Patient Protection and Affordable Care Act, HHS Notice of Benefit and Payment Parameters for 2027; and Basic Health Program

The 2027 benefit year HHS risk adjustment model factors resulting from the equally weighted (averaged) blended factors from separately solved models using the 2021, 2022, and 2023 benefit year enrollee- level EDGE data are shown in Tables 1 through 7. The HHS risk adjustment adult, child, and infant models are truncated to account for the high-cost risk pool payment parameters by removing 60 percent of costs above the $1 million threshold.\41\ Table 2 contains factors for each adult model, including the age-sex, HCC, RXC,\42\ RXC-HCC interaction, interacted HCC count, ACF, and enrollment duration coefficients. Table 3 contains the factors for each child model, including the age-sex, HCC, interacted HCC count, and ACF coefficients. Table 4 lists the HCCs selected for the interacted HCC count factors that would apply to the HHS risk adjustment adult and child models. Table 5 contains the factors for each HHS risk adjustment infant model. Tables 6 and 7 contain the HCCs included in the HHS risk adjustment infant models' maturity and severity categories, respectively.

\41\ As finalized in the 2020 Payment Notice (84 FR 17466 through 17468), we will maintain the high-cost risk pool parameters for the 2020 benefit year and beyond, unless amended through notice- and-comment rulemaking. We did not propose changes to the high-cost risk pool parameters for the 2027 benefit year. Therefore, we will maintain the $1 million threshold and 60 percent coinsurance rate for the 2027 benefit year.

\42\ In the 2026 Payment Notice, we incorrectly stated that for RXC eligibility (including medically administered injectable claims), a professional or outpatient medical claim does not need to have a risk adjustment eligible service code or bill type code. We subsequently updated the January 2025 version of the 2024 Benefit Year Risk Adjustment DIY software instructions (https://www.cms.gov/files/document/cy2024-diy-instructions-01072025.pdf) to reflect this erroneous statement. In the Final 2024 Risk Adjustment DIY software instructions updated April 9, 2025 (https://www.cms.gov/files/document/cy2024-diy-instructions-04092025.pdf), we corrected this error and clarified that the HCPCS-level file for RXC assignment can only be sourced from institutional inpatient and outpatient claims with RA-eligible bill types. We expect ACFs related to prescription drugs will be sourced using the same criteria as RXCs. We will announce changes to HHS risk adjustment claims filtering with regard to RXC or ACF eligibility in future guidance documents or notice and comment rulemakings, as appropriate.

We received several comments requesting additional changes to the HHS risk adjustment models that we did not consider or propose in the proposed rule. We respond to these comments below.

Comment: A commenter requested that HHS study the impact of calibrating the HHS risk adjustment models separately for the individual and small group markets due to differences in the characteristics of the enrollee population between the two markets.

Response: We did not propose and are not finalizing separate individual and small group market models. We acknowledge the commenter's request that HHS study the impact of calibrating the HHS risk adjustment models separately for the individual and small group markets. We have reviewed this potential approach and, through our analyses, have identified certain concerns with respect to recalibrating separate models for the individual and small group markets. First, we note that creating two separate risk adjustment models for the individual and small group markets for each of the age groups (adult, child, and infant) would significantly increase the complexity of the HHS-operated risk adjustment program by nearly doubling the number of HHS risk adjustment models.\43\ Furthermore, because both the small group and individual markets are subject to the same AV requirements, the only way to distinguish HHS risk adjustment models for the small group market from those for the individual market would be to separate the enrollee-level EDGE data used for HHS risk adjustment model recalibration into individual market and small group market datasets. Because the small group market is substantially smaller than the individual market,\44\ we have concerns that such a separation of datasets may result in sample size issues for models specific to the small group market, particularly for enrollees with high-cost and rare conditions.

\43\ Because there are no catastrophic plans in the small group market, separating the individual and small group markets would require the recalibration of 12 additional risk adjustment models, one for each metal level (except catastrophic) and age group (adult, child, infant). This would result in a total of 27 HHS risk adjustment models instead of the current 15.

\44\ For a State-level summary of enrollment by market, see, for example, CMS. (2025). Summary Report On Individual And Small Group Market Risk Adjustment Transfers For The 2024 Benefit Year, Appendix A. https://www.cms.gov/files/document/ra-report-by2024.pdf.

As an alternative to creating separate HHS risk adjustment models for the individual and small group markets, we previously considered \45\ adoption of enrollment duration factors by market. However, we did not find a meaningful distinction in relative costs between markets on average once we implemented the HCC-contingent enrollment duration factors presently included in the adult HHS risk adjustment models. Therefore, we determined it would not be necessary to introduce market- specific factors if HCC-contingent enrollment duration factors were implemented. Even though reasons for and patterns of partial-year enrollment differ by market, we concluded that the patterns most relevant for predicting cost (for example, how enrollment duration relates to costs conditional on the presence of HCCs) were the same for both markets. As such, whether we calibrate separate risk adjustment models or only specific HCCs to reflect individual and small group market-specific enrollment dynamics, we continue to believe that the factors for both markets would generally be very similar, meaning that such a separation would add little value to the models while adding additional complexity and raising sample size concerns in regard to the small group market.

\45\ See CMS. (2021). HHS-Operated Risk Adjustment Technical Paper on Possible Model Changes. https://www.cms.gov/files/document/2021-ra-technical-paper.pdf.

We also note that under the State payment transfer formula, we calculate HHS risk adjustment transfers separately for the individual and small group markets (except for in States with merged markets). As such, HHS risk adjustment transfers largely reflect differences in risk between individual and small group market plans separately, partially mitigating the potential for differences between the individual and small group markets to influence HHS risk adjustment transfers. Nevertheless, we intend to continue to analyze the differences in costs and utilization between the individual and small group markets to consider whether these types of changes would be necessary or appropriate in future benefit years.

Comment: A few commenters identified certain conditions and treatments that they believe are undercompensated in the risk adjustment models, including Graves' disease/hyperthyroidism and its related treatment Tepezza, autism spectrum disorder, and GLP-1 drugs. These commenters requested that HHS reconsider how these conditions and treatments and their associated costs are accounted for in the HHS risk adjustment models. Some commenters requested that HHS generally consider how to address emerging and expensive therapies such as cellular and gene therapies in the HHS-operated risk adjustment program. Another commenter noted a U.S. Senate report \46\

from the Committee on the Judiciary regarding some issuers' discretionary coding and other gaming mechanisms in the context of Medicare Advantage plans that may impact the transfers calculated in the HHS-operated risk adjustment program if similar behaviors are present in issuers' EDGE server submissions.

\46\ Grassley, C. (2026). How UnitedHealth Group Puts the Risk in Medicare Advantage Risk Adjustment Majority Staff Report. https://www.grassley.senate.gov/imo/media/doc/uhg_report_-_final.pdf.

Response: For the request to reconsider how Graves' disease/ hyperthyroidism and its related treatment Tepezza and their associated costs are accounted for in the HHS risk adjustment models, we note that we recently discussed the approach to the treatment of Tepezza and Graves' disease/hyperthyroidism in the HHS risk adjustment models in the 2025 Payment Notice (89 FR 26248 through 26249) and 2026 Payment Notice (90 FR 4445 through 4446), explaining that thyroid eye disease (thyrotoxicosis), the condition which Tepezza is approved to treat, is currently categorized in a condition category (Other Endocrine/ Metabolic/Nutritional Disorders) that is not a payment HCC in the HHS risk adjustment models. Further, all RXCs in the adult HHS risk adjustment models are associated with a payment HCC. We therefore generally have concerns about adding Tepezza to the HHS risk adjustment models at this time as it is currently not intended to treat or prevent the development of a condition included in a payment HCC. For these reasons, HHS did not propose and is not finalizing any changes for the treatment of Tepezza for thyroid eye disease in the 2027 benefit year HHS risk adjustment models. However, HHS intends to continue analysis of Graves' disease/hyperthyroidism and thyrotoxicosis and the use of Tepezza as more data becomes available and may consider potential changes to the treatment of this condition and drug in the HHS risk adjustment models for future benefit years if appropriate.

Regarding the request to modify how the HHS risk adjustment models account for autism spectrum disorder, GLP-1 drugs, and other emerging and expensive therapies such as cellular and gene therapies, we note that in comments on the 2026 Payment Notice (90 FR 4445), some commenters previously identified that it may be appropriate to address treatments associated with autism spectrum disorder (that is, adaptive behavior services), GLP-1 drugs for weight loss,\47\ and emerging and expensive therapies such as cellular and gene therapies using the ACF framework finalized in that rule. As stated in our response to comments about those conditions and treatments in that final rule, we will take these comments into consideration as we consider potential refinements to the HHS risk adjustment models in future benefit years. Although these treatments may be fair candidates for inclusion in the HHS risk adjustment models in the future, we have identified potential causes for concern with their inclusion. For example, we are concerned that there may be a selection issue in current enrollee-level EDGE data in that issuers may require enrollees to have considerable morbidity to be approved for the GLP-1 medications for weight loss. Thus, we continue to consider the utility of and concerns with adding model factors for GLP-1 drugs and weight loss to the HHS risk adjustment models.

\47\ GLP-1 drugs approved for the treatment of diabetes are presently included in RXC 7 (Anti-Diabetic Agents, Except Insulin and Metformin Only) to the extent that their positive predictive values (PPVs) with diabetes HCCs are high and do not indicate a large degree of over-prescribing. See, for example Table 10a of the 2025 Benefit Year risk Adjustment HHS-Developed Risk Adjustment Model Algorithm “Do It Yourself (DIY)” Software Technical Details, available at: https://www.cms.gov/marketplace/resources/regulations-guidance.

