The 2027 Part D Number That Matters Is Not the Average. It Is the Spread.

CMS ended the Part D Premium Stabilization Demonstration on July 28. The figure everyone is quoting, a 24 percent jump in the national average bid, is close to the least useful number in the announcement. Here is the one that will show up in your book.

You have already seen the headline. Probably some version of "Medicare drug subsidy ends, premiums expected to rise," and depending on where you read it, with a political frame attached on one side or the other.

None of that coverage told you the thing that changes how you should work your book between now and December 7. So here it is up front: the constraint that came off is not really the subsidy. It is the ceiling.

Three mechanisms, not one

Nearly every write-up of this announcement has described the demonstration as a subsidy. It was three things, and only one of them was subsidy-like.

One: the uniform base beneficiary premium reduction

$15 in 2025, reduced to $10 in 2026, gone in 2027. This is the piece getting the coverage. It is also the smallest piece of the story.

Two: the plan-level year-over-year total premium increase limit

$35 in 2025, raised to $50 in 2026, gone in 2027. This was plan-specific, not a market average. A participating plan simply could not move its total Part D premium more than that amount from one year to the next, regardless of what its own cost experience said it needed. This is the mechanism that matters most to you.

Three: narrowed risk corridors

In place for 2025, eliminated for 2026, gone in 2027. CMS took one additional downside risk during the first transition year. This one already left a year ago and its effect is partly baked into what we saw in the 2026 bids.

Think about what mechanism two was actually doing. A plan whose underlying economics called for a substantially higher premium in 2026 could not simply price to that number. It was held under a ceiling. That gap between what a plan needed to charge and what it was permitted to charge did not disappear. It was deferred.

For 2027, the ceiling is gone.

What the bid trajectory actually shows

The national average monthly bid amount is worth looking at as a series rather than a single-year jump:

The 2027 figure is roughly 24 percent above 2026, an increase of about $56.78 per month.

And it is still not a premium. NAMBA is an enrollment-weighted average of applicable plan bids for the basic Part D benefit, used in calculating the government’s direct subsidy to plans. Nobody pays it. Repeating that 24 percent figure to a client, or in a marketing piece, is the single fastest way for a tenured agent to be publicly wrong about something they should know better than.

What the trajectory does tell you is that the underlying cost of delivering the basic benefit has been climbing steeply since the redesign began phasing in. For two years, the demonstration absorbed part of the distance between that cost curve and what appeared on a client’s statement.

Why the average will understate the disruption

Here is the analytical point, and it is the reason this article exists.

When you release a plan-level ceiling, you do not shift the entire distribution evenly. You release the plans that were pressed against it. Plans with comfortable margins and stable books were never bound by the $50 limit in the first place and have no particular reason to move much. Plans that were bound by it now have room they did not have before.

Consider two markets with an identical average premium increase of $8:

• One where nearly every plan moves between $5 and $11.

• One where most plans move $3 and a handful move $40.

Same average. Completely different book of business. In the first market, a light-touch AEP is defensible. In the second, the clients concentrated in those few plans are the entire year.

CMS has publicly indicated that most beneficiaries should see modest premium increases in 2027, and that is entirely plausible as a statement about an average. It is not a statement about the client sitting across from you, and it is specifically not a statement about clients enrolled in the plans that were most constrained.

The useful expectation for 2027 is not "premiums up X percent." It is "wider dispersion than 2025 or 2026." Which means the comparison you run at the kitchen table carries more weight than it has in two years.

The Medicare Advantage blind spot

Everything above concerns standalone PDPs. The demonstration only ever applied to standalone PDPs, and that fact is being repeated in a way that leads to a reasonable but incomplete conclusion: MA-PD clients are unaffected, move on.

Scope your preparation that way and you may miss where this actually lands.

Part D economics feed MA-PD bids. Plans re-pricing the drug component for 2027, without the demonstration operating in the background of the standalone market, are making decisions in a different environment than they were a year ago. When drug benefit costs move, plans have more than one lever: the Part D portion of the premium, cost sharing, and the supplemental benefit package funded by rebates.

