential roofing work. |
CEU2023816008 | Average hourly earnings, production and nonsupervisory employees, roofing contractors (CES) | US dollars, not an index | January 1990 | Gross hourly pay on employer payrolls in NAICS 238160. |
CUUR0000SA0 | CPI-U, all items, US city average, not seasonally adjusted (CPI) | 1982-84 = 100 | January 1913 | The deflator. Used here to convert nominal changes into constant dollars. |
Series titles and index base dates read from the BLS time-series catalogue files on 3 September 2026.
A PPI number is meaningless without its base. WPU1361 stood at 354.8 in 2025, that is 354.8 against a 1982 average of 100, not against last year, and not against any other series on this page. The three PPI series here have three different bases (1982, June 1984 and December 2007), which is exactly why every chart below rebases them to a common year before plotting.
What the raw series show
Rebasing is not cosmetic here, it is the only way to put a 1982-based index, a June-1984-based index and a dollar figure on the same picture without lying about all three. Each line is that series divided by its own value in 2000.
Three things are visible immediately. The materials line and the earnings line move together until about 2005, then separate. The separation is not gradual: materials jump 12.3% in 2005, 9.7% in 2006, 26.5% in 2008 and a further 23.4% in 2009: the two largest annual moves in the whole series, and they happen through the deepest residential construction downturn in modern memory, when demand for roofing was collapsing.
The second separation begins in 2021: +11.5%, then +17.7% in 2022. That one at least coincides with an economy-wide inflation episode, and the CPI-U line moves too, though much less.
The third thing is easiest to miss and is the most useful. Since 2023 the materials line has flattened almost completely, up 4.8% across two and a half years, while the earnings line has kept climbing at 12.3% over the same window. For the first time in the series, labour is outrunning materials.
The same series in constant dollars
CUUR0000SA0, 1982-84 = 100), annual average of monthly values, retrieved 3 September 2026. Nominal indexes flatter everything, because everything rises. The honest comparison deflates by a general price index, and the arithmetic should be visible rather than asserted.
The deflator. CPI-U all items, US city average, not seasonally adjusted, series CUUR0000SA0, base 1982-84 = 100. Its 2000 annual average is 172.2 and its 2025 annual average is 321.9, so the price level rose by a factor of 321.9 ÷ 172.2 = 1.8696, or 87.0%.
Materials, real. WPU1361 rose from 100.0 to 354.8, a factor of 3.5475. Divide by the deflator: 3.5475 ÷ 1.8696 = 1.8975. Asphalt roofing materials cost 89.7% more in constant dollars in 2025 than in 2000: a real compound rate of 2.60% a year.
Labour, real. Hourly earnings rose from $15.68 to $32.76, a factor of 2.0895. Divide by the same deflator: 2.0895 ÷ 1.8696 = 1.1177. Roofing hourly earnings bought 11.8% more in 2025 than in 2000: a real compound rate of 0.45% a year, or about a nickel on the dollar per decade.
HyreRoof analysis: in real money, the material half of a roof has almost doubled in twenty-five years and the labour half has gained roughly a tenth. If your mental model of roofing inflation is "you cannot get crews any more", the federal payroll data does not support it. Whatever is happening to roofing crews, it is not showing up as real wage growth.
The two decompositions, side by side
| Measure | Series | 2000 → 2025 nominal | Compound annual | Real, deflated by CPI-U |
|---|---|---|---|---|
| Asphalt roofing materials | WPU1361 | +254.7% | 5.20% | +89.7% |
| Asphalt shingle industry output | PCU324122324122 | +229.1% | 4.88% | +76.0% |
| Roofing hourly earnings | CEU2023816008 | +109.0% | 2.99% | +11.8% |
| Softwood veneer and plywood | WPU0831 | +73.7% | 2.23% | −7.1% |
| Steel mill products | WPU1017 | +175.4% | 4.14% | +47.3% |
| Consumer prices | CUUR0000SA0 | +87.0% | 2.53% | by definition, 0% |
HyreRoof calculation from annual averages of BLS monthly data, retrieved 3 September 2026.
