The CMS CY 2027 Rate Announcement, published April 7, 2026, locked in a 2.48% average effective payment increase for Medicare Advantage, adding roughly $13 billion above CY 2026 levels. The January 2026 Advance Notice had projected 0.09%, or about $700 million. A 239 basis point swing of that magnitude requires plan-specific reconciliation before a single bid assumption can be set: the national average is derived from industry-aggregate FFS trends and carries none of the county benchmark structures, applicable percentages, Star quality bonuses, risk score normalization factors, or coding adjustments that determine what a specific plan actually receives per member per month.
The Advance Notice to Final Bridge
Three components explain the 239 basis point move. CMS updated the effective growth rate from 4.97% to 5.33% after incorporating FFS expenditure data through Q4 2025, which showed stronger inpatient utilization and Part B drug spend than the mid-2025 data underlying the January Advance Notice (CMS Rate Announcement, April 2026). Georgetown's Center for Health Insurance Reforms documented that 9 of 11 rate announcements since 2013 revised upward from the advance notice, with an average revision of 1.26 percentage points (Georgetown CHIR, April 2026). The CY 2027 revision of 2.39 points is nearly double that historical average.
The larger swing came from a single policy decision. CMS deferred the proposed recalibration of the CMS-HCC V28 risk adjustment model, which would have updated the model's underlying coefficients from a 2018/2019 calibration dataset to one based on 2023 diagnosis data linked to 2024 expenditures. That recalibration, combined with updated normalization factors, carried a combined advance notice impact of -3.32%. CMS retained the existing model; the normalization factor alone now produces -1.12%, preserving approximately 2.20 percentage points of plan payment relative to the advance proposal (CMS Rate Announcement, April 2026). The third component, a 0.25 point relaxation of the diagnosis-source exclusion, reflects a finalized exception for beneficiaries switching between MA organizations.
| Component | Advance Notice (Jan 2026) | Final (Apr 2026) | Change |
|---|---|---|---|
| Effective growth rate | +4.97% | +5.33% | +0.36 pp |
| Risk model revision + normalization | −3.32% | −1.12% | +2.20 pp |
| Diagnosis-source exclusion | −1.78% | −1.53% | +0.25 pp |
| Star Ratings | −0.03% | −0.03% | 0 |
| Rebasing/re-pricing | TBD | −0.17% | N/A |
| Net average change | +0.09% | +2.48% | +2.39 pp |
County Benchmarks and the Applicable Percentage
The 2.48% average conceals a wide county-level distribution. CMS sets each county's benchmark by multiplying a county-specific FFS per capita cost estimate by an applicable percentage that varies with market penetration quartile: 95% in the highest-cost Quartile 4 counties, 100% in Quartile 3, 107.5% in Quartile 2, and 115% in the lowest-cost Quartile 1 counties (CMS MA Ratebook, CY 2027). The national FFS per capita spending for CY 2027 is estimated at $1,293.23, multiplied by each county's geographic adjustment index to produce the local base before the quartile percentage applies (CMS Rate Announcement, April 2026).
Two dynamics amplify or dampen the 5.33% effective growth rate at the county level. Counties where per capita FFS spending ran above the national trend through 2025 see larger dollar benchmark increases than the aggregate implies. Counties near quartile boundaries face the possibility of a quartile reclassification that changes both the base benchmark and the applicable percentage simultaneously, a combined effect invisible in the national average. A county moving from Quartile 3 to Quartile 2 gains a higher applicable percentage on top of a higher FFS base. The Star quality bonus adds a further 5% to the benchmark for plans rated 4.0 Stars or higher, plus an additional 5% in qualifying bonus counties where MA enrollment exceeds 25% of Medicare beneficiaries. Plans concentrated in high-penetration Sun Belt markets can capture both tiers on the same membership base.
The practical implication: a plan with a multi-county service area should run the CY 2027 ratebook at the county level before assuming the 5.33% growth rate applies uniformly. The dollar variation across counties in a typical mid-size MA service area can span 200 to 400 basis points around the national average, depending on the local FFS utilization mix and penetration quartile.
