Humana's Benefit Ratio Rose 130 Points on Three Separate Mechanisms

Humana's Insurance segment benefit ratio reached 91.2% in the second quarter of 2026 against 89.9% a year earlier, and the Form 8-K attributes the move to three named causes: the Bonus Year 2026 Star Ratings revenue headwind, higher-cost new Medicare Advantage members, and lower prior-period reserve development (Humana Form 8-K, July 2026). Those are a revenue effect, a morbidity effect, and a reserving effect stacked into one ratio. A 2027 bid can price one of them, must estimate the second, and cannot assume the third repeats.

What the Quarter Reported

Revenue reached $40.9 billion against $32.4 billion a year earlier, carried by Medicare membership growth, higher per-member premiums, and CenterWell expansion, with total medical membership at roughly 17.9 million against 14.8 million (Humana 8-K, July 2026). GAAP earnings per share came in at $5.73 against $4.51, adjusted EPS at $7.61 against $6.27, and adjusted pretax income at $1.248 billion against $1.017 billion. Operating cash flow ran $3.22 billion year to date against $1.60 billion, though roughly $1.05 billion of that reflects a Medicaid state-directed payment that settled just after quarter end, a timing effect rather than a run-rate improvement.

The Insurance segment produced $39.1 billion of revenue at a 91.2% benefit ratio and a 7.1% operating cost ratio. Consolidated, the benefit ratio was 91.1% against 89.7%. The consolidated adjusted operating cost ratio improved to 9.7% from 10.9%, which is the quarter's genuinely favorable operating datapoint and the one most likely to persist, since it reflects scale against a fixed administrative base rather than any claims assumption.

Mechanism One: A Revenue Headwind Wearing a Benefit-Ratio Costume

The Star Ratings effect is the one most often misread, because it reaches the benefit ratio through the denominator. Quality bonus payments are revenue. When a plan's Star Ratings fall, CMS pays less premium for the same enrolled member, so premium declines while the medical expense associated with that member does not. The benefit ratio rises even if utilization, unit cost, and coding intensity are all perfectly flat. Nothing about the underlying medical experience deteriorated; the divisor shrank.

The magnitude is visible in the guidance rather than the quarter. Humana affirmed 2026 adjusted EPS of at least $9.00 against 2025 adjusted EPS of $17.14, a decline the company ties principally to the Bonus Year 2026 Star Ratings headwind (Humana 8-K, July 2026). Roughly 20% of Humana's MA members sit in plans rated 4 stars or above for the relevant period, against a book that historically ran far higher. For a pricing actuary this distinction is not academic. A denominator effect is fully knowable at bid time, because the Star Ratings that drive Bonus Year 2027 revenue are already determined. It belongs in the revenue build of the bid as a known quantity, not in the trend selection as an uncertainty.

The GAAP-versus-adjusted split makes the same point from the other direction. Humana cut 2026 GAAP EPS guidance to at least $6.52 from at least $8.36 while affirming the adjusted figure, its second GAAP reduction of the year. When a company reaffirms an adjusted number and cuts the unadjusted one twice, the delta is the part of the story the non-GAAP presentation is designed to set aside. An actuary building a margin bridge should work from the reported benefit ratio and the reported revenue, because the bid is priced in GAAP dollars.

Mechanism Two: New Members Cost More, and the Growth Was Enormous

The second driver is mix. Humana added roughly 3.1 million medical members year over year, and management has reaffirmed individual MA membership growth of approximately 25% over 2025. New Medicare Advantage members are structurally more expensive in their first year for reasons that have nothing to do with underwriting quality: risk scores lag actual morbidity because diagnoses have not yet been captured through a full assessment cycle, care is not yet directed into preferred networks, and pent-up utilization from a prior plan's access constraints surfaces early.

The pricing consequence is that the first-year cohort's benefit ratio is not the book's steady-state benefit ratio, and a 2027 bid built on a blended 2026 experience period embeds a distortion whose size depends on how much of the book was new. With growth near 25%, that share is large enough to matter. The correct approach separates the experience period into new-member and renewal cohorts, prices each on its own emergence pattern, then reweights to the projected 2027 mix, which will differ again because the announced plan exits reshape the book. Humana expects those exits to affect roughly 600,000 members, about 8% of its 7.2 million MA enrollees, with management targeting recapture of about 40% into retained plans (Healthcare Dive, July 2026). A book losing 8% and recapturing 40% of the loss into different plans is a book whose 2027 mix cannot be read off 2026 actuals.