Two of these are not roofing but belong on the page. Softwood veneer and plywood is the deck under the shingles: it is cheaper in real terms than in 2000, despite the lumber spike of 2021, which is not what most people would guess. Steel mill products, the input to a metal roof, rose fast in nominal terms but still less than asphalt.
What contractors actually charged
There is one federal series that measures the price of installed roofing work rather than of an input to it: PCU23816X23816X, the producer price index for roofing contractors. It carries a large qualifier, it covers nonresidential building work, not houses, and it only begins in December 2007, which is its index base.
Over the seventeen years from the 2008 annual average to the 2025 annual average, that index rose 107.4%, a compound 4.38% a year. Over exactly the same window, asphalt roofing materials rose 100.7%, roofing hourly earnings rose 75.6% and consumer prices rose 49.5%.
HyreRoof analysis: the price roofing contractors charge has tracked their materials cost far more closely than their labour cost. That is what you would expect if materials were the binding constraint on pricing. It is not what you would expect from a market where wage pressure was doing the work.
Read this series carefully, though. A commercial re-roof is a different product from a residential one: different membranes, different crews, different insulation, different scale. It is evidence about the roofing trade, not a substitute for a residential figure that does not exist.
A roof is not one input, so weight them
Neither series alone answers the question a homeowner is actually asking. The direct cost of an installed roof is a mixture, so the honest construction is a weighted index, and because the weight is an assumption rather than a measurement, it should be run at several values rather than presented as one number.
The formula. For material weight w, the HyreRoof Roofing Input Cost Index in year t, base 2000 = 100, is:
Index(t) = 100 × [ w × WPU1361(t)/WPU1361(2000) + (1 − w) × AHE(t)/AHE(2000) ]
This is a direct-cost basket only. It deliberately contains nothing for overhead, disposal, insurance, permits, warranty or margin, because no public series measures those for residential roofing. It is a floor on input inflation, not an estimate of a quoted price.
HyreRoof calculation, 2025 against 2000. At a 40% material weight the basket stands at 267.3 nominal and 143.0 in constant 2000 dollars. At 50% it is 281.9 and 150.8. At 60% it is 296.4 and 158.6. Across that whole range, the direct input cost of an asphalt roof rose 43% to 59% faster than general consumer inflation over twenty-five years.
How much of that is materials? Of the 181.9 index points the 50%-weight basket gained, materials contributed 127.4 and labour 54.5, so 70.0% of the increase is materials. At a 40% weight it is 60.9%; at a 60% weight it is 77.8%. Whatever reasonable split you assume between the two, materials are the majority of the story and usually the large majority.
The basket at three material weights
| Material weight | 2019 | 2020 | 2023 | 2025 | 2026 (part-year) | Materials share of the rise |
|---|---|---|---|---|---|---|
| 40% materials / 60% labour | 192.7 | 200.4 | 250.8 | 267.3 | 271.4 | 60.9% |
| 50% / 50% | 201.3 | 208.5 | 265.5 | 281.9 | 285.4 | 70.0% |
| 60% materials / 40% labour | 209.8 | 216.7 | 280.3 | 296.4 | 299.4 | 77.8% |
| Same, in constant 2000 dollars, 40% | 129.8 | 133.4 | 141.7 | 143.0 | 141.1 | , |
| Constant 2000 dollars, 50% | 135.6 | 138.8 | 150.1 | 150.8 | 148.4 | , |
| Constant 2000 dollars, 60% | 141.3 | 144.2 | 158.4 | 158.6 | 155.7 | , |
HyreRoof calculation. Base 2000 = 100. Constant-dollar rows divide the nominal row by the CPI-U ratio for that year.
Notice the last two columns of the constant-dollar rows. On every weighting, the real input cost of a roof is lower in the first seven months of 2026 than it was in 2025. That is the first sustained real decline in the series since 2016.