For the underprediction for autism spectrum disorder alleged by commenters and the potential remedy of adding adaptive behavior services for autism spectrum disorder and disorders with similar behavioral characteristics to the HHS risk adjustment models, we note that, unlike obesity/overweight, there is a payment HCC for autism spectrum disorder (HCC 102) presently included as a payment HCC in the HHS risk adjustment models. As such, some reflection of the risk associated with autism spectrum disorder is already included. We continue to explore whether adding a model factor for adaptive behavior services to the HHS risk adjustment models is appropriate. Finally, for emerging and expensive therapies such as cellular and gene therapies, we are limited from incorporating many of these treatments into the HHS risk adjustment models due primarily to inadequate data for developing appropriate model variables. Specifically, many of the gene therapies HHS reviewed while considering this issue were completely absent from the claims data for any enrollee in the currently available benefit years of enrollee-level EDGE data. Of those therapies that did appear, sample sizes were extremely low. Without cost data and sufficient sample size in the enrollee-level EDGE data, we are unable to estimate reliable coefficients for potential model factors associated with these therapies. Nevertheless, we reiterate that many of these drugs will be covered to some extent under the high-cost risk pool, which provides a coinsurance rate of 60 percent when an individual or small group market enrollees' claims exceed $1 million.\48\

\48\ See the 2018 Payment Notice (81 FR 94080 through 94082).

If we determine that model changes are necessary in regard to GLP-1 drugs, adaptive behavior services/treatments for autism spectrum disorder, or emerging and expensive therapies such as cellular and gene therapies, we would propose to make such changes through future notice- and-comment rulemaking.

In response to the comments raising discretionary coding and gaming concerns, we note that there are risk adjustment model specifications to mitigate the potential for upcoding. For example, we group subsets of payment HCCs into larger aggregate clusters, or HCC coefficient estimation groups.\49\ In these groups, the HCC estimates are constrained to be equal to each other, and each enrollee is only permitted to be credited with risk from any HCC coefficient estimate group once, regardless of the number of HCCs within the group appearing in their enrollee-level EDGE data. This approach serves multiple purposes, two of which are relevant to the commenters' concern. First, these HCC coefficient estimation groups limit diagnostic upcoding by severity within the HCC hierarchies to which they belong; and, second, they reduce additivity within disease categories (but not across disease categories) to decrease the sensitivity of the model to coding proliferation. Furthermore, as stated in the ten principles of risk adjustment \50\ and other HHS risk adjustment model development papers,\51\ in assessing whether to include specific HCCs in the HHS risk adjustment models, we assess whether HCCs are especially subject to discretionary diagnostic coding or

enhanced rates of diagnosis through population screening not motivated by improved quality of care and, if we determine this to be so, exclude the HCCs from the models. The potential for discretionary coding in the HHS risk adjustment models is an area that HHS is consistently monitoring and addressing as needed and will continue to monitor and address in the future as part of our ongoing efforts to continually improve the HHS risk adjustment models.\52\

\49\ See, for example, Kautter, J., et al. (2014). The HHS-HCC Risk Adjustment Model for Individual and Small Group Markets under the Affordable Care Act. Medicare & Medicaid Research Review (4)3. https://www.cms.gov/mmrr/Downloads/MMRR2014_004_03_a03.pdf. See also the 2014 Payment Notice (77 FR 73127 through 73129).

\50\ See the 2014 Payment Notice proposed rule (77 FR 73127 through 73130) and CMS. (2016). March 31, 2016, HHS-Operated Risk Adjustment Methodology Meeting Discussion Paper. https:// www.cms.gov/cciio/resources/forms-reports-and-other-resources/ downloads/ra-march-31-white-paper-032416.pdf.

\51\ See, for example, Kautter, J., Pope, G.,C., Ingber, M., et al. (2014). The HHS-HCC Risk Adjustment Model for Individual and Small Group Markets under the Affordable Care Act. Medicare & Medicaid Research Review, 4(3). https://www.cms.gov/mmrr/Downloads/MMRR2014_004_03_a03.pdf.

\52\ See, for example, CMS. (2019). Potential Updates to HHS- HCCs for the HHS-operated Risk Adjustment Program, Section 2.3. https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance/Downloads/Potential-Updates-to-HHS-HCCs-HHS-operated-Risk-Adjustment-Program.pdf. Note that “HCC Group constraints” is synonymous with “HCC coefficient estimation groups.”

As we consider potential refinements to the HHS risk adjustment models in the future, we will also continue to monitor the specific conditions and treatments identified by commenters, along with the structure of related model factors (including those that may be subject to discretionary coding and related gaming concerns), and the impact of recent interacted HCC counts and HCC-contingent enrollment duration model specification updates finalized in the 2023 Payment Notice (87 FR 27224 through 27230). As always, as part of our ongoing efforts to continually improve the precision of the HHS risk adjustment models, if we were to pursue changes to the risk adjustment models in the future, we would seek input from interested parties through notice-and-comment rulemaking or other appropriate vehicles.

Comment: One commenter requested that HHS disburse and collect risk adjustment transfer payments and charges over a longer period of time rather than in annual lump-sum transfers. The commenter suggested that installment payment arrangements and quarterly provisional calculation of transfers, among other mechanisms, would improve financial predictability for participating issuers, particularly in smaller or emerging markets.

Response: We did not propose modifying the timeline to collect and pay risk adjustment transfers and are not finalizing any such changes at this time. As discussed in the 2026 Payment Notice (89 FR 82308), unlike Medicare Advantage's risk adjustment program, under which CMS make risk-adjusted monthly payments to Medicare Advantage organizations during the coverage year (in advance of each month of coverage) using interim risk scores and then conducts a reconciliation to update risk scores after the final deadline for submission of all risk adjustment data, the HHS-operated risk adjustment program for the individual, small group, and merged markets uses a final data submission deadline 4 months after the end of the benefit year and calculates issuers' PLRSs and the State payment transfer amounts 2 months after that, resulting in State payment transfers being made 8 to 10 months after the end of the benefit year. HHS typically announces State payment transfer amounts no later than June 30 of the year following the benefit year, begins to collect charges in August of the year following the benefit year, and begins to make payments to issuers in the fall of the year following the applicable benefit year. In the 2026 Payment Notice, we solicited comments on the impact of the existing timeline for collection and disbursement of State payment transfers, including the impact of the “time value of money” on issuers' assessment of actuarial risk and the incentives for adverse selection, and what possible solutions or mitigating steps we should consider to address the impact of the time value of money on the HHS-operated risk adjustment program in future rulemaking. We continue to consider that feedback and will take this comment and those previous comments into consideration for future rulemaking, as applicable.

Comment: One commenter opposed risk adjustment in general on the basis of a belief that risk adjustment and the assignment of risk scores to enrollees based on health conditions is discriminatory.

Response: We do not agree that the use of factors based on enrollees' age, sex, and health conditions or utilization of services and treatments in risk adjustment is inappropriate or discriminatory. Consistent with section 1343 of the Affordable Care Act, the HHS- operated risk adjustment program reduces the incentives for issuers to avoid higher-than-average risk enrollees, such as those with chronic conditions, by using charges collected from issuers that attract lower- than-average risk enrollees to provide payments to health insurance issuers that attract higher-than-average risk enrollees. The Affordable Care Act limits issuers' ability to establish or charge premiums on the basis of age and prohibits issuers from doing so on the basis of sex or any individual health characteristic other than tobacco use.\53\ However, the cost of care for and actuarial risk of enrollees is, in part, correlated with their age, sex, health conditions (or severity thereof), and likelihood to utilize services and treatments. As such, without the inclusion of factors related to age, sex, health conditions, and use of services and treatments in the HHS risk adjustment models, some issuers would be incentivized to design plans that are less attractive to potential enrollees whose age-sex category, health conditions, or use of services and treatments is predicted to create a higher liability for the issuer. The various factors in the HHS risk adjustment models help alleviate this incentive by ensuring that the actuarial risk of an issuer's enrollee population in a State market risk pool, including issuers that enroll a higher-than-average proportion of enrollees who fall into a high-cost age-sex category or are likely utilizers of high-cost services, is appropriately assessed as part of the calculations under the State payment transfer formula. The use of factors associated with age, sex, health conditions, and the use of services and treatments in the HHS risk adjustment models is therefore necessary, appropriate, and helps reduce the likelihood that discrimination based on any of these factors will occur for health insurance coverage issued or renewed in the individual and small group (including merged) markets.

\53\ See section 2701 of the Public Health Service Act (42 U.S.C. 300gg) as amended by section 1201 of the Affordable Care Act. See also the Market Rules and Rate Review final rule (78 FR 13406, 13411-13).

Comment: A commenter stated that the CSR adjustment factors for Massachusetts' ConnectorCare Plan Type 1 and Type 2 are too low. A few other commenters stated concerns regarding footnote 27 in the proposed rule (91 FR 6304), which discussed the State-specific CSR adjustment factors for Arkansas' 94 percent AV Medicaid-expansion plans and Arkansas' 100 percent AV Medicaid-expansion plans. These commenters asserted that the 1.46 CSR adjustment factor--which HHS erroneously described as being applicable to Arkansas 100 percent AV Medicaid- expansion plans in footnote 79 of the 2025 Payment Notice (89 FR 26253)--was in fact more appropriate for these plans than the value we clarified in footnote 27 of the proposed rule (91 FR 6304) as being the correct value for these plans (that is, a CSR adjustment factor of 1.12).\54\

\54\ See also https://regtap.cms.gov/reg_librarye.php?i=4690 for an announcement and clarification of this error.

Response: We worked with the relevant State regulators in Massachusetts and Arkansas to identify the appropriate CSR adjustment factors for their State-specific CSR plan variants. Since Massachusetts

transitioned into the HHS-operated risk adjustment program beginning with the 2017 benefit year, the Massachusetts Health Connector, Massachusetts' Exchange, has consistently supported continued use of the 1.12 factor for their wraparound plans and has not indicated that a change is needed.\55\ Therefore, we do not believe that it is necessary to make changes to Massachusetts' wraparound CSR adjustment plan factor at this time and will continue to apply the 1.12 factor for these plans.

\55\ For examples, see for Massachusetts Health Connector's comments on the proposed 2025 Payment Notice at https://www.regulations.gov/comment/CMS-2023-0191-0081; also, see the Massachusetts Health Connector's comments on the proposed 2024 Payment Notice at https://www.regulations.gov/comment/CMS-2022-0192-0102.90.

