CMS released preliminary CY2027 Part D bid information on July 28. Three numbers, in the order they matter.

The national average monthly bid amount is $296.05, against $239.27 for 2026 — an increase of $56.78, or 23.7%. The base beneficiary premium is $41.33, against $38.99 — an increase of $2.34, or 6.0%. And the Part D Premium Stabilization Demonstration concludes at the end of CY2026, a year earlier than the program CMS said in 2024 could run for at least three years.

The first two numbers cannot both be describing the same underlying cost trend, and they aren't. Between 2024 and 2029 the IRA's premium stabilization provision caps the annual increase in the base beneficiary premium at 6% per year. The 2027 base premium did not rise 6.0% because bids rose 6.0%. The 2027 BBP rose by approximately 6%, reflecting the statutory annual limitation rather than a comparable increase in the NAMBA.

What the cap is holding back

The base beneficiary premium is the starting point for a plan-specific basic premium; the NAMBA is the enrollment-weighted average of plan bids used to calculate the government subsidy. When bids grow at 23.7% and the beneficiary-facing base is statutorily held to 6%, the gap is reflected in higher federal Part D subsidy costs and related demonstration payments, rather than being passed directly through to beneficiaries' base premiums. That is the mechanic, and it is unremarkable in isolation.

What is not unremarkable is that the 2026 figure of $38.99 was not the number a standalone PDP enrollee actually faced. Under the demonstration, CMS applied a uniform $10 reduction to the base beneficiary premium in 2026, following a $15 reduction in 2025, and capped the year-over-year increase in a plan's total Part D premium at $50 in 2026 and $35 in 2025. The 2025 design also narrowed risk corridors; that element was eliminated for 2026.

For participating PDPs, the demonstration reduced the $38.99 statutory BBP by $10 for the applicable premium calculation, producing a demonstration-adjusted figure of $28.99. The 2027 base is $41.33, with no comparable reduction applied. That is a gap of $12.34, or roughly 43% in the starting point for plan-specific premium construction — not the 6.0% headline.

MedPAC's March 2026 report to Congress puts the realized effect at an average monthly PDP premium reduction of $26 in 2025 and $16 in 2026. GAO puts the demonstration's total cost at $9.8 billion across 2025 and 2026. Both figures are consistent with a program that was doing substantial work.

Why CMS says it ended

CMS's stated rationale is narrow and worth reading literally: bid analysis indicated that Part D plan sponsors had sufficient experience under the redesigned benefit to support the assumptions in their PDP bids, so the agency will discontinue the demonstration to return the program to traditional market conditions in 2027.

That is a statement about sponsors' actuarial confidence in pricing. It is not a statement that costs have moderated. The 23.7% NAMBA increase is CMS's own arithmetic, drawn from the bids CMS just reviewed.

The demonstration was always framed as a bridge across IRA benefit redesign — the out-of-pocket cap and the shift of liability onto plan sponsors. It bridged. PDP enrollment rose from 22.8 million in 2024 to 24.9 million in 2026 rather than collapsing, which is what the demonstration was designed to prevent. Whether the underlying cost pressure it was bridging over has resolved is a separate question, and rising drug prices, GLP-1 utilization and specialty spend are not on a schedule that ended in December.

The structural spread nobody prices out of

Here is the number that frames 2027 competitively. In 2026 — with the demonstration in place — the enrollment-weighted average monthly standalone PDP basic premium was approximately $36, against approximately $8 for the drug-coverage component inside a Medicare Advantage prescription drug plan. More than four to one.

That gap is not a pricing accident. MA-PD sponsors can use rebate dollars generated on the medical side to buy down drug premiums. Standalone PDP sponsors have no such mechanism; a PDP has to price its drug benefit on the drug benefit's own economics. The demonstration was, functionally, a federal substitute for the rebate advantage MA-PDs enjoy structurally. Removing it does not level the field. It restores an asymmetry that the demonstration had been partially offsetting since 2025.

The migration question for 2027 is therefore not whether PDP enrollees will face larger increases than in recent years. It is how many of them, on seeing a plan-specific number in October, move to an MA-PD — and what that does to the risk composition of both books.

Note that the widely repeated projection of roughly 45% of PDP enrollees facing $11–20 monthly increases traces to information the administration provided to a news outlet, not to a published analysis.

On the MA side, Humana disclosed on its Q2 2026 call an intention to exit 2027 plans covering a reported ~600,000 members, described as concentrated in plans rated 3.5 stars or below for the 2027 bonus year, with an expectation of recapturing some of that membership.

The catalyst is a date

CMS is expected to release the 2027 MA and Part D landscape files in September, along with final plan-level premiums and related information. Annual Enrollment opens October 15.

Until the landscape files publish, every 2027 premium figure in circulation is a projection built on the NAMBA and the base premium. After they publish, plan-specific premiums are observable, the PDP-to-MA-PD spread is measurable at the plan level rather than the average, and the enrollment migration stops being a thesis. Three weeks from now, the first read on the redesigned benefit without a federal cushion is a data file rather than an argument.

Watch the spread, not the headline.

This is analysis, not investment advice. Continuum does not make recommendations to buy or sell securities.