For years, MSSP ACOs operated under a straightforward ceiling on positive HCC adjustment: risk scores could not grow more than 3% above the baseline year. Documentation programs were calibrated to it. Coding targets were set against it.
For any ACO entering a new agreement period in PY2024 or later, that ceiling no longer applies in its prior form. CMS replaced the flat 3% cap with a segment-level demographic ceiling that moves with each ACO's own population dynamics.
A single ceiling applied uniformly across all enrolled populations. One target. Every segment, every ACO.
A dynamic, population-specific ceiling per segment. Moves with demographic acuity. Not a fixed percentage above baseline.
The direction of impact is not uniform. If an ACO's demographic acuity grew between baseline years and PY, the new ceiling is more generous than the flat 3% cap. If demographic acuity declined -- or if the ACO was already coding aggressively in the baseline period -- the effective ceiling tightened materially. Two ACOs can run the same documentation program and land on opposite sides of this change.
Modeled across 135 mainstream MSSP ACOs entering new agreement periods in PY2024. Five LTC-centric outliers excluded (composite flat benchmark above $2,000 PMPM) to preserve comparability across the mainstream program population.
Operating under a tighter effective ceiling than the flat 3% methodology would have produced. Benchmark suppressed relative to prior rules.
Operating under a more generous ceiling than before. Demographic acuity growth between baseline years and PY created additional headroom.
Cap was never binding either way. These ACOs were not coding close enough to the ceiling under either methodology for the formula change to matter.
The asymmetry is stark. Being hurt by the new methodology costs a median $2.9M annually. Being helped by it delivers a median $196K lift. For every dollar of benchmark value the new formula adds to helped ACOs, it removes roughly fifteen dollars from hurt ones.
VBC Contract Performance Intelligence — RAF Ceiling Analysis — PY2024 — N=135The hurt cohort skews large. 41% of hurt ACOs are Large (15K-40K lives) and 18% are Very Large (40K+ lives). The Very Large hurt ACOs face a median annual suppression of $23M -- at scale, the formula change is a material financial event. Helped ACOs are overwhelmingly smaller, and the benefit is modest relative to the harm on the other side.
No Very Large ACO benefited from the new methodology. Every Very Large ACO affected was hurt. Among the 39 hurt ACOs, 72% carry benchmarks above $150M annually -- these are not marginal programs. The revenue distribution of helped ACOs is concentrated in the $50M-$150M benchmark range (63%), with modest dollar impact relative to the suppression experienced on the other side.
Every hurt ACO in the cohort falls into the Mid or High HCC tier. Not a single Low HCC ACO was hurt by the methodology change. The helped cohort is overwhelmingly Mid HCC with a small number of Low HCC ACOs. This means the new formula disproportionately affects ACOs managing more medically complex populations -- the ones whose benchmark integrity arguably matters most.
The benchmark suppression produced by the new methodology doesn't just reduce settlement size -- in five cases it eliminated it entirely. These ACOs generated real gross savings. Their clinical teams performed. Under the flat 3% cap, each would have cleared their MSR threshold and earned a shared savings payment. Under the new demographic ceiling, the suppressed benchmark moved the threshold out of reach.
Generated positive gross savings but missed MSR threshold solely due to the new ceiling methodology. Would have cleared under the flat 3% cap.
Implied lost shared savings payment across the five ACOs. Clinical performance was sufficient. Contract mechanics eliminated the payout.
ACOs that missed MSR under both methodologies but whose gap to threshold was materially widened by the new ceiling -- pushed further from a payout they were already close to.
62% of hurt ACOs cleared MSR despite the suppression. This means the benchmark impact is not visible in settlement outcomes for the majority of affected ACOs -- it is silently reducing the size of a payout they still received. The suppression is real whether or not it crosses the MSR threshold. For the five that crossed it, the consequence was total.
VBC Contract Performance Intelligence — MSR Clearance Analysis — PY2024AGND -- Aged, Not Dual-Eligible -- represents the dominant attributed population for most MSSP ACOs, typically comprising 85% or more of total attributed lives. It is also the segment where the new methodology bites hardest.