We have similarly worked with the Arkansas Department of Insurance to understand the structure of the funding for Medicaid-expansion plans in Arkansas, using this information to identify what CSR adjustment factors are appropriate for these unique State-specific Medicaid- expansion plans. The CSR adjustment factor of 1.12 will continue to apply for both the 94 percent and 100 percent AV Arkansas Medicaid- expansion plans because these plans are identical in their degree of plan liability to Federal CSR plan variants offered on the Exchange for which the CSR adjustment factor of 1.12 is applied. We will continue to align the CSR adjustment factor for both the 94 and 100 percent Medicaid-expansion plans in Arkansas with the 94 percent silver plan variant CSR adjustment factor for the 2027 benefit year and beyond unless changes occur to the Arkansas program. Furthermore, it would not be appropriate to assign Arkansas's 94 percent or 100 percent AV Medicaid-expansion plans the higher CSR adjustment factor for the AI/AN zero cost sharing (-02 variant) or limited cost sharing CSR plan variant (-03 variant) enrollees because the CSR adjustment factors for AI/AN plans are calibrated to reflect the increased plan liability for enrollees in these specific CSR plan variants in the absence of a Federal appropriation for CSRs. In particular, regardless of plan AV, enrollees in AI/AN CSR plans defined at Sec. 156.420(b)(1) and (2) do not experience cost sharing on any item or service that is an EHB furnished directly by the Indian Health Service, an Indian Tribe, Tribal Organization, or Urban Indian Organization (each as defined in 25 U.S.C. 1603), or through referral under contract health services. This feature and other unique characteristics of AI/AN CSR plans do not apply to the 94 and 100 percent Medicaid-expansion plans in Arkansas.

We therefore do not agree that the current State-specific CSR adjustment factors for Massachusetts' ConnectorCare Plan Type 1 and Type 2 are too low, nor do we agree that Arkansas 100 percent AV Medicaid-wrap plans should receive a CSR adjustment factor of 1.46 instead of the CSR adjustment factor of 1.12 that we have deemed appropriate for these plans. As such, we do not believe any State- specific CSR adjustment factors need further adjustment at this time. BILLING CODE 4120-01-P [GRAPHIC] [TIFF OMITTED] TR20MY26.001

[GRAPHIC] [TIFF OMITTED] TR20MY26.002

[GRAPHIC] [TIFF OMITTED] TR20MY26.003

[GRAPHIC] [TIFF OMITTED] TR20MY26.004

[GRAPHIC] [TIFF OMITTED] TR20MY26.005

[GRAPHIC] [TIFF OMITTED] TR20MY26.006

[GRAPHIC] [TIFF OMITTED] TR20MY26.007

[GRAPHIC] [TIFF OMITTED] TR20MY26.008

[GRAPHIC] [TIFF OMITTED] TR20MY26.009

[GRAPHIC] [TIFF OMITTED] TR20MY26.010

[GRAPHIC] [TIFF OMITTED] TR20MY26.011

[GRAPHIC] [TIFF OMITTED] TR20MY26.012

[GRAPHIC] [TIFF OMITTED] TR20MY26.013

[GRAPHIC] [TIFF OMITTED] TR20MY26.014

[GRAPHIC] [TIFF OMITTED] TR20MY26.015

[GRAPHIC] [TIFF OMITTED] TR20MY26.016

[GRAPHIC] [TIFF OMITTED] TR20MY26.017

[GRAPHIC] [TIFF OMITTED] TR20MY26.018

BILLING CODE 4120-01-C

As noted in the 2027 Payment Notice proposed rule (91 FR 6327), each benefit year, to evaluate HHS risk adjustment model performance, we examine each model's R-squared statistic and predictive ratio (PR). The R-squared statistic, which calculates the percentage of individual variation explained by a model, measures the predictive accuracy of the model overall. The PR for each of the HHS risk adjustment models is the ratio of the weighted mean predicted plan liability for the model sample population to the weighted mean actual plan liability for the model sample population. The PR represents how well the model does on average at predicting plan liability for that subpopulation.

A subpopulation that is predicted perfectly would have a PR of 1.0. For each of the final 2027 HHS risk adjustment models, the R-squared statistic and the PRs are in the range of published estimates for concurrent HHS risk adjustment models.\56\ Because we blend the coefficients from separately solved models based on the 2021, 2022, and 2023 benefit years' enrollee-level EDGE data, we are publishing the R- squared statistic for each model separately to assess model performance. The R-squared statistics for the 2027 benefit HHS risk adjustment models are shown in Table 8.

\56\ Hileman, G., & Steele, S. (2016). Accuracy of Claims-Based Risk Scoring Models. Society of Actuaries. https://www.soa.org/globalassets/assets/files/research/research-2016-accuracy-claims-based-risk-scoring-models.pdf. [GRAPHIC] [TIFF OMITTED] TR20MY26.019

We received public comments on the R-squared statistics for the 2027 benefit year HHS risk adjustment models. We respond to these comments below.

Comment: A few commenters referred to decreases in the R-squared values in their recommendations for HHS risk adjustment model improvements, to which we have responded elsewhere in this rule. One of these commenters suggested that we not recalibrate the HHS risk adjustment models based on these lower R-squared values.

Response: We note that the R-squared values for the 2027 benefit year HHS risk adjustment models as applied to the 2021 and 2022 benefit year enrollee-level EDGE data improved relative to the R-squared values for the 2026 benefit year HHS risk adjustment models as applied to these data years.\57\ Furthermore, for the adult models (which apply to the majority of enrollees in the enrollee-level EDGE data), the R- squared values for the 2027 benefit year HHS risk adjustment models as applied to the 2023 benefit year enrollee-level EDGE data are higher than the R-squared values for those models as applied to the 2022 enrollee-level EDGE data. Although we will continue to monitor the R- squared values in the future, the R-squared values for 2027 benefit year risk adjustment model recalibration remain high and within the expected range of R-squared values for the type of model used for the HHS-operated risk adjustment program.\58\ We remain confident the HHS risk adjustment models continue to operate effectively and appropriately predict plan liability for an average enrollee.

\57\ See Table 7 of the 2026 Benefit Year Final HHS Risk Adjustment Model Coefficients (January 13, 2025), available at https://www.cms.gov/files/document/2026-benefit-year-final-hhs-risk-adjustment-model-coefficients2025-01-13.pdf.

\58\ See Hileman, G. & Spenser S. (2016). Accuracy of Claims- Based Risk Scoring Models. Society of Actuaries. https://www.soa.org/globalassets/assets/files/research/research-2016-accuracy-claims-based-risk-scoring-models.pdf.

In part 2 of the 2022 Payment Notice (86 FR 24183 through 24186), we finalized the proposal to continue to use the State payment transfer formula finalized in the 2021 Payment Notice for the 2022 benefit year and beyond, unless changed through notice-and-comment rulemaking. We did not propose changes to the formula in the 2027 Payment Notice proposed rule (91 FR 6327). We therefore will continue to apply the formula as finalized in the 2021 Payment Notice (86 FR 24183 through 24186) in the States where HHS operates the risk adjustment program in the 2027 benefit year.

Additionally, as finalized in the 2020 Payment Notice (84 FR 17466 through 17468), we will maintain the high-cost risk pool parameters for the 2020 benefit year and beyond, unless amended

through notice-and-comment rulemaking. We did not propose changes to the high-cost risk pool parameters for the 2027 benefit year; therefore, we will maintain the $1 million threshold and 60 percent coinsurance rate.\59\ However, we received comments on the high-cost risk pool parameters and are responding to these comments below.

\59\ See, for example, the 2018 Payment Notice (81 FR 94081) and 2020 Payment Notice (84 FR 17467).

Comment: One commenter requested that HHS consider updating the high-cost risk pool parameters to account for inflation to appropriately reflect the evolving distribution of very high-cost claims that fall above the $1 million threshold (“attachment point”) of the high-cost risk pool, which would have the effect of also updating the plan liability for these claims in recalibration for calculating risk scores and HHS risk adjustment transfers.

Response: We did not propose and are not finalizing changes to the high-cost risk pool parameters for the 2027 benefit year.

The high-cost risk pool is a budget neutral aspect of the HHS- operated risk adjustment program in which payments are funded by a percent of premium charge on all risk adjustment covered plans within the respective national high-cost risk pool (one for the individual market and merged market plans, including catastrophic and non- catastrophic plans, and another for the small group market). In the 2018 Payment Notice (81 FR 94082), we estimated that the percentage of premium charge, which funds high-cost risk pool payments, would most likely not exceed 0.5 percent. When aggregating across the national risk pool markets, this statement has held true. For example, for the 2024 benefit year, the individual market high-cost risk pool charge as a percent of premium was 0.39 percent \60\ and the small group market high-cost risk pool charge as a percent of premium was 0.58 percent.\61\ Aggregated across both the individual and small group markets, the overall high-cost risk pool charge was 0.45 percent of premium. This consistently low percentage of premium charges prevents States and issuers with high-cost enrollees from bearing a disproportionate amount of unpredictable risk. Additionally, we believe that retaining the same parameters for the high-cost risk pool year- over-year promotes stability and predictability in our markets, and we have concerns that indexing the $1 million threshold for inflation would annually vary those parameters.

\60\ This number include individual market catastrophic, non- catastrophic plans, and merged market plans.

\61\ CMS. (2025). Summary Report On Individual And Small Group Market Risk Adjustment Transfers For The 2024 Benefit Year. https://www.cms.gov/files/document/ra-report-by2024.pdf.

With these considerations, we continue to believe a $1 million threshold and 60 percent coinsurance rate are appropriate to incentivize issuers to control costs while improving risk prediction under the HHS risk adjustment models. However, we continue to monitor the high-cost risk pool parameters to consider whether adjustments are needed and would propose such changes for future benefit years through notice-and-comment rulemaking if necessary. Comment Solicitation on Retaining Separate Risk Adjustment Transfer Calculations for Individual Catastrophic Plans and Individual Non-Catastrophic Plans under the State Payment Transfer Formula.

