Jackson Hammond is a Senior Policy Analyst at Paragon Health Institute. He has been active in the federal and state health policy space since 2017.
Prior to joining Paragon, Jackson was a health care policy analyst for American Action Forum (AAF). While at AAF, his work focused on payer issues including private insurance, Medicare, and Medicare Advantage. Furthermore, Jackson wrote extensively about the 340B Program and contributed to AAF’s research on a variety of drug pricing issues.
On September 26, 2025, the Centers for Medicare and Medicaid Services (CMS) released premium information on Medicare Part D, the prescription drug benefit used by seniors across the country. Part D has undergone significant changes in the past two years as a result of the Inflation Reduction Act (IRA). As Paragon has previously explained, the IRA’s Part D changes have resulted in large increases in premiums and federal subsidies, as well as several insurers dropping out of the stand-alone Part D market.1
These unsurprising negative IRA impacts caused the Biden administration to implement what is effectively a $5 billion bailout by giving taxpayer dollars to insurers to prevent them from raising premiums on seniors in an election year. The bailout, dubbed the Part D Premium Stabilization Demonstration program,2 was announced in July 2024 and took effect in 2025. On July 29 of this year, CMS notified the public about sensible changes it will be making to reduce the size of the bailout. In 2026, the average premium that enrollees will pay is projected to be lower than it was last year—but only because per-enrollee subsidies increased and CMS rejected certain bids from insurers. This brief will cover the updates in Part D and the impacts the IRA’s Part D redesign has had on the program.
Average 2026 enrollee premiums for Part D stand-alone prescription drug plans (PDPs, as opposed to combination Medicare Advantage + Part D plans) are lower than last year, with a projected decrease to $34.50 in 2026 from $38.31 in 2025.3 At least some of this decrease is likely because CMS decided to reject some PDP bids that “failed to address concerns regarding significant year-over-year premium increases” and were also market outliers relative “to similar plans in the same region.” However, the average premium decrease disguises a large increase in government subsidies to insurers for PDPs. Total subsidies to PDPs increased 31.2 percent from $185.75 per enrollee in 2025 to $243.78 per enrollee in 2026.4
These increases in government subsidies are part of a trend. Total subsidies per PDP enrollee were relatively constant between 2016 and 2023, ranging from $98.62 to $103.14.5 Then, in 2024, total subsidies jumped 30.1 percent to $134.16.6 In 2025, total subsidies increased another 38.5 percent to $185.75,7 and in 2026 total subsidies leaped 31.2 percent to $243.78.8
As explored below, these increases are the result of provisions in the IRA causing the average monthly plan bid (see text box) to increase 589 percent from $34.71 in 2023 to $239.27 in 2026.9 The IRA’s Part D redesign keeps the starting point used to calculate a portion of beneficiaries’ premiums artificially low, rising only 19.1 percent—from $32.74 to $38.99—in the same time period, which means beneficiaries’ actual premiums and subsidies must rise to cover the difference.
Every year, CMS announces what premiums will be for stand-alone prescription drug plans in the Part D program. Premiums are determined after CMS calculates a “base beneficiary premium” using a statutory formula. This base beneficiary premium is not what enrollees pay but is a starting point that plans must use to calculate how much they will need to charge enrollees in premiums. Each plan then submits a “bid,” which is in essence the price at which a plan will cover an enrollee. These bids are effectively averaged together by CMS to produce the National Average Monthly Bid Amount (NAMBA). CMS then compares the base beneficiary premium to the NAMBA to help determine the “direct subsidy” for each plan, which is essentially the per-enrollee subsidy the plan will receive upfront. The total subsidy includes both this direct subsidy and a back-end subsidy known as “reinsurance.” Based in part on plan bids’ assumptions regarding projected membership, CMS also publishes the “average monthly premium” for the next year, which is what enrollees can expect, on average, to pay personally in premiums for their Part D plan.
The sharp PDP premium and subsidy increases stem directly from structural changes enacted in the IRA. The IRA increased plan liability for costs while reducing enrollee liability for drugs throughout the year. Figures 1 and 2 show the basic drug payment structure changes that began in 2025 for the standard benefit, in particular the reductions in patient out-of-pocket liability.10 Where PDP enrollees previously had to cover 25 percent of the cost of drugs until hitting around $3,000 in direct out-of-pocket spending, enrollees will now pay nothing after spending $2,000 in out-of-pocket expenditures this year and $2,100 in 2026.11 As the figures below demonstrate, PDPs now must cover significantly more of the costs that subsidies do not pay, which means higher premiums.
However, premiums began increasing as a result of the IRA even before these changes took effect. Beginning in 2024, the IRA eliminated copayments and coinsurance for enrollees who reached the old catastrophic coverage phase and limited the starting amount that premiums are calculated by (see text box) to increases of only 6 percent annually through 2029 (a move done unilaterally by Democrats in Congress using reconciliation in an attempt to stave off immediate premium increases on enrollees after implementation). Because PDPs in 2024 now had—as a result of the IRA—more costs to cover in the catastrophic phase, their bids increased. But because the starting amount for premium calculations was limited to a 6 percent annual increase, direct government subsidies to plans had to increase to cover the difference. It should not have been a surprise, then, that total subsidies jumped considerably starting in 2024.
In addition to higher premiums and subsidies, the IRA caused PDPs to leave the market at an accelerated rate: The average enrollee has only 11 PDPs to choose from in 2026, compared to 21 in 2024 and 30 in 2021.12 Overall, the total number of PDPs decreased from 709 in 2024 to 360 in 2026. The reduction in plans is occurring at the same time that major insurers are eliminating commissions for brokers who enroll individuals in their plans,13 meaning seniors will have less choice in plans and less help in selecting those plans.
