Nine reform levers are on the table, from a five-minute rate-setting tweak to a full rebuild of how plans get paid. Here's what each one actually costs, saves, and requires — before taking a side.
High enrollment plus persistent overpayment estimates is why reform proposals keep clustering around the same handful of levers: risk adjustment, benchmarks, quality bonuses, audit and enforcement, and consumer-protection rules.
The biggest dollar figures come from payment-system reforms, not administrative tweaks — benchmark cuts, deeper coding-intensity adjustments, and quality-bonus changes dwarf everything else in scored savings. But the most actionable reforms are narrower: CMS has already started restricting which diagnoses count toward risk scores, and Congress has bipartisan bills moving on prior authorization and risk-score integrity. The sweeping proposals — large benchmark cuts, full competitive bidding — would require legislation and face much heavier resistance, because they more visibly threaten supplemental benefits and plan margins.
Each one below is collapsed to a single line by default — mechanism, feasibility, and the headline savings number. Expand any of them for the rationale, the pathway to enactment, and the real tradeoffs.
MA plans face stronger incentives than fee-for-service providers to code every remunerative diagnosis. Under-adjusted risk scores translate directly into overpayment. MedPAC says the newer V28 risk model helped, but MA coding intensity still generates excess payments and inequity across organizations.
The 5.9% floor is statutory, so a durable increase requires Congress — CMS can study coding patterns annually but can't raise the floor on its own. CBO modeled raising it to 8% (savings of $159B over 2025–2034) and to 20% ($1.049T over the same window). The financial hit falls on MA organizations directly; beneficiaries feel it indirectly through smaller rebates or thinner supplemental benefits. CBO expects average MA and FFS premiums to actually fall on net, since lower Part B spending outweighs plan-level premium effects.
Simple to implement and scores large savings, but it's a blunt instrument — an across-the-board haircut doesn't distinguish plans with modest coding from plans coding aggressively. MedPAC-style reformers and fiscal hawks support it; insurers argue large cuts would shrink benefits and enrollment.
This is the core anti-upcoding package, and it's the one with the strongest evidence behind it. OIG found that diagnoses reported only on chart reviews produced $6.7B in MA risk-adjusted payments for 2017 — $2.7B of that from chart reviews unlinked to any actual visit. Diagnoses reported only on HRAs and HRA-linked chart reviews produced $7.5B in payments for 2023, with in-home HRAs and linked chart reviews accounting for $4.2B of that. A peer-reviewed study separately tied chart reviews to a $2.3B increase in plan payments.
MedPAC has recommended this framework since 2016. Congress can legislate the full version — the No UPCODE Act (S. 1105) and H.R. 3467 both include it — but CMS can also narrow diagnosis sources through annual rate-setting on its own. It already has: the CY2027 rate announcement finalized the exclusion of audio-only-visit diagnoses and unlinked chart-review records (with a switcher exception), estimated to save $6.84B in CY2027 alone. The plans and vendors that perform in-home assessments and retrospective chart coding are the ones directly exposed.
Unusually strong case because it targets the specific diagnosis sources most associated with questionable payments, rather than a uniform cut. The catch is implementation complexity — distinguishing legitimate care-management assessments from coding-only encounters isn't trivial. The No UPCODE Act has bipartisan sponsorship, but the MA industry is mobilized against it; Better Medicare Alliance argues it would reduce benefits and enrollment.
CMS describes RADV (Risk Adjustment Data Validation) as its primary tool for addressing MA overpayments — this is enforcement of money already paid, not a new policy fiction. A parallel MedPAC track would set completeness and accuracy thresholds for encounter data, use payment withholds and refunds, and let CMS pull provider claims directly if a plan fails those thresholds.
The 2023 final RADV rule adopted extrapolation for payment year 2018 onward and rejected a fee-for-service adjuster insurers had pushed for, with CMS estimating $4.7B in recoveries over 2023–2032. It's regulatory, but contested: a federal district court vacated portions of the rule in 2025, and CMS says that ruling currently prevents extrapolated recovery while the case is under appeal. Plans, coding vendors, and risk-bearing provider groups are the ones directly exposed. MedPAC's broader encounter-data recommendations don't have a published official score.
The upside is that it's recovering money already overpaid, not a hypothetical. The downside is slow cash realization, litigation risk, and administrative burden — program-integrity rhetoric polls well, but industry resistance and court challenges are intense.
This is where the biggest money is. Reformers argue today's benchmark system is too generous and inequitable: MedPAC says the current county-quartile system creates payment "cliffs," distorts competition, and is skewed by including FFS beneficiaries who don't have both Part A and Part B. KFF's synthesis notes higher benchmarks and rebates are what finance MA's extra benefits — but they also limit how much of private-plan efficiency actually shows up as government savings.
Structural reform is legislative, and it reaches essentially every insurer and every MA enrollee. Proposals range widely: CBO modeled a uniform 10% cut (saving $392B over 2025–2032); MedPAC recommends blending local-area FFS spending with price-standardized national FFS spending, a uniform rebate percentage of at least 75%, and a discount rate of at least 2% (estimated at $10B over five years); Paragon has proposed capping most benchmarks at 100% of traditional Medicare spending (an estimated $385B over ten years); H.R. 3467 would cut the blended benchmark to 75% of one-twelfth of the current formula starting in 2028. CBO expects plans to respond partly by rebidding rather than simply cutting benefits outright.
