Download the Full Report (PDF)

Background

During the COVID-19 public health emergency, states received increased federal Medicaid matching rates in exchange for a commitment to maintain continuous coverage for those enrolled in Medicaid. With the public health emergency winding down, as part of the Consolidated Appropriations Act (CAA), 2023, Congress ended the continuous coverage requirement as of April 1, 2023, and enacted other provisions designed to promote a smoother renewal process for the approximately 94 million people who were enrolled in Medicaid at the end of 2022. This process is often called “Medicaid unwinding.” The CAA included a gradual phase down of the extra federal Medicaid funding over the course of calendar year 2023, requirements for states to report unwinding data to the Centers for Medicare and Medicaid Services (CMS), and new enforcement authorities for CMS to ensure that states were adhering to federal rules on the renewal process.

Congress also included a new federal requirement that all states provide 12 months of continuous eligibility for children in Medicaid and the Children’s Health Insurance Program (CHIP), beginning January 1, 2024. But it only applies to children after they have been successfully renewed or newly enrolled in Medicaid or CHIP within the past 12 months.

Under CHIP, states may cover some or all CHIP-eligible children in Medicaid (known as M-CHIP), create a separate program for CHIP-eligible children, or use a combination of the two approaches. Most children (67%) eligible for CHIP are covered through M-CHIP, and were therefore protected by the Medicaid continuous coverage requirement.¹ Some states also voluntarily paused separate state CHIP disenrollments during the pandemic. Fifty-seven percent of children no longer eligible for Medicaid were projected to transition to separate state CHIP programs, according to the Urban Institute.²

Medicaid/CHIP coverage is vital to children’s access to health care – during the continuous coverage protection, more than half of children nationwide were enrolled in Medicaid/CHIP and the uninsured rate for children dropped to one of the lowest levels in recent history.³ Children of color are disproportionately enrolled in Medicaid underscoring the unwinding’s potential impact on families of color.⁴ All states’ income eligibility levels for Medicaid/CHIP are substantially higher for children than for adults, and the median eligibility level is 255% of the federal poverty level ($65,841 annually for a family of three).⁵ Because so many children are eligible for Medicaid, and because families in the past have experienced high levels of administrative “churn,” federal researchers projected that nearly 3 out of 4 children who were disenrolled during the unwinding would remain eligible for Medicaid but would lose coverage nonetheless for procedural reasons.⁶

National data are not available to break out the number of children losing coverage due to procedural disenrollments (as opposed to a determination of ineligibility). However, overall, procedural terminations, as a share of total disenrollments to-date, are approximately 70%.⁷ Procedural disenrollments occur when families or individuals either don’t receive the renewal notice or understand what they need to do to stay enrolled, or they have difficulty navigating the state website or getting assistance through the call center. It is likely that many Medicaid beneficiaries, especially children, who were disenrolled for procedural reasons remain eligible.

This paper uses administrative data reported by the states to CMS to examine how children’s enrollment in Medicaid and CHIP was impacted in 2023 as the unwinding got underway. By the end of 2023, we find that enrollment in Medicaid/CHIP nationwide for children declined by 4.16 million compared to aggregate enrollment before unwinding began (as measured in the month prior to each state beginning its unwinding process). CHIP enrollment (which includes children in CHIP-funded Medicaid and those in separate CHIP programs) essentially remained flat despite expectations that it would grow.⁸ We include CHIP data in our analysis – in states with separate CHIP programs that saw CHIP enrollment increases that can offset a portion of the Medicaid enrollment declines.

As is customary in Medicaid, states had a number of important policy choices in how they would implement unwinding, including when to start and finish the unwinding process and whether renewals for specific groups would be prioritized. States also had wide variation in how well their systems could conduct automatic renewals based on reliable data available to the state which don’t involve any additional paperwork on the family’s part (known as “ex parte,” or administrative renewals). As a result, there are significant differences in ex parte renewal rates across states. Finally, states had a number of new optional flexibilities offered by CMS to smooth the process; some states adopted many of these temporary policy options while one state, Florida, did not take any.