In the 2027 Payment Notice proposed rule (91 FR 6327), we solicited comment on whether we should retain separate risk adjustment transfer calculations under the State payment transfer formula for individual catastrophic plans and individual non-catastrophic plans or whether we should calculate State transfers for these plans together in light of the September 4, 2025 guidance entitled “Guidance on Hardship Exemptions for Individuals Ineligible for Advance Payment of the Premium Tax Credit or Cost-sharing Reductions Due to Income, and Streamlining Exemption Pathways to Coverage.” \62\ This guidance expanded upon prior FFE hardship exemption policy by expanding eligibility for catastrophic plans starting with PY 2026.\63\ In the 2027 Payment Notice proposed rule (91 FR 6327), we sought comment on the impact of this policy on the HHS-operated risk adjustment program. We will take these comments into consideration as applicable.

\62\ See CMS. (2025). Guidance on Hardship Exemptions for Individuals Ineligible for Advance Payment of the Premium Tax Credit or Cost-sharing Reductions Due to Income, and Streamlining Exemption Pathways to Coverage. https://www.cms.gov/files/document/guidance-hardship-exemptions.pdf. We note that this final rule is also finalizing an amend to Sec. 155.605(d)(1) to codify the expansion of hardship exemption eligibility in this guidance document to individuals who are ineligible for APTC or CSR due to projected household income below 100 percent or above 250 percent FPL.

\63\ This guidance applies to consumers in FFE States and in SBE States that choose to have exemptions processed through the FFE, which currently include all SBEs except California, Connecticut, Maryland, and the District of Columbia. We note that in the 2027 Payment Notice proposed rule (91 FR 6353), we proposed to expand Sec. 155.605(d)(1) to codify the expansion of hardship exemption eligibility to consumers ineligible for APTC or CSRs due to projected household income below 100 percent or above 250 percent FPL in all States.

4. Risk Adjustment Data Validation Requirements When HHS Operates Risk Adjustment (HHS-RADV) (Sec. Sec. 153.350 and 153.630)

HHS conducts HHS-RADV under Sec. Sec. 153.350 and 153.630 in any State where HHS is operating risk adjustment on the State's behalf.\64\ The purpose of HHS-RADV is to ensure issuers are providing accurate high-quality information to HHS, which is crucial for the proper functioning of the HHS-operated risk adjustment program. HHS-RADV also ensures that risk adjustment transfers reflect verifiable actuarial risk differences among issuers, rather than risk score calculations that are based on poor quality data, thereby helping to ensure that the HHS-operated risk adjustment program assesses charges to issuers with plans with lower-than-average actuarial risk while making payments to issuers with plans with higher-than-average actuarial risk. HHS-RADV consists of an initial validation audit (IVA) and a second validation audit (SVA). Under Sec. 153.630, each issuer of a risk adjustment covered plan must engage an independent IVA entity. The issuer provides demographic, enrollment, and medical record documentation for a sample of enrollees selected by HHS to its IVA entity for data validation. Each issuer's IVA is followed by an SVA, which is conducted by an entity HHS retains to verify the accuracy of the findings of the IVA. Based on the findings from the IVA, or SVA (as applicable), HHS conducts error estimation to calculate an HHS-RADV error rate. The HHS- RADV error rate is then applied to adjust the PLRSs of outlier issuers, as well as the risk adjustment transfers calculated under the State payment transfer formula for the applicable State market risk pools, for the benefit year being audited.

\64\ Since the 2017 benefit year, HHS has operated the risk adjustment program in all 50 States and the District of Columbia.

In the 2027 Payment Notice proposed rule (91 FR 6327), we proposed to modify one intermediate step of the HHS-RADV error estimation methodology starting with 2025 benefit year HHS-RADV. In the 2026 Payment Notice (90 FR 4449 through 4452), we finalized excluding enrollees without

HCCs from IVA sampling beginning with 2025 benefit year HHS-RADV. We noted that this policy will impact the steps in the error estimation methodology during which HCC-associated error rates are applied to adjust issuers' PLRSs, and stated our intent to seek comments on potential modifications to the intermediate steps in the error estimation methodology to ensure that HCC-associated error rates continue to apply to only the proportion of total PLRSs that are associated with HCC-components of EDGE risk scores. As such, we proposed to add an additional scaling factor, ai, to the error estimation methodology to address this and capture the proportion of an issuer's total risk for the entire population that is associated with enrollees with HCCs. This scaling factor would be added to the final steps of error estimation in which HCC-associated error rates are applied to adjust issuers' PLRSs, and therefore, its addition would not impact the majority of the error estimation methodology, including the calculation of group failure rates, enrollee-level adjustments, or HCC- associated error rates.

The formula for the existing scaling factor HccPLRSweighti is the sum of sampled enrollees' stratum-weighted adjusted HCC-associated portion of EDGE risk scores divided by the sum of sampled enrollees' stratum-weighted total EDGE risk score. Because this formula is based on the issuer's sample, it depends on having enrollees with and without HCCs in the audit sample to appropriately estimate the proportion of the issuer's total PLRS that is HCC-related. However, as explained in the 2026 Payment Notice (90 FR 4452), when enrollees without HCCs are excluded from issuers' audit samples beginning with 2025 benefit year HHS-RADV, this formula will only estimate the proportion of enrollees' total EDGE risk scores that is HCC-related for the issuer's population of enrollees with HCCs.\65\ Therefore, as noted in the 2027 Payment Notice proposed rule (91 FR 6327), we proposed to create another scaling factor beginning with benefit year 2025 HHS-RADV that estimates the proportion of the issuer's total PLRS that is associated with enrollees with HCCs using the issuer's EDGE data. Together, these two scaling factors would capture the proportion of the issuer's total PLRS that is HCC-related.

\65\ In other words, this will factor out the contribution of demographic factors, enrollee RXCs, HCC-RXC interaction factors, CSR adjustment factors, HCC-contingent enrollment duration factors, and interacted HCC counts factors towards the EDGE risk scores of enrollees with HCCs. As previously explained, these factors are not included in the calculation of the HCC-associated error rate during HHS-RADV error estimation. See Section 13.3.1.3.3 Calculate Error Rates of the BY24 HHS-RADV Protocols, available at https://regtap.cms.gov/uploads/library/HHS-RADV_2024_Benefit_Year_Protocols_v1_5CR_060625.pdf.

Therefore, we proposed to introduce an additional scaling factor, ai, as follows: [GRAPHIC] [TIFF OMITTED] TR20MY26.020

Where:

meanRiskScorei,h is the average risk score for all enrollees in stratum h in issuer i's EDGE population; and strBMMi,h is the total stratum billable member months (BMM) for all enrollees in stratum h in issuer i's EDGE population.

We proposed to apply the scaling factor ai to the intermediate steps in the error estimation methodology as follows:

TotalERi = HccERi * HccPLRSWeighti * ai

The numerator in the formula for the scaling factor ai sums the product of each stratum's mean risk score and total BMM for strata 1 through 9, thereby creating an aggregate risk score for all enrollees with EDGE HCCs in an issuer's EDGE population.\66\ The denominator sums the product of each stratum's mean risk score and total BMM for strata 1 through 10. This includes all enrollees in the issuer's EDGE population including enrollees without HCCs, and thereby creates an aggregate risk score for the issuer.\67\ Overall, the scaling factor ai estimates the proportion of the issuer's total PLRS that is associated with enrollees with HCCs and, by combining it with the HCC PLRS weighting factor, we could continue to estimate the proportion of the issuer's total PLRS that is HCC-related after the removal of no-HCC enrollees from the IVA sample beginning with 2025 benefit year HHS- RADV. After leveraging EDGE data from the relevant benefit year to calculate the scaling factor ai and the total error rate TotalERi, we would continue to adjust issuers' PLRSs using the following formula:

\66\ An issuer's EDGE population only consists of enrollees in their risk adjustment covered plans. See Sec. Sec. 153.610(a) and 153.700(a).

\67\ Although enrollees without HCCs will be excluded from IVA sampling beginning with 2025 benefit year HHS-RADV, enrollees without HCCs on EDGE will be categorized into stratum 10 for these operational purposes.

AdjPLRSi = (1-TotalERi) * PLRSi

Without adding this additional scaling factor to the error estimation methodology beginning with 2025 benefit year HHS-RADV, the error rate would adjust elements of issuers' total PLRSs that are associated with enrollees without HCCs and are not intended to be adjusted during error estimation.\68\ We believe these adjustments would be inappropriate, and moreover, could result in double adjustments for any identified data errors of non-HCC components, such as demographic and enrollment factors, that are adjusted through separate processes.\69\ Therefore, starting with the 2025 benefit year of HHS-RADV, we proposed to add an additional scaling factor, ai, to the error estimation methodology to ensure that HCC-associated error rates continue to apply to only the proportion of total PLRSs that are associated with HCC-components of EDGE risk scores.

\68\ Enrollees without HCCs may contribute to the PLRS through demographic factors, enrollee RXCs, and CSR adjustment factors. As previously explained, these enrollees are not included in the calculation of the HCC-associated error rate during HHS-RADV error estimation.

\69\ While HHS-RADV also includes processes for validating RXCs and demographic and enrollment factors, any errors regarding these factors are treated as materially incorrect EDGE server data submissions. See 83 FR 16970 through 16971. Also see 84 FR 17501 and 85 FR 77002 through 77005.

We sought comments on this proposal.

After consideration of comments and for the reasons outlined in the proposed rule and this final rule, including our responses to comments, we are finalizing this policy as proposed. We summarize and respond to public comments received on the proposed addition of a scaling factor to the HHS-RADV error estimation methodology below.

Comment: Several commenters supported the addition of the proposed scaling factor for HHS-RADV. Several of these commenters noted that the addition of the scaling factor would more accurately reflect risk and capture the proportion of risk associated with enrollees with HCCs. One commenter

noted that this would better align the error rate calculation used with the basis of that calculation. One commenter noted that this proposal strengthens program integrity while maintaining the intent of HHS-RADV to target errors in the calculations of HCC-related risk.

