The 2027 Payment Notice final rule looks like an operations rule until the coefficient tables start moving. CMS finalized the rule in May 2026, made it effective July 20, 2026, and placed the risk adjustment recalibration on a new 2021, 2022, and 2023 enrollee-level EDGE data window. The user fee headline is easier to explain, especially the drop to a 1.9% FFE user fee and a $0.18 PMPM risk adjustment user fee, but the pricing work sits in the quieter sentence: the 2027 HHS-HCC coefficients are equally weighted blends of three separately solved models, trended forward to the 2027 benefit year.

That sentence belongs in every ACA rate filing workpaper for 2027. From reading annual Payment Notice risk adjustment sections over multiple filing cycles, the small coefficient and data-year changes often matter more to pricing teams than the headline policy provisions. A carrier can absorb a user-fee reduction through a denominator update; it cannot safely absorb a transfer-position shift without rerunning the member-level model, rechecking morbidity selection, and reconciling the actuarial memorandum to the new coefficients. The recalibration is technical. The rate effect is not.

CMS says the 2027 coefficients are similar to 2026 because two of the three EDGE years overlap. That is true at the aggregate level, and it is the reason CMS declined to publish a state transfer simulation. The pricing question is narrower: how much does the 2023 data year, replacing 2020 in the blend, change an issuer's own expected plan liability risk score, metal-level transfer estimate, and enrollment-duration adjustment under the actual 2027 mix it is pricing? A statewide median does not answer that question.

The Recalibration Method

The HHS risk adjustment model is not being redesigned for 2027. CMS finalized annual recalibration using enrollee-level EDGE data from 2021, 2022, and 2023, continuing the approach of solving separate models by data year and blending the resulting coefficients. The costs in each data year are trended forward to the 2027 benefit year, and the final adult, child, and infant coefficient tables reflect that trended, averaged result.

The distinction between a model redesign and a recalibration matters for pricing. A redesign changes the structural relationship between diagnoses, demographics, and model output. A recalibration leaves the structure intact but changes the weights assigned to the variables. For most ACA issuers, the operational burden of the 2027 rule is therefore not rebuilding the entire risk adjustment infrastructure; it is rerunning the same infrastructure with updated coefficients, new trend assumptions inside the expected claims projection, and new enrollment-duration factors that may behave differently by tenure band.

CMS also reviewed the 2023 EDGE data before incorporating it and stated that it did not identify systematic anomalies. That finding is important because 2023 is the new data year in the blend. The 2026 model used 2020, 2021, and 2022. The 2027 model drops 2020 and adds 2023. For an issuer with stable morbidity and stable coding patterns, the overlapping 2021 and 2022 years dampen volatility. For an issuer that changed footprint, plan mix, broker channel, metal distribution, or coding operations after 2020, the replacement of 2020 with 2023 may carry a larger issuer-specific effect than the national coefficient comparison suggests.

That is the first pricing task: isolate the effect of the coefficient update from the effect of the carrier's own enrollment forecast. A clean support file should run the current 2027 membership projection through the 2026 coefficients and the 2027 coefficients, holding everything else constant, then separately apply the expected 2027 morbidity and metal mix. Without that bridge, the actuarial memorandum cannot tell a regulator whether the transfer change came from CMS recalibration, member mix, coding intensity, or a planned portfolio shift.

What the 2021-2023 EDGE Window Changes

The 2021-2023 EDGE window is the first recalibration blend to sit fully on post-2020 experience while still reflecting the enrollment and utilization distortions that followed the pandemic coverage period. CMS addressed comments arguing that the 2021 and 2022 years remained affected by pandemic utilization and special enrollment period behavior, but the agency concluded those years were sufficiently similar to prior EDGE experience to retain them. CMS also found no anomaly in 2023 that justified exclusion.

For pricing actuaries, the issue is not whether CMS had enough evidence to finalize the blend. The issue is what the blend implies for 2027 rate adequacy when the market being priced may not resemble any one of those years. The 2021, 2022, and 2023 EDGE years include expanded premium tax credit effects, an enlarged low-income enrollee population, higher special enrollment period activity, and care patterns that were still normalizing after 2020. The 2027 market may face subsidy expiration pressure, verification tightening, and renewed churn among low-income members. The historical window and the forecast window point in different directions.

