On October 14, 2026, the Department of Health and Human Services stood before reporters in Washington to announce what it described as a major victory for fiscal discipline. Alongside Centers for Medicare & Medicaid Services Administrator Mehmet Oz, HHS Secretary Robert F. Kennedy Jr. unveiled the immediate termination of Affordable Care Act premium tax credits for roughly 189,000 people enrolled through the federal exchange. HHS projected that the maneuver would prevent $738 million in improper subsidy outlays in 2026 and generate annualized savings exceeding $1.5 billion.

The rhetoric surrounding the announcement followed a familiar script. Kennedy declared that taxpayers “should not foot the bill for duplicate coverage or subsidies for people who have died.” Oz struck an even more aggressive posture, claiming that the previous administration was “sleeping on the job” by pausing automated data sweeps, which allegedly left a massive backlog of improper enrollments. In Oz’s telling, cutting these subsidies was a victimless correction: individuals flagged for having duplicate coverage would simply “default to Medicaid,” and “no one is losing health insurance coverage.”

As an economic proposition, this defense falls apart under basic arithmetic. As a portrait of how social insurance functions in the real world, it is entirely fictional.

The Anatomy of a Purge

To understand what actually happened on October 14, one has to examine the administrative breakdown of the 189,000 enrollees that federal officials presented to the press. Approximately 11,000 were identified as deceased in federal records through July 2026, with another 3,500 scheduled for termination by October 31. That accounts for under 6 percent of the total purge. Roughly 58,000 were individuals flagged for overlapping Medicare files. The overwhelming majority—approximately 120,000 people, or nearly two-thirds of the entire cohort—were living, working-class Americans flagged for overlapping Medicaid enrollment.

This composition is critical because it exposes the enduring power of a classic zombie idea: the myth that social welfare programs are fundamentally compromised by widespread, fraudulent double-dipping by ordinary citizens. A zombie idea, in economics, is a theory or narrative that has been thoroughly disproved by empirical evidence, yet continues to shamble on because it serves a convenient political purpose. In this case, the purpose is justifying public spending cuts under the righteous veneer of program integrity.

In reality, low-income citizens do not derive financial gain from possessing both an active Medicaid file and an ACA marketplace policy. As Kennedy himself conceded in an unscripted moment during the press conference, “People who are double enrolled . . . are not gaining anything from that double enrollment, and they don’t even know it in most cases, that they’re double enrolled, there’s no advantage to them of being double enrolled.”

Why, then, do these overlaps exist? The answer lies not in opportunistic fraud, but in the structural friction of low-wage labor markets and the administrative fragmentation of American health care.

The Math of Low-Wage Churn

In states that expanded Medicaid under the Affordable Care Act, the program covers adults earning up to 138 percent of the federal poverty level. For a single individual in 2026, that threshold sits at roughly $21,000 a year. Just above that threshold, individuals become eligible for subsidized private plans on the ACA exchange, assisted by Advance Premium Tax Credits.

Millions of American workers—gig workers, seasonal farmworkers, retail clerks, home health aides, and restaurant staff—do not earn steady, predictable salaries. Their hours fluctuate from week to week and month to month. A worker might earn $1,400 in February, putting them within Medicaid limits, and $2,100 in July, pushing them over the statutory line into exchange eligibility. When workers take on more hours or find a new job, they dutifully report their income and enroll in an exchange plan to maintain care.

Under normal circumstances, state Medicaid agencies are supposed to process changes in circumstance and close inactive files. But state bureaucracies are chronically underfunded and administratively overwhelmed. This friction grew exponentially worse following the unwinding of pandemic-era continuous enrollment between 2023 and 2024, when state agencies had to reevaluate more than ninety million Medicaid recipients. That massive dislocation created backlogs that persist to this day. A state Medicaid office may take four, six, or nine months to formally close a file after a worker has transitioned to an unsubsidized or partially subsidized private plan.

When CMS runs an automated data match comparing federal marketplace enrollments against state Medicaid databases, it captures these administrative processing delays and classifies them as duplicate enrollments. The worker did not commit fraud; they simply attempted to follow the rules while two legacy computer databases failed to communicate in real time.

The Paperwork Guillotine

Having identified these paper overlaps, CMS did not deploy caseworkers to confirm whether individuals actually had active, accessible Medicaid coverage. Instead, the agency instituted an administrative cliff. Under the enforcement protocol described by federal officials, enrollees flagged by data matches are mailed a written notice giving them a strict thirty-day window to respond, gather documentation, and navigate an arcane dispute process.

If an enrollee fails to satisfy the bureaucracy within thirty days, CMS automatically terminates their Advance Premium Tax Credit. Peter Nelson, the federal official overseeing the ACA marketplace, confirmed that only about 50 percent of the enrollees sent notices during the initial notification wave responded within the thirty-day window. The remaining half had their subsidies zeroed out immediately.

To treat a 50 percent non-response rate as proof of ineligibility is an act of willful institutional blindness. Low-income workers move far more frequently than the general population due to housing instability. They live in rental units with inconsistent mail delivery, work multiple shifts that prevent them from spending hours on the phone with call centers, and often lack reliable internet access to upload identity verification documents. A thirty-day deadline is not a program integrity filter; it is a procedural guillotine designed to depress enrollment through attrition.