Response: We agree with the commenters that this proposal would more precisely capture the risk associated with enrollees with HCCs, better align the error rate calculation with the basis of that calculation, and strengthen program integrity. Specifically, without adding this additional scaling factor to the error estimation methodology beginning with 2025 benefit year HHS-RADV, the error rate would adjust elements of issuers' total PLRSs that are associated with enrollees without HCCs and are not intended to be adjusted as part of this calculation.\70\ We believe these adjustments that would result from the absence of the scaling factor would be inappropriate, and moreover, could result in double adjustments for any identified data errors of non-HCC components, such as demographic and enrollment factors, that are adjusted through separate processes.\71\ Therefore, starting with the 2025 benefit year of HHS-RADV, we are finalizing the incorporation of an additional scaling factor, ai, into the error estimation methodology to ensure that HCC-associated error rates continue to apply to only the proportion of total PLRSs that are associated with HCC components of EDGE risk scores, as proposed.

\70\ Enrollees without HCCs may contribute to the PLRS through demographic factors, enrollee RXCs, and CSR adjustment factors. As previously explained, these enrollees are not included in the calculation of the HCC-associated error rate during HHS-RADV error estimation.

\71\ While HHS-RADV also includes processes for validating RXCs and demographic and enrollment factors, any errors regarding these factors are treated as materially incorrect EDGE server data submissions. See 83 FR 16970 through 16971. Also see 84 FR 17501 and 85 FR 77002 through 77005.

Comment: A commenter suggested that HHS should provide clear operational details regarding application of this scaling factor to help reduce avoidable uncertainty affecting premium development and issuer participation, and noted that a predictable HHS-RADV process is critical for market stability.

Response: We agree that a stable and predictable HHS-RADV process plays an important role in market stability and that the inclusion of this scaling factor in the error estimation methodology is appropriate to more precisely account for the relevant part of an issuer's population in the calculation of HHS-RADV error rates. We intend to update the HHS-RADV Protocols to reflect the adoption of the new additional scaling factor.\72\ We also provided analysis in the proposed rule (91 FR 6444) on the estimated impact of the proposed policy. Specifically, in the Regulatory Impact Analysis of the proposed rule, we simulated the impact of the additional scaling factor and found that HHS-RADV adjustments to risk adjustment State transfers decreased in magnitude by 11.7 percent in the individual non- catastrophic market (going from $148 million to $139 million) and by 13.8 percent in the small group market (from $81 million to $69.8 million). Because the HHS-operated risk adjustment program, including HHS-RADV adjustments to State transfers, is budget neutral, we would see the same impact on negative risk adjustment State transfers (in other words, risk adjustment charges) and HHS-RADV adjustments, in that both would decrease in magnitude. When examining the impact of the additional scaling factor on HHS-RADV adjustments over premium, our analysis found that there was only a 0.01 percent change in positive HHS-RADV adjustments in both markets when comparing results without the additional scaling factor with results reflecting the adoption of the additional scaling factor. This corresponds with a percentage point (PP) change of -0.02. Because our simulation showed a decrease in premium impact in both the individual non-catastrophic and small group markets, we do not believe that this scaling factor will create uncertainty in premium development and impact issuer participation in the HHS-operated risk adjustment and HHS-RADV programs.

\72\ Each benefit year of HHS-RADV, HHS releases the HHS-RADV Protocols, which provide details on the operations and workflows associated with HHS-RADV, including error estimation, and are available on REGTAP.

As noted in the 2027 Payment Notice proposed rule (91 FR 6327), we proposed an HHS risk adjustment user fee for the 2027 benefit year of $0.20 per member per month (PMPM). Under Sec. 153.310, if a State is not approved to operate, or chooses to forgo operating, its own risk adjustment program, HHS will operate risk adjustment on its behalf. For the 2027 benefit year, HHS will operate risk adjustment in every State and the District of Columbia. As described in the 2014 Payment Notice (78 FR 15416 through 15417), HHS' operation of the risk adjustment program on behalf of States is funded through a risk adjustment user fee. Section 153.610(f)(2) provides that, where HHS operates a risk adjustment program on behalf of a State, an issuer of a risk adjustment covered plan must remit a user fee to HHS equal to the product of its monthly billable member enrollment in the plan and the PMPM risk adjustment user fee specified in the annual HHS notice of benefit and payment parameters for the applicable benefit year.

OMB Circular No. A-25 established Federal policy regarding user fees, and specifies that a user charge will be assessed against each identifiable recipient for special benefits derived from Federal activities beyond those received by the general public.\73\ The HHS- operated risk adjustment program provides special benefits as defined in section 6(a)(1)(B) of OMB Circular No. A-25 to issuers of risk adjustment covered plans, because it mitigates the financial instability associate with potential adverse risk selection.\74\ The HHS-operated risk adjustment program also contributes to consumer confidence in the health insurance industry by helping to stabilize premiums across the individual, merged, and small group markets.

\73\ See Circular No. A-25 Revised. https://www.whitehouse.gov/wp-content/uploads/2017/11/Circular-025.pdf.

\74\ Id.

For the 2027 benefit year, HHS proposed to calculate the Federal administrative expenses of operating the HHS risk adjustment program. This calculation resulted in a risk adjustment user fee rate of $0.20 PMPM based on our estimated costs for HHS risk adjustment operations and estimated billable member months (BMM) for individuals enrolled in risk adjustment covered plans. As noted in the 2027 Payment Notice proposed rule (91 FR 6324), these costs cover development of the models and methodology, collections, payments, account management, data collection, data validation, program integrity and audit functions, operational analytics, interested parties training, operational support, and administrative and personnel costs dedicated to HHS- operated risk adjustment program activities. To calculate the risk adjustment user fee, we divided HHS' projected total costs for administering the program on behalf of States by the expected number of BMM in risk adjustment covered plans in States where the HHS-operated risk

adjustment program will apply in the 2027 benefit year.

We estimated that the total cost for HHS to operate the risk adjustment program on behalf of all States and the District of Columbia for the 2027 benefit year would be slightly more than $65 million, which is similar to the 2026 benefit year budget.\75\

\75\ We estimated that the total costs for HHS to operate the risk adjustment program on behalf of States for the 2026 calendar year would be approximately $65 million. See the 2026 Payment Notice (90 FR 4424 at 4448).

Similar to prior benefit years, we projected risk adjustment enrollment scenarios for the 2027 benefit year. Based on our estimates, for the 2027 benefit year, we noted that we do not expect enrollment changes to significantly impact collections under this user fee rate.

We stated in the 2027 Payment Notice proposed rule (91 FR 6327) that our intention was to reconsider the enrollment estimates for this final rule. If these newer enrollment estimates are too low in comparison to our proposed user fee estimates, we noted that the final user fee may be higher. We also noted that if these enrollment estimates are higher in comparison to our proposed user fee estimates, the final user fee may be lower. In addition, we noted that if any events result in a deviation from our expectations of current conditions that would significantly change our estimates around costs, enrollment projections, or the finalization of proposed risk adjustment policies between the proposed rule and this final rule, we may modify the proposed HHS risk adjustment user fee rate in the final rule. Because we projected a similar budget to operate the HHS-operated risk adjustment program for the 2027 benefit year as for the 2026 benefit year, we proposed an HHS risk adjustment user fee of $0.20 PMPM for the 2027 benefit year. We sought comment on the proposed HHS risk adjustment user fee rate of $0.20 PMPM for the 2027 benefit year.

As in all years, we consider updated information regarding estimates of costs, enrollment projections, and finalization of risk adjustment policies between the proposed rule and the final rule to determine the final risk adjustment user fee rate. Specifically, between the proposed and final rules, we updated our enrollment projections based on the availability of interim 2025 risk adjustment data, and our projections in enrollment in the off-Exchange individual and small group markets are higher than the estimates that we used for proposed rule.

Because some of our enrollment estimates were higher than previously anticipated, and the costs for the HHS-operated risk adjustment program for the 2027 benefit year have not changed, the PMPM cost of the HHS-operated risk adjustment program for the 2027 benefit year is lower than the proposed HHS risk adjustment user fee rate.

After consideration of comments and our additional data analyses between the proposed and final rules, and for the reasons outlined in the proposed rule and this final rule, including our responses to comments, we are finalizing an HHS risk adjustment user fee rate for the 2027 benefit year of $0.18 PMPM. We summarize and respond to public comments received on the 2027 HHS risk adjustment user fee rate below.

Comment: Several commenters stated support for maintaining the proposed HHS risk adjustment user fee rate of $0.20 PMPM for the 2027 benefit year.

Response: We believe that when costs, risk adjustment policies, and projections of enrollment in the HHS-operated risk adjustment program are constant, the HHS risk adjustment user fee rate should remain constant. However, based on more recently available data, we are finalizing an HHS risk adjustment user fee rate for the 2027 benefit year of $0.18 PMPM.

Comment: Some commenters requested that HHS consider the pricing impacts of the HHS risk adjustment user fee rate.

Response: We note that we must set the HHS risk adjustment user fee for a given benefit year to secure adequate funding for the HHS- operated risk adjustment program. We are finalizing an HHS risk adjustment user fee rate for the 2027 benefit year of $0.18 PMPM, which is lower than the rate for the 2026 benefit year. We anticipate that the overall impact on pricing of an $0.18 PMPM HHS risk adjustment user fee rate will be minimal, since it is lower than the user fee rate for the 2026 benefit year and because $0.18 PMPM is a very small percent of overall premium.

C. Part 154--Health Insurance Issuer Rate Increases: Disclosure and Review Requirements

1. Submission of Rate Filing Justification (Sec. 154.215) a. CSR Reimbursement

Section 1402 of the Affordable Care Act requires issuers to provide cost-sharing reductions (CSRs) to increase the actuarial value of coverage for consumers with incomes between 100 and 250 percent of Federal Poverty Level (FPL) who enroll in silver-level QHPs in the individual market, as well as eligible American Indian/Alaska Native consumers who enroll in QHPs at any metal level. Section 1402 of the Affordable Care Act also states that HHS will reimburse issuers for the cost of providing CSRs to eligible enrollees but does not include a valid appropriation to make such payments.\76\

\76\ Until October 2017, HHS relied on the permanent appropriation at 31 U.S.C. 1324 as the source of funds for Federal CSR reimbursement to issuers.