That mismatch creates a familiar ACA pricing problem: the risk adjustment model measures relative plan liability under past data, while rate development prices absolute expected claims under a future population. If healthier partial-year enrollees entered during the low-income SEP period and then leave or churn under 2027 eligibility and subsidy changes, the issuer's expected transfer position can change even when the coefficient table is stable. The coefficient is only one half of the calculation. The exposed population is the other half.

The practical response is scenario testing at the cell level. A carrier should not merely run a best estimate risk score and transfer value. It should run a baseline mix, a subsidy-expiration attrition mix, a higher-morbidity retention mix, and a coding-intensity sensitivity tied to diagnosis capture operations. The output should show plan liability risk score, induced demand adjustment, geographic cost factor, allowable rating factor, transfer PMPM, and the residual premium change required to keep the target loss ratio intact. These are ordinary rate filing exhibits, but the 2027 recalibration makes them more visible.

Enrollment-Duration Factors

The most actuarially interesting passage in the final rule is not the broad statement that coefficients are similar. It is CMS's comparison of year-to-year coefficient movement. CMS reported that 2027 age-sex coefficients varied from 2026 by a median of plus or minus 3.7%, compared with a 9.1% median across prior recalibrations. Adult enrollment-duration factors moved by a median of plus or minus 10.3%, compared with 9.8% since the HCC-contingent enrollment-duration factors were introduced for the 2023 benefit year. The age-sex change is quiet. The enrollment-duration change is where partial-year membership enters the pricing conversation.

Enrollment-duration factors are a response to a known ACA issue: partial-year enrollees do not always have the same risk-score-to-cost relationship as full-year enrollees. Some partial-year members enroll because they have immediate care needs. Others enroll because of an income, job, household, or subsidy change and may be lower morbidity than the full-year population. The coefficient assigned to an enrollment-duration category is therefore doing more than a mechanical tenure adjustment. It is compressing selection, diagnosis capture, and utilization timing into a model factor.

CMS received a comment arguing that special enrollment periods from 2021 through 2023, especially the SEP for enrollees at or below 150% of the federal poverty level, brought lower-risk individuals into the market as partial-year enrollees and deflated the enrollment-duration factors. CMS finalized the coefficients anyway, citing its review of the data and the absence of anomalies that would justify exclusion. The policy answer is settled for 2027. The pricing answer is still issuer-specific.

An issuer with heavy low-income SEP volume should test whether its own partial-year experience resembles the EDGE years now embedded in the coefficients. If partial-year members in its 2024 and 2025 experience were healthier than the full-year book, a lower enrollment-duration factor may make sense. If its partial-year members were higher morbidity because the carrier attracts members with immediate treatment needs, the national coefficient may understate its plan liability unless the morbidity assumption compensates elsewhere. The actuarial work is not to argue with the model. It is to understand where the issuer's actual risk pool diverges from the model's average.

This is also where pricing and risk adjustment operations meet. Diagnosis capture is weaker for short-duration enrollees because the carrier has fewer months of claims and encounter history. A short-duration member who has diabetes, chronic kidney disease, or major depression may generate claims before the diagnosis signal is fully captured in EDGE submissions. If a carrier improves concurrent coding and encounter completeness between 2023 and 2027, the same morbidity may produce a different risk score. That is a coding-intensity assumption, not a morbidity trend, and it should be documented separately.

No State Transfer Simulation

CMS declined a request to publish state payment transfer simulations for the 2027 recalibration. The agency's reasoning was direct: it has historically released such simulations when proposing or considering major model changes, and for 2027 it finalized annual recalibration only. CMS also noted that the 2027 and 2026 models share two EDGE years, that the coefficients are generally similar, and that issuers can use prior transfers, proposed coefficient tables, and the HHS-developed risk adjustment algorithm software to assess impact.

That leaves carriers with more responsibility, not less. A state transfer simulation gives an external benchmark for direction and magnitude. Without it, the issuer's internal model becomes the primary support for any transfer assumption that materially affects rate development. State reviewers will still ask why projected transfers changed from 2026 to 2027. A response that points to the final rule without showing the issuer's own coefficient bridge will be thin.

The absence of a CMS simulation also changes peer comparison. In prior model-change cycles, a statewide simulation could help issuers separate marketwide movement from carrier-specific movement. For 2027, carriers need to build their own market view from public rate filings, prior-year transfer reports, and competitor premium behavior. That is less precise, but it is still necessary because risk adjustment is a zero-sum transfer within the applicable state market risk pool. One issuer's expected payment improvement is another issuer's charge deterioration.