A cross-check the basket does not pass cleanly
What the inputs predict
Rebased to 2008, the input basket reaches 185.6 by 2025 at a 40% material weight, 188.1 at 50% and 190.7 at 60%. The three are close together because materials and labour both roughly doubled over that window, so the weighting barely matters.
That is a prediction: if roofing contractors passed through their direct input costs and nothing else, their output price should have risen by about 86% to 91% since 2008.
What contractors actually charged
The roofing contractor output PPI reached 207.4 on the same 2008 base. That is about 17 to 22 index points above what the input basket predicts: roughly a tenth more than input costs alone would explain.
Four ordinary explanations fit. The output series covers nonresidential work while the inputs are generic. General liability and workers’ compensation sit inside the output price and inside neither input. Disposal and landfill costs do the same. And a basket of two inputs is a crude model of a business.
Why asphalt behaves like a fuel, because it is one
The biggest single-year moves in the roofing materials index are 2008 (+26.5%) and 2009 (+23.4%). Residential construction was in freefall in both years. A materials index that rises by a quarter, twice, while its main end market collapses is not being driven by roofing demand.
The BLS commodity classification hints at why. Asphalt itself is filed under group 05, fuels and related products and power: item 8102, "Asphalt". The finished roofing products made from it are filed somewhere else entirely, under group 13, nonmetallic mineral products, item 61. The raw material for your shingles sits in the same statistical family as diesel and jet fuel; the shingle sits with cement and glass.
HyreRoof analysis: that is the most useful single thing on this page for anyone trying to anticipate roofing prices. The dominant input to an asphalt shingle is priced by refinery economics, not by how many roofs are being replaced. Roofing demand affects the shingle price at the margin. The crude oil complex sets the level.
It also explains the shape of the recent flattening. Since 2023 the materials index has risen 4.8% while consumer prices rose 8.7%: a real fall of 3.6%. Over the same period real hourly earnings in roofing rose 3.3%. The two lines have crossed for the first time in the series.
The reversal since 2023, in four numbers
HyreRoof analysis: if this holds, the composition of roofing cost inflation has genuinely changed. For twenty-three years the pressure came from the material. Since 2023 it has come from the wage. A homeowner deciding whether to replace now or wait is facing a different question in 2026 than in 2021, waiting no longer protects you from a shingle price spike, and it no longer costs you nothing on the labour side either.
The caveat is size. Two and a half years is short, 2026 is a part-year, and monthly PPI figures are revised. This is a change worth watching, not a trend worth betting a re-roof on.
The full series, 2000 to 2026
| Year | Roofing materials WPU1361 · 1982=100 | Shingle industry PCU324122324122 · Jun 1984=100 | Contractor output PCU23816X23816X · Dec 2007=100 | Hourly earnings CEU2023816008 · $ | CPI-U CUUR0000SA0 · 1982-84=100 | Employment CEU2023816001 · 000s |
|---|---|---|---|---|---|---|
| 2000 | 100.0 | 108.0 | not published | $15.68 | 172.2 | 185.1 |
| 2001 | 103.3 | 111.4 | not published | $16.11 | 177.1 | 181.2 |
| 2002 | 106.6 | 114.9 | not published | $16.60 | 179.9 | 177.6 |
| 2003 | 110.6 | 120.3 | not published | $16.99 | 184.0 | 180.7 |
| 2004 | 111.3 | 121.6 | not published | $17.30 | 188.9 | 189.4 |
| 2005 | 125.0 | 133.9 | not published | $17.16 | 195.3 | 203.2 |
| 2006 | 137.1 | 147.5 | not published | $17.23 | 201.6 | 204.5 |
| 2007 | 139.7 | 149.2 | not published | $17.65 | 207.3 | 197.7 |
