The Geography of Remodeling Cost Burden

Household income, home values, and the uneven financial weight of major home repairs across 147 U.S. markets.

Published: August 2026 Sample: 147 U.S. Markets MARKET ANALYSISNATIONALRB-2026-002
−0.827
Income vs. income-burden — the dominant factor
3.95×
Burden spread: 9.56% (Frisco, TX) → 37.76% (Detroit, MI)
−0.682
Home value vs. cost-to-home-value — a distinct signal
+0.071
5-yr appreciation vs. burden — not a strong predictor
10.4 pp
Dallas–Frisco within-metro gap, same costs, same HPA series

Executive Summary

This brief presents a cross-sectional analysis of the financial burden that a standardized major home repair places on households across 147 U.S. cities and suburbs. Burden is measured as a standardized remodeling project cost relative to either household income or home value — not as a measure of homeowner cash flow, credit access, or actual repair behavior.

The central finding is that household income is the dominant correlate of remodeling cost burden across the sample (Pearson r = −0.827, income vs. roof cost-to-income, n = 147). The income-burden ratio ranges from 9.56% in Frisco, TX to 37.76% in Detroit, MI — a 3.95× spread driven almost entirely by income variation. Project costs are relatively uniform across markets (interquartile range $14,065–$16,095 for a mid-tier roof replacement), so income is the effective variable.

The Zillow typical home value (ZHVI) provides a distinct and partially independent burden signal (r = −0.682, ZHVI vs. roof cost-to-home-value). The income-based and home-value-based measures are correlated with each other (r = +0.747) but not interchangeable: some markets that appear heavily burdened on an income basis carry comparatively low home-value burden, and vice versa.

Five-year home price appreciation (2019–2024, FHFA All-Transactions HPI) is not a strong cross-sectional predictor of current remodeling cost burden (r = +0.071, HPA5 vs. income burden; r = −0.143, HPA5 vs. home-value burden). Markets with similar appreciation trajectories span nearly the full distribution of burden levels. This does not establish that appreciation has no effect on remodeling economics — it shows that cross-sectional appreciation levels alone are not a reliable proxy for current burden.

Finally, within-metropolitan variation in income burden can be as large as cross-regional variation. Within the Dallas MSA, Frisco (9.56%) and Dallas proper (19.94%) differ by 10.4 percentage points despite sharing the same FHFA appreciation series and nearly identical project costs. Metro-level averages conceal variation of this magnitude.


Research Question & Data

Research question: What factors are most strongly associated with differences in the financial burden of major home repairs across U.S. housing markets? Three explanatory variables are examined — household income, home value, and recent home price appreciation — against two burden measures analyzed in parallel: cost-to-income and cost-to-home-value. The analysis is cross-sectional and observational; it describes associations, not causal mechanisms.

Sample

147 U.S. cities and suburbs drawn from the NumeralQ active market footprint. The sample is not a probability sample of U.S. housing markets, and no claim of national representativeness is made.

Remodeling Cost Estimates — NumeralQ V2 Model

Two standardized projects: Roof replacement — 2,000 sq ft asphalt-shingle residential reroof, base mid-tier $14,500. HVAC replacement — mid-tier gas furnace + central A/C split-system, existing ductwork, base mid-tier $11,000. Localized cost = base × market multiplier. All analysis uses mid-tier costs.

Limitation: model outputs, not observed transaction prices. Actual costs vary by scope, materials, contractor, and permit requirements.

Household Income — ACS 2024 5-Year (Table B19013)

Median household income, Census place geography, reference period 2020–2024. Four markets carry margins of error exceeding 10%: Virginia Beach, VA (24.9%), West Palm Beach, FL (18.2%), Aurora, IL (13.2%), Naperville, IL (11.8%).

Limitation: household income is not disposable income; it excludes taxes, debt service, and other obligations.

Home Value — Zillow ZHVI, June 2026

ZHVI, All Homes, middle tier (33rd–67th percentile), smoothed, seasonally adjusted. This is Zillow’s “typical home value,” not the median. The Woodlands, TX (CDP) is absent from Zillow’s city-level file; Houston metro series substituted ($308,933).

Limitation: an asset-value estimate, not liquid financial capacity or accessible home equity.