These factors, in turn, contributed to large differences in child disenrollment rates across states. Reflecting concerns about these state differences, in December 2023, Health and Human Services Secretary Xavier Becerra sent letters to 9 Governors expressing concern about the high numbers or rates of child disenrollment from Medicaid. These states were: Arkansas, Florida, Georgia, Idaho, Montana, New Hampshire, Ohio, South Dakota, and Texas.⁹

Pace and Approach to Child Disenrollment Varied Significantly by State

Under the 2023 CAA, states could not end Medicaid coverage for anyone enrolled in Medicaid prior to March 2023, so April 1^st was the first day people could lose their Medicaid/CHIP coverage.¹⁰ Six states began the renewal process in March – Arizona, Arkansas, Idaho, New Hampshire, Oklahoma, and South Dakota. Sixteen states began in April. A handful of states delayed the process for either all Medicaid enrollees (Oregon) or children (Kentucky and North Carolina) and had very few child disenrollments during 2023.

Of the 4.16 million fewer children enrolled in Medicaid/CHIP, one out of four live in Texas and 14% live in Florida. These two states alone account for over one-third of the net Medicaid/CHIP enrollment decline among children (1.6 million) nationwide. As Figure 1 shows, half the children (50%) disenrolled nationwide reside in Texas, Florida, Georgia, and California, and disenrollments are continuing in 2024. These four states account for one-third (31%) of the total child population in the U.S.¹¹

As Figure 2 shows, children’s Medicaid/CHIP enrollment losses quickly accelerated in the second half of 2023 after hitting the one million mark in June.

The top ten states with the greatest Medicaid and CHIP child enrollment losses in 2023 are listed in Table 1. Texas, Florida, and Georgia top the list with Texas exceeding one million children. Please see Appendix 1 for a list of all states.

Due to variation in state size, we also examined the percentage loss in child enrollment in Medicaid/CHIP, compared to pre-unwinding enrollment. As noted above, most children losing Medicaid likely remain eligible, which is undoubtedly the family’s most affordable coverage option with children receiving comprehensive child-focused benefits generally not subject to premiums and cost-sharing. Table 2 shows the states with the largest percent declines in their child Medicaid/CHIP enrollment with South Dakota, Montana and Utah at the top of the list, which experienced declines of one-quarter or more compared to enrollment prior to unwinding. Texas, Idaho, Arkansas, and New Hampshire saw at least one in five fewer children in their Medicaid/CHIP programs by the end of 2023. Please see Appendix 1 for a list of the percentage drop in child enrollment in Medicaid and CHIP for all states.

Finally, we examined which states today have fewer children enrolled in Medicaid/CHIP than they had prior to when the pandemic-related continuous coverage provision took effect, not just prior to when the unwinding began. Some have claimed they are “right sizing” their program to get back to pre-pandemic enrollment. This is a flawed assumption given that all states had children and adults who were eligible for Medicaid but not enrolled prior to the pandemic, and the fact that the number and rate of uninsured children was increasing while Medicaid/CHIP participation was declining before the pandemic .¹² Historically, Medicaid enrollment increases in most years, although pre-pandemic projections by the Congressional Budget Office reflected considerably slower annual growth of about 0.1% for children over the 10-year period ending 2030. Still, the fact that Medicaid and CHIP enrollment is notably lower than pre-pandemic in some states suggests we are losing ground in the nation’s success in covering children.¹³

Table 3 shows which states already are nearing or are below their enrollment prior to the pandemic and the states’ participation rates. It is only if a state had very high participation rates of eligible children enrolled in Medicaid prior to the pandemic and child enrollment grew substantially during the pandemic would there likely be a relatively high number of children who would now be ineligible. Among these states, only Vermont and Iowa had participation rates of over 95% prior to the pandemic. This suggests that the other states listed have moved aggressively to a resumption of administrative practices that inhibit retention of coverage at renewal.

Are Children Moving to Separate CHIP or the Marketplace?

CHIP covers children whose incomes are just above Medicaid. As noted earlier, many states use federal CHIP dollars to cover children in Medicaid and operate under Medicaid rules; in fact, about two-thirds of children whose benefits are funded through CHIP are covered in Medicaid.¹⁴ These children were also protected during the pandemic and are now being redetermined in the unwinding process. Separate CHIP programs, however, should be seeing growth in enrollment as children in families whose income increased during the pandemic are found ineligible and transitioned to CHIP. Federal data used for the analysis above includes both Medicaid and CHIP.