On October 11, 2017, the Attorney General of the United States provided a legal opinion stating that HHS and the Department of the Treasury could no longer use the permanent appropriation at 31 U.S.C. 1324 to fund CSR reimbursements to issuers. In accordance with that opinion, HHS directed CMS to discontinue CSR reimbursements to issuers until Congress provides an appropriation.\77\ In response to the termination of CSR reimbursement, State Departments of Insurance allowed or instructed issuers to increase (or “load”) premiums either primarily, or only, on silver-level QHPs to offset the issuers' cost of providing CSRs.\78\

\77\ See Letter from Acting HHS Secretary Eric Hargan to CMS Administrator Seema Verma Regarding Payments to Issuers for Cost- Sharing Reductions (CSRs) (Oct. 12, 2017), conveying legal opinion of Attorney General Jefferson B. Sessions III (Oct. 11, 2017). Available at https://www.hhs.gov/sites/default/files/csr-payment-memo.pdf.

\78\ For purposes of this preamble, we use the term “CSR loading” to refer to any rating practices to increase premiums to offset amounts of unreimbursed CSRs whether that is “silver loading” or “broad loading.”

There are several ways that issuers have determined the CSR load factor. For example, issuers have loaded silver plans offered on- Exchange to recover expected lost CSRs based on experience, they have loaded silver plans on-Exchange based on an assumed distribution of enrollment and enrollee utilization (sometimes a set amount mandated by the State), or they have loaded all plans at all metal levels by the same amount to spread lost CSRs across the entire individual market (broad loading). These differing approaches may result in loads that produce additional revenue collected from the applied CSR load that exceeds the amount of unreimbursed CSRs by assuming an enrollment and claims distribution that significantly diverges from what actually occurs. Based on our review of actuarial memoranda submitted by issuers for PY 2026, HHS believes these excessive loads on silver plans in particular (and in some cases as mandated by State law) lead to inflated premiums for silver plans, further distort pricing for bronze and

gold plans relative to silver plans, limit consumer choice, and significantly increase the cost of the second lowest-cost silver plan available to a consumer, which in turn increases PTC amounts and Federal expenditures. b. Rate Filing Justifications Regarding CSRs

Section 2794 of the PHS Act directs the Secretary, in conjunction with the States, to establish a process for the annual review of premium increases for health coverage prior to the implementation of the increase. HHS has historically interpreted this requirement for premiums as referring to the underlying rates that are used to develop premiums. Section 154.215 requires issuers to submit rate filing justifications to CMS and the applicable State (76 FR 29964, 29969).

Under Sec. 154.215(b)(1) through (3), the rate filing justification has three parts. The Unified Rate Review Template (URRT) (Part I of the rate filing justification) is required for all single risk pool products, including new products. It is intended to capture information needed to monitor premium increases of health insurance coverage offered through and outside the Exchanges in the individual and small group markets and ensure compliance with the single risk pool methodology, including allowable market level index rate adjustments to reflect risk adjustment payments and charges, and other Federal rating requirements.\79\

\79\ See Unified Rate Review Instructions at https://www.cms.gov/files/document/unified-rate-review-instructions.pdf.

Part II of the rate filing justification is the Written Description Justifying the Rate Increase (Consumer Justification Narrative). Part II is required only for rate increases in single risk pool products that are subject to review (that is, a plan within the product that has a rate increase of 15 percent or greater or a State-specific threshold). Part II is a consumer-friendly narrative that provides the justification for the rate increase, describes the relevant Part I data, the assumptions used to develop the rate increase, and an explanation of the most significant factors causing the rate increase.\80\

\80\ Id.

An actuarial memorandum (Part III of the rate filing justification) is required for any rate increase in a single risk pool plan. It is also required for any rate filing containing QHPs or whenever a State requires it to be submitted. Further, an actuarial memorandum is required for all plans in States that do not have an Effective Rate Review Program and for which CMS is responsible for reviewing the rate filing.\81\ The Part III actuarial memorandum includes the actuarial reasoning and assumptions, justifications, and methodologies that support the entries in the URRT.\82\ The actuarial memorandum must also capture appropriate actuarial certifications related to the development of the index rate in accordance with Federal regulations, and the development of plan specific premium rates using allowable modifiers to the index rate.\83\ The issuer is required to provide an explanation of how these modifiers are developed and applied to the market-wide adjusted index rate to derive the plan-adjusted index rate.

\81\ CMS reviews rate filing justifications from issuers in States without an Effective Rate Review Program--currently Oklahoma, Tennessee, and Wyoming.

\82\ See Unified Rate Review Instructions at https://www.cms.gov/files/document/unified-rate-review-instructions.pdf.

\83\ Id.

In a Bulletin issued on May 2, 2025 (PY26 Rate Filing Guidance),\84\ we instructed issuers that make permitted plan-level adjustments to account for CSR amounts provided to eligible enrollees for which the issuer does not otherwise receive reimbursement \85\ (that is, issuers that load premiums) to specify the actual CSRs the issuers paid for PY 2024 in the actuarial memorandum submitted with their PY 2026 rate filing. Issuers report plan-level adjustments when they submit a rate filing justification to the State or CMS for review. States or CMS review those rate filing justifications to ensure compliance with the Federal rating rules, including 45 CFR 156.80.\86\

\84\ CMS Insurance Standard Bulletin Series: Plan Year 2026 Individual Market Rate Filing Instructions. (May 2, 2025) https://www.cms.gov/files/document/py-26-individual-market-rate-filing-instructions.pdf.

\85\ See 45 CFR 156.80(d)(2)(i).

\86\ If the rate filing contains a proposed increase that meets or exceeds the threshold at Sec. 154.200(a)(1) (currently 15 percent), then the State or CMS also reviews the proposed increase to determine if it is an unreasonable rate increase. Also see 45 CFR 154.205(a). When CMS reviews a rate increase subject to review under Sec. 154.210(a), CMS will determine that the rate increase is an unreasonable rate increase if the increase is an excessive rate increase, an unjustified rate increase, or an unfairly discriminatory rate increase.

Specifically, through the PY26 Rate Filing Guidance, and under our authority under Sec. 154.215(a)(2) and (3), we directed issuers that loaded for unreimbursed CSRs to: (1) specify the actual CSRs the issuer paid on behalf of enrollees for PY 2024 (in dollars); (2) specify the CSR load factor for PY 2026 and explain how it was determined; and (3) explain how the additional revenue to be collected from the applied CSR load compares to the expected amount of CSRs that will be provided to enrollees in PY 2026.

Following issuance of the PY26 Rate Filing Guidance, CMS issued additional guidance entitled, “Frequently Asked Questions on Plan Year 2026 Individual Market Rate Filing Instructions,” on May 27, 2025.\87\ We noted that if an issuer was not able to calculate the precise amount of actual CSRs paid for enrollees for PY 2024 by the applicable rate filing deadline, CMS would accept an estimate developed using a reasonable methodology that enables it to estimate the value of CSRs provided for PY 2024 as accurately as possible, detailed in the actuarial memorandum submitted with the PY 2026 rate filing justification.\88\

\87\ CMS Frequently Asked Questions on Plan Year 2026 Individual Market Rate Filing Instructions. https://regtap.cms.gov/reg_librarye.php?i=5894.

\88\ Under Sec. 156.80(d)(2)(i), an issuer may vary premium rates for a particular plan from its market-wide index rate for a relevant State market based on the actuarial value and cost-sharing design of the plan, including accounting for, if permitted by the applicable State authority, CSR amounts provided to eligible enrollees under Sec. 156.410, provided the issuer does not otherwise receive reimbursement for such amounts. Therefore, if there is a valid appropriation such that HHS and the Department of the Treasury resume making advance payments of CSRs, issuers may not apply any CSR load to QHPs receiving advance CSR payments. In addition, in the event that advance payments of CSRs are made to issuers to reimburse them for CSRs provided, HHS will calculate these monthly advance payments using the formula finalized in the 2015 Payment Notice and using the standard methodology as set forth in 45 CFR 156.430(c)(2) for reconciliation of CSR amounts. See 79 FR 13804-13808. Also see 90 FR 4424, 4488.

As noted in the 2027 Payment Notice proposed rule (91 FR 6327), starting with rate filings for the 2027 plan year, we proposed to continue to require issuers that load premiums to submit certain information as specified in the PY26 Rate Filing Guidance in their URRTs and actuarial memoranda for each plan year in which CSRs are not funded. Specifically, in the URRT for the upcoming plan year, issuers would report actual CSR amounts paid on behalf of enrollees and the additional revenue collected from the previously applied CSR load using the most recent annual data that is available prior to the applicable filing year, using the standard methodology set forth in Sec. 156.430(c)(2). In most cases, the most recent annual CSR data would reflect the plan year that is two years before the upcoming plan year (for example, CSRs paid for eligible enrollees and the additional revenue collected from the CSR load applied in PY 2025 would be reported during the 2026 filing year for PY 2027 rate filings).

As described by the Secretary in applicable guidance,\89\ under the CMS standard methodology, issuers re-adjudicate the actual complete set of claims incurred by an enrollee in the cost-sharing reduction plan variation as if they had been enrolled in the associated standard plan to determine the difference the enrollee would have paid in deductibles, copayments, coinsurance, and other out-of-pocket expenses for EHBs (other than premiums and balance billing). The difference equals the amount of CSRs provided by the issuer.\90\ As stated in the 2016 Payment Notice, we believe that the standard methodology is the most accurate method for calculating the actual value of CSRs that the issuer has provided on behalf of enrollees in a plan year.\91\ Additionally, we believe that most issuers are familiar with that methodology, which was required to calculate CSRs paid on behalf of enrollees for the 2017 plan year, the most recent year in which CMS provided advance CSR payments to issuers.