CMS launched the three-year Part D demonstration program in July 2024. The demonstration had three components: (1) It reduced base beneficiary premiums by $15 from what they otherwise would have been without the demonstration, (2) it capped total premium increases at $35, and (3) it adjusted risk corridors—which in essence limit both losses and profits of plans—in Part D to increase the government liability for PDPs that face losses (see Figure 3), including those resulting from (1) and (2).
The statutory authority for this demonstration came from Section 402 of the Social Security Act,14 which provides broad authority for CMS to waive requirements for payment in Medicare and Medicaid in order to create demonstration programs that “increase the efficiency and the economy” of providing services, reduce costs, and improve the provision and utilization of services.
However, this “demonstration” program demonstrates absolutely nothing, let alone efficiency and economy. There is no control group because 99 percent of plans participated in it.15 The timing of the decision to create this demonstration—just before the 2024 presidential election—raises questions about whether this was an election year ploy meant to buoy the chances of a second Democratic presidential term and avoid fallout from a policy that was passed on a purely partisan basis.
As a result of the Part D demonstration, the actual average monthly premium significantly decreased in 2025 even as the starting point used to calculate that premium slightly increased from 2024. This average monthly premium—what enrollees can actually expect to pay after PDPs factor in additional benefits and coverage beyond what is required by Part D—dropped 29 percent from $53.95 in 2024 to $38.31in 2025.
By CMS’s own admission in 2024, the significant increase in pre-demonstration premiums was expected as plan liability increased per the IRA.16 The Part D demonstration nevertheless kept premiums from rising—but only because of the “demonstration” infusing $5 billion in taxpayer funds, according to the Congressional Budget Office (CBO).17 Total government subsidies for PDPs have increased from $27.5 billion in 2023 to $51.7 billion in 2025—an increase of roughly 88 percent in just two years. Viewed as a whole, this timeline is a troubling set of machinations undertaken by the Biden administration: pass a bill in Congress on a partisan basis that purports to help Medicare beneficiaries by capping their payments at pharmacies throughout the year but also inevitably increases these beneficiaries’ costs in premiums and taxpayer debt in subsidies, then—just before an election—disguise the increase in premiums and shift additional costs to taxpayers, resulting in fewer choices and less competition and therefore even greater costs to all in the long term.
In July 2025, CMS announced that it would be adjusting the demonstration to facilitate “the program’s return to operating under regular market conditions.”18 CMS is reducing the base beneficiary premium reduction from $15 to $10, increasing the limit on total Part D premium increases from $35 to $50, and eliminating the altered risk corridors from the first year. CMS reasons that, because PDPs have a better understanding of how “IRA benefit changes will impact utilization and costs,” the PDPs will be better able to adapt their plans. The IRA has thrown Medicare’s prescription drug market into turmoil, unmooring it from its original market-based foundations. The Biden administration’s demonstration has made the problem worse by increasing taxpayer exposure to cover the rising costs from the IRA. The sensible phase-down by CMS creates a glide path to more normal Part D market conditions over the long run.
These more normal market conditions are responsible for the increase in federal subsidies in 2026 (when the IRA also requires CMS to limit enrollee premium increases). The direct subsidy increases due to increased bid amounts were inevitable as a mathematical consequence of the IRA, but the effects on enrollees were masked by the Biden administration’s demonstration program, which merely delayed the IRA’s costs from becoming apparent. This phase-down is a pragmatic and necessary decision to reduce unsustainable taxpayer exposure without leaving a cliff where Part D premiums will spike. In other words, if and when premiums and subsidies inevitably go up, that is a result of the IRA and not the phase-down by CMS.
As noted above, CMS also limited premium increases by rejecting PDPs with higher bids that they considered regional outliers. While CMS has authority to require certain contractual “terms and conditions” and regulations to reject bids with “significant increases in cost sharing,”19 it has never been used in this way before this year. Moreover, it is unclear how and whether this authority applies to increases in standalone Part D premiums, and there are no regulations or guidance from CMS explaining what significant means or the approach it will take in making these decisions in this context narrowly focused on premiums. Regardless, this rejection of plan bids raises questions about CMS setting a precedent to reject bids the agency considers to be too high. By banning plans that CMS views as too expensive, overall subsidies are somewhat lower, but CMS is substituting its judgment for those of enrollees. If the plans are not worthwhile, PDP enrollees will simply not sign up for them, and competition will keep prices low more effectively and sustainably than government decision-making and distortion would.
CMS has taken the correct course in phasing down the Part D demonstration, and it should continue to phase down the demonstration in 2027 and allow it to expire thereafter as planned by the Biden administration, if not sooner. However, without legislative correction, the Part D market will remain distorted—driving higher costs for taxpayers and fewer choices for seniors. Restoring cost discipline and consumer choice to Part D will require legislation to align with the original market-based vision that made the program successful for two decades.
To improve Part D in this manner, Congress should fix the many flaws in the IRA. And to deal with the immediate issue that gave rise to the Part D demonstration, Congress should consider gradually raising the out-of-pocket limit to $3,100, which was the limit in the original bipartisan Part D redesign.20 Most seniors would be unaffected. Former CBO director Douglas Holtz-Eakin has estimated that less than 3 percent of Medicare enrollees have more than $2,000 in out-of-pocket costs.21 This will enable a smoother exit from the demonstration without subjecting enrollees to significant out-of-pocket cost increases. It could also be a catalyst for a broader set of market-based changes to Part D that empower beneficiaries with more choices from less government micromanagement regarding benefit structure, as well as limit current and future taxpayers’ payments to ensure greater competition within a capitated amount that drives efficiency.
Subscribe
Sign up now for your health policy updates.
