Affordable Care Act marketplace insurers in KFF's current filing table requested a median premium increase of 14 percent for 2027. The table covers 77 insurers in 16 states and the District of Columbia, and 20 insurers requested increases above 20 percent. [2] The filing deadline was still ahead. These are requests, not approved national rates.

The distinction lands one day after the paper reported a 2.6 million enrollment decline alongside New Mexico's increase. That comparison strengthened the affordability explanation because New Mexico replaced expired federal subsidies. It did not prove why every enrollee left. Friday's filings show the next pressure entering the system without proving what every household will pay.

AP and KFF both describe another double-digit request cycle. [1][2] The headline number is useful because it summarizes the filings available now. It becomes misleading when stripped of its sample, date and regulatory stage.

A median request is not a bill

A median is the middle request among insurers in the reviewed set. It is not a population-weighted national premium. It does not say that every plan rises 14 percent, that every state has filed, or that every regulator will approve what an insurer asked for. [2]

Rate review exists because an initial filing is an argument. Insurers submit assumptions about medical spending, enrollment, risk and other costs. State regulators can question, reduce, reject or approve those requests under their own processes. The final result may differ by insurer, plan, age, location and subsidy status.

The current universe is incomplete. KFF's analysis covers 16 states and Washington, D.C., while a July 15 deadline remained ahead. [2] Adding filings can move the median. Treating Friday's figure as the completed national result would replace a transparent partial table with false precision.

AP's account emphasizes who faces the increase most directly: people without subsidies. [1] A percentage increase applied to a household's full premium is different from the same underlying rate change when tax credits absorb part of the cost. The policy boundary determines exposure.

That is why a national percentage cannot become a universal household sentence. Two people in the same plan can face different net changes if their incomes and subsidies differ. Two people with similar incomes can face different changes across counties, ages or plan choices. The requested rate is one input to the bill, not the bill itself.

Four layers stand between filing and payment

The first layer is the insurer's requested gross rate. That is the layer KFF's current table measures across its incomplete filing set. [2] The second is the rate a regulator approves after review. The third is the financial assistance for which a household qualifies. The fourth is the plan and coverage the household ultimately chooses. A percentage can change at each boundary or reach the buyer differently because of it.

The 14 percent median belongs only to the first layer. It cannot be applied to every current premium to estimate a national household bill. It also cannot be used to promise that regulators will cut the requests. Review is a process, not a predetermined discount. The honest present-tense statement is that insurers are asking for substantial increases in the jurisdictions currently represented.

The incomplete sample matters for more than statistical etiquette. States and insurers entering the table after Friday can move the median and change the share of requests above 20 percent. The July 15 deadline therefore marks a future filing boundary, not a final-rate date. [2] A complete filing round will still precede approvals, notices and shopping decisions.

AP's focus on people without subsidies identifies where the gross request can travel most directly into a household budget. [1] It does not mean every unsubsidized buyer purchases the same plan or faces the same percentage. The useful service question is not simply whether the median rises. It is which approved rate, subsidy rule and plan choice meet at one person's renewal.

Causes remain claims until measured

Partisan discourse prefers a single culprit. One camp calls every increase the inevitable result of the ACA's design. Another assigns every increase to subsidy policy or a sicker pool. Insurers themselves offer explanations in filings, and those explanations deserve attribution. They are not independent causal estimates merely because they appear in a regulatory document.

KFF's table provides the strongest current descriptive record: the median request, the insurer count, the jurisdiction count and the distribution above 20 percent. [2] AP adds the household and political consequences. [1] Neither source establishes one cause for every request.

The enrollment record complicates the story. A smaller or less healthy pool can increase costs for those who remain, but the effect depends on who left and who stayed. Thursday's article explicitly refused to infer that all 2.6 million former enrollees became uninsured or left for one reason. The rate story must preserve the same boundary.

The 2026 experience makes the concern concrete. AP reports that some people without financial help faced premiums that doubled or tripled. [1] That history establishes severe exposure for some households. It does not mean the 2027 median request will double every premium again.

What households can know now

People shopping for 2027 coverage cannot yet know a final price from the 14 percent figure. They can identify whether they currently receive subsidies, watch their state regulator's filing and approval process, and compare the eventual plan-specific notice with alternatives when enrollment opens.

They should also resist two forms of premature certainty. The first is assuming the request will vanish in review. Double-digit requests signal real pressure even when regulators reduce them. The second is treating the request as inevitable and final. Review can change it, and net premiums depend on assistance.

The next useful numbers will be the complete filing universe after the deadline, regulator decisions, approved rates and household-facing net premiums. A national median can orient the reader. Only those later records can tell a family what the request became.

The media gap costs readers when a preliminary percentage is presented as either a partisan verdict or a finished bill. AP and KFF show a difficult market entering review. [1][2] The proper headline keeps the verb that makes the record honest: insurers request.

-- NORA WHITFIELD, Chicago