\89\ See Manual for Reconciliation of the Cost-Sharing Reduction Component of Advance Payments for Benefit Year 2017 (March 29, 2018) at https://www.cms.gov/cciio/resources/forms-reports-and-other-resources/downloads/final-csr-reconciliation-guidance-by2017.pdf.

\90\ Id.

\91\ HHS Notice of Benefit and Payment Parameters for 2016 Final Rule, 80 FR 10750, 10842 (February 27, 2015), available at: https://www.govinfo.gov/content/pkg/FR-2015-02-27/pdf/2015-03751.pdf.

We note that although CMS proposed that issuers use the standard methodology to calculate CSR amounts paid on behalf of enrollees and to submit an aggregate amount of CSRs provided at the plan level on the URRT, this data submission would not require issuers to use the CSR reconciliation process implemented by CMS, as described in Sec. 156.430(c)(2). We believe this proposal would result in a lower burden on issuers as compared to the burden that would be associated with submitting policy-level CSR data to CMS through the CSR reconciliation process. Submission of policy-level CSR data would require direct electronic submissions of data that must conform to our business rules, data element validations, and required file formats. When HHS previously collected such data, it resulted in issuers attempting submissions multiple times before the submission could be accepted by the system. The proposed process will leverage the existing URRT submission process and be much less burdensome.

We also proposed that issuers would include in the URRT the applicable CSR load factor for each plan that would be applied to the market adjusted index rate to calculate the calibrated plan adjusted index rate for the upcoming plan year. We proposed to collect the CSR load factor, if any, to fulfill our responsibility to ensure compliance with Sec. 156.80(d), which requires all permitted plan-level adjustments to be “actuarially justified.” In addition, we proposed that issuers would include in the actuarial memorandum an explanation of the methodology used to determine the load factor. We proposed that issuers would also include in the URRT the additional revenue expected to be collected from the applied CSR load factor and the expected amount of CSRs that will be paid for enrollees for the upcoming plan year. We also proposed that issuers would include in the actuarial memorandum an explanation comparing these amounts. This explanation would allow the State or CMS, as applicable, to determine whether the load factor is actuarially justified and not excessive in relation to the amount expected to be paid for unreimbursed CSRs.

In the 2026 Payment Notice, we stated our expectation that CSR loading practices, to the extent permitted by State regulators, are intended to account for unpaid CSRs. We also noted that, while there is no requirement that a State permit CSR loading, in States that have an Effective Rate Review Program, the State has the responsibility to determine whether an issuer's adjustments to the market-wide index rate for plan-specific factors (including accounting for CSR amounts) are actuarially justified.\92\ We further proposed that an actuarially justified CSR load factor is one that is calibrated on actual experience and that only accounts for the projected revenue loss of unreimbursed CSR payments without materially exceeding that amount. As such, collecting through the URRT and actuarial memoranda information on paid CSRs, additional revenue collected from the previously applied CSR load, CSRs expected to be paid, the CSR load factor and expected resultant additional revenue for the upcoming plan year, the underlying methodology for determining the CSR load factor that would be applied for the upcoming plan year, and an explanation of how the expected additional revenue compares to the amount of CSRs expected to be paid, will benefit State regulators (and CMS in States where CMS functions as the primary reviewer of rates) by providing regulators the data necessary to determine whether CSR load amounts are actuarially justified plan-level adjustments to the index rate under Sec. 156.80.

\92\ HHS Notice of Benefit and Payment Parameters for 2026 Final Rule, 90 FR 4424, 4489 (January 15, 2025) available at https://www.govinfo.gov/content/pkg/FR-2025-01-15/pdf/2025-00640.pdf.

While we recognize the additional burden on issuers to provide this information, given the significant impact of CSR loading on Federal expenditures through additional premium tax credit (PTC) spending, we believe collection of this information is an important program integrity measure that will help ensure that CSR loads are appropriate to recover lost CSR payments and are not inappropriately inflating Federal expenditures or undermining Federal rating rules.

Therefore, we proposed to collect information on adjustments to the index rate to account for unreimbursed CSRs as described above. We did not propose changes to any regulation text as the collection of these data is already captured under Sec. 154.215(d)(1), which states that historical and projected claims experience must be included in the URRT. Additionally, Sec. 156.80(d)(2)(i) states that plan-level adjustments to account for unreimbursed CSR payments provided to eligible enrollees are permissible only if actuarially justified and permitted by the applicable State authority. We requested comment on all aspects of our proposal to require issuers that intend to load premium rates to account for unpaid CSRs for the upcoming plan year to submit this information in their URRTs and the actuarial memoranda for each plan year in which CSRs are not funded, beginning with PY 2027 rate filings.

After consideration of comments and for the reasons outlined in the proposed rule and this final rule, including our responses to comments, we are finalizing this policy as proposed. As we stated in the proposed rule, we intend to release guidance on the submission of the required URRT and actuarial memoranda information as part of revised Unified Rate Review Instructions, as we have historically provided detailed guidance to issuers on how to complete each field of the URRT and satisfy the criteria for the actuarial memorandum in the Unified Rate Review Instructions.\93\ We summarize and respond to public comments received on the proposed policy below.

\93\ The information collection described in this section will be submitted as a revision to the currently approved PRA package CMS-10379 (OMB Control Number 0938-1141) for OMB review under the Paperwork Reduction Act.

Comment: A few commenters supported the proposed CSR data collection, stating that they are in favor of increased transparency and

standardized reporting to better understand CSR loading practices and ensure CSR loads align with actual costs. Many other commenters expressed concern that the proposed CSR data reporting requirements would impose significant operational and financial burdens on issuers and may lead to higher premiums.

Response: We acknowledge that issuers will incur operational and financial burdens as a result of these reporting requirements. However, collection of this information is an important program integrity measure that will help ensure that CSR loads are appropriate to recover lost CSR payments and are not inappropriately inflating Federal expenditures, excessively increasing silver plan premiums paid by unsubsidized enrollees, or undermining Federal rating rules. This information will also benefit State regulators by providing necessary data to determine if their issuers' CSR load amounts are actuarially justified plan-level adjustments to the index rate under Sec. 156.80. We also note that when this information was previously collected through the policy-level electronic data collection process in place for determining the actual amount of CSRs provided under Sec. 156.430, that process imposed additional burden on issuers associated with submitting multiple electronic files through that data collection process. Collecting this CSR information through the URRT will minimize the burden on issuers.

Comment: Many State insurance departments and Exchanges stressed the importance of preserving State-directed CSR loading methodologies and rate review processes. They opposed Federal standardization that might override State-specific actuarial soundness standards or impose a single CSR loading methodology. Moreover, some commenters stated that CSR load data may vary significantly by State and issuer, making direct comparisons misleading without proper context. They suggested that, to avoid misinterpretation, detailed discussions of CSR load data should be included in actuarial memoranda rather than in standardized templates.

Response: We agree with commenters that the actuarial memorandum, not the URRT, is the appropriate place to include detailed discussions of CSR load data. We clarify that we are not finalizing a single CSR loading methodology at this time. As stated in the proposed rule, the plan-level adjustments that result from CSR loading must be actuarially justified and calibrated on actual market experience that accounts only for the projected revenue loss of unreimbursed CSR payments without materially exceeding that amount.\94\ CSR loading is meant to recover lost CSR reimbursement and should not be seen as a premium adjustment factor to achieve a desired premium subsidy level. Specifically, silver-plan premiums should not be loaded in a way that purposefully collects more in premium than the expected amount of unreimbursed CSRs to increase subsidy amounts and lower premiums at other metal levels. This practice goes beyond CSR loading and is an inappropriate use of the CSR load factor. We are finalizing this description of actuarially justified plan-specific adjustment factors as proposed.

\94\ We note that a plan-level adjustment to account for unreimbursed CSRs is not considered actuarially justified under Sec. 156.80 solely because the adjustment or methodology on which it is based was prescribed by a State.

States with an Effective Rate Review Program are responsible for reviewing proposed rate increases subject to review to determine whether those increases are unreasonable by, in part, examining the reasonableness of the assumptions used by the health insurance issuer to develop the proposed rate increase and the validity of the historical data underlying the assumptions.\95\ Such States are also responsible for reviewing all proposed rate changes for compliance with Federal rating rules. This includes determining whether an issuer's adjustments to the market-wide index rate for plan-specific factors (including accounting for unreimbursed CSR amounts) are actuarially justified and compliant with the single risk pool regulations at Sec. 156.80.

\95\ See 45 CFR 154.301(a)(3)(i).

A loading practice that results in additional premium that materially exceeds what is projected to account for the impact of unreimbursed CSRs does not appropriately account for CSRs provided to eligible enrollees as required for permitted plan-specific adjustments to the index rate under Sec. 156.80(d)(2)(i).\96\ As such, in States without an Effective Rate Review Program, CMS may exercise its authority in Sec. 154.205(b) to determine a rate increase is unreasonable if such increase includes a CSR load that was developed based on a methodology that does not reasonably account for the projected amount of unpaid CSRs. In accordance with section 1311(e)(2) of the Affordable Care Act and Sec. 155.1020, CMS may also recommend that the Exchange not make available through the Exchange plans with such an unreasonable rate increase based on patterns or practices of excessive or unjustified premium increases. In evaluating whether a State has an Effective Rate Review Program, as defined in Sec. 154.301, CMS will similarly evaluate whether the State rate review process adheres to the requirements set forth in Sec. 154.301(a)(3) for issuers' assumptions about the premium increases that are necessary to account for unreimbursed CSRs. We do not believe the requirement to submit additional information regarding how unreimbursed CSRs are accounted for within the plan-level adjustments to the index rate interferes with State-directed CSR loading methodologies or rate review processes where these processes comply with the requirements under Sec. 154.301.

\96\ This should not be construed to prevent broad loading, as issuers are permitted to spread these costs evenly across all plans in the single risk pool. See for example, 78 FR 39870, 39878 (July 2, 2013) (administrative costs generally may be spread across an issuer's entire risk pool).

