Did the uninsured rate go up or down?
Each bar is the % of people under 65 with no health insurance in a term’s last year, minus the same share in the year before the term began.
Went up in 2 terms, went down in 3 terms, up in one part and down in the other in 1 term, measured in part in 2 terms, no bar in 4 terms.
Change in the uninsured rate in each term, in percentage points.
Bush (W.) 1: from 17.0% in 2000 to 16.6% in 2004 is a change of −0.4 points.Clinton 1: from 17.5% in 1994 to 16.1% in 1995 is a change of −1.4 points.Clinton 2: from 17.5% in 1997 to 17.0% in 2000 is a change of −0.5 points.Bush (W.) 1: from 17.0% in 2000 to 16.6% in 2004 is a change of −0.4 points.Bush (W.) 2: from 16.4% in 2004 to 16.8% in 2008 is a change of +0.4 points.Obama 1: from 16.8% in 2008 to 16.9% in 2012 is a change of +0.1 points.Obama 2: from 16.9% in 2012 to 10.3% in 2016 is a change of −6.6 points.Trump 1: 10.3% in 2016 to 11.0% in 2018 on the older survey, then 12.0% in 2019 to 11.5% in 2020 on the new one.Biden: from 11.5% in 2020 to 9.9% in 2024 is a change of −1.6 points. Measure: the National Health Interview Survey, run by the CDC’s National Center for Health Statistics (NCHS). It changed three times, and no bar spans two versions. Source: CDC/NCHS, National Health Interview Survey (Health, United States 2020-2021, Table HINone). Later years come from newer NCHS reports, named under How we counted.
Clinton 1: NCHS’s table skips years here. So this bar compares 1994 with 1995 only.
Clinton 2: NCHS redesigned the survey in 1997. In NCHS’s words, “Data before 1997 are not strictly comparable with data for later years due to the 1997 questionnaire redesign.” So this bar starts in 1997, on the new survey.
Bush (W.) 1: In 2004, the survey began to ask people with no reported coverage about Medicaid by name. In NCHS’s words, “Beginning in quarter 3 of the 2004 NHIS, people under age 65 with no reported coverage are asked explicitly about Medicaid coverage.” NCHS printed two figures for that year, 16.6% without the new question and 16.4% with it. This bar ends on the first. The next term starts from the second.
Trump 1: NCHS redesigned the survey again in 2019. In NCHS’s words, “In 2019, the NHIS questionnaire was redesigned and other changes were made to weighting and design methodology. Data for 2019 have not been fully evaluated for comparability with earlier years.” So this term has two bars, never joined: 2016 to 2018 on the older survey, then 2019 to 2020 on the new one. For 2020 (interviews by telephone only from 19 March 2020), NCHS adds: “Though NHIS survey weights account for changes in respondent characteristics due to overall changes in the 2020 data collection, differences observed in estimates between 2020 and earlier time periods may still be impacted by these changes.”
These numbers show what happened during each term. Congress, the courts, the states, the Federal Reserve and events abroad also move them.
How we counted
Each bar is the rate in a term’s last year minus the rate in the year before it began. NCHS gives one rate a year. A year counts for the president who took office in it, so a term’s own years are its first four calendar years. For Bush (W.) 1, that is 2001 to 2004, and its bar starts from 2000.
The survey’s parts
NCHS changed the survey in 1997, 2004 and 2019. We use each part on its own, so no bar compares two parts. Each line below gives our name for a part, then the NCHS report it comes from.
- People under 65 with no health insurance, NHIS, 1984-1995 (before the 1997 questionnaire redesign) CDC/NCHS, National Health Interview Survey (Health, United States 2020-2021, Table HINone); HINone, Percent of population, Total under 65 years; years 1984 and 1989, then 1994 and 1995
- People under 65 with no health insurance, NHIS, 1997-2004 (before the 2004 Medicaid question) CDC/NCHS, National Health Interview Survey (Health, United States 2020-2021, Table HINone); HINone, Percent of population, Total under 65 years; years 1997-2003 and the column 2004(1)
- People under 65 with no health insurance, NHIS, 2004-2018 (with the added Medicaid question) CDC/NCHS, National Health Interview Survey (Health, United States 2020-2021, Table HINone); HINone, Percent of population, Total under 65 years; the column 2004(2) and years 2005-2018
- People under 65 with no health insurance, NHIS, 2019-2024 (after the 2019 questionnaire redesign) CDC/NCHS, National Health Interview Survey (Long-term Trends in Health Insurance Coverage, 1968-2024, Table 1)
Terms with no bar
- Reagan 1: NCHS’s table has one year in this term, its fourth year, at 14.5%. It has no figure for the year before the term, so no change can be shown.
