US Health Insurance Statistics 2026
By Axis Intelligence Research
Co-author: Jennifer Miller | Last updated: October 5, 2026 | License: CC BY 4.0
26.7 million Americans (7.9%) were uninsured for all of 2025, per the Census Bureau. But federal enrollment records already show 7.45 million fewer people in Medicaid, CHIP and ACA Marketplace plans than a year earlier — a decline Axis Intelligence Research measures at 27.9% of the entire uninsured population, through its Survey Lag Gap (SLG™).
Quick Answer
The official U.S. uninsured rate is 7.9% (26.7 million people) for 2025, per the Census Bureau’s P60-291 report; the CDC’s survey puts it at 8.3% (28.0 million). Both describe 2025. In 2026, Medicaid and CHIP enrollment fell 4.55 million year over year (CMS, May 2026) and Marketplace effectuated enrollment fell from 22.1 million to 19.2 million. Axis Intelligence Research places the 2026 uninsured exposure band at 9.2% to 10.1% of the population.
Key Findings
- According to Axis Intelligence Research, administrative records show a 7.45 million-person decline in Medicaid, CHIP and Marketplace enrollment in 2026, equal to 27.9% of the 26.7 million people the Census Bureau counted as uninsured in 2025 (Survey Lag Gap, SLG™).
- Axis Intelligence Research estimates the U.S. uninsured exposure band for 2026 at 9.2% to 10.1%, against the official 2025 rate of 7.9% — the first band computed from 2026 enrollment records rather than survey answers.
- Axis Intelligence Research finds that 78.9% of Medicaid and CHIP disenrollments in May 2026 were procedural, not eligibility-based, a figure it publishes as the Paperwork Attrition Ratio (PAR™), from CMS renewal data.
- Medicaid enrollment fell by 4.4 million (6%) between May 2025 and May 2026, to 66.4 million, according to CMS’s May 2026 enrollment snapshot.
- Working-age adults in the 10 states that had not expanded Medicaid were uninsured at 18.1% in 2025, twice the 9.0% rate in expansion states, per CDC NHIS data.
What Is the U.S. Uninsured Rate in 2026?
No federal survey has measured 2026 yet. The latest official numbers describe 2025, and they disagree by design.
| Measure | 2025 value | What it counts | Source |
|---|---|---|---|
| Uninsured all year (CPS ASEC) | 7.9% / 26.7M | No coverage at any point in 2025 | Census Bureau, Sept. 15, 2026 |
| Uninsured at interview (NHIS) | 8.3% / 28.0M | No coverage on the day surveyed | CDC NCHS, May 28, 2026 |
| Uninsured at interview, 2021 | 9.2% / 30.0M | Same NHIS concept | CDC NCHS |
| Gap between the two 2025 surveys | 1.3M | Axis calculation: 28.0M − 26.7M | Axis Intelligence Research |
The Census Bureau says the 2025 rate was not statistically different from 2024. The CDC reached the same verdict for its own series. A full-year measure counts someone as insured if they held coverage for a single month, which is exactly why it reacts last to a wave of terminations. The Census Bureau’s own America Counts analysis makes the point: people who lost Medicaid in 2024 still appeared as Medicaid-covered that year, and the loss only surfaced in the 2025 estimates.
Uninsured Rate by Age
| Group | Uninsured rate, 2025 | Change | Source |
|---|---|---|---|
| Adults 18–64 | 11.6% | Unchanged from 2024; 13.5% in 2021 | CDC NHIS |
| Children | 5.6% (4.0M) | From 5.1% (3.7M) in 2024, not significant | CDC NHIS |
| All ages | 8.3% (28.0M) | From 9.2% (30.0M) in 2021 | CDC NHIS |
Uninsured Rate by Race and Ethnicity (Adults 18–64)
| Group | 2025 | Source |
|---|---|---|
| Hispanic | 21.9% | CDC NHIS |
| Black, non-Hispanic | 11.3% (14.1% in 2021) | CDC NHIS |
| White, non-Hispanic | 8.5% | CDC NHIS |
| Asian, non-Hispanic | 4.9% | CDC NHIS |
Commentary, Jennifer Miller: The flat 7.9% will be quoted all autumn as proof that coverage held. Read the footnote first. A full-year survey answers “were you ever covered in 2025?” — a question designed to be slow. It is a reliable baseline and a poor early-warning signal, and the people losing coverage this year are, almost by definition, the ones it cannot see yet.
