If you have been following the debate about Medicaid cuts, you know it’s a giant program, America’s largest in terms of the number of people it covers. But you may be confused about how many people Medicaid actually covers. Is it 71 million Americans, the number most commonly used in The New York Times and many press reports? Or 83 million Americans, a number Medicaid experts like us and some others frequently use? There are other numbers too, and other ways of thinking about the reach of the program.
On the chance that others are confused by the different Medicaid counts, let me clear this up.
The 71 million number comes from CMS administrative data (meaning program data reported from the states, not self-reported survey data). It doesn’t include the closely connected CHIP program in the 71 million count. It’s also the most recent data (2025). However, this count excludes people who get Medicaid coverage for some—but not all—Medicaid-covered services, such as for family planning. During the pandemic and the unwinding of the continuous enrollment provision, these data were updated regularly and helpful to track changes during a time of dynamic enrollment changes.
You likely already figured out that the 83 million number then includes the people who have partial Medicaid coverage for significant services but not everything covered by Medicaid. A large group of them have coverage for family planning and breast and cervical cancer.
This group includes, for example, an estimated 1.3 million people who would lose Medicaid help with Medicare premiums and cost sharing but keep Medicare, if the Medicaid cuts proposed by Republicans in Congress go into effect. This data set also allows for breaking out the expansion population, which is obviously important in the current debate. It’s also based on administrative data, but from 2024, not 2025.
On balance, the 83 million count is probably the best number to use now. It best represents the larger universe of people who have a stake in the current fight about Medicaid and the policy decisions being debated. However, there is no technically right or wrong choice about which count to use.
Then, there is a third number. It’s 69 million. It comes from federal surveys. The survey data is highly useful for other reasons, but it’s the weakest way to count the Medicaid total because people often don’t know what kind of coverage they have. That’s true in surveys of health insurance status whether people have public or private coverage. With all of the different names state Medicaid programs use, as well as county and private plans people may enroll in, plus people who may be dually eligible for Medicaid and Medicare, some Medicaid beneficiaries are not sure what kind of coverage they have. As a result, these surveys undercount Medicaid enrollees. The data are also from 2023. On the plus side, these data are especially useful for analysis. They allow comparative analysis of all the sources of coverage people have and changes in the uninsured rate, as well as of disparities by race, income, and other factors. For example, we were able to document an increase in Medicaid and non-group coverage and decreases in the uninsured during the 2019 to 2023 period using the survey data.
An even more important number politically is the number of people who have some reason to care about Medicaid. They may have been covered by Medicaid themselves or had a family member covered by it. They may think they may need the program in the future for themselves or an elderly family member. We can’t calculate that number, but we can say it’s larger than the number on Medicaid at any one point in time, and we do know that half of the American people say they or a family member have been on the program at some time. That isn’t a “count,” it’s a measure of Medicaid’s “reach.” It speaks to why the public resists cutting Medicaid even more powerfully than the traditional “counts” do.
Politically, it probably doesn’t matter much whether 71 million or 83 million beneficiaries are attributed to Medicaid, both are huge numbers. For liberals a bigger number is helpful; it underscores the importance of the program. For conservatives, it underscores their belief that Medicaid is simply “too big.” For the public and in politics, anything in the millions can be made to sound big and consequential, and understandability is often more important than sheer size. Still, it’s not an insignificant difference. It’s also one that happens to make Medicaid even that much bigger than Medicare (67 million), if politically still slightly less hefty. It matters too when calculating how many people will be affected by Medicaid cuts and who they are. And possibly it’s useful to explain why there are different numbers out there about what seemingly is an all-time simple question: how many people are on Medicaid?
