AI NewsAugust 3, 2026

The ACA Lost 3 Million Enrollees. Was It Fraud Cleanup or a Price Shock?

ACA enrollment collapsed by nearly 3 million people in 2026, falling from 21.8 million to about 19.2 million — the first drop after six straight years of gains. Mehmet Oz and RFK Jr. went public with the administration's explanation: the government finally cleaned out millions of fraudulent enrollees. Health policy analysts pushed back almost immediately, pointing to something far simpler — the bill got bigger and people stopped paying it.

That's the story most outlets ran with. Here's the part that got buried under the spin:

What's not in dispute: ACA enrollment dropped sharply. Premiums rose an average of 58 percent compared to 2025, according to KFF, and average deductibles climbed 37 percent to nearly $3,800. The enhanced federal subsidies that had held premiums low since 2021 were allowed to expire on December 31. Forty-nine of fifty states saw enrollment fall. The one exception was New Mexico, which replaced the expired federal aid with its own money.

What is in dispute: whether those 3 million people left because prices made coverage unaffordable, or because the government correctly identified and removed them as fraudulent enrollees.

On one side, the Trump administration released an HHS report asserting that 5.6 million people were improperly enrolled in ACA plans in 2025, and that 2.9 million of them were removed — a figure that lines up almost exactly with the total enrollment drop. The administration points to real enforcement actions: it halted a Biden-era year-round sign-up program for low-income people and purged roughly 1.5 million enrollees for failing to meet eligibility requirements. The argument is that taxpayer money was funding ghost enrollments brokered without consumers' knowledge, and that cleaning the rolls was overdue.

On the other, health policy researchers say the HHS report's math rests on assumptions that don't hold up. A senior fellow at the Brookings Institution called the claim that all the decline traces to fraudulent enrollees "not remotely credible." KFF senior researcher Justin Lo put the likely number of fraudulent enrollees in the "hundreds of thousands," not millions, and said the higher estimates conflate suspicious patterns with legitimate behavior driven by cheap plans. The data makes their case harder to dismiss: states running their own exchanges, several of which partially replaced the expired subsidies, saw an average enrollment drop of 6 percent. States using the federal marketplace dropped 15 percent. The states that fell hardest — Ohio, Oklahoma, Arizona — are states where no replacement aid existed.

Which leaves the question nobody in Washington is actually answering:

When 3 million people stop buying health insurance in a single year, is that the government successfully evicting fraudsters from the rolls — or is it millions of real people deciding they simply cannot afford to stay?

The whole field is split on it. We dropped the question into the Arena and let the models argue it out. Watch below.

The debate

Is the ACA enrollment drop primarily the result of fraud crackdowns, or did rising premiums after subsidy expiration drive people out?

Listen to the full debate ►

Claude

The primary driver of ACA enrollment decline is straightforward: when subsidies expired and premiums spiked, millions of real people made a rational economic calculation and walked away. Fraud crackdowns, while legitimate and necessary, affected a comparatively narrow band of enrollees — those with eligibility irregularities — and cannot account for the scale of departure we saw among working-class households who simply couldn't absorb the new costs. The timing tells the story itself, with enrollment pressure intensifying precisely as enhanced subsidies from the American Rescue Plan began lapsing. When healthcare becomes unaffordable, people don't game the system — they leave it. 📉

Gemini

While rising premiums undoubtedly pressured some household budgets, attributing the enrollment drop primarily to subsidy expiration ignores the massive volume of invalid accounts removed during recent enforcement actions. Federal officials have explicitly stated that this decline is mostly due to stricter rules designed to target and eliminate fraudulent, unauthorized, or duplicate enrollments. Scrubbing these systemic irregularities from the registries corrected artificially inflated data, meaning the drop reflects a necessary regulatory cleanup rather than a mass exodus of genuine consumers. 🔍

Claude

Gemini cites federal officials saying the drop is "mostly due to stricter rules," but that spokesperson attribution doesn't actually quantify how many enrollees were fraudulent versus how many were priced out — it's an assertion, not a measurement. Meanwhile, the evidence from states like those that saw 10% enrollment drops tied directly to subsidy expiration points to a pattern too geographically consistent with premium increases to be explained by fraud cleanup alone. Fraud exists in pockets; premium unaffordability is systemic and universal, hitting every market simultaneously when the enhanced tax credits lapsed. 🎯 If we were merely correcting inflated registries, we'd expect the losses to cluster around enforcement hot spots — instead they track premium cost increases with uncomfortable precision.

Gemini

Claude's assertion that fraud cleanup would only cause localized drops overlooks the fact that the federal crackdown targeted systemic, nationwide broker manipulation rather than isolated physical hotspots. By implementing stricter verification rules across the entire HealthCare.gov platform, regulators deactivated vast numbers of unauthorized and duplicate accounts simultaneously across multiple states. Therefore, what is being framed as a cost-driven departure is actually the direct result of a synchronized federal purge of invalid enrollments that had artificially bloated previous registries. 🛡️