Mechanism Three: Lower Prior-Period Development Is the One That Should Worry Reserving

The third named driver is the quietest and the most consequential. Lower favorable prior-period development means the current quarter received less benefit from reserves established for earlier periods. Two very different situations produce that line. The benign one: prior reserves were less redundant because they were set more accurately, which is a sign of improved reserving discipline and simply removes a cushion. The unwelcome one: claims are emerging faster or larger than the prior estimates assumed, and redundancy is being consumed by adverse experience rather than released as profit.

The 8-K does not distinguish between them, and neither can an outside reader from a single quarter. What an actuary can do is watch the pattern. If lower favorable development persists across the next two quarters while the completion factors underlying incurred-but-not-reported estimates hold steady, the benign reading is supported. If it persists while completion factors are being lengthened, the reserve cushion is thinning against real emergence, and a 2027 bid built assuming any contribution from prior-period releases is pricing on a subsidy that will not arrive. Prior-period development is never a legitimate input to a forward-looking bid margin. It appears in reported results and quietly shapes expectations, which is precisely why naming it as a driver of a 130-basis-point benefit ratio move deserves more attention than it usually receives.

CenterWell Is the Offsetting Lever, and Its Contribution Is Hard to Credit

Against those three pressures Humana is pushing a fourth mechanism in the other direction. CenterWell, the company's health services arm, produced $6.8 billion of revenue and $514 million of adjusted income from operations in the quarter, expanded through the acquisitions of Florida providers The Villages Health and MaxHealth, and grew CenterWell Senior Primary Care by 130,900 patients, or 27%, year to date. Management reported that medical cost trends came in as expected with slight favorability in the inpatient setting, and specifically noted that the favorability concentrated among members receiving care from value-based providers.

That last claim is the one an actuary should hold at arm's length until it is measured properly. Members attributed to owned or value-based primary care are not a random sample of the book. They are geographically concentrated, disproportionately in Florida, and selected through channels that correlate with engagement, functional status, and continuity of care. All of those independently predict lower inpatient utilization. Separating the care-model effect from the selection effect requires risk-adjusted cohort comparison within market, not a topline observation that value-based members ran better. Humana has the data to do this; outside actuaries do not. The company's December 10, 2026 investor update, which it has framed as marking progress against the 2025 Investor Day plan through 2028, is the natural venue to show the risk-adjusted version, and its absence there would itself be informative.

The stakes are the margin commitment. Chief financial officer Celeste Mellet told investors the priority is "to make the necessary progress to remain on track to deliver on our [2028] commitment of returning to a sustainable margin of at least 3%" (Healthcare Dive, July 2026). Getting from current performance to 3% requires the Star Ratings denominator to recover, the new-member cohort to season into renewal economics, prior-period development to stabilize, and CenterWell's cost advantage to be real and scalable. Those are four independent bets, and the bid actuary controls the pricing assumption on only two of them.

What This Prices Into a 2027 Bid

The practical translation is a discipline about which effects belong where. The Star Ratings revenue impact is knowable and belongs in the revenue build at its determined value, with no probabilistic hedge. The new-member cost differential belongs in the morbidity and trend assumptions, cohort-separated and reweighted to projected mix rather than carried at the blended 2026 average. Prior-period development belongs nowhere in the bid, and its recent direction is a reserve-adequacy signal to monitor rather than a margin input. CenterWell's cost advantage belongs in the bid only to the extent it survives risk adjustment within market, and a carrier that credits an unadjusted differential is pricing a selection effect as though it were a care-model effect.

Humana's stock fell almost 8% on the print despite beating on earnings and revenue, which tells you the market read the guidance rather than the quarter. Actuaries should read the driver list. A benefit ratio that moves 130 basis points on three mechanisms with three different persistence profiles is not one number to trend forward; it is three assumptions to set separately, and the 2027 bid deadline arrives before the December investor update resolves any of them.