A useful filing support exhibit separates three values: the modeled transfer under 2026 coefficients and 2027 projected membership, the modeled transfer under 2027 coefficients and the same projected membership, and the final priced transfer under 2027 coefficients and final expected membership. The first bridge measures the pure coefficient update. The second measures the combined effect of coefficient update and business-plan mix. The third ties to the premium rate. That structure gives reviewers a clean way to see what changed.

User Fees and Operating Cost

The risk adjustment recalibration does not stand alone in the final rule. CMS finalized a 2027 FFE user fee of 1.9% of monthly premiums and an SBE-FP user fee of 1.5%, both lower than the 2026 rates. CMS also finalized a risk adjustment user fee of $0.18 per member per month for the HHS-operated program, down from the 2026 rate. These are small numbers in isolation, but ACA pricing is built from small loads that compound through the gross premium formula.

The 1.9% FFE user fee is a percentage-of-premium item, so it sits in the premium denominator. The $0.18 risk adjustment user fee is a PMPM item, so it sits with administrative expense or non-benefit expense. The first affects the grossing-up algebra. The second affects the numerator. Mixing those two treatments is a common filing error because both are called user fees, but they do not enter the rate calculation in the same place.

The existing actuary.info analysis of the CMS 2027 NBPP user fee cut and ACA rate filing mechanics walks through the premium formula. The recalibration article adds a second layer: once the user fee reduction lowers the administrative load, the expected transfer assumption can still move the needed premium in the opposite direction. A carrier with an adverse recalibration or mix effect may file a rate increase even after the user fee reduction. That is not inconsistent. It is how the ACA premium build works.

For 2027, pricing teams should keep the risk adjustment user fee separate from the expected transfer PMPM in exhibits and internal review decks. One is the cost of operating the HHS program. The other is the carrier's expected payment or charge under the state market transfer formula. Combining them obscures the source of rate movement and makes variance analysis harder after final transfers are released.

Medical Trend Interaction

PwC's 2027 Behind the Numbers outlook projects 9.0% group medical cost trend and 8.5% individual market trend, with the 2026 individual trend restated to 8.5%. The individual market figure is especially relevant because it sits next to, not inside, the risk adjustment recalibration. Risk adjustment changes relative payment between issuers. Trend changes the absolute claim cost that all issuers must fund.

The interaction matters in two ways. First, EDGE data costs are trended to the 2027 benefit year for coefficient development, but a carrier's rate filing trend selection may differ from CMS's model calibration assumptions. Second, high medical trend magnifies the dollar effect of transfer PMPM changes. A 2% movement in expected transfer on a $450 allowed-claims base is smaller than the same percentage movement on a $520 base. The coefficient movement may look modest as a factor change and still be meaningful as a premium PMPM.

High trend also reduces the margin for loose morbidity assumptions. If an issuer is already pricing 8.5% individual market trend, GLP-1 pharmacy pressure, provider contract increases, and behavioral health utilization, an understated transfer charge can push the rate below adequacy faster than it would in a lower-trend year. The 2027 filing season therefore needs combined scenario testing: medical trend, risk adjustment transfer, subsidy attrition, and enrollment duration should be stressed together, not as independent sensitivities that never meet in the same case.

The most useful scenario is not the most extreme one. It is a plausible adverse case: individual medical trend 100 basis points above best estimate, healthier subsidized members disproportionately lapsing, partial-year members retaining higher morbidity than the EDGE-derived factor implies, and transfer charges moving 2 to 4 PMPM against the issuer. That case tells management whether the filed rate can survive normal forecasting error. It also gives the appointed actuary a defensible basis for the risk margin.

Rate Filing Support

The actuarial memorandum for a 2027 ACA filing should not treat recalibration as a footnote. At minimum, the support should identify the coefficient vintage used, the date the coefficients were loaded into the risk adjustment model, the membership data used for simulations, and the reconciliation from prior-year transfer estimates to the 2027 priced assumption. A reviewer should be able to trace the number from model output to URRT input without asking for a separate explanation.

Patterns we've seen in recent ACA filing cycles suggest three pressure points. First, the best-estimate transfer often changes late when membership projections are updated after preliminary rate work. Second, the actuarial memo sometimes describes morbidity and risk adjustment in separate sections even though the same enrollment shift drives both. Third, carriers occasionally disclose the expected transfer amount but not the sensitivity around it. The 2027 recalibration makes all three weaknesses more visible.

