The 82-Year Gap: How Building-Stock Age Maps Trade Demand Across America

Building age is a starting point for local demand research

If you want to know where replacement and retrofit work could come from, don't stop at population growth. Look at how old the buildings are, and treat the answer as a hypothesis to test rather than destiny. Old HVAC systems fail. Old panels get upgraded. Old roofs get torn off. Old pipe gets replaced. But building age is not equipment age: many older buildings have had systems replaced since construction, and owners on thin budgets defer work when cash is tight. Level's own building-level study of 25,912 commercial buildings found that age moved replacement revenue in only 3 of 12 segments, inverted in retail, and did not move routine service in office buildings. That study is HVAC and mechanical only, so it does not test roofing, plumbing or electrical demand.

So we pulled the year built for tens of millions of parcels from national tax-assessor records and computed the median recorded year built across parcels by metro. The spread is enormous.

The oldest and newest metros in America

Method: public county tax-assessor parcel records, current assessment snapshot. Median year built over parcels with a valid year built (1700-2026), by metro (top 34 by parcel count). n per metro in the hundreds of thousands. Median age is 2026 minus median year built. Population note: these are parcels, not housing units or buildings; one parcel can hold many buildings or units, and a 1700-2026 range filter does not remove placeholder years that fall inside that range. County coverage per metro and the parcel types included are not shown here, so treat metro figures as provisional. Census ACS "median year structure built" counts housing units, a different population, and can differ materially from these parcel medians.

There is an 82-year gap between the oldest metro in this dataset (Pittsburgh, median build year 1917) and the newest (Austin, 1999). In Pittsburgh, 90.8% of parcels with a valid year built predate 1980, the highest share among the 34 metros we measured. In Austin, barely a quarter (25.9%) does.

Nationally, the housing stock is old overall: Census American Community Survey estimates, which we have not linked or re-verified on this page, put the national median home age around 42 to 43 years, with roughly half of U.S. homes built before 1980. Those estimates count housing units, not parcels, so they are context rather than a check on the table above. The point is how unevenly age is distributed, and what that might mean for where you work.

Two Americas for the trades

The geography sorts into two very different demand profiles:

The replacement / retrofit belt, Northeast and Rust Belt. In the table, Pittsburgh (1917), the NY metro (1936), Philadelphia (1947), Detroit (1960) and Cleveland (1963) carry median build years from 1917 to 1963, with 67.9% to 90.8% of parcels built before 1980. Our dataset also places DC, Pennsylvania, New York, West Virginia, and New Jersey in this group at the state level; those state figures are not shown on this page. A building this old has likely been through at least one system replacement already, so parcel age alone does not say when the next one is due. Treat these metros as candidates for HVAC replacement, re-piping, panel upgrade, and re-roofing demand, to test against local permit and job data, not as a guaranteed well of work.

The new-install belt, Sun Belt and Mountain West. In the table, Dallas-Fort Worth (1988), Orlando (1989), Phoenix (1990) and Austin (1999) carry the newest median build years. Our dataset also places San Antonio and, at the state level, Nevada, Utah, Arizona, Texas, and Idaho in this group, with median build years from 1987 to 1999; those rows are not shown on this page. Our hypothesis is that demand here leans toward new-construction rough-in and first-cycle replacement rather than deep retrofit, with a cash-flow profile more tied to new construction, which tends to be more cyclical. That is untested on this page.

Old doesn't mean cheap, and that's the sweet spot

A more useful question than age alone may be age paired with value. Our dataset puts Washington DC's median vintage at 1942 (not shown in the table above), but this page does not show property values for DC or any other market, so we are not making a value claim here. Operator check: pull median assessed or market values for the counties you are considering from the same assessor files, and compare them with median year built before you rank a market.

The hypothesis: where old stock is also high-value, more owners may renovate and upgrade rather than tear down, which could support premium, spec-grade retrofit work such as high-end HVAC replacements, whole-home re-wires and architectural roofing. The counterexample to watch for is an old, high-value building whose owner runs a thin maintenance budget and defers. Test the idea with your own win rate and average ticket by market before treating it as a pricing signal.

How to use this

• Targeting a market to expand into? Weight toward old, high-value stock if you do replacement/retrofit work; weight toward new-install metros if your model is new construction.
• Forecasting demand durability? Treat it as a hypothesis that replacement-belt demand is less tied to the construction cycle than new-install demand. At the building level, replacement revenue is episodic and lumpy, and owners defer when budgets are thin, so an aging system does not guarantee a booked job in a downturn. New-install-belt demand is likely to move more with housing starts. Check it against your own replacement job counts by quarter across at least one slow period.
• Pricing? Building age alone is not a pricing signal. If your own data shows you winning at higher average tickets in old, high-value markets, that supports premium positioning there; this page does not show that for DC, the NY metro or anywhere else.

Structural demand can tell you where to look for the work. What separates the contractors who profit from it is whether they know their unit economics well enough to price and staff for it.

Related reading

• How Many Competitors Do You Really Have?
• When Do Contractors Actually Need a CFO?

FAQ

Which U.S. metro has the oldest buildings? Among the 34 metros in this parcel dataset, Pittsburgh, median build year 1917, with 90.8% of parcels built before 1980. New York (1936) and Philadelphia (1947) follow. These are parcel medians from assessor records, not Census housing-unit medians, and county coverage per metro is not yet disclosed.

Why does building age predict trade demand? It is a structural indicator worth testing, not a proven predictor. Building systems, HVAC, plumbing, electrical, roofing, have finite useful lives, so older stock can generate replacement and retrofit demand that may be less tied to the construction cycle than new-install work. But equipment is often replaced on a different schedule from the building, owners defer, and Level's building-level HVAC and mechanical study found that age moved replacement revenue in only 3 of 12 segments and inverted in retail.

Where does this data come from? Public county tax-assessor property records (year built), aggregated to state and metro medians. National context comes from U.S. Census / American Community Survey housing-age estimates, which count housing units rather than parcels; we have not published a metro-by-metro comparison against ACS. Machine-readable data: levelcfo.com/data/benchmarks/permit-market.json. No address or individual parcel is identified.