- Bush (H.W.): NCHS’s table has one year in this term, its first year, at 15.6%. It has no figure for the year before the term, so no change can be shown.
- Reagan 2: NCHS’s table has no year in this term.
- Trump 2: NCHS has no final figure for this term yet. Its early release is named under the next heading.
NCHS’s early release
For the term still running, NCHS has only an early release, which it labels preliminary. In NCHS’s words: “All estimates in this report are based on preliminary data from the 2025 National Health Interview Survey (NHIS).” NCHS published it in May 2026. Early figures can still change, so this page prints none of them, and the running term has no bar. Source: CDC/NCHS, National Health Interview Survey (Health Insurance Coverage: Early Release of Estimates, 2025, Table I).
NCHS adds about the same report:
- 2025: “A new sample design was implemented with the 2025 NHIS and aligns with the 2020 decennial census.”
- 2025 (October to December 2025 fieldwork): “Due to a partial federal government closure from October 1 to November 12, 2025, data collection for NHIS was paused.”
Other notes from NCHS
- 2000 to 2002 (Table HINone footnote 1): “Estimates for 2000–2002 were calculated using 2000-based sample weights and may differ from estimates in other reports that used 1990-based sample weights for 2000–2002 estimates.”
How others count it
The Census Bureau counts people with no health insurance in two other surveys. They ask other questions and count people of all ages, so their levels differ from NCHS’s. Compare the direction of each change.
The Current Population Survey (CPS) counts people with no health insurance for the whole year. Census changed it three times. Where a change falls inside a term, the table shows each side on its own. Source: US Census Bureau, Current Population Survey ASEC (health insurance historical tables).
The American Community Survey (ACS) has no figure for 2020, because Census didn’t publish its usual yearly figures that year. Source: US Census Bureau, American Community Survey 1-year estimates (health insurance historical tables, HIC-4_ACS).
| Term | NCHS (this chart) | Census CPS | Census ACS |
|---|---|---|---|
| Reagan 1 | one year only | no data | no data |
| Reagan 2 | no data | +0.5 from 1987 | no data |
| Bush (H.W.) | one year only | +1.6 | no data |
| Clinton 1 | −1.4 1994 to 1995 | +0.6 | no data |
| Clinton 2 | −0.5 from 1997survey changed | −0.1 to 1999−0.5 from 1999survey changed | no data |
| Bush (W.) 1 | −0.4survey changed | +1.2 | no data |
| Bush (W.) 2 | +0.4 | +0.6 | one year only |
| Obama 1 | +0.1 | +0.5 | +0.2 |
| Obama 2 | −6.6 | −4.5 from 2013survey changed | −6.2 |
| Trump 1 | +0.7 to 2018−0.5 from 2019survey changed | +0.7 from 2017survey changed | +0.6 to 2019 |
| Biden | −1.6 | −0.6 | −0.4 from 2021 |
| Trump 2 | not measured yet | −0.1 so far, to 2025 | not measured yet |
Each cell is the change over the term, in points, with years shown where a count covers only part of a term. In Clinton 1 and Bush (W.) 1, the counts go in different directions. The chart shows NCHS’s count only.
Recessions
The recession months follow the dates of the National Bureau of Economic Research (NBER), the group that dates US recessions. NBER counts a recession from the month after the economy’s high point (the peak) through the month of its low point (the trough). Each month counts for the term it falls in, so a recession that runs across two terms counts in both, each with its own months. Source: National Bureau of Economic Research (NBER), Business Cycle Dating Committee.
What this chart asks
The uninsured rate is the share of people with no health insurance, from a yearly NCHS survey. Each bar shows how far that share moved during one term, up or down.