Moreover, Oz’s assertion that disenrolled workers “default to Medicaid” is flatly inaccurate. CMS’s own regulatory notices specify that the agency is terminating federal premium tax credits, not the underlying commercial insurance policies. When CMS cuts off the tax credit, the private insurance company does not automatically cancel the plan or redirect the patient to Medicaid. Instead, the insurer bills the policyholder for the full, unsubsidized monthly premium—amounts that frequently run between $400 and $800 a month. A retail worker earning $2,000 a month cannot pay an $800 premium. When the payment fails, the commercial carrier cancels the policy for nonpayment, leaving the worker with no insurance whatsoever.

A Well-Worn Historical Playbook

This tactic of using procedural hurdles to strip low-income families of benefits while claiming to combat dependency has a long, documented history. We have seen this exact experiment conducted before, and we know precisely how it ends.

In 2018, Arkansas became the first state in the nation to implement mandatory Medicaid work reporting under an administrative waiver granted by the first Trump administration. The policy required beneficiaries aged thirty to forty-nine to log eighty hours of work or qualifying activities each month through an online portal. Proponents claimed the mandate would encourage self-sufficiency and purge individuals who refused to work.

The empirical results were devastating. Within six months, more than 18,000 eligible Arkansans lost their health coverage. A landmark study published in the New England Journal of Medicine by Benjamin Sommers and colleagues examined the outcome and found that the policy did not increase employment by a single percentage point. Instead, thousands of working adults lost coverage because the state required them to navigate a confusing, desktop-only website that shut down every night for maintenance. They were eligible workers defeated by paperwork.

The parallel between Arkansas in 2018 and the CMS purge of 2026 is exact in its institutional mechanics: both policies exploit bureaucratic complexity to achieve disenrollment while publicly framing the outcome as an administrative success. The Arkansas comparison breaks down only in scope. While Arkansas targeted Medicaid recipients in a single state, CMS is applying automated data purges across all states using the federal exchange, threatening nationwide coverage stability.

We saw the same dynamic during the 2023–2024 post-pandemic Medicaid unwinding. Nationwide, state agencies disenrolled more than twenty-five million individuals. According to comprehensive monitoring by health researchers, more than 70 percent of those terminations occurred not because people were proven to be over-income, but for “procedural reasons”—a returned envelope, an unanswered letter, or a clerical backlog. CMS is now importing the worst practices of the Medicaid unwinding directly into the federal health insurance marketplace.

The Fallacy of Cost Reduction

Behind the administration’s celebratory press releases lies a fundamental economic error: confusing budget cuts with cost reduction. Stripping a health care subsidy from a low-income family does not eliminate the cost of their medical care. It merely transfers that cost from the federal balance sheet to the least efficient, most expensive corners of the American health care system.

When uninsured people become sick, they do not stop needing care. They defer primary care, forgo maintenance medications for chronic conditions like hypertension and diabetes, and avoid diagnostic screenings. When an untreated chronic condition inevitably metastasizes into an acute crisis, they arrive at hospital emergency rooms. Federal law mandates that emergency rooms stabilize patients regardless of ability to pay, but emergency rooms are the single most expensive venue in modern medicine to deliver care.

Hospitals absorb these expenses as uncompensated care, which reached tens of billions of dollars nationally even before the current round of cuts. Hospitals do not simply absorb these losses; they pass them along. They shift costs onto commercial insurers through higher reimbursement rates, which insurers pass on to employers and consumers through increased private premiums. Local governments step in with municipal tax dollars to prevent safety-net public hospitals from closing. In basic macroeconomic terms, terminating a $300 monthly preventive subsidy does not save $3,600 a year; it frequently creates a $30,000 emergency bill that taxpayers and insured workers ultimately foot through indirect channels.

The timing of this administrative crackdown is equally telling. The purge of 189,000 enrollees on October 14 came just one day before the launch of the annual open enrollment period for the 2027 plan year, and three weeks before the midterm congressional elections. It follows a prior sweep on September 22, in which CMS abruptly dropped 315,000 marketplace accounts covering over 760,000 individuals labeled as “unauthorized,” claiming $2.2 billion in savings.

These aggressive actions arrive as total effectuated enrollment in the ACA individual market has already plummeted by nearly three million people in 2026, dropping to 19.2 million. That contraction was driven primarily by the expiration of enhanced premium subsidies originally enacted under the Inflation Reduction Act, compounded by double-digit premium rate increases across private insurers. When health policy experts pointed out this steep decline, administration officials claimed it was proof that their fraud enforcement was working. But as Matthew Fiedler, a senior fellow at the Brookings Institution’s Schaeffer Center for Health Policy, observed, the administration’s claim that this enrollment collapse reflects improper sign-ups leaving the market “is not remotely credible.”

The Regulatory Road Ahead

Rather than pausing to evaluate the human fallout, federal regulators are accelerating the machinery of disenrollment. CMS has announced plans to implement a centralized national monthly data-matching clearinghouse in early 2027 to police overlap continuously. Furthermore, starting in January 2027, states will be required to verify address information on a regular basis, creating another recurring procedural hurdle that will systematically trip up housing-insecure workers.

The fundamental conflict here is between two competing models of public administration. One model views social insurance as a public trust whose primary duty is ensuring that vulnerable people have uninterrupted access to necessary medical treatment, accepting that minimal friction in administrative databases is preferable to leaving a sick person without a doctor. The other model treats social insurance as a corporate ledger to be trimmed through automated algorithms, viewing high dropout rates as evidence of efficiency rather than administrative failure.

When government officials announce that they have saved hundreds of millions of dollars by eliminating duplicate lines on an electronic spreadsheet, they are presenting half of an equation. The other half will not appear in federal press releases. It will appear in uncollected medical debt, overburdened county hospitals, and families forced to choose between groceries and insulin. That is not fiscal stewardship; it is administrative neglect disguised as arithmetic.