Medicare Advantage Mass Exit for 2027

The cost curve moved up as the premium factors moved down. Everything else in the Medicare Advantage exits follows from that.

By RaeAnn, Founder and Chief Executive Officer, HLTHWORKS

1.3 MILLION LIVES

Medicare Advantage lives with coverage impacted by announced plan exits across the 2026 and 2027 enrollment cycles, and the 2027 count is not final.

The headline is that more than a million Medicare Advantage lives have their coverage impacted. The story is that no insurance carrier can actuarially mitigate these moves. CMS raised the benchmark well above what it proposed, and the carriers cut anyway. That single fact tells you this is arithmetic rather than negotiation.

FIGURE	WHAT IT IS
600,000	Humana, 2027 plan exits
600,000	UnitedHealth, 2026 exits across 109 counties
80,000	Molina, the entire MAPD product retired for 2027
7 percent	Share of surveyed health plans calling Medicare Advantage currently profitable

What was announced

On its second quarter earnings call, Humana told investors that its 2027 Medicare Advantage plan exits will affect approximately 600,000 members. It is the second consecutive year of exits, following a 2026 cycle in which the company left three states and 194 counties. Leadership tied the decision directly to a goal of reaching a sustainable pre tax margin of at least three percent in 2028.

Humana is the largest number announced. It is neither the first nor the most severe.

CARRIER	WHAT HAS BEEN DISCLOSED
Humana	2027 exits affecting approximately 600,000 members. The majority involve plans rated 3.5 stars or lower for the 2027 bonus year, though leadership said the decisions were not primarily driven by star ratings, and that the portfolio was prioritized toward plans with greater value based care penetration. The criterion was what the company called the lower tail of profitability and return. On the call the company said it expects to recapture a significant portion of that volume, and that the figure following the 2025 exits ran at just over forty percent. Recapture is investor language. Inside a plan the work is called selective retention: keeping preferred members in preferred service areas.
UnitedHealth	Exited more than 100 plans across 109 counties for 2026, affecting roughly 600,000 members. Medicare Advantage membership is down about 965,000 since year end. Revenue essentially flat while operating earnings rose fifty five percent.
Molina	Retiring the traditional individual Medicare Advantage prescription drug product entirely for 2027, roughly one billion dollars in premium and about 80,000 members, to concentrate on its dual eligible book.
System owned plans	Presbyterian Health Plan is exiting most of its Medicare Advantage plans for 2027, about 30,000 members. Providence Health Plan is exiting as well. In both cases the parent system continues delivering care. What is withdrawn is the insurance product.
Others signaling	CVS and Aetna have stated they are prioritizing margin over membership for 2027. Elevance has warned it will continue leaving unprofitable geographies. UCare, Samaritan Health, and Blue Cross Vermont exited in the prior cycle.

THE NUMBER THAT EXPLAINS ALL OF IT

In a January 2026 survey of Medicare Advantage leaders at more than thirty five health plans, spanning nationals, Blues, and provider sponsored plans, just seven percent said their Medicare Advantage business was currently profitable. That is not a cycle. That is a segment operating at a loss and deciding, all at once, to shrink its way back.

The math changed. That is the whole story.

The cost curve moved up as the premium factors moved down. Those factors are the benchmark, the risk adjustment factor, the quality bonus, and the rebate. The benchmark is only one input. The risk adjustment factor is the multiplier applied to it, and that is where the damage occurred.