| 2008 | 176.7 | 185.4 | 105.5 | $18.66 | 215.3 | 191.4 |
| 2009 | 218.1 | 219.2 | 113.8 | $19.26 | 214.5 | 170.9 |
| 2010 | 218.9 | 220.4 | 111.4 | $20.06 | 218.1 | 168.3 |
| 2011 | 226.0 | 231.4 | 113.3 | $20.64 | 224.9 | 169.9 |
| 2012 | 221.2 | 232.2 | 117.6 | $20.59 | 229.6 | 164.6 |
| 2013 | 229.7 | 237.7 | 119.7 | $20.89 | 233.0 | 165.9 |
| 2014 | 224.5 | 233.3 | 124.5 | $20.77 | 236.7 | 169.9 |
| 2015 | 223.2 | 231.5 | 127.3 | $21.56 | 237.0 | 184.1 |
| 2016 | 221.6 | 229.6 | 129.2 | $22.46 | 240.0 | 189.8 |
| 2017 | 220.5 | 228.7 | 132.2 | $23.45 | 245.1 | 199.5 |
| 2018 | 233.8 | 242.0 | 133.4 | $24.46 | 251.1 | 210.7 |
| 2019 | 243.9 | 251.4 | 136.9 | $24.87 | 255.7 | 219.9 |
| 2020 | 249.2 | 255.8 | 141.3 | $26.33 | 258.8 | 214.6 |
| 2021 | 277.7 | 282.1 | 149.8 | $28.38 | 271.0 | 223.0 |
| 2022 | 326.9 | 329.2 | 177.9 | $29.25 | 292.7 | 229.5 |
| 2023 | 339.3 | 341.4 | 204.4 | $30.07 | 304.7 | 243.6 |
| 2024 | 350.0 | 350.8 | 209.9 | $31.08 | 313.7 | 252.1 |
| 2025 | 354.8 | 355.5 | 218.8 | $32.76 | 321.9 | 254.8 |
| 2026 part-year | 355.4 | 357.2 | 226.8 | $33.77 | 331.2 | 245.9 |
Annual averages of monthly observations, computed by HyreRoof from BLS data retrieved 3 September 2026. 2026 covers January to July for the price series and January to June for the payroll series, so it is not comparable to a full year without that caveat.
The contractor output column is blank before 2008 because the series does not exist before December 2007, that is its base month. A blank here means "not published", never "zero", and it is drawn differently for that reason.
The workforce behind the wage line
Hourly earnings are only half of what a roofer earns and none of what the trade looks like. Two more series from the same payroll survey fill it in.
Weekly earnings. Average weekly hours in roofing were 35.15 in 2000 and 36.46 in 2025. Multiplying by hourly earnings gives $551.12 a week in 2000 and $1,194.47 in 2025, a nominal rise of 116.7%, slightly more than the hourly figure, because the working week lengthened a little. Against 87.0% consumer inflation, that is a real gain of about 16%. Twenty-five years of it.
Headcount. Employment at roofing contractors ran at 185,100 in 2000, peaked at 204,500 in 2006, fell to 164,600 by 2012: a 19.5% collapse from the peak: and then recovered to 254,800 in 2025, the highest in the series. The first half of 2026 averages 245,900, below the 2025 average.
HyreRoof analysis: a trade that shed a fifth of its workforce and then grew it by 55% from the trough, while real hourly pay moved by roughly a tenth across the whole period, does not look like a labour market clearing through price. It looks like one clearing through headcount. That is consistent with the wage line on the charts above and inconsistent with the shortage narrative as usually told.
How to use this if you are about to buy a roof
- Do not convert these percentages into a price
None of them is a price. If you want a figure for your own house, start from an estimate built on your measurements: the roof cost calculator shows its assumptions on the page, and treat this study as context for why the number is where it is.
- Ask a quote to separate materials from labour
A quote that gives one number tells you nothing about which half moved. A quote that splits them can be checked against the series above, and against the next quote.
- Treat a "material price increase" surcharge as checkable
Monthly PPI data for asphalt roofing products is public and free. If a surcharge is attributed to shingle prices in a period when WPU1361 was flat, that is a fair question to ask, politely, before signing.
- Do not assume waiting saves money
It did not between 2004 and 2009, or between 2020 and 2023. Real material prices have eased since 2023, but the labour component is now the one rising, and a failing roof generates costs that no index tracks.