ACOs whose AGND segment benchmark was suppressed by the new demographic ceiling relative to the flat 3% cap. AGND is the highest-lives segment for most ACOs -- suppression here compounds at scale.
Only 4 of 135 ACOs saw AGND benchmark lift from the new methodology. The formula change is structurally unfavorable for AGND at the population level.
For every ACO that benefited in AGND, six were suppressed. In the segment that matters most by volume, the new methodology is a net negative by a wide margin.
Why AGND is the highest-risk segment under the new formula: AGND populations tend to have stable demographic acuity with meaningful HCC coding history. ACOs that coded well in baseline years built a high BY3 HCC score -- which the new formula uses as the denominator in the ceiling calculation. A high BY3 HCC means a lower ceiling ratio, which means less room for HCC growth in PY before hitting the cap. Strong historical coding becomes a structural headwind under the new methodology.
All findings are derived from the CMS MSSP Public Use File for the most recent performance year. Coding intensity signals and peer cohorts referenced elsewhere in VBC CPI analysis are constructions of VBC Contract Performance Intelligence -- they are not published by CMS. The benchmark comparison methodology below is original analytical work.
Most recent CMS performance year. N=140 ACOs entering new agreement periods. 5 LTC-centric outliers excluded (composite flat benchmark above $2,000 PMPM). Clean analysis cohort: N=135.
Segment-level adjusted benchmark PMPM under the flat 3% methodology:Adj Bench = Unadj Bench × (min(HCC_PY, HCC_BY3 × 1.03) / HCC_BY3)
Composite benchmark computed as lives-weighted average across ESRD, DIS, AGDU, AGND.
The unadjusted benchmark PMPM reflects the ACO's historical expenditure base after trend update and regional adjustment -- the final step before risk adjustment is applied. This is the correct base for modeling RAF sensitivity: regional adjustment is included, risk adjustment is the variable of interest. The delta between the two ceiling methodologies isolates the ceiling effect on the risk adjustment step alone.
Segment-level adjusted benchmark PMPM under the demographic ceiling per 42 CFR 425.610:Ceiling = (DMGR_PY / DEMOG_HCC_BY3) × 1.03Adj Bench = Unadj Bench × (min(HCC_PY, Ceiling) / HCC_BY3)
Composite benchmark computed as lives-weighted average across all four segments.
Because regional adjustment is held constant in the unadjusted benchmark base across both scenarios, the delta between methodologies captures only the difference in how the two ceiling formulas constrain the HCC ratio. All other benchmark construction factors are identical between scenarios.
Delta PMPM = Composite Benchmark (Demo) minus Composite Benchmark (Flat).
Annual dollar impact = Delta PMPM × total attributed lives × 12 months.
Positive delta = helped. Negative delta = hurt. Zero = cap not binding under either methodology.
Because both scenarios use the same unadjusted benchmark base -- historical expenditure updated for trend and regional adjustment -- the delta is not contaminated by regional blending differences. It measures the ceiling constraint effect on risk adjustment and nothing else.
Counterfactual gross savings = actual gross savings minus benchmark delta (subtracting a negative delta increases savings). Counterfactual MSR threshold = counterfactual total benchmark × MSR rate. ACO classified as having lost MSR clearance if actual savings missed threshold but counterfactual savings cleared it.
Size buckets based on attributed lives: Small under 5K, Mid 5K-15K, Large 15K-40K, Very Large 40K+. Revenue proxy = composite benchmark PMPM × total lives × 12 months. HCC tier classifications are CMS-published. All other segmentation is derived by VBC Contract Performance Intelligence.
Independence disclosure: VBC Contract Performance Intelligence is not affiliated with any vendor, platform, or health plan. This analysis is not commissioned work. All data sourced from CMS public use files. Derived metrics -- including the benchmark comparison methodology, coding intensity signals, and peer cohorts referenced in other VBC CPI publications -- are constructions of VBC Contract Performance Intelligence and are not CMS-published figures.
The model is already built. The conversation can start with your segment-level ceiling and dollar impact on the table -- before your next performance year closes.