How Do Americans Get Health Insurance? Coverage by Type
Employment-based insurance covered 53.5% of the population for some or all of 2025, per the Census Bureau, followed by Medicare (20.1%), Medicaid (17.1%), direct-purchase coverage (10.5%), TRICARE (2.8%) and VA/CHAMPVA (1.2%). Shares overlap because people can hold more than one plan in a year.
| Coverage type | Share of population, 2025 | Change vs 2024 | Source |
|---|---|---|---|
| Employment-based | 53.5% | — | Census Bureau |
| Medicare | 20.1% | +0.6 pp | Census Bureau |
| Medicaid | 17.1% | −0.5 pp | Census Bureau |
| Direct purchase | 10.5% | — | Census Bureau |
| TRICARE | 2.8% | — | Census Bureau |
| VA and CHAMPVA | 1.2% | — | Census Bureau |
| Any public coverage | 35.8% | Not significant | Census Bureau |
How Many People Have Each Type of Health Insurance?
| Coverage type | 2025 (people) | 2024 (people) | Source |
|---|---|---|---|
| Total population | 337.9M | 336.8M | Census Table 1 |
| Any health plan | 311.2M | 309.9M | Census Table 1 |
| Any private plan | 222.0M | 222.1M | Census Table 1 |
| Employment-based | 180.7M | 180.5M | Census Table 1 |
| Direct purchase | 35.53M | 36.26M | Census Table 1 |
| Marketplace coverage | 14.26M | 14.48M | Census Table 1 |
| TRICARE | 9.29M | 9.59M | Census Table 1 |
| Any public plan | 121.1M | 120.8M | Census Table 1 |
| Medicare | 67.91M | 65.74M | Census Table 1 |
| Medicaid | 57.80M | 59.22M | Census Table 1 |
| VA and CHAMPVA | 4.21M | 3.87M | Census Table 1 |
| Uninsured all year | 26.67M | 26.94M | Census Table 1 |
Only two year-over-year changes were statistically significant, per the Census Bureau: Medicare rose by about 2.2 million and Medicaid fell by about 1.4 million. People hold more than one plan, so types do not sum to the total.
Public Coverage for Working-Age Adults
Public coverage for adults ages 19 to 64 fell 0.5 percentage points to 17.3% in 2025, driven by a 0.6-point drop in Medicaid, per the Census Bureau. Among adults 65 and older, private coverage fell 2.8 points to 41.5%.
Marketplace Share of Under-65 Coverage
Exchange-based private plans reached 6.3% of Americans under 65 in 2025, up from 4.3% in 2021, per CDC NHIS. That growth is the cohort now under the most pressure. What those plans cost each household, channel by channel, is covered in our health insurance cost breakdown for employer and Marketplace plans.
How Many People Lost Medicaid in 2026?
Medicaid enrollment fell by 4.4 million (6%) between May 2025 and May 2026, and CHIP by 150,000 (2%), per the CMS May 2026 Medicaid and CHIP enrollment snapshot. Total enrollment stood at 73.5 million: 66.4 million in Medicaid and 7.1 million in CHIP.
| Segment | May 2026 enrollment | Change vs May 2025 | Source |
|---|---|---|---|
| Medicaid + CHIP total | 73.5M | — | CMS |
| Medicaid adults | 38.3M | −2.6M | CMS |
| Medicaid children + CHIP | 35.2M | −2.0M | CMS |
| Medicaid | 66.4M | −4.4M (−6%) | CMS |
| CHIP | 7.1M | −150,000 (−2%) | CMS |
| Basic Health Program / related | 1.6M | — | CMS |
Applications rose 15% over the same period, to 2.7 million in May 2026. People are still knocking; fewer are staying in. One reporting caveat matters: California removed limited-benefit enrollees from its counts in March 2026 without revising earlier months, per CMS, so part of the year-over-year decline reflects a definitional change in the largest state.
Medicaid Renewal Outcomes, May 2026
Of the 5,657,983 people due for renewal in May 2026, 68% were renewed (48% ex parte, without paperwork), 19% were disenrolled, 15% were disenrolled for procedural reasons, and 13% remained pending, per CMS.
