Demographic & Economic Intelligence

Between April 2020 and mid-2025 the United States added 10.2M residents, and 48.9% of that net growth landed in just five states. This report maps where the people, the jobs, the paychecks and the homes actually went: all 51 states ranked on 16 indicators, the 20 largest metros compared, 2,462 PUMS areas screened, and a two-page profile of every state.

341.8M

U.S. population, 2025

+10,206,753 since 2020

1.43M

Housing units permitted, 2025

annual, all structure types

4.4%

Unemployment, 2026-06

4.3% across 2025

13.3%

Workers fully remote, 2024

down from 17.9% in 2021

Eight findings organize what follows. First, growth stayed concentrated: the five fastest-growing states account for 48.9% of net national population growth since 2020. Second, migration, not births, is now the growth engine almost everywhere; more than half the states buried more residents than they birthed in at least one recent year. Third, remote work has settled rather than snapped back: 13.3% of workers were fully remote in 2024, half again the pre-pandemic norm, and in California the share fell from 21.3% to 14.1% without ever approaching it.

Fourth, homebuilding did not follow the people evenly: the states attracting the most new residents split sharply into those permitting homes for them and those falling behind, which is the subject of the Growth–Supply Mismatch section. Fifth, the largest interstate commuting corridor in America is Prince George's County, MD → District of Columbia, DC at 108,471 workers, and the top of that list is dominated by two capital regions. Sixth, K-12 chronic absenteeism remains far above its pre-2020 levels in most reporting states even as graduation rates hold. Seventh, the wage map and the cost map disagree: several high-wage states rank near the bottom once rents are held against incomes. Eighth, the gap inside states now rivals the gap between them; in the widest case, the best-paid PUMS area out-earns the poorest by nearly seven to one.

The national series below are the backdrop for every state ranking in this report. Population is the Census Bureau's Vintage-2025 estimate; permits are annual Building Permits Survey totals summed across the states; unemployment is the annual average of state labor forces; output is nominal state GDP summed (real GDP is compared only within states, where its chained-dollar arithmetic is valid).

+3.1%

Population growth, 2020–2025

10.6M

Units permitted, 2019–2025

8.1%

Unemployment at the 2020 peak

3.6% at the 2023 trough

+42.9%

Nominal GDP growth, 2019–2025

not inflation-adjusted

Every figure in this report is computed from the same published files and PUMS database that power this site's free explorers, by a deterministic pipeline whose output is committed and diffed. No number is typed into the text by hand; the prose interpolates the same values the tables show. The sources, their vintages, and every known limitation:

1. Population, components, age structure: Census Population Estimates Program, Vintage 2025 (2020–2025). A single vintage throughout: Vintage 2025 restates every year back to 2020, and mixing vintages (or splicing to the 2010s series, which is benchmarked to the 2010 census) fabricates trends. The 2020 estimates year covers April–December and is excluded from component sums.
2. Microdata (remote work, PUMA profiles, occupation shares): American Community Survey PUMS, 2024 one-year, with 2021–2023 for the remote-work series; weighted with exact integer person weights. The remote-work universe is workers reporting a means of transportation to work; “remote” = worked from home. PUMA geographies are the 2020-vintage definitions. PUMS figures are survey estimates and carry sampling error; PUMA-level medians especially so.
3. Building permits: Census Building Permits Survey, annual totals 2019–2025, all structure types. Permits are authorizations, not completions; the gap matters in high-abandonment markets. Multifamily = units in structures of 5+.
4. Unemployment, labor force: BLS Local Area Unemployment Statistics: annual averages 2019–2025, monthly through 2026-06. The ACS also publishes an unemployment rate from a different survey with a different reference period; this report quotes LAUS everywhere except inside the ACS-internal income section, and never mixes the two in one comparison.
5. Wages: BLS Quarterly Census of Employment and Wages, 2019-Q4 → 2025-Q4. Figures are average weekly wages across all covered employment; annualize at ×52. Nominal throughout; growth reflects composition shifts as well as raises. “Best-paying major sector” requires the sector to employ ≥1% of the state's covered workforce.
6. Output: BEA state GDP, 2019–2025. Real growth uses each state's chained-2017$ series, which is valid within a state; chained dollars are never summed across states here, so the national dashboard line is nominal and labeled as such.
7. Commuting flows: LEHD LODES, 2023, job type JT01 (primary jobs). Two states (AK, MI) did not publish 2023; their balances use their latest published year, labeled in the table; a missing year is never treated as zero. Connecticut reports as planning regions, not its legacy counties, and is never county-joined to sources keyed the old way. LODES is jobs-based administrative data; it counts jobs, not commuters' survey responses.
8. Rents and affordability: HUD Fair Market Rents, FY2023→FY2026. The affordability measure is 12× the 2BR FMR divided by ACS median household income: a cost-to-income ratio for a standardized unit, not a statement about what any individual household pays.
9. Incomes, poverty, attainment, rent burden: ACS 2020–2024 five-year estimates, with published margins of error shown where quoted. Five-year estimates trade currency for precision; differences within the MOE are noise and the text avoids claiming them.
10. Sale prices: Redfin county market data, June 2026 vs June 2025. County grain only, never aggregated to states. Cells with too few sales are suppressed at source and shown as dashes.
11. K-12: NCES Common Core of Data and state files, latest year 2024. Stats publish on different lags per state; each figure carries its reporting year. The chronic-absence series uses only the 2022–2024 collection (2022, 2023, 2024); the pre-2022 collection used a different definition and is never spliced. Non-reporting states show dashes, not zeros.
12. The Opportunity Index: eight pillars (population growth ↑, domestic migration ↑, unemployment ↓, real GDP growth ↑, wage growth ↑, permits per 1K ↑, 2BR-rent share of income ↓, bachelor's attainment ↑), each z-scored across the 51 states, winsorized at ±2.5σ, equally weighted, rescaled to 50±10. Equal weights are an editorial choice; all pillar scores are published so the index can be re-weighted by the reader.
13. The Growth–Supply Mismatch: cumulative 2020–2025 permitted units divided by net 2020–2025 population change, growing states only. The ratio is undefined for shrinking states (a negative denominator would rank them as infinitely supplied), so they are classed separately rather than scored.

Note. Ranks always run 1 = “best” in the direction stated in each table (lowest unemployment, lowest poverty, most affordable, and so on); direction is part of the definition, not a value judgment. Where a state is missing from a source, its cell is a dash and it is excluded from that ranking; it is never assigned zero. This report was assembled in August 2026; the data vintages above, not the assembly date, are what each figure refers to.