Affordable Care Act marketplace enrollment fell by roughly 2.6 million people from a year earlier, according to the first complete 50-state analysis of effectuated February enrollment. New Mexico moved the other way. It increased enrollment by about 14 percent after becoming the only state to fully replace the expired federal subsidies. [1]
The new count supplies what Wednesday's account of rising premiums and record deductibles lacked. That article argued that the realized average premium increase looked smaller than the cost of keeping the same plans partly because people bought cheaper coverage or left the marketplace. Thursday's data put a national number on that attrition. They do not identify what happened to each person who departed.
The state pattern is striking. Ohio and Oklahoma each lost more than 32 percent of their marketplace enrollment. Florida lost about 443,000 people, though nearly 4 million remained enrolled. New Mexico alone fully replaced the federal assistance and alone posted an increase. [1] That comparison strengthens the argument that price mattered. It does not turn a comparison into proof that subsidies explain every change.
HHS attributes a significant share of the decline to removing fraudulent or phantom enrollments. Analysts emphasize the expiration of enhanced subsidies and the resulting affordability shock. Both mechanisms can operate inside a 50-state total. The national number does not reveal how much of each state's movement came from verification, price, plan choice or the removal of ineligible records. A slogan can choose one cause. The data cannot.
Nor does leaving a marketplace plan necessarily mean becoming uninsured. Some former enrollees may have obtained employer coverage, Medicaid or another plan. Others may now have no coverage. AP quoted KFF's Cynthia Cox describing the marketplace as coverage of last resort for many users, which makes loss of insurance a serious concern, but concern is not an individual coverage record. [1] The next useful evidence will track where departing enrollees went.
New Mexico is therefore a counterfactual with limits. It offers a real policy difference: one state replaced the expired money while the others did not. It offers a real outcome difference: that state's enrollment rose while the national total fell. It cannot isolate subsidy policy from every demographic, administrative and market difference among states. Evidence earns its authority by stopping where proof ends.
That boundary makes the comparison more useful, not less. A state moving against a broad national decline gives investigators a concrete place to test explanations. Enrollment records can be compared with subsidy amounts, verification removals and later coverage outcomes. The current analysis identifies the question and its strongest contrast; it does not finish the causal work.
The divergence rewards people who erase that boundary. The fraud-cleanup frame treats a smaller roll as a cleaner roll. The affordability frame treats the entire decline as people priced out. The New Mexico comparison gives the latter argument more weight because money and enrollment moved together in the opposite direction there. But it still cannot tell us whether every removed name represented fraud or every departing household lost insurance because of price.
The honest consequence is narrower and more important. The marketplace now covers 2.6 million fewer people than it did a year ago, and the one state that replaced the expired federal subsidies broke from the pattern. [1] That is not a controlled experiment. It is a strong lead for the next investigation: follow the former enrollees, separate administrative cleanup from lost coverage and test whether other states can reproduce New Mexico's result.
-- SAMUEL CRANE, Washington