We intend to review the submitted CSR data. If issuers provide a State with proposed rate increases which include a CSR load that does not appear to be developed based on a methodology that reasonably accounts for the amount of unpaid CSRs and the State does not deem such rate increases as “unreasonable,” CMS may use its authority under Sec. 154.301(d) to evaluate whether the State has ceased to satisfy the criteria set forth Sec. 154.301(a) and (b) and no longer has an Effective Rate Review Program.

The regulations at Sec. 155.1000 provide Exchanges with broad discretion to certify QHPs that otherwise meet the QHP certification standards specified in Part 156 and afford Exchanges discretion to deny certification of QHPs that meet minimum QHP certification standards, but are nonetheless not in the interest of qualified individuals. As stated earlier in this preamble and in the proposed rule, CSR loading practices that result in QHP premium adjustments that materially exceed the projected revenue loss from unreimbursed CSR payments, or excessively load plans at the silver metal tier beyond expected unreimbursed CSR payments to lower premiums for other metal tiers, do not align with requirements for actuarially justified plan-level adjustments to the index rate permitted under Sec. 156.80(d). Such loading practices can result in significantly higher premiums that harm unsubsidized consumers enrolled in Exchange plans. Accordingly, CMS may determine that denial of an issuer's request for certification on the FFE is warranted if an issuer engages in CSR loading practices that result in premiums that are significantly higher than would reasonably be expected

when accounting for the amount of unreimbursed CSRs in violation of Sec. 156.80(d).

Comment: Some commenters indicated that changes in CSR loading and premium adjustments could cause shifts in enrollment, raising concerns about market stability. A few commenters noted that such shifts in enrollment could negatively affect American Indian/Alaska Native populations and Tribal health programs reliant on stable silver plan pricing.

Response: As some commenters stated, when CSR costs are embedded into silver premiums, benchmark premiums rise and Federal PTC spending increases--contributing to fully subsidized bronze and gold plans and leading to unsubsidized enrollees paying higher silver plan premiums. Therefore, it is imperative that regulators ensure that issuers are setting CSR load factors at a level no higher than necessary to fully compensate for unreimbursed CSR amounts provided to enrollees. Imposing a higher load exacerbates shifts in enrollment and decreases market stability.

We appreciate that American Indian/Alaska Native populations and Tribal health programs are especially reliant on stable silver plan pricing. We affirm that CMS engages in tribal consultation for all rulemakings and has worked with the appropriate parties to ensure that any such potential impact is mitigated.

Comment: Several commenters stated that the proposed definition of “actuarially justified” CSR load factors, which emphasizes calibration on actual CSR experience, may conflict with Affordable Care Act provisions requiring a single risk pool and prohibiting rates based on plan-specific experience. Those commenters recommended that CMS should allow issuers to consider the expected membership distribution and unreimbursed CSR costs at the overall single risk pool or market level when developing CSR load factors, rather than requiring issuers to develop CSR load factors that are specific to their enrolled population. Commenters also recommended that CMS clarify whether “actuarially justified” implies an issuer/plan-level “matching” requirement or whether only experience-calibrated methodologies are acceptable.

Response: We proposed that an actuarially justified CSR load factor is one that is calibrated based on actual CSR experience rather than an unreasonable projection or assumption. We agree that the Affordable Care Act provisions require a single risk pool and prohibit rates based on plan-specific experience. An actuarially justified CSR load factor should not be based on plan-specific experience; it must be set to only recover expected unpaid CSRs, which we referred to as “actual CSR experience.” We did not propose, and do not intend to imply, that an issuer/plan-level “matching” is required or that only experience- calibrated methodologies are acceptable. We acknowledge that the CSR load amount will be based on a projection of CSR enrollment distribution and costs, but we emphasize that the projection should reflect realistic assumptions. It may be appropriate to consider CSR enrollment across the entire market rather than only considering the specific enrollment of one particular issuer when projecting CSR enrollment and determining the CSR load. For example, some States may require all issuers to use the same CSR load factor on all silver plans, and this practice is permissible so long as the CSR load is based on a reasonable projection of CSR enrollment and utilization of services. In this case, the CSR load factor should be set to only recover expected unpaid CSRs. The CSR load factor should not be used to inflate silver premiums beyond expected unpaid CSR amounts to reduce premiums on other metal tiers and further distort pricing between metal levels.

Comment: Some commenters requested that CMS clarify the goals and intended uses of the CSR data collection and recommended enhanced communication to consumers about how CSR loading affects premiums and subsidies to improve transparency and understanding.

Response: As we stated in the proposed rule, the purpose of this proposal is to allow the State or CMS, as applicable, to determine whether the CSR load factor is actuarially justified and not excessive in relation to the amount expected to be paid by an issuer in unreimbursed CSRs. We are concerned that excessive CSR loads on silver plans lead to inflated premiums for silver plans, further distorting pricing for bronze and gold plans relative to silver plans; limit consumer choice; and significantly increase the cost of the second lowest cost silver plan, which in turn increases PTC amounts and Federal expenditures. The additional data will be used to ensure that CSR loads are actuarially justified to compensate for the amount of unreimbursed CSRs. To the extent they exist, CSR loads should be limited to account for the amount of the actual unpaid CSRs. Silver plan premiums should not be increased significantly to achieve a desired subsidy level rather than as a reflection of unpaid CSRs resulting from CSR plan variant enrollees in those plans. If, for example, an issuer projects a 20 percent premium load to silver plans is needed to recover unpaid CSRs, but the issuer applies a 30 percent load to silver plans, that would presumably lower premiums for bronze and gold plans that are unaffected by unpaid CSRs. That load amount greatly exceeds the projected amount of unpaid CSRs and goes beyond the goal of recovering unreimbursed CSR costs. The data collection is intended to ensure that CSR loading is not being used as a tool to achieve a target price point that does not reflect the allowable rating factors.

For the comments recommending communication to consumers about how CSR loading affects premiums and subsidies, these comments are out of scope for this final rule because they do not relate to the specific proposals included in the proposed rule. We note, however, that Healthcare.gov, an official consumer-facing CMS website, includes information for consumers regarding CSRs, including how to qualify and an explanation of how CSRs can lower out-of-pocket costs.\97\

\97\ Healthcare.gov, “Saving Money on Health Insurance: Cost- sharing Reductions”, available at https://www.healthcare.gov/lower-costs/save-on-out-of-pocket-costs/.

Comment: Several commenters requested CMS to delay implementation to allow for system updates. Many commenters noted that rate filings are due in April and that issuers' current systems cannot generate the requested data. Several commenters anticipated technical difficulties in using the standard methodology set forth in Sec. 156.430(c)(2) to report CSR amounts paid on behalf of enrollees as proposed.\98\ Other commenters requested CMS allow a safe harbor for good faith efforts to calculate actual amounts.

\98\ Under the CMS standard methodology, issuers re-adjudicate the actual complete set of claims incurred by an enrollee in the CSR plan variation as if they had been enrolled in the associated standard plan to determine the difference the enrollee would have paid in deductibles, copayments, coinsurance, and other out-of- pocket expenses for EHBs (other than premiums and balance billing). The difference equals the amount of CSRs provided by the issuer. See Manual for Reconciliation of the Cost-Sharing Reduction Component of Advance Payments for Benefit Year 2017 (March 29, 2018) at https://www.cms.gov/cciio/resources/forms-reports-and-other-resources/downloads/final-csr-reconciliation-guidance-by2017.pdf.

Response: We recognize that some updates to issuers' systems will be necessary, but we do not agree that a delayed applicability date is warranted. The proposed data collection process is intended to leverage the existing rate

filing process by using the URRT and actuarial memorandum. We proposed that issuers use the standard methodology to determine the amount of CSRs paid by the issuer because we believe it is the most accurate method for such calculation and because of issuers' familiarity with this methodology. While we recognize that many issuers might have to update their systems to calculate the actual value of CSRs provided, we believe that issuers should already be calculating the actual CSRs paid for enrollees, as specified in the PY26 Rate Filing Guidance. That guidance instructed issuers to report the actual CSRs paid for enrollees for PY 2024 in the actuarial memorandum submitted with the 2026 rate filing. While CMS provided flexibility in the May 27, 2025 Frequently Asked Questions on Plan Year 2026 Individual Market Rate Filing Instructions for issuers that were unable to timely calculate the precise amount of actual CSRs paid for enrollees for PY 2024 to provide an estimate developed using a reasonable methodology detailed in the actuarial memorandum, that flexibility was limited to PY 2026 rate filing justifications.

As stated in the proposed rule, the approach of collecting an aggregate amount of CSRs provided at the plan level on the URRT is less burdensome than the collection of policy-level CSR data through the CSR reconciliation process that was required for all issuers that received advance payment of CSRs prior to the cessation of these payments in October 2017. Therefore, while issuers may incur some additional costs for actuarial analysis, issuers are generally already familiar with the process of calculating the actual amount of CSRs provided using the standard methodology, and the limited additional burden issuers might incur does not warrant a delayed implementation.

We acknowledge that in some States initial rate filings must be submitted in April. We note, however, that those are proposed rate changes, and States typically provide issuers with the opportunity to adjust their rate filings before finalizing them several months later. Likewise, the deadlines for submitting proposed rate changes to CMS are June 1 (for States without an Effective Rate Review Program) or July 15 (for States that have an Effective Rate Review Program). Final rate filings are not due to CMS until August 12 (for filings that contain a QHP in States with Exchanges served by the HealthCare.gov platform) or October 15. Issuers in States with earlier deadlines should refer to the 2026 Rate Review Timeline Bulletin \99\ for deadlines for revising their initial rate submissions.

\99\ Bulletin: Timing of Submission of Rate Filing Justifications for the 2026 Filing Year for Single Risk Pool Coverage Effective on or after January 1, 2027 (February 23, 2026) available at https://www.cms.gov/files/document/2026-rate-review-timeline-bulletin.pdf.