For issuers with multistate footprints, the support should also distinguish state-level market effects from issuer-level coefficient effects. A carrier may have a favorable recalibration in one state because its morbidity is concentrated in HCCs that gained coefficient value, and an unfavorable effect in another state because its growth came through low-risk partial-year enrollment. A national average transfer bridge is not enough when rates are filed state by state.

Documentation should also connect the model to ASOP-style data reliance and model governance, even when the filing does not cite a particular standard at length. The HHS-HCC software, internal diagnosis mapping, EDGE extracts, and enrollment-duration logic are a model chain. The workpaper should show data source controls, version control for coefficient tables, reasonableness checks against historical transfers, and a management review sign-off for any manual adjustment. If the transfer assumption is one of the largest non-claim items in the rate build, it deserves the same control discipline as trend selection.

Transfer Scenarios for 2027 Pricing

The most useful recalibration exhibit is a compact transfer scenario grid built from actual projected membership, not an illustrative population. Start with the filed 2026 transfer estimate, restate it on the 2027 rating area, plan, and metal distribution, then isolate three changes: the new coefficient table, the new enrollment-duration distribution, and the 2027 morbidity selection assumption. That ordering prevents the common error of attributing all movement to CMS when part of the change comes from the issuer's own business plan.

A carrier entering new counties should run the bridge separately for incumbent and expansion membership. Expansion membership often has weaker diagnosis history, different broker sourcing, and different metal selection than the renewal book. Combining the two can make the total transfer assumption look reasonable while hiding a large adverse transfer on the new block. The same logic applies to a carrier that is exiting bronze products or leaning into low-premium silver plans. Product strategy changes the risk adjustment denominator before medical trend ever touches the claim projection.

The bridge should also show the direction of transfer movement as a premium percentage and as PMPM. Regulators tend to read rate increases in percentage terms, while finance teams manage margin in PMPM terms. A $4 PMPM adverse transfer shift may look modest against a $600 premium, but it can consume a meaningful share of target margin if the issuer prices to a 2% to 3% underwriting gain. The rate filing should make that arithmetic visible rather than burying the transfer inside a net claims line.

For carriers using external risk adjustment vendors, the 2027 recalibration creates a vendor-control question. The actuarial team should confirm when the vendor loaded the final coefficients, whether the software applies the correct adult, child, and infant tables, and how the tool treats short-duration enrollees with incomplete diagnosis history. A late coefficient refresh can produce a different transfer estimate from the one used in the filed rates. That difference may be explainable, but it should not be discovered after state review questions arrive.

The final scenario should tie back to the pricing decision. If the best estimate assumes favorable selection from tighter SEP verification, the adverse scenario should remove that favorable assumption and show whether the filed rate still clears the minimum margin threshold. If the best estimate assumes enhanced diagnosis capture offsets a lower enrollment-duration coefficient, the adverse case should cap the coding improvement at a conservative level. The purpose is not to create an exhaustive stress-testing library. It is to prove that the rate is adequate under the most credible ways the 2027 recalibration can disappoint.

Why This Matters

The 2027 recalibration is a narrow regulatory action with a broad pricing footprint. It updates the coefficient blend, leaves the model structure intact, and gives CMS a defensible reason not to publish a state transfer simulation. It also lands in a year when individual market actuaries are already pricing subsidy uncertainty, elevated medical trend, GLP-1 credibility gaps, and tighter enrollment controls. The result is a filing environment where small model changes can produce meaningful rate support questions.

For ACA rate filing actuaries, the immediate work is practical. Rerun the member-level model with 2027 coefficients. Bridge the transfer movement. Test partial-year enrollment. Keep user fee mechanics separate from transfer assumptions. Reconcile the result to medical trend and morbidity scenarios. Then write the memorandum so a regulator can see each step without reconstructing the model from scratch.

Carriers that treat the 2027 Payment Notice as a compliance update will miss the pricing signal. The coefficient tables are not a policy press release; they are rate inputs. The actuaries who can explain how 2021-2023 EDGE data changes the expected transfer for their own projected 2027 membership will have the better filing, the clearer management conversation, and the earlier warning if rate adequacy starts to slip.

Further Reading

Sources