Home Price Appreciation — FHFA HPA5

5-year appreciation = (Index Q4 2024 − Index Q4 2019) ÷ Index Q4 2019 × 100. FHFA All-Transactions HPI, MSA/Metropolitan Division level; 147 markets map to 125 unique FHFA series.

Limitation: window includes pandemic-era dynamics and is not a current-market indicator.

Descriptive Statistics

VariableMeanMedianSDMinMax
Roof mid cost ($)$15,316$14,790$1,946$12,470$23,200
Household income ($)$75,623$70,518$22,895$39,938$165,576
Zillow typical home value ($)$414,161$367,620$243,103$77,245$1,538,073
FHFA HPA5 (%)55.0%54.9%13.0%14.3%88.5%
Roof cost-to-income (%)21.69%21.00%5.79%9.56%37.76%
Roof cost-to-home-value (%)4.67%4.14%2.45%1.22%19.52%

Income varies roughly 2.4× more than project costs (coefficient of variation 30.3% vs. 12.7%) — the primary reason income dominates the burden distribution.


Finding 1  —  Household Income Is the Dominant Correlate of Remodeling Cost Burden

Household income is the strongest observed correlate of remodeling income burden across the 147-market sample (r = −0.827). The relationship is partially mechanical — income is the denominator — but the empirical strength confirms income as the first-order predictor. The income-burden spread (9.56% to 37.76%) is driven almost entirely by income variation, not cost variation.

The bottom of the distribution is occupied by high-income suburbs: Frisco, TX (9.56%), The Woodlands, TX (10.10%), Cary, NC (10.21%), Sugar Land, TX (10.43%), Bellevue, WA (11.03%). The top is occupied by low-income legacy cities: Detroit, MI (37.76%), Trenton, NJ (36.43%), Cleveland, OH (36.25%), Hartford, CT (36.24%), Newark, NJ (36.21%). All highest-burden markets carry incomes below $53,000; none have unusually elevated project costs.

Key arithmetic — columns: roof_local_mid ÷ household_income Frisco, TX       14,355 ÷ 150,212 = 9.5565%
Detroit, MI      15,080 ÷  39,938 = 37.7585%
Bellevue, WA     18,270 ÷ 165,576 = 11.0342%
Trenton, NJ      19,140 ÷  52,537 = 36.4315%

Spread: 37.7585% ÷ 9.5565% = 3.95×
9.1% 15.2% 21.3% 27.4% 33.5% 39.6% $39K $65K $91K $118K $144K $171K Frisco, TX Detroit, MI Bellevue, WA Trenton, NJ The Woodlands Figure 1 — Income vs. Income Burden (n=147) r = −0.827 Median Household Income ($000s, ACS 2024 5-yr) Roof Cost as % of Income

Figure 1. Household income vs. roof cost-to-income burden, all 147 markets (r = −0.827). Burden falls steeply as income rises and flattens above roughly $120,000.


Finding 2  —  Home Value Provides a Distinct and Partially Independent Burden Signal

The Zillow typical home value is strongly and negatively correlated with the cost-to-home-value burden measure (r = −0.682): markets with higher home values show lower cost burden relative to home value, because project costs grow far more slowly than home values across the sample.

The income-based and home-value-based measures are correlated but not interchangeable (r between ZHVI and household income = +0.747, but they diverge materially in individual markets). Detroit leads both distributions — highest income burden (37.76%) and highest home-value burden (19.52%) — because both its income and home value are depressed. Miami carries high income burden (28.09%) but comparatively moderate home-value burden because its ZHVI ($582,620) is high relative to its household income ($62,462).

Key arithmetic — columns: roof_local_mid ÷ zillow_typical_home_value Detroit, MI     15,080 ÷   77,245 = 19.52%   (income burden also #1)
Miami, FL       17,545 ÷  582,620 =  3.01%   (income burden = 28.09%)
San Jose, CA    18,995 ÷ 1,413,804 =  1.34%
Bellevue, WA    18,270 ÷ 1,470,486 =  1.24%

Detroit – Miami income-burden gap: 9.67 pp   |   home-value-burden gap: 16.51 pp
1.2% 5.0% 8.9% 12.8% 16.6% 20.5% $75K $377K $679K $981K $1K $2K Detroit Miami San Jose Figure 2 — Home Value vs. Home-Value Burden (n=147) r = −0.682 Zillow Typical Home Value ($000s, June 2026) Roof Cost as % of Home Value

Figure 2. Zillow typical home value vs. roof cost-to-home-value burden, all 147 markets (r = −0.682). High-ZHVI coastal markets sit at the low-burden extreme on this measure even where income burden is elevated.


