GrowthMap USA scores 3,144 U.S. counties on future housing-demand pressure. Every score is a national percentile built from dated public records. Missing data is never treated as zero — a category is scored only on the metrics that exist for that county, and gaps lower its confidence score instead.
| Block | Weight | What it measures |
|---|---|---|
| Demographics | 18% | 2020–24 and 2014–24 population growth, net migration, density, prime-age (25–44) shift, household growth |
| Physical growth | 25% | Building permits per 1k residents, permit momentum 2022–24, housing-unit stock growth, multifamily share |
| Housing market | 22% | 12-month price and rent growth, for-sale supply per household, days-to-pending, permit shortfall vs. population growth |
| Future demand | 25% | Migration momentum vs. 2015–19 baseline, employment and establishment growth 2020–23, unemployment rate (BLS annual), 5-year price trend |
| Price gap | 10% | Price-to-income ratio and price level (inverse — cheaper counties score higher) |
The 3-year view leans hardest on permits, multifamily construction and market momentum; the 10-year view leans on demographics, migration and long-run price trends. The Future Housing Pressure model drops the price block entirely and re-normalizes weights. Growth vs. Price starts from the opportunity score and shifts it up for cheap counties, down for expensive ones.
Counties under 5,000 people have sparse, noisy data. Their composite scores are pulled toward the national median in proportion to population, so a 400-person county cannot top the rankings on one lucky metric.
We rebuilt scores using only data available at a historical cut-off — 14 signals including 2014–19 population growth, 2015–19 migration and migration acceleration, 2019 permits, 2015–19 employment growth, 2019 unemployment, Zillow price/rent momentum, affordability, inventory tightness, and days-to-pending as of 2019 — ranked every fully-covered county, then measured actual home-value change over the next five years. Model weights are tuned on a random half of counties and verified on the other, untouched half. The bars show realized appreciation by score decile — worst-ranked counties on the left, best-ranked on the right. “Accuracy” below means: of the counties the model ranked in its top tier, what share actually outperformed the national median over the following five years.
| Source | Used for | Vintage |
|---|---|---|
| Census Population Estimates (PEP) | Population, growth, net migration, age structure | 2020–2024 + 2014/2015–19 baselines |
| Census ACS 5-year, table-based SF | Income, households, housing units, median value fallback | 2023 (+ prior-year trend) |
| Census Building Permits Survey | Permitted units, permit momentum, multifamily share | 2022–2024 annual county files |
| Census County Business Patterns | Employment and establishment growth | 2020 → 2023 |
| Zillow Research | Typical value (ZHVI), rent (ZORI), for-sale inventory, days-to-pending | — |
| BLS Local Area Unemployment | County unemployment rate (annual average) | — |
| World Bank Open Data | Country population, net migration, urbanization, GDP, unemployment, inflation, FDI (World scope) | 2014–2024 |
| BIS residential property prices (via FRED) | Real house-price growth, 1 yr and 5 yr, for 52 countries (World scope) | quarterly, latest |
| UCDP/PRIO Armed Conflict Dataset | Countries at war intensity are excluded from world scoring, refreshed from the latest annual dataset | 2025 release |
| Australian Bureau of Statistics | State/territory population + net overseas/interstate migration (3101.0), mean dwelling price (6432.0) — Regions scope | quarterly, latest |
| Statistics Canada | Provincial population (17-10-0009) + New Housing Price Index (18-10-0205) — Regions scope | quarterly/monthly, latest |
Sometimes a region beats its whole country. The Regions tab scores 21 first-level sub-national units — 8 Australian states/territories and 13 Canadian provinces/territories — on population growth, net migration (Australia), and new-home price momentum, as percentiles within the region universe. Their percentiles are not comparable to county or country scores. More countries join as keyless official sources allow; the UK Land Registry HPI is next when its endpoint is reachable.
World mode scores 193 sovereign countries from World Bank indicators, BIS real house prices and FDI flows, using the same percentile architecture as US counties. Countries under comprehensive US sanctions or in active conflict are excluded from scoring and shown gray with the reason; conflict status is cross-checked against the latest UCDP/PRIO Armed Conflict Dataset (currently 2025, 14 war-intensity countries), not headlines. Countries with foreign-ownership restrictions (CAN, NZL, CHE and others) are scored but flagged Restricted in their detail card, with the specific rule named.