Children's Health Insurance Program, CHIP (United States, 1997 to 2025)

Congress created a federal-state health program for uninsured children in lower-middle-income families in 1997; the share of children without insurance fell from 15 percent to 9 percent by 2012, but funding was repeatedly short-term, two reauthorizations were vetoed in 2007, immigrant children faced a five-year bar until 2009, and researchers still disagree about how much it replaced private coverage.

The record

Creation and design

• The Balanced Budget Act of 1997 (P.L. 105-33) created the State Children's Health Insurance Program (SCHIP, later CHIP) as a new Title XXI of the Social Security Act, effective August 5, 1997. (Congressional Research Service, "State Children's Health Insurance Program (CHIP): A Brief Overview," R40444; CRS legislative history, R40229; both opened)
• It targets children under 19 who are uninsured and above the pre-1997 Medicaid limits. States may set income limits up to 200 percent of the federal poverty level, or 50 percentage points above their prior Medicaid level, and can go higher with approval. New Jersey's 350 percent was the highest reported in 2010. (CRS R40444)
• States choose a Medicaid expansion, a separate program, or both. The federal government pays an enhanced share: in fiscal year 2009 the CHIP federal share ranged from 65 percent to 83.09 percent, against 50 percent to 75.84 percent for Medicaid. (CRS R40444)
• Unlike Medicaid, CHIP federal money is a capped allotment, not an open-ended entitlement. The 1997 law appropriated about $40 billion for fiscal years 1998 to 2007, at just over $4.2 billion a year at first, under $3.2 billion a year for 2002 to 2004 (called the "CHIP dip"), and about $5.0 billion by 2007. Actual federal spending from 1998 to 2007 was about $34.9 billion. (CRS R40444)

Timeline

The 2007 vetoes and the 2009 law

• CRS reports two 2007 vetoes (H.R. 976 on October 3, H.R. 3963 on December 12). Both bills provided $61.4 billion in national allotments over five years, a $36.2 billion increase over a $25.2 billion baseline, paid for with a 61 cent federal cigarette tax increase. (CRS R40130, "The Children's Health Insurance Program Reauthorization Act of 2009"; opened)
• A Boston University Washington Center report says the override vote fell 13 votes short of two-thirds, and that Republican opponents argued the program covered children whose parents earned well above the poverty line. The report quotes President Bush only as willing to "sit down" about adding funding. (BU Washington Center, Oct. 18, 2007; opened. It did not state the President's stated objections.)
• The 2009 law set national allotments of $68.9 billion for FY2009 to FY2013, raised the federal cigarette tax from $0.39 to about $1.00 a pack, and gave states an option to cover lawfully residing immigrant children and pregnant women without the five-year waiting bar. (CRS R40130; R40444)
• Before 2009, legal immigrants who arrived after August 22, 1996 were barred from Medicaid and CHIP for their first five years. States had to opt in to waive it. (CRS R40444)
• CBO estimated that of 6.5 million added enrollees by FY2013 under the 2009 law, 2.4 million (37 percent) would otherwise have had private coverage and 4.1 million (63 percent) would have been uninsured. (CRS R40130)

Coverage trends

• A federal evaluation required by the 2009 law found that the uninsured rate for all children fell from 15 percent (11.4 million) in 1997 to 9 percent (6.6 million) in 2012. For children below 200 percent of poverty it fell from 25 percent to 13 percent. Public coverage rose from 20 percent to 35 percent of all children. (HHS ASPE, "CHIPRA Mandated Evaluation of CHIP: Final Findings"; summary page; both opened)
• The same report found 8.13 million children enrolled at some point in FFY 2013, total spending of $13.2 billion ($9.2 billion federal, $4.0 billion state), and that 68 percent of remaining uninsured children were eligible but not enrolled. (ASPE)
• Participation in Medicaid or CHIP among eligible children rose from 82 percent in 2008 to 88 percent in 2012. (ASPE)
• Census Bureau reports a 2023 uninsured rate of 5.8 percent for children under 19, up 0.5 points from 2022. The page did not say whether that change is statistically significant. (Census, Health Insurance Coverage in the United States: 2023; opened)
• MACPAC reports CHIP spending of $28.2 billion in fiscal year 2024. (MACPAC, MACStats release, Feb. 2026; opened. It gave combined Medicaid and CHIP enrollment of 78.0 million in July 2025, not CHIP alone.)

Access and unmet need

• The ASPE evaluation found CHIP enrollees had better access, fewer unmet needs and more financial protection than uninsured children, with the largest gaps for dental care. It also found about 25 percent of enrollees still had some unmet need, most often dental care (12 percent). Emergency room use and hospital stays were similar for CHIP and uninsured children. (ASPE)
• Janet Currie, Sandra Decker and Wanchuan Lin (2008) found public insurance eligibility "unambiguously improves current utilization of preventive care" for children, with little effect on current health status and some evidence of future health gains from early childhood Medicaid. Their abstract refers to public insurance and Medicaid, not CHIP by name. (NBER Working Paper 14173; opened)

Crowd-out: contested

Crowd-out means families dropping private insurance because a public option is available.