- Remember what an index cannot see
Regional differences, steep-slope premiums, tear-off of multiple layers, deck repair, code upgrades and disposal fees are all invisible to a national producer price index and all present on your invoice.
Method
- 1 Series selection
Candidate series were identified from the BLS time-series catalogue files rather than by searching for a number that fitted.
pc.industryandpc.productwere read to find every PPI industry covering roofing manufacture or roofing contracting;wp.itemandwp.groupfor every commodity index covering asphalt roofing products;ce.seriesfor every payroll series published for NAICS 238160. Base dates were taken frompc.seriesandwp.series. - 2 Retrieval
Monthly observations for 2000 to 2026 were pulled from the BLS Public Data API v2 in three requests (2000–2009, 2010–2019, 2020–2026) on 3 September 2026. Values the API returns without a numeric figure were skipped rather than interpolated or zeroed.
- 3 Annualisation
Each annual figure is the unweighted mean of the monthly observations published for that year. 2026 has seven months of PPI and CPI data and six months of CES data, and is labelled as a part-year everywhere it appears.
- 4 Rebasing
Because the three PPI series carry three different bases, every chart divides each series by its own value in the chart’s base year. No series is ever plotted against another series’ base, and no two bases are ever spliced together.
- 5 Deflation
Real figures divide the nominal ratio by the CPI-U ratio for the same years. Example: materials 354.8 ÷ 100.0 = 3.5475; CPI-U 321.9 ÷ 172.2 = 1.8696; 3.5475 ÷ 1.8696 = 1.8975, i.e. +89.7% real. Every real figure on this page is that arithmetic with different inputs.
- 6 Compound rates
Compound annual rates use (end ÷ start) raised to the power 1/n, minus one, with n = 25 for 2000→2025 and n = 17 for 2008→2025.
- 7 The weighted basket, run at three values
The material weight in an installed roof is an assumption, not a measurement, so it is run at 40%, 50% and 60% and all three are published. The conclusion that materials are the majority of the increase holds at every one of them, which is the only reason it is stated as a conclusion.
- 8 What was deliberately not done
No producer price change is converted into a homeowner price. No national dollar figure for an installed roof is published, because no federal series measures one. No survey of quotes is used, and no contractor pricing data is used.
Limitations
- Producer price indexes are not homeowner prices, and no adjustment makes them so
Producer price indexes measure what producers receive. They exclude sales tax, and they exclude every element of an installed roof that is not the manufactured good itself. The weighted basket is a direct input-cost index and nothing more.
- The planned OEWS wage series could not be retrieved
The production plan named BLS OEWS for occupation 47-2181 (Roofers). On 3 September 2026 the OEWS occupational profile at bls.gov returned "Access Denied" to automated retrieval, no national OEWS series for that occupation was located in the BLS time-series catalogue files, and the unregistered API quota was exhausted before an alternative series identifier could be tested. No OEWS figure is published here. The substitute, CES
CEU2023816008, is arguably the better instrument in any case: OEWS is a three-year rolling cross-section not designed for comparison over time, while CES is a monthly time series built for exactly that. - Average hourly earnings is not a wage rate
CES average hourly earnings is gross payroll divided by hours for production and nonsupervisory employees. It moves when the occupational mix changes, when overtime changes and when the geographic mix of employment changes, not only when anybody gets a raise. It also excludes self-employed roofers entirely, and roofing has a large self-employed segment.
- The only installed-price series covers nonresidential work
PCU23816X23816Xis titled "Roofing contractors, nonresidential building work". There is no residential equivalent. Every statement made from it on this page is qualified in the same sentence, and none of it should be read as a house price. - 2026 is a part-year and PPI is revised
The 2026 figures average seven months of price data and six months of payroll data. PPI and CES data are subject to revision after first publication, so figures near the end of the series are less firm than figures near the start.