MECHANISM	WHAT IT DID
V28 at full weight	Payment year 2026 is the first year risk scores are calculated one hundred percent from V28, ending a three year phase in that had cushioned the impact. Roughly two thousand diagnosis codes removed. Published analysis puts average risk score compression near three percent, with a Trust Fund impact around eleven billion dollars.
Recalibrated coefficients	Diabetes is the clearest illustration. Under V28 the diabetes categories share a single coefficient near 0.166. Under V24, diabetes with complications carried roughly 0.302 to 0.368 depending on severity. The patient did not change. The payment for that patient did.
Unlinked chart review excluded	Finalized April 6, 2026. Diagnoses from unlinked chart review records no longer count toward risk scores without a qualifying encounter, subject to a narrow exception for members switching between Medicare Advantage plans. CMS projected roughly seven point two billion dollars, on the order of a one and a half percent payment impact. Plans submitted about 88.8 million unlinked chart review diagnoses in 2023, and roughly eighty five percent could not be matched to an encounter.
Star ratings misses	Humana lowered full year GAAP earnings guidance from at least $8.36 to at least $6.52 as lower star ratings cut into quality bonus payments. That single line is roughly the size of a small plan’s entire margin.

AND THE BENCHMARK WENT UP ANYWAY

CMS proposed a 0.09 percent increase in January. More than one hundred organizations objected. In April CMS finalized 2.48 percent, over thirteen billion dollars, and 4.98 percent once risk score trend is counted. CMS also declined to finalize a further risk model update, leaving V28 coefficients stable for the year.

The rate improved and the carriers cut anyway. Sequence matters here. The V28 compression landed in payment year 2026 and does not reverse. It is the baseline the 2027 bid was built on, not a cushion. On top of that baseline, 2027 adds the chart review exclusion at roughly a percent and a half, against a benchmark up two and a half. Plans were bidding a modest increase over a base that had already been reset downward, into a cost trend running well above both.

AND THE COST CURVE MOVED THE OTHER WAY

Members are more complex. Utilization has stayed elevated well past the point it was expected to normalize, and the acuity plans report has continued to rise. Which produces the fact nobody in this industry says out loud: the risk model was recalibrated downward at the same moment the population it measures became harder to care for.

And pharmacy became a different business. Under the Part D redesign, plan liability in the catastrophic phase rose from twenty percent to sixty percent of gross costs, while Medicare reinsurance fell from eighty percent to twenty percent for brands. Consider a member on a GLP-1 costing roughly a thousand dollars a month, now available for a broader set of indications on 2026 formularies. That member reaches the out of pocket cap around month three to five and pays essentially nothing after that. The drug does not become free. The plan carries months four through twelve at sixty percent, against a bid filed June 1 of the prior year that cannot be reopened.

No insurance carrier can actuarially mitigate these moves. When the premium factors fall and the cost curve rises in the same cycle, the only lever left is subtraction.

II Four kinds of exit, and only two remove coverage

Coverage of this subject treats every announcement as the same event. Four distinct decisions are being made, and conflating them produces the wrong conclusion for everyone reading.

TYPE	WHAT IT MEANS	WHO
Geographic	Leave counties, keep the product. A judgment about specific markets. Reversible.	Humana, UnitedHealth, Aetna, Elevance
Product	Retire the product line entirely. A judgment about the product itself. Very hard to reverse.	Molina
Insurance risk	A health system withdraws its own plan while continuing to deliver care in the same community. It stops owning the premium. It does not stop treating the patient.	Presbyterian, Providence, Samaritan Health
Network	The provider leaves the plan, or the plan terminates the provider. The provider keeps treating Medicare patients under Traditional Medicare.	Twenty five health systems and counting

Only the first two remove coverage from a market. The other two are judgments about who bears risk and on whose terms, and they are routinely misreported as withdrawals.

A health system that terminates a Medicare Advantage contract has not left Medicare. It continues treating those same patients under Traditional Medicare and retains its Medicaid managed care and commercial relationships. It declined one payer’s terms for one product. Read that for what it is: a preference for Traditional Medicare economics over that plan’s Medicare Advantage economics.

MOLINA IS THE ONE WORTH STUDYING

Molina did not trim counties. It concluded the product was the problem, retiring roughly a billion dollars of premium and about eighty thousand members to concentrate on a dual eligible book of roughly five billion. It announced in early February, days after CMS proposed a near flat rate in the Advance Notice. CMS later softened that considerably. Molina did not reverse course.

