Which American homes need repairs?

Key findings

• The Philadelphia Fed estimates $198.4 billion of outstanding repair needs from 2023 housing conditions, priced using 2024 contractor costs. Source

  Link to finding

• In all four regions, homes built before 1980 had a higher estimated prevalence of repair needs than later homes. The descriptive gaps ranged from 10.4 to 13.3 percentage points. Source

  Link to finding

• In the Northeast, estimated repair needs affected 43.6% of rented homes and 34.6% of owned homes, a 9.0-percentage-point descriptive gap. Source

  Link to finding

• The South’s estimated repair bill was $80.4 billion. Its 36.0% prevalence shows why a large total bill should not be read as a high probability for an individual home. Source

  Link to finding

What does the repair estimate measure?

The model connects reported housing problems to possible repairs and prices them using contractor-cost data. It covers 110 repair scenarios, with rules to avoid counting overlapping work twice. A home with no estimated cost may still have problems the survey or model does not capture.

This is not an annual maintenance budget, an amount already spent, or a quote for bringing every home to an ideal standard. It also cannot identify all hazards: the research does not price every unobserved problem, including lead-paint and water-contamination risks.

Age and tenure reveal more than the total bill

Across each of the four regions, the published pre-1980 group has a higher prevalence of repair needs than the later-built group. Our subtraction puts those gaps between 10.4 and 13.3 percentage points. That pattern warrants attention, but it does not show that age alone caused the difference.

Rented homes also have higher point estimates than owned homes in each region. Income, building type, maintenance history and other factors can differ between these groups. The chart and table describe the published comparisons; they do not adjust for those factors or rank regions by statistical significance.

The Room Study subtractions from Philadelphia Fed data appendix, December 2025, pages 2–5. Descriptive gaps, not significance tests. Method and limits

Download chart (SVG)

A bigger bill does not mean a worse typical home

The South has the largest total bill among the four regions, at $80.4 billion. Its estimated prevalence of repair needs is 36.0%, compared with 37.9% in the Northeast. Population size and the mix and cost of repairs all contribute to a regional total.

The four rounded regional cost figures sum to $198.3 billion, while the report’s national estimate is $198.4 billion. That rounding difference should not be “corrected” by changing a region’s number. Likewise, the report’s mean and median costs cover only homes with positive estimated repair costs.

What should come before a room refresh?

The useful distinction is between improving how a room looks and understanding what the building needs. A national estimate cannot tell us what is wrong with a particular home. It can remind us to establish the condition of the building before allocating the whole budget to furniture and finishes.

Our housing-age page provides a separate 2024 view of where older occupied homes are concentrated. The two pages support different questions. Their records have not been matched, and we do not multiply the 2024 age shares by these repair rates to manufacture a new national estimate.

Methodology & sources

The source is the Philadelphia Fed’s December 2025 brief and regional data appendix, pages 2–5. Conditions come from the 2023 AHS public-use file; prices come from 2024 Gordian RSMeans contractor-cost data. The table denominator is all occupied homes in each named group, not only homes that need repairs.

The Room Study calculated the age gap as the published pre-1980 percentage minus the published 1980-or-later percentage. The tenure gap is rented minus owned. These simple subtractions were independently reproduced; they use one-decimal source estimates. The CSV preserves those inputs, the source page and units.

We have not reproduced the Fed’s full repair-cost model or licensed cost database. Sampling, reporting and model uncertainty remain. The regional appendix does not supply confidence intervals for the displayed comparisons, and this page adds no significance test. No region is labeled “worst,” and age is not treated as a cause.

Cooling repairs were added in this release and account for about 9% of the total. Other methodological changes also affect comparisons. This page therefore does not treat the raw headline change from an earlier report as a like-for-like trend. The source has a separate consistent-needs comparison; our scope here is the four-region cross-section.

The report covers selected metropolitan areas as well as regions, but it is not a complete 50-state estimate. We publish no unsupported state ranking. Selected numerical facts and our comparisons are shared with attribution; the source reports and licensed repair-cost inputs remain with their publishers.

We review these resources when the underlying releases change. A review date does not make a figure expire: its measurement period stays visible until a checked revision replaces it.