CMS 2027 Medicare Advantage Rate Reversal: What 2.48% Means for Plan Actuaries

On April 6, 2026, CMS released the final CY 2027 Medicare Advantage and Part D Rate Announcement at a net average payment increase of 2.48%, worth over $13 billion relative to CY 2026, or 4.98% including estimated risk score trend.

The January Advance Notice had projected 0.09%, roughly $700 million. Almost all of the 239 basis point swing comes from one decision, and it is not a data refresh.

Key Takeaways

• 220 of the 239 basis points come from CMS declining to recalibrate the CMS-HCC risk model, which turned a proposed 3.32% payment reduction into a 1.12% normalization drag.
• The retained model is calibrated on 2018 diagnoses and 2019 expenditures, a five-year gap that now spans post-COVID utilization normalization, GLP-1 and oncology cost growth, and changed coding density.
• The unlinked chart review exclusion finalized at 1.53% rather than the proposed 1.78%, after CMS added an exception preserving CRR diagnoses for members switching between MA organizations.
• On 100,000 members at a $1,200 monthly benchmark, the 2.39 point swing is roughly $34 million of additional annual revenue to allocate before the June 1, 2026 bid deadline.
• Dual-focused books may see 0.8% to 1.5% rather than the 2.48% headline, because V28 moves HCC weight away from conditions concentrated in dual populations and FFS growth runs below average in their counties.

What Moved Between January and April

Three components account for essentially the whole swing, and they are unequal in size by an order of magnitude.

The effective growth rate contributed 36 basis points, reflecting Part A and Part B per-capita spending data through Q4 2025: stronger inpatient utilization in the winter respiratory season and continued Part B drug spending growth in oncology and metabolic therapies.

The chart review switching exception added 25 basis points by narrowing the diagnosis exclusion's scope. Unlinked chart review records are diagnosis codes submitted without an associated beneficiary encounter, typically from retrospective health risk assessments and chart audits. CMS kept the exclusion but will still count CRR diagnoses for beneficiaries moving from one MA organization to another, on continuity-of-care grounds: a receiving plan has no prior-year encounter data through its own network, and a payment cliff there would discourage enrolling members with complex conditions.

The risk model decision contributed roughly 220 basis points on its own. Georgetown's Medicare Policy Initiative analysis puts the avoided reduction at 3.32%, replaced by a 1.12% drag from the normalization factor alone. Georgetown documents an average upward advance-notice-to-final revision of 1.26 percentage points over eleven years; the 2.39 point CY 2027 swing sits well outside that.

A Model That Still Prices 2019

The mechanism worth following is not the size of the increase but which model the risk scores come out of, because that determines who inside the aggregate gets it.

CMS proposed recalibrating V28 using 2023 diagnosis and 2024 expenditure data from the fee-for-service population. It kept the existing calibration base: 2018 diagnoses and 2019 expenditures. The stated reason was to allow the market time to adjust after V28 reached full 100% weighting in CY 2026, ending the transition from V24. Layering a data recalibration onto a model plans had run at full weight for one year would have compounded the uncertainty in this bid cycle.

The consequence for a plan actuary is distributional rather than aggregate. Normalization calibrates to total FFS spending, so retaining the older model does not change what CMS pays in total. It changes the relative prediction across HCC categories, and therefore across plans by population composition. A book weighted toward oncology utilization, metabolic disease or behavioral health, where spending relationships have moved most since 2019, will be predicted by a model fitted before that movement. Whether that runs rich or lean is a plan-specific question the 2.48% headline does not answer.

That is also why the same rate year lands unevenly across market positions. Broad-market HMOs with encounter-based coding in counties running above-average FFS growth capture close to the full increase. Dual-focused plans may land at 0.8% to 1.5%: V28 redistributes HCC weight away from conditions prevalent in dual populations, dual enrollees concentrate in counties where FFS growth runs below the national average, and look-alike D-SNP and integrated care requirements add compliance cost against the revenue.

The bid clock makes the difference operational. Plans that built preliminary models against the 0.09% baseline now have 2.39 additional points of per-member-per-month revenue and roughly eight weeks to place it. At 100,000 members and a $1,200 monthly benchmark that is about $34 million a year, allocated across benefit richness, premium, provider rates and margin, with Oliver Wyman noting heightened documentation standards on SSBCI offerings.

Investors read it as revenue: UnitedHealth Group, Humana and CVS Health each rose more than 8% the session after the release, per STAT News. The starting positions differ, though. UnitedHealthcare cut enrollment roughly 9% from its October 2025 peak through county exits and benefit trims, to about 9.4 million members, while Humana expanded through 2026.

Two Adjustments That Do Not Go Away

The relief is real and both of the things it defers keep accruing.

The chart review exclusion is the nearer one, and its plan-level distribution is skewed rather than average. A plan capturing most diagnoses through primary care, specialist, emergency and inpatient encounters may lose 0.5% or less of risk-adjusted revenue. A plan where 5% to 15% of risk-adjusted revenue derives from CRR-sourced diagnoses loses proportionally more, with the range running from near zero to 3% or more for the most CRR-dependent organizations.

Crowell & Moring notes CMS expects the impact to be greater for organizations that rely heavily on unlinked chart review, and MedPAC's March 2026 report documents how far that reliance has grown. CMS also finalized exclusion of audio-only encounter diagnoses, and aligned both rules across Part D alongside the segmented RxHCC models for MA-PD plans and standalone PDPs that Avalere flagged as materially altering Part D risk adjustment.

Replacing CRR revenue with prospective coding is not a free substitution. It requires providers to document every active condition at each visit, which means EHR template changes, workflow redesign and often coding quality bonuses written into provider contracts. Those costs sit in the medical loss ratio, so a plan that fully replaces the lost diagnoses still books a worse ratio than before the rule.

The recalibration is the further one. Every year of deferral adds a year of divergence between 2018/2019 spending relationships and current ones, and Georgetown's warning is that "the longer the agency waits to update the V28 model, the larger the year-over-year bottom line repercussions." If CMS proposes it in the CY 2028 Advance Notice, the accumulated adjustment arrives in a single year and is larger than the 3.32% proposed for CY 2027, because it carries one more year of drift. The 2.48% is a deferral being recognized as revenue, not a permanent reset of the baseline.