THE QUESTION THAT LEAVES FOR EVERYONE ELSE

Molina did not leave Medicare. It left the half of Medicare that does not come with a Medicaid contract attached. Duals carry higher revenue, real rather than documented acuity, a member already known through Medicaid, and a state contract that a benefit war cannot easily dislodge.

If a carrier that specializes in government programs concludes that individual Medicare Advantage does not work without a Medicaid anchor, what is the durable advantage in the general market? The honest answers are scale, provider alignment, and Stars performance. Any organization that cannot name which of the three it holds is competing on benefits, and benefits are the first thing cut in a compressed cycle.

Where the members actually go

Every projection assumes displaced members choose another Medicare Advantage plan. In most counties they will. In a growing number, that option is narrowing, and in some it will not exist.

When no Medicare Advantage plan remains available, the member returns to Traditional Medicare, and Traditional Medicare is not comprehensive coverage. Part B generally pays eighty percent of the approved amount for covered services. Part A carries its own deductible for each benefit period and daily coinsurance once a stay runs long. The exposures are structured differently and do not aggregate neatly. What they share is the only thing that matters here: Traditional Medicare has no annual out of pocket maximum.

That is a difference in kind rather than degree. A Medicare Advantage plan is required to cap what a member can be asked to pay in a year. Traditional Medicare is not. A member coming off a capped product with an integrated drug benefit is moving from a benefit with a defined worst case to one without.

WHAT A DISPLACED MEMBER IS ACTUALLY FACING

The Medicare Supplement window is narrow and timed. Federal protections give a member whose plan leaves the service area a right to buy certain Medicare Supplement plans without medical underwriting, within a limited window after coverage ends. It does not extend to every plan letter, and outside a handful of states requiring broader access, a member who misses it can be underwritten and declined. A member with multiple chronic conditions is precisely the member underwriting is designed to price.

The cost structure changes in kind. A zero premium Medicare Advantage plan is replaced by a Part B premium, a Medicare Supplement premium, and a standalone Part D premium. Standalone Part D is also consolidating, with plan counts down more than twenty percent for 2026 and roughly half what they were in 2024.

Medicare Advantage grew for two decades on a real proposition: richer benefits, a capped exposure, one card instead of three, at little or no premium. That proposition was financed by premium factor and quality bonus models that are now eroded. The 2027 cycle is the first in which the reverse migration becomes visible at scale.

An enrollment file records that a member left. It does not record whether they landed anywhere.

What to do, by seat

IF YOU ARE RECEIVING DISPLACED MEMBERS

Start from the constraint. Bids for 2027 were filed June 1, 2026. Premium, benefit design, and cost sharing are locked. Nothing you learn between now and January changes 2027 revenue. Every lever still available to you is operational.

Build the prospective encounter strategy now. First ninety days, provider engagement contracted before January, outreach designed to produce a qualifying visit rather than a chart. Unlinked chart review no longer carries a diagnosis on its own.
Treat the cohort as selected risk in care management design rather than in pricing, because pricing is closed. These are members a carrier with full claims history examined and declined. Staffing, outreach intensity, and care management capacity are the variables you still control.
Carry what you learn into the 2028 bid. That is the first cycle in which any of this can actually be priced, and the assumptions have to be built from the cohort you are about to receive rather than from the book you had.

IF YOU ARE THE PLAN EXITING

Build the selective retention strategy before the Annual Notice of Change goes out, not after. Preferred members, preferred service areas, and the product you intend to move them into all have to be defined in advance. This is the second half of the same decision that produced the exit, not a separate recovery effort.
Selective retention strategies are a marketing and steering question before they are a growth question. Transition communications, broker instructions, and plan design will all be read against those rules. Exits are the most scrutinized member communication a plan sends.
Whatever retention you achieved last cycle, assume it has to be earned again against competitors running the same play in the same counties.