Finding 3  —  Five-Year Appreciation Is Not a Strong Cross-Sectional Predictor of Current Burden

Five-year home price appreciation (FHFA All-Transactions HPI, Q4 2019–Q4 2024) shows near-zero correlation with income-based burden (r = +0.071) and a weak negative correlation with home-value burden (r = −0.143). Markets with similar appreciation levels span nearly the full range of burden outcomes — Frisco (54.9% HPA5, 9.56% burden) and Detroit (50.8% HPA5, 37.76% burden) sit at opposite ends of the burden distribution despite comparable appreciation.

This finding should not be read as showing that appreciation has no economic relevance to remodeling. Appreciation affects accessible home equity, refinancing capacity, and willingness to invest in repairs — channels not captured in this cross-sectional cost-burden measure. The finding applies specifically to the cross-sectional association between cumulative 2019–2024 appreciation and current income-burden ratios across this 147-market sample.

Key correlations — source: locked correlation matrix, n = 147 HPA5 vs. roof cost-to-income:           r = +0.071   (near zero)
HPA5 vs. roof cost-to-home-value:      r = −0.143   (weak negative)
Income vs. roof cost-to-income:         r = −0.827   (11.7× stronger, for comparison)
9.1% 15.2% 21.3% 27.4% 33.5% 39.6% 14% 29% 45% 60% 76% 91% Frisco Detroit Miami Figure 3 — HPA5 vs. Income Burden (n=147) r = +0.071 FHFA 5-Year HPI Appreciation, Q4 2019–Q4 2024 (%) Roof Cost as % of Income

Figure 3. FHFA 5-year appreciation vs. roof cost-to-income burden, all 147 markets (r = +0.071). The near-flat trend line is the finding: appreciation level does not reliably predict current burden.


Finding 4  —  Within-Metro Variation Can Equal or Exceed Cross-Regional Variation

Metro-level burden averages obscure material within-metro income heterogeneity. The Dallas–Fort Worth complex illustrates this most sharply: Dallas proper (income $70,518, burden 19.94%) and Frisco (income $150,212, burden 9.56%) share the same FHFA appreciation series and nearly identical project costs. Their burden difference — 10.39 percentage points — is driven entirely by income.

The same pattern holds in the Houston complex: Houston proper (21.70% burden) vs. The Woodlands (10.10% burden) — an 11.60 pp gap within the same metro, sharing the same FHFA series. These within-metro gaps are comparable in size to the difference between entire regional groups (e.g., Rust Belt average burden ~33% vs. national mean ~22%).

MarketIncomeRoof/IncomeHPA5FHFA Series
Dallas, TX$70,51819.94%54.9%MSAD 19124
Frisco, TX$150,2129.56%54.9%MSAD 19124 (same)
Gap$79,69410.39 pp
Houston, TX$64,81321.70%43.6%MSA 26420
The Woodlands, TX$140,70110.10%43.6%MSA 26420 (same)
Gap$75,88811.60 pp

Texas Suburb Cluster (n = 6) — Supporting Reference

Six markets within the Dallas and Houston metro complexes, ranked by income burden. Cary, NC is a comparable high-income suburb shown separately below for context — it is not part of this Texas grouping or its group mean.

MarketIncomeRoof/IncomeZillow Typical HV
Frisco, TX$150,2129.56%$656,145
The Woodlands, TX$140,70110.10%$308,933 (metro fallback)
Sugar Land, TX$136,21710.43%$452,251
Plano, TX$112,25312.79%$504,391
Round Rock, TX$99,28714.17%$407,038
McKinney, TX$72,53719.79%$482,599
Group mean (n = 6)$118,53512.81%
Cary, NC (shown for comparison, not part of group)$134,90510.21%$631,158

This grouping was assembled from the observed results rather than a pre-specified sampling method; it is supporting descriptive material, not a headline finding.