The most powerful budget lever on the table, and the one with the most visible distributional consequences for supplemental benefits. It's also the reform a majority of eligible beneficiaries — who are now enrolled in MA — would feel most directly, which is exactly why it's politically the hardest.
The current system weakens the link between plan efficiency and taxpayer savings, because benchmarks are tied to administratively-set FFS rates rather than actual competition. Bidding is meant to force plans to compete more on price and less on regulatory arbitrage — this is the "big rethink" version of MA payment reform, most associated with Brookings/USC and BPC proposals.
A major legislative redesign, not a rate-setting tweak. It would affect every MA plan and, depending on design, could shift relative attractiveness between MA and traditional Medicare. It's usually paired with standardized benefit designs and better plan-comparison tools, since bidding only works if beneficiaries can actually compare plans. Brookings/USC's earlier modeling (on 2015 data, assuming full phase-in) estimated roughly $10B/year in savings and a $1.4B cut to Part B premiums; BPC's more recent estimate puts a competitive-bidding benchmark at $400–500B over 2024–2033. CBO has not published a recent score for a current bidding proposal.
The most elegant long-run fix in the expert literature, because it directly changes the incentive architecture instead of patching around it. Also the hardest to build politically — standardization, regional bidding design, and beneficiary-protection rules all become contentious the moment they're specified.
The Quality Bonus Program is expensive and, per MedPAC, poorly targeted: administratively burdensome, cost-increasing, and not especially informative to beneficiaries. KFF estimates federal spending tied to the QBP will hit at least $13.4B in 2026; MedPAC separately estimates the risk-adjusted benchmark increase from the QBP adds about $16B to 2026 Medicare spending.
Congress controls the statutory benchmark bonuses; CMS controls parts of the star-rating methodology. CBO estimates eliminating benchmark increases for quality bonuses saves $94B over 2021–2028, and eliminating double bonuses saves $18B over 2019–2028. MedPAC's proposed MA Value Incentive Program (MA-VIP) — which would require legislation — is estimated to save more than $10B over five years. CRFB estimates a fully budget-neutral QBP redesign at $115–170B over 2024–2033. Notably, CMS's own CY2027 final rule moved the other direction: its finalized Star Ratings methodology changes are estimated to increase Medicare spending by $18.56B over 2027–2036 — a reminder that "star reform" can cut either way depending on design.
One of the cleaner savings options on paper, since the current program already looks weak on policy-design grounds. The counterargument is that shrinking QBP could dilute incentives for measured quality improvement and reduce rebate-funded benefits at highly-rated plans. Narrow changes like ending double bonuses poll better than full elimination; AHIP has publicly supported legislation to raise the benchmark cap so more earned bonuses get paid in full — the opposite direction from most of these proposals.
Even with current risk adjustment, very high-cost cases can still be undercompensated — which can give plans a reason to avoid certain enrollees or skimp on access for medically complex patients. A well-designed stop-loss layer is meant to close that gap without adding net federal cost.
Would need new legislative authority broad enough for CMS to implement it. H.R. 3467 authorizes the Secretary to establish stop-loss payments starting in 2028, based on auditable encounter data, explicitly budget-neutral by design. It would mainly affect plans with unusually high-cost enrollees — especially Special Needs Plans and other medically complex populations.
Done well, this improves fairness and reduces subtle risk-selection incentives. Done poorly, it partly socializes catastrophic costs and weakens plan incentives to manage expensive care tightly. Reads as reasonable inside a larger reform package; weak as a stand-alone policy on its own.
Unlike most of the other items here, this isn't primarily a savings play — it's aimed at reducing provider and beneficiary friction and limiting opaque utilization management. CMS says its 2024 interoperability rule is meant to cut payer/provider burden and improve patient access to their own health information.
Much of this is already administrative. CMS's 2024 rule (requiring MA plans and other payers to build Prior Authorization APIs and modern data-exchange standards) is already final. The bipartisan Improving Seniors' Timely Access to Care Act of 2025 would go further for MA specifically: electronic PA programs by 2028, public reporting of approval/denial rates and appeal reversals, response-time reporting, and disclosure of AI or other decision-support tools used in determinations. As of early July 2026 it had been forwarded to full committee and calendared for markup.
One of the most politically feasible reforms on this list, precisely because it's bipartisan and framed around patient access rather than payment cuts. The tradeoff is that it mostly improves process and transparency rather than directly addressing MA overpayment.
MedPAC has repeatedly said policymakers still lack robust information on how MA plans actually spend money and use supplemental benefits. These bills don't fix overpayment by themselves — they're the data infrastructure that makes future reforms easier to target and defend.
All legislative, all introduced in a cluster in June–July 2026. H.R. 9392 would require encounter data to include allowed amounts, cost-sharing, and indicators for at-home HRAs. H.R. 9644 would require plan-level publication of revenue, claims spending, non-claims spending, and MLR components in consumer-friendly form starting in 2029. H.R. 9395 would require MA organizations to report whether enrollees were signed up through an agent or broker, plus related compensation data, starting in 2028. H.R. 5243 separately advances supplemental-benefit data transparency.