IF YOU ARE A HEALTH SYSTEM OR MEDICAL GROUP

Map the exits against your service area now. Volume, payer mix, and referral patterns move before you feel them.
Separate the three decisions in your planning. Declining a payer’s contract, sponsoring a health plan, and serving the population are independent choices. A system can walk away from one carrier’s terms, or from owning insurance risk entirely, and remain the largest provider of Medicare care in its market.
Model the shift toward Traditional Medicare where options thin out. It pays differently, authorizes differently, and carries no supplemental benefit structure, and your bad debt profile changes with it.

IF YOU ARE AN INVESTOR

Enrollment growth in this cycle is not automatically a good sign. Ask where the members came from and why they were available.
Ask any portfolio plan how it intends to document acuity on an inherited population now that unlinked chart review no longer counts.

The calendar between now and January

DATE	WHAT HAPPENS
September 30	Annual Notice of Change must reach members. Displacement becomes real to the person holding the card.
October 15	Annual Enrollment opens. Broker behavior in the first two weeks indicates where displaced volume is heading.
December 7	Annual Enrollment closes and you know who you have. Begin compiling new member profiles and building the onboarding strategy now, not in January.
January 1	Coverage begins and the ninety day encounter window opens. Plans without a prospective model have already lost it.

The line worth holding onto

The industry spent fifteen years treating Medicare Advantage membership as the scoreboard. In a single bid cycle, every major carrier decided it was not. Revenue held flat and earnings rose fifty five percent at the largest player in the country while it shed nearly a million members.

The cause is not strategy. The cost curve moved up as the premium factors moved down, and no insurance carrier can actuarially mitigate both at once. CMS improved the rate substantially over its own proposal and the carriers cut anyway. Seven percent of surveyed plans call this business profitable. The other ninety three percent are doing subtraction, and the subtraction lands on counties, on products, on networks, and eventually on a person who opens a letter in September and learns their plan is no longer available where they live.

Sources

Carrier disclosures. Humana second quarter 2026 earnings call and results release, July 29, 2026; UnitedHealth Group second quarter 2026 results and Form 8-K, July 16, 2026; Molina Healthcare fourth quarter and full year 2025 results, February 5, 2026, and 2025 Form 10-K; Becker’s Payer Issues; Healthcare Dive; Fierce Healthcare; Modern Healthcare; Forbes; CNBC.
Rate environment and payment policy. CMS calendar year 2027 Advance Notice of January 26, 2026 and final Rate Announcement of April 6, 2026, including the finalized exclusion of unlinked chart review record diagnoses and the decision not to finalize an updated risk adjustment model; CMS fact sheet; Crowell and Moring; American Hospital Association; Georgetown Medicare Policy Initiative.
Risk model mechanics. CMS-HCC V28 model documentation and phase in schedule; published actuarial analysis of full weight V28 risk score compression and Trust Fund impact.
Part D and pharmacy economics. Milliman analysis of Inflation Reduction Act changes to Part D plan liability; KFF snapshot of the Medicare Part D prescription drug benefit; MedPAC status report on Part D.
Market context. HealthScape Advisors Medicare Advantage health plan outlook for 2027, based on a January 2026 survey of leaders at more than thirty five health plans; KFF analysis of Medicare Advantage options following 2025 plan terminations; Payer Perspectives on the 2027 bid submissions; Becker’s Hospital Review running list of health systems dropping or narrowing Medicare Advantage contracts.

Figures reflect company disclosures and published analysis available at the time of writing. Plan year 2027 service areas and benefits are not final, and additional exits are expected. Members should consult the Annual Notice of Change and official plan materials. This article is analysis provided for general informational purposes and is not legal, actuarial, investment, or benefits advice.

ALSO IN THE STANDARD

Article 11. UnitedHealth Group: the Hemsley approach. A turnaround built on subtraction.

Article 12. Centene’s turnaround, powered by Sarah London.

Article 13. The Health Plan AI Maturity Model survey and findings.

HLTHWORKS builds artificial intelligence governance, scalability, sustainability, oversight, audit readiness, and a mature enterprise solution for medical groups, health systems, health plans, payers, and the vendors, plus investors serving them.