The Paperwork Attrition Ratio (PAR™): How Much Medicaid Loss Is Procedural?
The Paperwork Attrition Ratio (PAR™) is an Axis Intelligence Research metric: the share of Medicaid and CHIP disenrollments at renewal that happened for procedural reasons — a missing form, an unanswered notice — rather than a determination that the person was ineligible.
PAR = procedural disenrollment share ÷ total disenrollment share × 100
Nationally, PAR = 15 ÷ 19 × 100 = 78.9 for the May 2026 renewal cohort. Applied to the cohort size, Axis Intelligence Research estimates about 1,075,000 disenrollments, of which roughly 848,700 were procedural, in one month.
Paperwork Attrition Ratio by State (May 2026)
| State | Disenrolled at renewal | Procedural | PAR™ | Renewals due |
|---|---|---|---|---|
| District of Columbia | 11% | 11% | 100.0 | 13,652 |
| Indiana | 47% | 46% | 97.9 | 98,256 |
| California | 20% | 19% | 95.0 | 995,076 |
| Ohio | 20% | 19% | 95.0 | 233,842 |
| Hawaii | 10% | 9% | 90.0 | 38,958 |
| New Mexico | 27% | 24% | 88.9 | 32,658 |
| Maryland | 26% | 23% | 88.5 | 87,728 |
| Louisiana | 25% | 22% | 88.0 | 111,446 |
| Texas | 32% | 23% | 71.9 | 264,602 |
| New York | 25% | 16% | 64.0 | 376,823 |
| Florida | 19% | 11% | 57.9 | 172,329 |
| Pennsylvania | 15% | 7% | 46.7 | 199,158 |
| Kansas | 15% | 6% | 40.0 | 38,980 |
| South Carolina | 29% | 5% | 17.2 | 65,756 |
| United States | 19% | 15% | 78.9 | 5,657,983 |
Source: Axis Intelligence Research PAR™ calculation from CMS Eligibility Processing Data, May 2026. All 50 states and DC are in the CSV. CMS rounds state percentages to whole points, so states with small disenrollment shares carry wider rounding error.
PAR and the disenrollment rate must be read together. Indiana combines a 47% disenrollment rate with a near-total procedural share — the worst pairing in the country. South Carolina disenrolled 29% but mostly on eligibility grounds. The District of Columbia’s 100.0 sits on an 11% disenrollment rate, so the absolute exposure is small.
Commentary, Jennifer Miller: A procedural termination is not evidence that someone was ineligible; it is evidence that the renewal didn’t complete. Some of those people return within months — “churn” in program language — and some do not. The useful question for a state is not whether PAR is high, but whether it is high alongside a high disenrollment rate. That combination is where coverage loss looks administrative rather than intended.
How Many People Are Enrolled in ACA Marketplace Plans in 2026?
19.2 million people had effectuated Marketplace coverage in February 2026, per HHS ASPE’s ACA Exchange Enrollment in 2026 brief, down from 22.1 million a year earlier, a 13% drop per KFF. Effectuated means the person paid a premium, which is why this number fell more than sign-ups did.
| Metric | Value | Source |
|---|---|---|
| Effectuated enrollment, Feb 2026 | 19.2M | HHS ASPE |
| Effectuated enrollment, Feb 2025 | 22.1M | KFF, citing HHS |
| Year-over-year change | −13% | KFF |
| Avg. premium payment increase to keep the same plan | 114% | KFF |
| Avg. net premium payment increase after plan switching | 58% | KFF |
| Projected 2026 average enrollment | 17.5M | KFF |
Why Marketplace Enrollment Fell: Two Explanations
The administration and independent researchers read the same 2.9 million drop differently, and we publish both rather than pick one. HHS ASPE attributes it to program-integrity work, estimating 2.9 million improper or phantom enrollments were removed and 2.6 million remain, after a peak it estimates at 5.6 million in 2025. KFF attributes it to the expiration of the enhanced premium tax credits, which raised premium payments 114% on average for enrollees keeping the same plan. Both estimates are model-based; neither has been independently audited. KFF’s own state analysis uses a slightly different 2025 base (21.8 million), and both bases are logged in our methodology.
The Survey Lag Gap (SLG™): How Much Coverage Loss Hasn’t Reached the Official Rate Yet?