Using separate CHIP enrollment data from CMS and state administrative sources where available, we are tracking enrollment trends in separate CHIP programs which were expected to be the predominant source of coverage for children losing Medicaid during the unwinding. Unfortunately, enrollment growth in most separate CHIP programs is limited with some states experience a decline in CHIP enrollment as well as Medicaid. When comparing Medicaid child enrollment losses to separate CHIP enrollment growth, we found that only 10.3% of Medicaid enrollment declines were offset by CHIP gains.¹⁵

Enrollment in federal and state subsidized marketplaces has increased sharply, reaching a new enrollment milestone of 21.45 million enrollees following the 2024 open enrollment period, driven by enhanced subsidies and aggressive efforts by the Biden Administration to promote enrollment.¹⁶ The marketplaces, however, are a relatively less important source of coverage overall for children than for their parents and other adults – due to higher child Medicaid/CHIP eligibility than adults as described above. Children account for approximately 10% of Marketplace enrollment and this share has remained in the same range for a number of years – even though approximately 600,000 more children selected a plan during the marketplace 2024 open enrollment compared to the year prior. But those marketplace child enrollment gains offset only about 14% of child Medicaid/CHIP enrollment declines through December 2023.

Conclusion

States have approached the Medicaid unwinding differently with respect to ensuring that eligible children remain enrolled and whether they intend to improve their eligibility systems and renewal processes going forward. This could increase participation among eligible children and help reverse unwinding-related child enrollment declines. For example, a growing number of states (12, as of this writing) are seeking Section 1115 demonstration authority to extend multi-year periods of continuous coverage for children – especially young children – such as Oregon’s approach, which will establish continuous eligibility for young children from birth to their 6^th birthday. After that, the state will conduct renewals every two years. For states leading the way on assuring continuous access to health care for children, the lessons of the continuous coverage provided during the pandemic have spurred them to innovate more streamlined ways to keep eligible children enrolled in Medicaid and avoid inappropriate coverage loss at renewal with a particular emphasis on protecting young children.

Other states, however, have prioritized the hasty disenrollment of adults and children, despite projections that many still remain eligible. These include some of the largest non-expansion states where children make up a larger share of Medicaid enrollment – including Texas, Florida, and Georgia. Troubling new data is starting to emerge on gaps in coverage and access to care¹⁷ that will be especially pronounced in states that moved quickly and had high rates of procedural terminations as well as low rates of data-driven automatic or ex parte renewals.

New enrollment of children in the marketplaces and separate CHIP programs appears to be offsetting only a modest share of the net decline in total child Medicaid and CHIP enrollment. No data is available on how many of these children are now covered by employer-sponsored dependent insurance and whether coverage is affordable. A new survey of adults losing Medicaid coverage found that only 16% moved to employer-sponsored coverage.¹⁸ However, the costs of dependent coverage have been rising steadily¹⁹ so children are even less likely to move to employer coverage than adults who had been enrolled in Medicaid.

If a substantial share of the children losing Medicaid remain uninsured, as is likely, this means that a large number of low-income children are going without needed care, which will lead to poorer health in the short and long term while their families are exposed to high medical costs and medical debt.

Appendix

Methodology

Data Sources

This report from the Georgetown University Center for Children and Families (CCF) uses data from the Centers for Medicare & Medicaid Services (CMS) State Medicaid and CHIP Applications, Eligibility Determinations, and Enrollment Data to analyze child enrollment in Medicaid and the Children’s Health Insurance Program (CHIP) by state. All states and the District of Columbia are required to report monthly Medicaid and CHIP enrollment data to CMS as part of the Performance Indicator Project. CCF uses updated rather than preliminary enrollment data for all months except December 2023; data for December 2023 was only available as preliminary enrollment. These data show changes in net Medicaid and CHIP child enrollment, which take in to account any enrollment declines (such as from unwinding-related disenrollments of children covered by Medicaid and CHIP) offset by any enrollment increases (such as from new enrollments among uninsured but eligible children and re-enrollments among eligible children whose families reapplied following disenrollments). Seven states (Colorado, Kentucky, Missouri, New Jersey, Rhode Island, Virginia, and West Virginia) also cover pregnant individuals through CHIP; these data do not allow us to separate coverage of pregnant individuals from child coverage at the state level, so those individuals are included in this analysis. In December 2023, approximately 123,000 pregnant individuals nationwide had CHIP coverage.

Arizona does not separately report child enrollment to CMS, so CCF substitutes state administrative data, which Arizona reports quarterly, for Arizona’s child enrollment numbers in all months.

Where noted, CCF includes data from CMS’s December 2023 Medicaid and CHIP Enrollment Snapshot, as these snapshots provide additional breakouts not available in the above dataset. CMS, however, uses preliminary data in these snapshots, which may differ from the updated numbers used in CCF’s analyses.

