Immigration cushions but rarely fully offsets native-born decline in shrinking US counties

Summary

Over the last decade, native-born population fell in 1,656 of 3,209 U.S. counties (52%) in this analysis. In most of those counties, rising immigration cushioned but did not reverse the decline: foreign-born population grew in 935 of the 1,656 (56%), but the county's total population still shrank in 839 of those. Only 96 counties (5.8%) saw immigration growth large enough to fully offset the native-born loss and keep total population flat or growing. Nationally, across the 1,656 native-declining counties, native-born population fell by about 2.53 million while foreign-born population rose by about 0.70 million in the same counties — immigration made up roughly 28% of the native-born loss, not the whole of it. This matches the pattern found in Census Bureau and outside research: immigration is a real, measurable stabilizer for shrinking places, but as international migration has fallen sharply since 2024, fewer counties are being saved from outright population loss than in earlier years of the decade.

Step 1 — What the published literature already shows

Several independent analyses of the same underlying Census Bureau county estimates converge on the same qualitative finding, at different points in the decade:

• Economic Innovation Group (2022), analyzing 2016-2021 county estimates: dozens of counties, from Miami-Dade to small Nebraska agricultural counties, would have grown had immigration kept pace with its 2016 level; nationally the U.S. would have had roughly 2 million more residents had 2016-level immigration continued through 2021.
• American Immigration Council (2022), on the 2020-2021 vintage: 71% of counties had net positive international migration, and immigration fully prevented what would otherwise have been a population loss in 40 counties, including New Haven CT, Salt Lake City UT, and Durham-Chapel Hill NC — but international migration that year (245,000 nationally) was less than half its 2011-2018 average, so it fell short of offsetting domestic out-migration and natural decrease in many large urban counties (NYC, LA, Miami).
• Population Reference Bureau (2026), analyzing the newest 2024-2025 vintage: net international migration prevented population loss in only 254 counties (down 40% from the prior year) and helped offset — but not fully reverse — decline in another 882 counties; 377 counties lost population despite immigration gains, the most since 2021. This directly reflects the Census Bureau's reported 54% year-over-year drop in net international migration (2.7 million to 1.3 million) between the 2024 and 2025 estimates.
• The Census Bureau itself (press release, Jan. 27, 2026) attributes the national growth slowdown almost entirely to the international-migration decline, not to a comparable change in births/deaths, and separately projects (via CBO) that without immigration the total U.S. population would begin shrinking by 2033.

All of these sources rely on the Bureau's county-level "Components of Change" series (births, deaths, net international migration, net domestic migration), a genuinely finer breakdown than what this connector's tables carry. That table is published only as a large multi-megabyte CSV (co-est*-alldata.csv) and as a key-gated Census API dataset (pep/components); this session's fetch tool could not parse the raw CSV as text (returned as an unrecognized binary content type despite a text/csv header) and the API call requires a Census API key not available here — so the components-of-change breakdown itself is a genuine tool/access gap, disclosed rather than worked around. The literature above is the closest substitute: it analyzes that exact Census series and its headline county counts are quoted directly from it.

Step 2 — An independent county-level check with connector data

To add an original, decade-spanning computation (rather than only relaying the county counts above, which cover shorter 1-5 year windows each), this analysis used census.acs_nativity — ACS 5-Year survey counts of native-born and foreign-born population by county — comparing the 2009-2013 vintage against the 2019-2023 vintage, a roughly decade-apart pair of overlapping 5-year windows (centered near 2011 and 2021 respectively). Native-born population was computed as total population minus foreign-born population per county (the table's own native_born column is null in the 2013 vintage due to a Census negative-value suppression sentinel propagating through its three summed components, so the subtraction avoids that gap).

Of 3,209 counties with data in both vintages:

• 1,656 counties (51.6%) had a native-born population decline over the decade.
• Within those 1,656: in 96 counties foreign-born population rose enough that total population still grew or held flat — immigration fully did the job.
• In 839 counties foreign-born population rose, but not enough — total population fell anyway. Immigration helped, but did not "keep the place going."
• In the remaining 721 counties foreign-born population was flat or also declining — no immigration offset was present at all.
• In aggregate across all 1,656 native-declining counties, native-born population fell by about 2.53 million while foreign-born population rose by about 0.70 million in those same counties — immigration replaced about 27.6% of the native-born loss, cushioning the decline without reversing it for the group as a whole.

This is consistent with, not contradictory to, the published literature: it confirms that immigration offsets are the exception (full offset in under 6% of native-declining counties) rather than the rule, and that even where immigration is rising, it is more often a partial cushion than a full replacement.

Step 3 — Caveats and what would sharpen this further

Metric limits: ACS 5-year vintages are rolling averages over overlapping periods, not point-in-time counts, so this is a proxy for "the last ten years," not an exact calendar-decade comparison; margins of error are non-trivial for small counties, so a handful of the 96/839/721 splits near zero net change are within noise. The Census Bureau's own annual components-of-change file (net international migration, net domestic migration, and natural change reported separately, year by year) is the more precise and standard instrument for this question and is the one the cited literature uses directly — this analysis substitutes a coarser but genuinely decade-spanning nativity comparison because that annual file could not be parsed by the tools available in this session (see Step 1). A fully-resourced follow-up should fetch co-est2025-alldata.csv directly and decompose growth into births, deaths, net international migration, and net domestic migration per county.

Direction of the finding does not depend on this substitution: every literature source and the independent nativity computation here agree that immigration measurably slows, but infrequently fully reverses, population decline in native-born-shrinking counties, and that this stabilizing effect has weakened noticeably since 2024 as net international migration itself has fallen.