1 of 7
How to read one bar
Bush (W.) 1: from 17.0% in 2000 to 16.6% in 2004 is a change of −0.4 points.Reagan 1: NCHS’s table has one year in this term and none for the year before it, so it has no bar. This step shows another term’s bar. Bush (W.) 1: from 17.0% in 2000 to 16.6% in 2004 is a change of −0.4 points.Reagan 2: NCHS’s table has no year in this term, so it has no bar. This step shows another term’s bar. Bush (W.) 1: from 17.0% in 2000 to 16.6% in 2004 is a change of −0.4 points.Bush (H.W.): NCHS’s table has one year in this term and none for the year before it, so it has no bar. This step shows another term’s bar. Bush (W.) 1: from 17.0% in 2000 to 16.6% in 2004 is a change of −0.4 points.Clinton 1: from 17.5% in 1994 to 16.1% in 1995 is a change of −1.4 points.Clinton 2: from 17.5% in 1997 to 17.0% in 2000 is a change of −0.5 points.Bush (W.) 1: from 17.0% in 2000 to 16.6% in 2004 is a change of −0.4 points.Bush (W.) 2: from 16.4% in 2004 to 16.8% in 2008 is a change of +0.4 points.Obama 1: from 16.8% in 2008 to 16.9% in 2012 is a change of +0.1 points.Obama 2: from 16.9% in 2012 to 10.3% in 2016 is a change of −6.6 points.Trump 1: 10.3% in 2016 to 11.0% in 2018 on the older survey, then 12.0% in 2019 to 11.5% in 2020 on the new one.Biden: from 11.5% in 2020 to 9.9% in 2024 is a change of −1.6 points.Trump 2: NCHS has no final figure for this term yet, so it has no bar. This step shows another term’s bar. Bush (W.) 1: from 17.0% in 2000 to 16.6% in 2004 is a change of −0.4 points. Mark another term and this example follows your mark.
2 of 7
The rows under each name
Under each name: the term’s own years, the rate in the year before the term began and in its last year, and the months of recession inside the term, by NBER’s dates (the standard US recession dates). The recession row gives dates only.
3 of 7
Terms with no bar
Reagan 1 and Bush (H.W.): one year in NCHS’s table and none before the term, so no change to show.
Reagan 2: no year of the term in NCHS’s table.
Trump 2: still running. NCHS has only an early release, which can still change. How we counted names it.
4 of 7
Terms with a note below
Clinton 1 and Clinton 2: the bar covers only part of the term, so it’s striped.
Bush (W.) 1: the survey changed inside the term. The note says which version each end is on.
Trump 1: the survey changed inside the term, so two bars, never joined.
5 of 7
Up or down, term by term
The line under the question counts the bars. Went up in 2 terms, went down in 3 terms, up in one part and down in the other in 1 term, measured in part in 2 terms, no bar in 4 terms.
6 of 7
What the chart can’t say
These numbers show what happened during each term. Congress, the courts, the states, the Federal Reserve and events abroad also move them. How we counted, under the chart, gives the rule in full, the survey’s parts, NCHS’s notes and how two Census surveys count it.
7 of 7
Each term’s numbers
The numbers behind the chart above, as a table for every term, in time order and measured the same way. The median is the middle value of the terms measured.
| Term | On the chart | Period counted | Start | End |
|---|---|---|---|---|
| Reagan 1 | one year only | one year only | ||
| Reagan 2 | no year in this term | no year in this term | ||
| Bush (H.W.) | one year only | one year only | ||
| Clinton 1 | −1.4 | 1994 to 1995 | 17.5% | 16.1% |
| Clinton 2 | −0.5 | 1997 to 2000 | 17.5% | 17.0% |
| Bush (W.) 1 | −0.4 | 2000 to 2004 | 17.0% | 16.6% |
| Bush (W.) 2 | +0.4 | 2004 to 2008 | 16.4% | 16.8% |
| Obama 1 | +0.1 | 2008 to 2012 | 16.8% | 16.9% |
| Obama 2 | −6.6 | 2012 to 2016 | 16.9% | 10.3% |
| Trump 1 | +0.7then−0.5 | 2016 to 2018then2019 to 2020 | 10.3%then12.0% | 11.0%then11.5% |
| Biden | −1.6 | 2020 to 2024 | 11.5% | 9.9% |
| Trump 2 so far | not measured yet | not measured yet | ||
| Median of the terms measured | −0.5 | |||
How we counted, in short
- Uninsured rate: the chart shows the change from the value before the term to its last value. Series: People under 65 with no health insurance, NHIS, 1984-1995 (before the 1997 questionnaire redesign); People under 65 with no health insurance, NHIS, 1997-2004 (before the 2004 Medicaid question); People under 65 with no health insurance, NHIS, 2004-2018 (with the added Medicaid question); People under 65 with no health insurance, NHIS, 2019-2024 (after the 2019 questionnaire redesign). Source: CDC/NCHS, National Health Interview Survey (Health, United States 2020-2021, Table HINone); CDC/NCHS, National Health Interview Survey (Long-term Trends in Health Insurance Coverage, 1968-2024, Table 1).
Download this topic’s data (CSV)Every row of every series these tables read, from the first period a term counts, as each agency publishes it.