The Survey Lag Gap (SLG™) is an Axis Intelligence Research metric measuring how much coverage loss federal administrative records already show that the official survey-based uninsured count has not yet captured.
SLG = (Medicaid decline + CHIP decline + Marketplace effectuated decline) ÷ latest Census full-year uninsured count × 100
| Component | Value | As of | Source |
|---|---|---|---|
| Medicaid year-over-year decline | 4.4M | May 2026 | CMS |
| CHIP year-over-year decline | 0.15M | May 2026 | CMS |
| Marketplace effectuated decline | 2.9M (22.1M → 19.2M) | Feb 2026 | HHS ASPE / KFF |
| Administrative decline (sum) | 7.45M | — | Axis calculation |
| Census uninsured, full year 2025 | 26.7M | 2025 | Census Bureau |
| SLG™ | 27.9 | Oct 2026 baseline reading | Axis Intelligence Research |
This is the baseline reading. An SLG of 27.9 means administrative coverage declines equal to 27.9% of the existing uninsured population are already on the record.
The 2026 Uninsured Exposure Band
Not every lost enrollment becomes an uninsured person. Some people move to a job-based plan; HHS argues some Marketplace enrollments were never real. So Axis Intelligence Research publishes a band, not a point, using the Census Bureau’s total population of 337.9 million (Table 1, P60-291):
- Lower bound, 9.2%: (26.7M + 4.55M Medicaid/CHIP decline) ÷ 337.9M — treats the entire Marketplace drop as non-coverage loss, the HHS reading.
- Upper bound, 10.1%: (26.7M + 7.45M) ÷ 337.9M — treats every lost enrollment as a newly uninsured person.
Both bounds sit above the official 7.9%. We decline to publish a single 2026 point estimate: the share of Medicaid leavers who found other coverage is not measured in any current federal release, and inventing it would be exactly the kind of precision this dataset exists to avoid.
Commentary, Jennifer Miller: Treat the band the way a clinician treats a reference range. It says where the true value can plausibly sit, not where it does. The lower bound already assumes the most favorable reading of the Marketplace data, and it still lands 1.3 points above the headline rate.
Does Medicaid Expansion Change the Uninsured Rate?
Working-age adults in non-expansion states were uninsured at 18.1% in 2025, against 9.0% in expansion states, per CDC NHIS — a ratio of 2.0 to 1, per Axis Intelligence Research. As of January 1, 2025, ten states had not expanded Medicaid: Alabama, Florida, Georgia, Kansas, Mississippi, South Carolina, Tennessee, Texas, Wisconsin and Wyoming, according to the Census Bureau’s America Counts analysis. Seven are in the South, where Medicaid coverage fell 0.6 points in 2025 to 14.9%, the lowest of any region.
How Many More Americans Will Be Uninsured by 2034?
The Congressional Budget Office estimates the Medicaid provisions of the 2025 reconciliation law (P.L. 119-21) will increase the number of people without health insurance by 7.5 million in 2034, per its October 2025 cost estimate. That projection is separate from the Marketplace subsidy changes discussed above, and it phases in: the work-reporting requirements it includes have not yet taken full effect in 2026 enrollment data.
Health Insurance Coverage Questions, Answered
Why does the Census Bureau report 7.9% uninsured while the CDC reports 8.3%?
They measure different things. Census (CPS ASEC) counts people with no coverage at any point in the year; the CDC’s NHIS counts people uninsured on the day they were interviewed. The NHIS figure is higher because someone covered for part of the year counts as insured in the Census measure.
Will the 2026 uninsured rate be higher than 2025?
Federal surveys for 2026 have not been released. Enrollment records show Medicaid, CHIP and Marketplace enrollment down a combined 7.45 million, which Axis Intelligence Research translates into a 9.2% to 10.1% exposure band, above 2025’s 7.9%.
What does “procedural disenrollment” mean in Medicaid?
It means coverage ended because the renewal process was not completed — a form not returned or a notice not answered — rather than because the state found the person ineligible. In May 2026, 15% of people due for renewal were disenrolled procedurally, per CMS.
Which state has the highest share of procedural Medicaid disenrollments?
Among states with substantial disenrollment, Indiana stands out: 47% of people due for renewal in May 2026 were disenrolled, and 46% procedurally, a Paperwork Attrition Ratio of 97.9 per Axis Intelligence Research. The District of Columbia scores 100.0 on a much smaller 11% disenrollment rate.