Risk Score Projection: Normalization and Coding Adjustments
The county benchmark sets the ceiling; the plan's projected risk score and the bid-to-benchmark ratio determine how much of that ceiling becomes actual revenue. Risk score projection for CY 2027 has three distinct adjustment layers, each with a different plan-specific profile.
The normalization factor reduces raw plan risk scores by 1.12%, applied uniformly across all MA plans (CMS Rate Announcement, April 2026). This factor adjusts for the observed tendency of MA coding to produce higher average risk scores than FFS coding patterns would generate for the same population. The plan-specific dollar impact scales with the starting risk score: a population coding at an average risk score of 1.20 before normalization arrives at approximately 1.187 post-adjustment, a 0.013-point reduction that, multiplied by the applicable benchmark, produces the PMPM revenue haircut.
The diagnosis-source exclusion operates differently. Diagnoses from chart reviews not linked to a specific beneficiary encounter are excluded from CY 2027 risk score calculations, producing a -0.25 point industry average change from the advance notice position (CMS Rate Announcement, April 2026). The plan-level exposure varies considerably: organizations that relied on retrospective unlinked chart review programs for HCC capture can face per-member revenue reductions in the 2% to 3% range on affected subpopulations, while plans that shifted to encounter-based coding programs see minimal impact. The finalized plan-switching exception offsets the haircut only for members who changed MA organizations between service periods.
The MA coding pattern adjustment, set at the statutory minimum of 5.90%, produces a -0.59% flat effective rate offset applied to all plans (CMS Rate Announcement, April 2026). Unlike the normalization factor or the diagnosis-source exclusion, this adjustment does not vary with the plan's coding completeness or population mix. It represents the CMS-mandated floor correction for the observed coding intensity differential between MA and FFS, and it has held at the 5.90% statutory minimum for nine consecutive years despite recurring MedPAC recommendations to increase it.
From Revenue PMPM to Bid Economics
Once a plan has projected county-level benchmarks and expected population risk scores, the bid mechanics translate those inputs into actual payment and supplemental benefit budget. The core relationship: CMS pays the plan its submitted bid amount multiplied by each enrolled member's individual risk score, plus a rebate when the bid falls below the applicable benchmark.
The rebate percentage depends on Stars (KFF MA Payment Primer, 2026): plans with contracts rated below 3.5 Stars retain 50% of the bid-to-benchmark spread as supplemental benefit dollars; contracts rated 3.5 to 4.0 Stars retain 65%; contracts rated 4.5 Stars or above retain 70%. A plan bidding $950 against a $1,000 applicable benchmark generates $32.50 PMPM in rebate dollars at 3.5 to 4.0 Stars ($50 spread multiplied by 65%), or $35.00 PMPM at 4.5 Stars or above ($50 multiplied by 70%). That $2.50 difference between Stars tiers may look narrow, but across 100,000 members it represents $3 million annually in supplemental benefit capacity or reduced member cost-sharing.
| Stars level | Rebate retention | Rebate on $50 spread (PMPM) | Annual impact at 100K members |
|---|---|---|---|
| Below 3.5 | 50% | $25.00 | $30M |
| 3.5 to 4.0 | 65% | $32.50 | $39M |
| 4.5 and above | 70% | $35.00 | $42M |
The rebate dollars fund supplemental benefits (dental, vision, hearing, fitness, reduced cost-sharing, over-the-counter allowances), Part D premium subsidies, and Part B buy-down. Plans allocate rebates across these uses through the standardized bid submission tools CMS provides; the split is a pricing decision that carries competitive consequences in AEP and compliance consequences in benefit design documentation. A plan that reallocates rebate dollars from OTC allowances to reduced Part B premium faces a different member retention profile than one that shifts in the other direction, and both face documentation requirements tied to the actuarial certification that the benefit package is actuarially equivalent.