So the tenured version of this is not "check your PDP list." It is: watch for your MA-PD exposure to appear in the benefit summary rather than the premium line. Read the Annual Notice of Change documents on that book as carefully as you read the premium, because a stable premium with a thinner ancillary package is still a disrupted client.

What still constrains plans

Two things did not go away, and leaving them out would overstate the case.

The statutory 6 percent cap remains

The Inflation Reduction Act limits annual growth in the national base beneficiary premium to 6 percent per year through 2029. The 2027 BBP is $41.33, which is exactly 6 percent above the 2026 figure of $38.99. That provision is entirely separate from the demonstration and it stays. It also does not do what many agents think it does, since it governs the national figure rather than any individual plan premium.

CMS still reviews and negotiates bids

CMS retains the ability to reject or require revision of bids it considers problematic, and in the 2026 cycle it signaled it would push back on standalone PDP bids that were significant outliers. A plan freed from the $50 ceiling is not free to price anywhere it likes.

Which is precisely why the defensible framing is "greater variation is possible" rather than "premiums are about to spike." The first is analytically supportable and keeps you credible. The second is a prediction you cannot back and may have to walk back.

Working the book: August through December

Now through mid-September, before the landscape

• Segment the book into three lists: standalone PDP, MA-PD, and LIS or dual eligible. Three different conversations, three different levels of exposure.

• Identify concentration. If a meaningful share of your PDP book sits in one or two plans, your actual exposure is those plans’ 2027 pricing, not the market’s average. That is a specific thing you can check in September rather than a general worry.

• Do not run a proactive premium-warning campaign. You have no plan-level data and neither does anyone else. This is the discipline that separates a calm December from a defensive one.

• Do runlight-touch AEP review outreach that makes no claim about direction. "I want to review your coverage before AEP" is always true and never wrong.

• Get anyone on your team who answers a phone fluent on the difference between the base beneficiary premium, the national average bid, and a plan premium. Those three numbers are circulating in the press interchangeably and the confusion will arrive as client questions.

Mid to late September, when CMS publishes the landscape

• Pull the landscape and look at dispersion in the states you actually write, not just the national average.

• Map your concentration list against the new pricing first. That tells you the size of your year in an afternoon.

• Reconcile against Annual Notice of Change letters as clients start bringing them in.

October 15 through December 7

• Re-run every standalone PDP comparison. All of them, including the clients who have been in the same plan for four years. Two quiet years are exactly what makes that feel unnecessary, and that is the trap.

• On MA-PD, compare benefit summaries with the same rigor you apply to premiums.

• Document the comparison you ran. In a year with wider variation, the file note matters more.

Talking about this without stepping on compliance

The public coverage of this decision is unusually political, with a bailout framing on one side and a benefit-cut framing on the other. Neither belongs in your client conversations, your social posts, or your seminar deck.

Stay on mechanics and dates:

• What the program was and that it ends after 2026.

• That plan-level numbers arrive with the CMS landscape in mid to late September.

• That the client’s own plan will send them a notice about changes.

• That comparing options matters more this year than it did the last two.

Avoid naming specific carriers, predicting specific dollar amounts, characterizing CMS’s decision as good or bad, and implying anything about a client’s plan before you have that plan’s numbers. Route every conversation to plan-specific documents.

The bottom line

For two years the standalone Part D market has been operating under a ceiling. The subsidy piece is what is getting the headlines. The ceiling is what will show up in your book.

2027 is the first year since the redesign in which a plan’s premium reflects its own bid without an artificial limit on how far it can move. That argues for one thing above all others, and it is not a warning email in August. It is re-running every comparison in October, and not letting two quiet years talk you out of it.

This article is for licensed insurance professionals and is provided for educational purposes. It is not intended for distribution to Medicare beneficiaries. Figures reflect CMS information published as of August 5, 2026 and are subject to change. Plan-specific premiums and benefits for 2027 are not final until CMS publishes the 2027 landscape.