Where noted, CCF also reports enrollment in separate CHIP programs based on state administrative data and data reported by CMS as part of their unwinding data reporting. Thirty-nine states with separate CHIP programs are included in the analysis; three states (Kentucky, Rhode Island, and West Virginia) cover pregnant individuals through CHIP and eight states (California, Illinois, Maryland, Michigan, Minnesota, Nebraska, Oklahoma, and Rhode Island) cover individuals under the From-Conception-to-End-of-Pregnancy option. These data do not allow us to separate these groups from child coverage. The most recently reported CHIP data in each state ranges from September 2023 to March 2024.

Timing

This paper uses a “pre-unwinding baseline” to measure changes in net Medicaid/CHIP enrollment through December 2023 during the Medicaid unwinding. The baseline in each state is the month before the month state began the unwinding process; the baseline for the U.S. total is the sum of all states’ baseline enrollment. Because of differences in when states began unwinding, some states may have more or fewer months of data analyzed in this report; states that began unwinding earlier may show larger enrollment declines compared to states that started later.

---

Endnotes

¹ Medicaid and CHIP Payment and Access Commission (MACPAC), MACStats: Medicaid and CHIP Data Book, Exhibit 32. Child Enrollment in CHIP and Medicaid by State, FY 2022 (thousands) (MACPAC, December 2023), available here. For a list of states and how they cover children, see T. Brooks, et al., “Medicaid and CHIP Eligibility, Enrollment, and Renewal Policies as States Prepare for the Unwinding of the Pandemic-Era Continuous Enrollment Provision” (KFF and Georgetown University Center for Children and Families, April 2023), available here.

² M. Buettgens and A. Green, “What Will Happen to Unprecedented High Medicaid Enrollment after the Public Health Emergency?” (Urban Institute, September 2021), available here.

³ J. Alker and A. Osorio, “Child Uninsured Rate Could Rise Sharply if States Don’t Proceed with Caution” (Georgetown University Center for Children and Families, February 2020), available here.

⁴ M. Guth, et al., “Medicaid and Racial Health Equity” (KFF, June 2023), available here.

⁵ Brooks, op. cit.

⁶ Assistant Secretary for Planning and Evaluation (ASPE), “Unwinding the Medicaid Continuous Enrollment Provision: Projected Enrollment Effects and Policy Approaches” (ASPE, August 2022), available here.

⁷ See “What is happening with Medicaid renewals in each state?” (Georgetown University Center for Children and Families), available here, and “Medicaid Enrollment and Unwinding Tracker” (KFF), available here.

⁸ Centers for Medicare and Medicaid Services, “December 2023 Medicaid and CHIP Enrollment Trends Snapshot,” available here.

⁹ Centers for Medicare and Medicaid Services, “Biden-Harris Administration Releases New Medicaid and CHIP Renewal Data Showing the Role State Policy Choices Play in Keeping Kids Covered” (December 2023), available here.

¹⁰ Most states pay a monthly capitation rate to cover enrollees in managed care so enrollment tends to occur on a monthly basis.

¹¹ Georgetown University Center for Children and Families analysis of U.S. Census Bureau 2022 American Community Survey, Table S0101: Age and Sex.

¹² J. Alker and A. Corcoran, “Children’s Uninsured Rate Rises by Largest Annual Jump in More than a Decade,” (Georgetown University Center for Children and Families, October 2020), available here.

¹³ Congressional Budget Office, “Medicaid—CBO’s Baseline as of March 6, 2020,” (Congressional Budget Office, March 2020), available here.

¹⁴ MACPAC, op. cit.

¹⁵ Georgetown University Center for Children and Families analysis of state- and CMS-reported data on separate CHIP programs. These numbers reflect cumulative totals between March 2023-March 2024, but timing varies by state depending on when states started unwinding and how quickly they post enrollment data. See methodology for more information.

¹⁶ E. Park, “Child Enrollment in the Marketplaces Rose by Nearly 40 Percent During 2024 Open Enrollment but Increase Offsets Only Modest Share of Child Medicaid Unwinding Enrollment Losses” (Georgetown University Center for Children and Families, March 2024), available here.

¹⁷ P. Shin, et al., “One Year After Medicaid Unwinding Began, Community Health Centers, Their Patients, and Their Communities are Feeling the Impact” (George Washington University Geiger Gibson Program in Community Health, April 2024), available here.

¹⁸ L. Lopes, et al., “KFF Survey of Medicaid Unwinding” (KFF, April 2024), available here.

¹⁹ KFF, “2023 Employer Health Benefits Survey” (October 2023), available here.