Your Revenue Plan Rests on a Number That Moved Five Points Between Draft and Final
In January 2026, CMS proposed an effective growth rate of 0.09% for Medicare Advantage — essentially flat. Health insurers lost 14–20% of their market value in a single session. In April, the final number came in at 5.33%. The year before, the same number moved from 5.93% to 9.04% between notice and final. If a revenue projection in your building still carries Advance Notice numbers as final, it is betting the plan on a draft. Six things from building an instrument that decomposes rate projections.
1. The machine has four dials. An MA benchmark is county FFS spending × a quartile percentage (95–115%, with the lowest-spending counties paid the highest percentage) × any star bonus. Bid below it, keep 50–70% of the savings by star tier; bid above it, the member pays the excess. Everything else arrives through the risk score that multiplies this number — which is where the trouble lives.
2. The yardstick is public and unflattering. MedPAC's 2026 estimate: MA payments run about 114% of what the same members would cost in fee-for-service — $76 billion above — roughly 11 points from favorable selection and 4 from coding intensity that survives the statutory 5.9% adjustment, which has sat at its legal minimum since 2019. Any MA revenue story gets told against that number, not around it.
3. v28 measures the plan, not the model. CMS priced the risk-model transition at −3.12%. The peer-reviewed insurer spread runs from −17.9% to roughly zero — and it correlates with how aggressively each plan coded under the old model. Two identical-size plans will feel utterly different pain, and the difference is a measurement of coding posture. No plan has published its own number. Unmeasured is not small; it is unknown with a known distribution.
4. The new exclusion is already in the walk. The CY2027 announcement removes diagnoses from audio-only encounters and chart reviews unlinked to service records — a −1.53% line. Revenue booked from those channels is revenue the payer has already announced it will not pay.
5. On the Medicaid side, the margin story is now. Industry-standard tracking puts Medicaid MCO underwriting margins at −0.1% in 2025, with roughly half of plans underwater — after the fattest years in the program's history. Every national carrier claims rates lag post-unwinding acuity. Notably: no independently quantified rate-vs-acuity gap exists anywhere. The argument is real as negotiation posture and unproven as a number.
6. Nobody has checked the bids in thirteen years. The only independent bid-versus-actual evidence is GAO's 2008 and 2013 work: plans projected 4.1% margins and earned 6.6%; special-needs plans earned 8.6% against 6.2% projected. Since then — nothing. Every margin projection in the industry stands on a thirteen-year evidence vacuum, and the last audit found the errors ran systematically in the plans' favor.
We built these into an instrument: the benchmark-bid-rebate arithmetic, the published walk beside the plan's assumptions, the v28 posture check, the Medicaid certification mechanics — every threshold cited, every declared absence stated. Browser-only; nothing entered leaves the machine.
A rate assumption nobody re-derived is a wish with a decimal point.
What number in your current forecast is still, technically, a proposed number?