Disclosure reforms are consistently easier to advance than payment cuts. Their obvious limit: transparency alone doesn't guarantee lower spending or better outcomes — it just makes the next argument easier to have with real data.
| Idea | Estimated savings | In legislation or regulation |
|---|---|---|
| Raise coding-intensity adjustment | $159B (to 8%) or $1.05T (to 20%), 2025–2034, CBO | No enacted package; CBO option menu, MedPAC agenda |
| Two years of data + exclude HRA/chart-review diagnoses | $124B (2025–2034, CBO); $6.84B in CY2027 alone (CMS, narrower version) | No UPCODE Act, H.R. 3467, CY2027 Rate Announcement |
| RADV / encounter-data integrity | $4.7B (2023–2032, 2023 RADV rule); no score for encounter-data package | CMS 2023 RADV final rule; MedPAC recommendation |
| Benchmark cut or redesign | $392B (10% cut, CBO); $10B/5yrs (MedPAC); $385B/10yrs (Paragon) | H.R. 3467, MedPAC blueprint, CBO option |
| Competitive bidding | ~$10B/yr (Brookings/USC); $400–500B/10yrs (BPC) | No current bill; expert proposal only |
| QBP / star-rating reform | $94B (end quality bonus increases); $18B (end double bonuses); CMS's own CY2027 star changes cost $18.56B | H.R. 3467, CMS CY2027 final rule, MedPAC blueprint |
| Budget-neutral stop-loss | No official score — neutral by design | H.R. 3467 |
| Prior-authorization reform | No federal savings score — burden/interoperability focus | H.R. 3514 / S. 1816, CMS-0057-F |
| Cost/MLR/broker transparency | No official score — informational | H.R. 9392, H.R. 9395, H.R. 9644, H.R. 5243 |
| Bill / rule | Status | What it does |
|---|---|---|
| S. 1105 — No UPCODE Act | Introduced Mar. 2025 | Two years of diagnosis data; excludes chart-review and HRA-only diagnoses; requires the coding adjustment to reflect actual MA/FFS coding differences |
| H.R. 3467 | Introduced May 2025 | Broad restructuring: capitated-only payments for most plans, 75% benchmark reduction formula, claims-only risk adjustment, ends quality benchmark increases, budget-neutral stop-loss, auto-enrollment, 3-year lock-in, hospice integration |
| H.R. 3514 / S. 1816 — Timely Access to Care Act | Introduced May 2025; calendared for markup Jul. 2026 | Electronic prior authorization, public PA metrics, appeal/reversal reporting, AI-use disclosure, MedPAC reporting |
| H.R. 9392 — MA Cost Transparency Act | Introduced Jun. 2026 | Encounter data must include allowed amounts, cost-sharing, at-home HRA indicators |
| H.R. 9395 — Transparency in MA Steering Act | Introduced Jun. 2026 | Reports agent/broker enrollment and compensation data starting 2028 |
| H.R. 9644 — MA MLR Transparency Act | Introduced Jul. 2026 | Plan-level revenue, claims, non-claims cost, and MLR disclosure starting 2029 |
| H.R. 5243 | 2026 committee activity | Expands supplemental-benefit data transparency |
| CMS 2023 RADV final rule | Issued Jan. 2023; partly vacated 2025, under appeal | Extrapolated audit recoveries, no FFS adjuster |
| CMS-0057-F | Issued Jan. 2024 | Prior Authorization API and interoperability requirements |
| CY2027 MA & Part D final rule | Issued Apr. 2026 | Star Ratings methodology changes (net +$18.56B), SSBCI/debit-card transparency |
| CY2027 Rate Announcement | Issued Apr. 2026 | Excludes audio-only and unlinked chart-review diagnoses from risk scores (est. $6.84B CY2027 savings) |
Money flows: benchmark → risk score → bid & rebate → quality bonus → oversight
If the goal is deficit reduction with the strongest official scores, the leading reforms are benchmark cuts or redesign, coding-intensity adjustments, and risk-adjustment tightening. If the goal is high-feasibility, near-term reform, the strongest options are narrower diagnosis-source exclusions, stronger audit recovery, and prior-authorization transparency. If the goal is full system redesign, competitive bidding and full benchmark reform matter most — analytically strongest, politically hardest.
The cleanest bipartisan opening is the anti-upcoding lane. It already has a legislative vehicle (No UPCODE Act), an administrative complement CMS has already started implementing (the CY2027 diagnosis-source exclusions), and unusually strong backing from OIG, MedPAC, and the broader overpayment literature. Even this narrower package is now central enough to MA business models that insurer opposition is intense.
Transparency bills and prior-authorization reform are much easier to move politically, but they're second-order: they improve oversight and beneficiary experience without closing the main spending gap by themselves. The most coherent bundle, on the evidence here, is anti-upcoding reform plus stronger audits plus benchmark/QBP reform plus operational transparency and prior-authorization safeguards.