- Annual averages hide within-year movement
Averaging twelve monthly observations smooths away the peaks. The 2022 monthly high in WPU1361 is above the 2022 annual average shown here; anyone who bought at that point paid more than this table implies.
- National indexes have no geography
Every series here is a national aggregate. Regional shingle pricing, regional wage levels, regional disposal costs and regional code requirements are all invisible to them.
- This is not financial or purchasing advice
Historical price behaviour is not a forecast. The reversal since 2023 is two and a half years long and could reverse again in either direction.
Questions
How much have roof prices actually gone up since 2000?
Which BLS series measures shingle prices?
WPU1361 is the PPI commodity index for prepared asphalt and tar roofing and siding products, base 1982 = 100, published since 1926. PCU324122324122 is the PPI industry index for asphalt shingle and coating materials manufacturing (NAICS 324122), base June 1984 = 100, published since 1975. The first follows a product group, the second follows an industry’s output. Over 2000 to 2025 they rose 254.7% and 229.1% respectively.Why does a PPI number need its base period stated?
WPU1361 stood at 354.8 in 2025 and PCU23816X23816X stood at 218.8 in the same year, but the first is measured against a 1982 average of 100 and the second against December 2007 = 100. Comparing 354.8 with 218.8 directly would be meaningless. Every chart on this page rebases each series to a common year before plotting it, for exactly that reason.Is a producer price index the same as what I pay?
Did materials or labour drive roofing cost inflation?
Are roofing wages rising because of a labour shortage?
Why did shingle prices rise so sharply in 2008 and 2009?
WPU1361 rose 26.5% in 2008 and a further 23.4% in 2009: the two largest annual moves in the whole 2000 to 2026 window, while US residential construction was in a severe downturn. A materials index rising a quarter twice while its main end market contracts is not being driven by roofing demand. BLS classifies asphalt itself under fuels and related products, and finished asphalt roofing products under nonmetallic mineral products, which is a strong hint about where the pressure came from. We have not attempted to quantify the crude oil link and do not claim it.Have roofing material prices come down?
WPU1361 rose 4.8% while CPI-U rose 8.7%: a real fall of about 3.6%. Over the same window roofing hourly earnings rose 12.3% nominally, or about 3.3% in real terms. That is the first period in the series where labour outran materials in real terms. It is two and a half years of data, one of them a part-year, so treat it as a change to watch rather than an established trend.Is there a federal index for the price of an installed residential roof?
PCU23816X23816X, the PPI for roofing contractors, and its title says "nonresidential building work". It rose 107.4% between the 2008 and 2025 annual averages, against 49.5% for consumer prices. That is useful evidence about the roofing trade, but it is a commercial series and it should not be quoted as a house figure.What deflator did you use, and why that one?
CUUR0000SA0, base 1982-84 = 100. It is the standard general-purpose measure of the consumer price level and it is what a homeowner’s money actually competes against. Its 2000 annual average is 172.2 and its 2025 annual average is 321.9, a factor of 1.8696. Every real figure on this page is a nominal ratio divided by that year’s CPI-U ratio, and the arithmetic is shown in the Method section rather than asserted.What happened to plywood and steel, for the deck and for metal roofing?
WPU0831) rose 73.7% between 2000 and 2025, which is less than consumer inflation: a real fall of 7.1%, despite the well-publicised 2021 lumber spike that took the annual average to 449.6 before it fell back. Steel mill products (WPU1017) rose 175.4%, a real gain of 47.3%. Both are meaningful context: the deck under an asphalt roof has become cheaper in real terms, and the main input to a metal roof rose less than asphalt did.How often will this page be updated?
Can I reuse these figures?
Written and audited by
HyreRoof Research
Primary-source research, data analysis and fact checking
We are a research desk, not a sales floor. We read the statute, the licensing board’s own pages, the code section or the federal dataset ourselves, and we publish the figure with the document it came from and the date we retrieved it. Where a number cannot be traced to a primary source, we publish the shorter page and say what we could not verify. On our first study that rule removed a Minnesota exam statistic and left two states blank. Those gaps are on the page, not in a file somewhere.