Supplemental Benefit Compression Under Cost Trend Pressure
The bid-to-benchmark spread that generates rebate dollars depends on the plan's medical cost trend relative to benchmark growth. Multiple forecasters have placed health plan medical cost trends for 2027 in the 6% to 8% range. Against the 5.33% effective growth rate benchmark, a plan projecting 7% medical cost trend faces a 1.67 point cost-revenue gap. On an 85% medical loss ratio baseline, that gap compresses operating margin by approximately 1.42 points, reducing supplemental benefit capacity by a corresponding dollar amount or forcing a bid-to-benchmark ratio increase that cuts into available rebate.
The supplemental benefit trade-off is not symmetric across plan types. Plans with D-SNP and dual-eligible populations carry a compounding headwind: the V28 model phase-in continues to redistribute payment weight away from HCC categories common in dual populations, including certain chronic behavioral health, peripheral vascular disease, and some diabetes complications, producing a larger effective risk score reduction than the 1.12% normalization factor implies for these books. Combined with the diagnosis-source exclusion's disproportionate impact on retroactively coded chronic conditions common in dual populations, dual-focused plans often see a net revenue uplift from the 2.48% headline that, at the plan level, runs 0.8 to 1.5 percentage points below the average, even before accounting for the supplemental benefit obligations embedded in D-SNP integrated care requirements.
Stress Testing Coding Intensity Assumptions
The coding intensity assumption embedded in a plan's projected risk score is among the most consequential inputs in the revenue build. Two adjustments interact here, and conflating them produces revenue overstatement that persists through the bid submission into the actuarial certification.
The first adjustment is the CMS-mandated coding pattern factor at 5.90%, fixed and non-negotiable for CY 2027. The second is the plan's internal projection of its own coding completeness relative to prior experience, embedded in the risk score trend assumption. If a plan's HCC capture program was providing 3% of coded diagnoses through unlinked chart reviews, and the CY 2027 exclusion removes that source, the effective risk score trend carries a headwind that is additive to the 5.90% CMS adjustment. A revenue model that does not explicitly decompose the unlinked-review HCC contribution will overstate projected PMPM by the full amount of the excluded diagnoses, scaled by their average payment weight.
Stress tests should run at three scenarios: a base case where all HCC diagnoses survive the exclusion (encounter-linked coding, no unlinked reviews), a downside where 10% of current HCC capture was unlinked-review-dependent, and a bear case at 20%. The per-member dollar impact at each scenario, multiplied across the membership base, sets the floor for contingency reserves embedded in bid pricing. Plans that have not audited their HCC source attribution by encounter type should treat the bear case as the conservative estimate until the audit is complete.
Bid Documentation Tying Revenue to the Rate Announcement
CMS actuarial certification requirements expect each material revenue assumption in the bid submission to be supportable by reference to CMS publications, plan-specific experience, or recognized industry data. A revenue build that cites the "2.48% rate increase" as the basis for projected PMPM will not meet this standard, because the 2.48% is an industry average from which the plan has made multiple plan-specific adjustments, each of which requires its own documentation trail.
The documented revenue build should layer from the county ratebook upward: the applicable benchmark for each county in the service area, the Star quality bonus calculation with the expected QBP Stars level, the risk score projection with normalization factor, HCC coefficient source, coding intensity decomposition, and diagnosis-source exclusion estimate, and finally the bid-to-benchmark ratio and rebate calculation. Each layer should reference the specific section of the CMS Rate Announcement that governs the parameter, not the summary Fact Sheet. The certifying actuary is affirming that the projected revenue reflects the actual payment methodology, not a reasonable approximation of it.
Plans that built preliminary bids against the January Advance Notice assumptions need targeted revision, not ground-up rebuilds, but the revisions must be documented with the same granularity as the original assumptions. The 2.20 point release from the deferred risk model recalibration is the largest single change; its plan-specific impact depends on the plan's exposure to V28-sensitive HCC categories, which the bid documentation should quantify explicitly. The 2.48% headline is the starting calibration point. The path from that number to plan-specific bid revenue runs through county, Stars, risk score, coding source, and benefit design. Each step carries an actuarial assumption that the certifying actuary has signed off on.