• Jonathan Gruber and Kosali Simon report a central estimate of about 60 percent and argue that tight anti-crowd-out rules may cut enrollment more than they cut crowd-out. (NBER Working Paper 12858; abstract opened)
• Eric Seiber and Curtis Florence report low crowd-out in small firms and much higher in large firms. They note that adding age-by-year controls reduced their estimates to zero, and that their text and table disagree on one large-firm figure (49 percent versus 59 to 73 percent). (Health Services Research, 2010, PMC2813446; opened)

Funding instability

• Federal funding expired on September 30, 2017. States kept programs running on leftover money, and some made contingency plans that could have reduced coverage and sent notices to families. Congress passed a six-year extension on January 22, 2018. (KFF, "Summary of the 2018 CHIP Funding Extension"; opened)
• Earlier, from 2005 to 2007, Congress passed separate shortfall payments after some states ran out of allotments, and in 2007 passed a series of continuing resolutions with termination dates of November 16, December 14, December 21 and December 31. (CRS R40229)

What this case shows

1. A program defined by who counts as a child worth covering. The income line, the Medicaid overlap, and the five-year bar on lawfully present immigrant children were all set in statute. Until 2009, a child could be legally present and still excluded. States had to opt in afterward, so coverage depends on the state. This is a plain example of who was left out by a vote.
2. Open vote, open veto. Two bills passed Congress in 2007 and were vetoed on dated, recorded actions. The House override failed by 13 votes. Voters could see the tally and the arguments, which is the thesis in its visible form.
3. Stop-and-go money made the program hard to plan around. Capped allotments, a "dip" in the early 2000s, shortfall bills, and a funding lapse in 2017 meant that states and families had to guess whether coverage would continue. Ordinary people cannot fix a deadline they cannot see.
4. Measured results exist, and they are partial. A federal evaluation recorded falling uninsured rates and better access than for uninsured children. It also recorded that a quarter of enrollees still had unmet needs, mostly dental, and that emergency use did not differ. Good public record, mixed picture.
5. The cost question is disputed in the open. Estimates of crowd-out run from about 4 percent to about 60 percent depending on method. A voter can see the dispute, but only by reading research papers.

The strongest objections, and our answers

1. Objection: Coverage gains are not all due to CHIP. The ASPE source itself says CHIP "together with Medicaid" fueled the decline, so crediting CHIP alone would overstate it. Mitigation: We state the decline is shared with Medicaid and that national trends also moved with the economy. We have not isolated CHIP's own effect.
2. Objection: CHIP replaced private coverage (crowd-out). Mitigation: We show four estimates with their methods. The range is wide and we do not pick one. Different methods answer different questions: direct substitution at enrollment (ASPE) versus population-wide shifts (Gruber and Simon).
3. Objection: The vetoes reflected real concerns about program size and income limits. Mitigation: The BU report records that opponents argued the bill reached children well above poverty. We did not find the President's full stated reasons and say so.
4. Objection: Immigrant restrictions were a national policy, not a CHIP flaw. Mitigation: The five-year bar came from the 1996 welfare law, not Title XXI. We describe it as a rule that applied to CHIP, and note that the 2009 law let states waive it.
5. Objection: Uninsured rates are still falling, so the program works. Mitigation: Census shows the rate rose 0.5 points to 5.8 percent in 2023. That may reflect other policy changes, such as the end of pandemic-era Medicaid protections, which we have not checked.

Open questions: what would change our minds

• The Balanced Budget Act text and the original sponsors (the 1997 bill's Senate leaders and its estimated cost, often cited as $24 billion) were not verified. The CRS reports we opened do not mention them. The $24 billion figure is unverified and omitted.
• The 2018 ten-year extension through FY2027 is known only from search-result titles. We did not open the law or a CRS summary. Unverified.
• ASPE's own report ran past the fetch limit; the final 10,669 characters were not read.
• The report's survey covers only ten states. National generalization is uncertain.
• The Gruber and Simon and Currie, Decker and Lin papers were read at abstract level only.
• Bush's stated reasons for the vetoes were not found. Primary source: veto messages (not opened).
• Why the 2023 child uninsured rate rose, and whether that is statistically significant, was not checked.
• Combined Medicaid and CHIP enrollment (78.0 million in July 2025) is known, but a CHIP-only count was not found.
• What would change our minds: an analysis that isolates CHIP from Medicaid and private-market changes.

Sources

Found an error or a better source? We keep every correction on the record with its date and reason. Nothing here is quietly rewritten.