Market Rankings — Roof Cost-to-Income Burden

Full sample: 147 markets. National mean = 21.7%. Spread = 9.56%–37.76%.

20 Lowest-Burden Markets

RankMarketStateRoof/Inc %IncomeRoof CostHPA5 %
1FriscoTX9.56%$150,212$14,35554.9%
2The WoodlandsTX10.10%$140,701$14,21043.6%
3CaryNC10.21%$134,905$13,77564.4%
4Sugar LandTX10.43%$136,217$14,21043.6%
5BellevueWA11.03%$165,576$18,27041.9%
6GilbertAZ11.71%$122,551$14,35567.7%
7Overland ParkKS11.96%$114,009$13,63056.0%
8RockvilleMD12.74%$128,649$16,38540.2%
9PlanoTX12.79%$112,253$14,35554.9%
10San JoseCA12.97%$146,427$18,99534.6%
11ScottsdaleAZ13.08%$110,886$14,50067.7%
12ChandlerAZ13.28%$108,095$14,35567.7%
13San FranciscoCA13.89%$140,970$19,57514.3%
14Round RockTX14.17%$99,287$14,06544.8%
15Virginia Beach *VA14.48%$103,125$14,93553.0%
16SeattleWA14.52%$123,860$17,98041.9%
17CharlestonSC14.91%$92,414$13,77576.6%
18AustinTX15.17%$93,658$14,21044.8%
19IrvineCA15.47%$120,919$18,70553.7%
20ChesapeakeVA15.66%$95,373$14,93553.0%

20 Highest-Burden Markets

RankMarketStateRoof/Inc %IncomeRoof CostHPA5 %
147DetroitMI37.76%$39,938$15,08050.8%
146TrentonNJ36.43%$52,537$19,14068.3%
145ClevelandOH36.25%$40,801$14,79056.2%
144HartfordCT36.24%$46,411$16,82061.7%
143NewarkNJ36.21%$52,060$18,85058.0%
142RochesterNY35.63%$47,213$16,82069.9%
141SpringfieldMA34.97%$52,656$18,41558.0%
140EriePA34.59%$46,113$15,95057.5%
139BuffaloNY32.74%$50,041$16,38561.0%
138DaytonOH32.37%$45,247$14,64557.2%
137West Palm Beach *FL31.88%$53,668$17,11077.8%
136AkronOH30.46%$48,076$14,64555.2%
135JacksonMS29.64%$42,071$12,47037.7%
134AllentownPA29.00%$55,494$16,09563.1%
133BirminghamAL28.65%$46,051$13,19549.9%
132MiamiFL28.09%$62,462$17,54588.5%
131CincinnatiOH27.95%$52,909$14,79060.5%
130New York CityNY27.93%$80,483$22,47543.3%
129AuroraIL27.86%$64,528$17,98052.6%
128AlbanyNY27.60%$61,986$17,11054.1%

* Income estimate carries margin of error >10% (see Limitations). Ranking position for this market should be interpreted with additional caution.


Limitations

  • Cross-sectional design. Associations are observational; no causal claims are warranted. Time-series or panel analysis would be required to establish causation.
  • Non-probability sample. 147 markets are drawn from the NumeralQ active footprint, not a random or stratified sample of U.S. housing markets.
  • Model-based cost estimates. NumeralQ V2 costs are model outputs, not observed transaction prices. Actual costs vary by scope, materials, contractor, and permit requirements.
  • Household income is not disposable income. It excludes taxes, debt service, and other obligations.
  • ZHVI is not liquidity. The Zillow typical home value is an asset-value estimate, not accessible equity or cash.
  • HPA5 window effect. The 2019–2024 window includes pandemic-era price dynamics; a different window could produce different results.
  • ACS margins of error. Four markets carry income MOE >10%: Virginia Beach, VA (24.9%), West Palm Beach, FL (18.2%), Aurora, IL (13.2%), Naperville, IL (11.8%).
  • The Woodlands, TX ZHVI. CDP absent from Zillow’s city-level file; Houston metro series substituted ($308,933).
  • HVAC results are co-linear with roof results (same multiplier structure); shown for completeness, not as an independent signal.

This brief is a research communication prepared by NumeralQ Research. It is not investment advice, a commercial appraisal, or a regulatory filing. Analysis reflects data available as of August 2026.