CHIP runs on a capped budget. That's what makes its flat line strange.
Summary
The Children's Health Insurance Program spent $17.4 billion in federal money covering 8.9 million children in FY2023 — under a funding cap Medicaid doesn't have. In the past year CHIP enrollment barely moved, down just 2%, while Medicaid fell 5% and shed 3.7 million people. The capped program is proving sturdier than the open-ended one.
Follow the dollar
is a joint federal-state program, but the federal government pays the larger share. In fiscal year 2023, states and territories spent $23.4 billion combined on -funded coverage — $17.4 billion of it federal money, per MACPAC's MACStats Data Book, which tracks state-by-state spending using 's own expenditure records. Most of that money — $15.3 billion — pays for children whose coverage counts as "Medicaid-expansion ": states use dollars to cover kids inside their Medicaid programs. Another $7.1 billion runs states' free-standing separate programs, plus coverage for pregnant women in seven states that use funds that way. Just over $1 billion — about 4 cents of every dollar — pays for running the program itself: eligibility systems, outreach, staff.
View data as table
| Federal share | $17,388.8M | revenue |
|---|---|---|
| State share | $6,006.5M | revenue |
| Medicaid-expansion CHIP | $15,262.4M | of total spending |
| Separate CHIP & pregnant women | $7,122.8M | of total spending |
| State program administration | $1,010.2M | of total spending |
The federal share isn't a flat match rate — it's an "enhanced" version of each state's regular Medicaid match, calculated under a formula in Section 2105(b) of the Social Security Act: take a state's regular Medicaid match rate, close 30% of the remaining gap to 100%, and cap the result at 85%. For fiscal year 2026, that formula sends wealthier states — California, New York, Massachusetts, and seven others — to the statutory floor of 65%, while Mississippi's lower per capita income pushes its rate to 83.83%, the highest of any state. The four U.S. territories with a program are set by statute at the maximum, 85%.
A budget with a ceiling
That enhanced match rate is generous, but it only applies to a fixed pot. Each year, per CRS's overview of CHIP financing, every state gets a capped federal allotment — sized off its recent spending and adjusted for health-cost and child-population growth. If a state's costs run ahead of its allotment, it can draw on a national Child Enrollment Contingency Fund or on money redistributed from states that under-spent their own allotments in the prior two years. If those sources still don't cover it, the state is on its own. Medicaid has no equivalent ceiling; has always had one. The Consolidated Appropriations Act, 2023 extended 's federal allotments, its redistribution and contingency-fund mechanisms, and a "maintenance of effort" rule barring states from tightening children's eligibility below current levels — all through fiscal year 2029. The money and the eligibility floor are both locked in for another three years.
The same system, counted in children
None of that financing structure explains why enrollment has held up better than Medicaid's over the past year — the two programs' caseloads move for administrative and behavioral reasons, not funding-formula ones. But the contrast is real and it's recent. Between February 2025 and February 2026, per CMS's monthly Medicaid and CHIP enrollment snapshot, enrollment slipped from 7.4 million to 7.2 million — down 130,000 people, or 2%. Medicaid, over the same twelve months, fell from 71.3 million to 67.7 million — down 3.7 million people, or 5%.
View data as table
| CHIP | -2% | 7.4M -> 7.2M, Feb 2025-Feb 2026 |
|---|---|---|
| Medicaid | -5% | 71.3M -> 67.7M, Feb 2025-Feb 2026 |
Nobody changed the income rules that decide who qualifies for either program during this period — both are still means-tested at the same thresholds they were a year ago, as the Georgetown Center for Children and Families has documented. What's moved instead is who successfully makes it through a renewal or stays enrolled through paperwork churn — and, researchers suspect, a "chilling effect" among mixed-status immigrant families wary of enrolling citizen children in any public program. The larger cuts to Medicaid eligibility rules passed in 2025's reconciliation law haven't fully phased in yet; 's steadier line so far measures the period before that law starts to bite, not immunity from it.
The takeaway
- runs on a leash Medicaid doesn't wear. Federal money is capped by statute in a fixed annual allotment; federal Medicaid money is an open-ended match with no ceiling. That's the opposite of what you'd guess from the last year of enrollment data.
- Almost none of the dollar is overhead. Of $23.4 billion spent in FY2023, roughly 4 cents on the dollar paid for running the program; the rest was somebody's coverage.
- 's flat line is a snapshot, not a verdict. Enrollment held closer to steady than Medicaid's did over the last twelve months, but the eligibility rules that will test that steadiness haven't fully taken effect yet.
The 8.9 million figure counts children only (MACPAC's Exhibit 32, child enrollment in Medicaid-expansion plus separate ); the $7.1 billion separate- spending figure in the money-flow chart also includes coverage for pregnant women in seven states, so the two totals aren't directly comparable population counts. The February 2025–2026 enrollment comparison uses 's monthly Performance Indicator Data, a different and broader snapshot definition that includes pregnant women; advises against comparing figures across its own data sets, so the FY2023 exhibits and the monthly snapshot are presented here separately rather than combined into one series.
Sources
- Medicaid and Payment and Access Commission (MACPAC), MACStats: Medicaid and Data Book, December 2024, Exhibit 33 — spending by funding source and benefit category, FY2023, the source for all dollar figures in the money-flow chart. macpac.gov
- MACPAC, MACStats: Medicaid and Data Book, December 2024, Exhibit 32 — -funded child enrollment by state, FY2023, the source for the 8.9 million figure. macpac.gov
- U.S. Department of Health and Human Services, Federal Financial Participation in State Assistance Expenditures; Federal Matching Shares for Medicaid, , and Aid to Needy Aged, Blind, or Disabled Persons for October 1, 2025, Through September 30, 2026 (Federal Register notice, Nov. 29, 2024) — the enhanced FMAP formula and state-by-state rates for FY2026. govinfo.gov
- Congressional Research Service, Federal Financing for the State Children's Health Insurance Program () (R43949) — how 's capped allotments, growth-factor rebasing, contingency fund, and redistribution mechanism work. everycrsreport.com
- Georgetown University Center for Children and Families, Consolidated Appropriations Act, 2023: Medicaid and Provisions Explained — the extension of allotments, redistribution, the contingency fund, and the children's maintenance-of-effort rule through FY2029. ccf.georgetown.edu
- Centers for Medicare & Medicaid Services, February 2026: Medicaid and Eligibility Operations and Enrollment Snapshot (released May 29, 2026) — national monthly Medicaid and enrollment, the source for the February 2025–February 2026 comparison. medicaid.gov
- Georgetown University Center for Children and Families, Drop in Child Medicaid and Enrollment Even Before H.R. 1 Policies Take Full Effect Is Troubling Sign — context on unchanged eligibility rules, chilling effects, and the pending 2025 reconciliation-law cuts. ccf.georgetown.edu
Comments
Always open. Logged-in readers can annotate paragraphs in place.
Medicaid is an open-ended entitlement: whatever a state spends on eligible people, the federal government matches, with no ceiling. , its smaller sibling for children just above the Medicaid income line, is the opposite — a capped block of federal money, set once a year, that runs out if a state overspends it. On paper that should make the more fragile program. Over the last twelve months it hasn't worked out that way: enrollment barely moved while Medicaid enrollment fell hard.