- 36
- primary sources read and cited
- 16
- federal and state government domains
- 36
- citations carrying a retrieval date
- 3
- researched pages published
How this desk works
- Primary sources only. Statutes from the legislature’s own publishing system, licensing rules from the board that issues the licence, datasets from the agency that collected them. Never a directory, an aggregator or another guide.
- Three states, not two. A requirement is recorded as verified present, verified absent, or not verified. Most comparisons collapse the third into the second, which turns an unchecked cell into a factual claim.
- Retrieval dates on everything. Regulation changes. A citation without the date it was read is not a citation.
- Failures are published. When a source blocks automated retrieval we record the failure and leave the row empty, rather than filling it from a secondary summary.
- Authorship is organisational. Research is attributed to this desk, never to an invented expert. Outside commentary, where used, is attributed to named and verifiable people.
Data as of 3 September 2026. Authorship on this site is organisational: the analysis belongs to the desk rather than to a named individual, and we do not publish credentials we do not hold. Our editorial policy sets out how we source, date and correct what we publish.
Sources & retrieval dates
- BLS Public Data API v2 , Monthly observations for WPU1361, WPU136, WPU1362, PCU324122324122, PCU23816X23816X, CEU2023816008, CEU2023816007, CEU2023816001, CUUR0000SA0, WPU0831 and WPU1017, 2000 to 2026 Retrieved 3 September 2026.
- BLS time-series catalogue, pc.series , PPI industry series titles and index base dates, including PCU23816X23816X (base 200712) and PCU324122324122 (base 198406) Retrieved 3 September 2026.
- BLS time-series catalogue, pc.industry , PPI industry codes, confirming 23816X "Roofing contractors, nonresidential building work" and 324122 "Asphalt shingle and coating materials manufacturing" Retrieved 3 September 2026.
- BLS time-series catalogue, pc.product , PPI product codes under NAICS 324122, including 3241221 and 3241222 Retrieved 3 September 2026.
- BLS time-series catalogue, wp.series , PPI commodity series titles and base dates, including WPU1361 (base 198200) and WPU0831 Retrieved 3 September 2026.
- BLS time-series catalogue, wp.item , PPI commodity item structure: asphalt itself at 05-8102, prepared asphalt roofing products at 13-61 Retrieved 3 September 2026.
- BLS time-series catalogue, wp.group , PPI commodity groups, confirming group 13 is "Nonmetallic mineral products" Retrieved 3 September 2026.
- BLS time-series catalogue, ce.series , CES series titles for NAICS 238160 roofing contractors, including CEU2023816008 (average hourly earnings, production and nonsupervisory, from 1990) Retrieved 3 September 2026.
- BLS time-series catalogue, oe.occupation , OEWS occupation codes, confirming 472181 "Roofers" Retrieved 3 September 2026.
- BLS time-series catalogue, oe.area and oe.areatype , OEWS area coding, confirming areatype N / area 0000000 is the national aggregate. Searched in support of the OEWS retrieval attempt recorded in Limitations Retrieved 3 September 2026.
- BLS Producer Price Index programme , Programme documentation. Returned "Access Denied" to automated retrieval on 3 September 2026; no text from it is quoted on this page Retrieved 3 September 2026.
- BLS OEWS occupational profile, 47-2181 Roofers , The wage source named in the production plan. Returned "Access Denied" to automated retrieval on 3 September 2026; no OEWS figure is published on this page Retrieved 3 September 2026.
Want a number for your own roof rather than an index?
The index series explain the direction of travel. They cannot price your house. Both of these work from your measurements, publish their assumptions, and ask for nothing from you.
HyreRoof does not manufacture or sell roofing materials, does not perform roofing work, and has no commercial relationship with any manufacturer, distributor or contractor. This study analyses public Bureau of Labor Statistics data, which HyreRoof did not collect. Every calculation labelled "HyreRoof analysis" is ours and is reproducible from the series identifiers above. If a figure is wrong, tell us and we will correct it in place and note the correction.