Here’s What Happens When 7.5 Million Americans Lose Medicaid Coverage in the Largest Healthcare Rollback in History

The reconciliation law President Trump signed on July 4, 2025 is the largest single Medicaid rollback on record. The law contains roughly a trillion dollars in gross Medicaid and CHIP cuts over the coming decade, and the Congressional Budget Office’s detailed August 2025 estimates put the resulting coverage losses at 7.5 million people losing health insurance through Medicaid and CHIP alone, rising to 10 million when marketplace changes are factored in. Larry Levitt of KFF has called it “the biggest rollback in federal support for healthcare ever.” The package layers new work requirements onto state Medicaid programs, restricts how states can tax hospitals to draw down federal matching dollars, and tightens eligibility for lawfully present immigrants. Congress also allowed enhanced Affordable Care Act subsidies to expire, further shrinking marketplace enrollment.

To understand what this means at a kitchen table in Little Rock or Memphis, consider what Medicaid actually does for the economy. In the first quarter of 2026, the program delivered $1.05 trillion in transfer income to American households, making it the third-largest federal transfer program after Social Security and Medicare. Pulling roughly a trillion dollars out of that pipeline over several years lands directly on household balance sheets already under pressure. The personal savings rate stood at 3.9% in the first quarter of 2026, down from 6.2% in the first quarter of 2024. Consumer sentiment, as measured by the University of Michigan, fell to a record low of 44.8 in May 2026, partially recovered to 55.2 in July, and then dropped again to 51.7 in August as inflation and foreign-policy uncertainty weighed on households. That August reading sits at the first percentile of the index’s entire history.

The Arkansas Tell

The strongest reason to take the CBO estimate seriously is that this experiment has already been run. When Arkansas imposed work requirements in 2018, more than 18,000 people lost coverage before a federal court halted the program in April 2019. The New England Journal of Medicine study of those events found that more than 95% of the target population appeared to meet the requirements or qualify for an exemption, yet thousands still lost coverage. The losses were driven not by ineligibility but by confusion: many beneficiaries were unaware of the reporting requirement entirely, and those who knew about it often could not navigate the monthly online portal. According to Arkansas Advocates for Children and Families, 97% of those removed were actually compliant or had valid exemptions. The coverage losses were, in the words of the lead Harvard researcher, a consequence of “red tape” rather than policy design.

Tennessee’s 2005 TennCare retrenchment removed approximately 190,000 beneficiaries from coverage in a matter of months. Research published in Health Affairs found that the disenrollment led to a 27.6% greater increase in county-level eviction filings in Tennessee relative to other southern states in the years that followed. A separate NBER analysis found that hospital service utilization declined after the policy and that the share of uninsured patients among hospital admissions rose sharply. The pattern repeated at the federal level in the early 1980s, when Reagan-era budget cuts pushed both the uninsurance rate and the poverty rate higher during a recession, precisely when the safety net was supposed to catch people falling through.

The downstream costs are predictable. Dr. Adam Gaffney of Harvard Medical School expects “a surge in uninsurance, a rising number of Americans who are not going to the doctor, growing numbers who are not taking the medications they need or avoiding the emergency room because they don’t want to get hit with medical bills.” Rural hospitals, which depend disproportionately on Medicaid reimbursement and on the state provider taxes Congress just restricted, will absorb the first wave. Household healthcare spending, already running at $3,700.1 billion at an annual rate in April 2026, up from $3,494.0 billion a year earlier, will shift from insured care toward out-of-pocket bills and unpaid emergency room visits.

What To Watch

The early signals are already arriving. Nebraska became the first state in the nation to implement the new work requirements, rolling them out on May 1, 2026, eight months ahead of the January 1, 2027 federal deadline. The first round of disenrollments, completed in late July 2026, removed roughly 200 Nebraskans from coverage. Montana followed with its own implementation on July 1, 2026, and Iowa is set to begin on December 1, 2026. The federal deadline applies to all 41 expansion states. Whether the early-mover results will scale to those larger, more complex state systems is the central empirical question of the next 18 months.

With unemployment at 4.1% as of August 2026 and the personal savings rate having fallen sharply since 2024, the financial cushion available to absorb a coverage shock is thin. The CBO projects coverage losses will grow each year through 2034, reaching 10 million uninsured at peak. That trajectory makes the disenrollment data emerging from states in 2026 and 2027 the most consequential public health signal in a generation. If the Arkansas pattern holds at scale, the share of losses attributable to paperwork rather than genuine ineligibility will be the number that reveals whether this is deliberate policy or administrative attrition.

Editor’s note: This article has been updated to reflect Nebraska’s May 1, 2026 launch as the first state to implement the new work requirements, with the first disenrollment round removing roughly 200 enrollees, and to note Montana’s July 1, 2026 implementation and Iowa’s December 1, 2026 start date. The unemployment figure has been updated to 4.1% as of August 2026, per the Bureau of Labor Statistics, and consumer sentiment data has been extended through the August 2026 final reading of 51.7, which sits at the first percentile of the index’s recorded history.

Contact [email protected] for any questions or corrections.