Is ACA Marketplace enrollment really down, or were those enrollments fraudulent?
Effectuated enrollment fell from 22.1 million to 19.2 million. HHS ASPE attributes 2.9 million of removals to improper or phantom enrollment; KFF links the drop to a 114% average increase in premium payments after enhanced tax credits expired. Both are model-based estimates.
How many people are on Medicaid right now?
66.4 million people were enrolled in Medicaid and 7.1 million in CHIP in May 2026, for a total of 73.5 million, per CMS. Medicaid alone fell 4.4 million from a year earlier.
Are children losing health coverage?
Medicaid child and CHIP enrollment fell by about 2.0 million between May 2025 and May 2026, per CMS. The CDC’s 2025 child uninsured rate was 5.6%, up from 5.1% in 2024, a change it did not find statistically significant.
When will official data show 2026 coverage losses?
The CDC’s NHIS early release for the first half of 2026 and the Census Bureau’s September 2027 report on 2026 coverage. Monthly CMS enrollment snapshots will arrive sooner.
Methodology
Scope. U.S. health insurance coverage status and enrollment, 2021–2026, from six primary sources read on October 5, 2026: Census Bureau P60-291 and its America Counts story (CPS ASEC 2026, coverage in calendar 2025); CDC NCHS NHIS Early Release 2025; CMS Medicaid and CHIP Eligibility Operations and Enrollment Snapshot for May 2026; HHS ASPE Issue Brief on 2026 exchange enrollment; CBO cost estimate for P.L. 119-21 Medicaid provisions. KFF is used for the 2025 Marketplace base, premium-payment changes and enrollment projection.
Formulas. SLG™ = (4.4M + 0.15M + 2.9M) ÷ 26.7M × 100 = 27.9. Population = 337.9M (Census Table 1). Exposure band: lower = (26.7M + 4.55M) ÷ 337.9M = 9.2%; upper = (26.7M + 7.45M) ÷ 337.9M = 10.1%. PAR™ = procedural disenrollment % ÷ total disenrollment % × 100, per state and nationally (15 ÷ 19 = 78.9). Estimated May disenrollments = 5,657,983 × 19% ≈ 1,075,000; procedural ≈ 848,700. All arithmetic was run twice, in floating point and exact rational arithmetic.
Scope of the inputs. The SLG mixes a May 2026 Medicaid comparison with a February 2026 Marketplace comparison, the latest available for each. California’s March 2026 reporting change inflates its share of the Medicaid decline. CMS state renewal percentages are rounded to whole points. The exposure band does not net out people who moved from Medicaid to Marketplace or to employer coverage.
What we declined to compute. A single-point 2026 uninsured rate; it would carry precision the sources do not support.
About This Dataset
health-insurance-statistics.csv contains 306 rows: national coverage counts (Census Table 1, 2024–2025), coverage shares and uninsured rates (Census, CDC), Medicaid/CHIP enrollment and renewal outcomes (CMS, May 2026, all 50 states and DC), Marketplace effectuated enrollment (HHS ASPE, KFF), CBO projections, and every Axis calculation with its formula in method_note. Columns: metric, value, unit, as_of_date, geography, segment, source_org, source_document, source_url, retrieved_date, is_primary, axis_calculated, method_note, license. License: CC BY 4.0. Citation line: Axis Intelligence Research, Health Insurance Statistics 2026, 2026.
Cite This Page
- APA: Axis Intelligence Research & Miller, J. (2026). Health insurance statistics 2026: Who is covered, who lost coverage, and what the 7.9% uninsured rate misses. Axis Intelligence. https://axis-intelligence.com/health-insurance-statistics/
- MLA: Axis Intelligence Research and Jennifer Miller. “Health Insurance Statistics 2026.” Axis Intelligence, 5 Oct. 2026, axis-intelligence.com/health-insurance-statistics/.
- Chicago: Axis Intelligence Research and Jennifer Miller. “Health Insurance Statistics 2026.” Axis Intelligence, October 5, 2026. https://axis-intelligence.com/health-insurance-statistics/.
Related: telehealth statistics and the Medicare policy picture · digital health statistics · healthcare data breaches exposing insurance records · what retirees will spend on health care
