Humana

IThe business

Humana is a health insurer in Louisville, Kentucky, with about 67,060 employees. At the end of 2025 it covered about 15 million people in medical plans, 4.6 million of them military families whose care it only administers, and by 30 June 2026 the total had risen to 17.9 million.²⁶ Its business has two parts. The Insurance segment sells Medicare Advantage, Medicare drug plans, Medicaid and a few smaller products. The CenterWell segment runs primary-care clinics for seniors, a mail-order and specialty pharmacy, and home health, and sells most of its services to the insurer.²

The way I think about it: the federal government pays Humana a fixed amount each month for every senior who joins one of its Medicare Advantage plans. The amount is set by CMS, the agency that runs Medicare, and is adjusted for how sick the person is. Humana then pays that person's doctors, hospitals and drug bills and keeps what remains. If bills run ahead of the payment, profit disappears fast. If they run behind it, profit is large. That is why a business with $129.7B of revenue earned $2.7B of operating income in 2025, a margin of 2.1% C (calculated).²

Where the money comes from

The 2025 10-K gives revenue by product. Individual Medicare Advantage premiums were $90.4B, 70.3% of premiums and services revenue. Group Medicare Advantage, sold to employers and public bodies for their retirees, was $9.0B. Stand-alone Part D drug plans were $6.8B, state Medicaid contracts and other $14.5B, Medicare Supplement $1.1B and specialty products $1.0B.² In all, 83% of 2025 revenue came from federal government contracts, and 14% came from one place: the 1.0M individual Medicare Advantage members in Florida, who produced $17.8B.² The pharmacy benefit manager, Humana Pharmacy Solutions, serves Humana's own members and is eliminated in the accounts.⁴

The shape of the last year matters. Humana ended 2025 with 412,500 fewer individual Medicare Advantage members than a year earlier, 5.25M against 5.66M, because it left plans and counties that lost money.² Then in the 2026 enrollment season it did the reverse. It added about 1.0M members in the annual election period, and 1,204,400 in the first half of the year, to reach 6.45M.⁴⁹⁶ Over 70% of the new sales were people switching from rival plans, and nearly 30% were former Humana members who had left.⁴ Retention, the share of members who stay, improved by more than 500 basis points, beating the 350 to 450 basis points the company had set as a goal for 2028.⁴

I read this as a deliberate swing. In 2025 Humana chose margin over members. In 2026 it chose members, after the government's quality scores, the Star Ratings, cut what it earned on each one. New members cost more than members who have stayed for years, so the 2026 benefit ratio is guided to 92.75% for the Insurance segment, up from 90.4% in 2025.³² The bet is that those members become profitable in 2027 and 2028. That bet is the whole case, and the rest of this page tests it.

The insurer

Two statistics describe an insurer. The benefit ratio is medical and drug costs as a share of premiums. The operating cost ratio is administrative cost as a share of revenue. For Humana the benefit ratio was 90.2% in 2025 and the operating cost ratio 12.0% on the filing's definition.² In the second quarter of 2026 the Insurance segment's benefit ratio was 91.2% and its operating cost ratio 7.1%, a fall of 1.2 points from a year earlier as revenue grew faster than costs.⁹ The company guides the whole-year Insurance benefit ratio to 92.75%, plus or minus a quarter of a point, and the third quarter to slightly above 94%.³⁹ The ratio rises through the year because members reach their deductibles and use more care, and because the drug benefit changed under the Inflation Reduction Act.² That is also why 2026 earnings are almost all in the first half: adjusted EPS was $17.91 in the first six months, and the third quarter is guided to about a $1.00 loss.⁵⁹

Medicare Advantage pays by risk adjustment. CMS gives each member a risk score built from the diagnoses doctors submit, and pays more for sicker members. That makes coding of diagnoses a profit lever and a legal exposure. In 2024 CMS began a three-year phase-in of a new model, called V28, which pays less for some diagnoses; 2026 is the final year of the phase-in.⁶ The 10-Q says Humana's actuarially certified bids through 2027 keep its position that CMS is required to apply an adjustment for fee-for-service coding differences in any audit of past payments.⁶

CenterWell

CenterWell had $22.5B of revenue in 2025, of which $17.7B came from Humana's own insurance plans, and income from operations of $1.34B, about the same as in 2024.² Its three lines are these.

• Primary care. Clinics for seniors under the Conviva and CenterWell Senior Primary Care names. At 30 June 2026 there were 398 centers and about 622,000 patients, up 27% since the start of the year, partly through the purchase of The Villages Health in the fourth quarter of 2025 and MaxHealth in Florida in February 2026 for about $908M in cash.⁹⁶ The clinics take on the cost of care for each patient in return for a fixed payment, and V28 cut that payment, so the company is working to offset it with productivity.⁹
• Pharmacy. CenterWell Pharmacy fills prescriptions by mail and runs a specialty pharmacy, with $11.7B of 2025 revenue from Humana's own plans and $1.2B from others.² Mail order is a way to steer members to the cheapest source and to keep the margin inside the group.
• Home. CenterWell Home Health in 37 states, a value-based home care unit called OneHome, and a 35% minority stake in Gentiva Hospice that Humana has agreed to sell, valued at about $900M.²⁹

Humana does not own all of the clinics it runs. 146 clinics were held in a partnership with the private-equity firm Welsh, Carson, Anderson & Stowe at the end of 2025, with put and call options on the stake.² The options are valued each quarter and swing the reported profit by hundreds of millions: they were $513M of expense in 2025 and are one reason GAAP EPS of $9.84 is so far below the adjusted $17.14.² On 13 October 2026 Humana announced that BlackRock, through HPS, will buy the WCAS stake, valued at about $2.8B, and form a new joint venture covering nearly 150 clinics with no put option. Humana can buy clinics from it from 2028, and BlackRock can force a sale of what is left after 2033.¹⁵ The company expects a significant fall in the liabilities on its balance sheet. I can't judge the price Humana will eventually pay to buy the clinics back, or the return protections it gave BlackRock, because the terms are not in the release.

Medicaid

Humana has state Medicaid contracts in Florida, Kentucky, Illinois, Indiana, Louisiana, Ohio, Oklahoma, South Carolina, Virginia and Wisconsin, and says its footprint spans 13 states with Georgia and Texas expected to start in 2027.²⁴ Membership was 1.60M at 30 June 2026. The company was awarded a statewide Illinois contract that starts in January 2027, and its Florida contract, its largest, was extended to 31 January 2035. It says the Texas and Georgia awards face procurement challenges.⁹⁶ Medicaid is a small profit pool today but the company expects its margin to improve as new contracts mature, and I treat it as an option, not a pillar.

IIThe moat

A durable advantage is a cost or a habit that a rival with money and talent can't copy in a reasonable time. Humana has some real ones. It also has one big dependency, the government's rate and its quality score, and neither is in its control. I try to separate the two.

The pattern is that the advantages are real in care delivery and weak in the part that sets the profit. Humana can steer a patient into its own clinic and pharmacy, which lowers costs. It can't set the rate. In 2027 CMS pays Medicare Advantage plans on average 2.48% more, 4.98% when it adds the growth it expects from changes in coding and the mix of members, against medical and drug costs the company expects to rise at a high single-digit rate.¹⁶⁹ The gap between the two is what the company has to close by trimming benefits, better care and lower costs.

Evidence it is widening

• The 2027 Star Ratings put 95% of members in plans of four stars or more, and the company says it exceeded its goal of Top Quartile results, defined as Stars revenue per member 10% above the median of its peers.⁷⁸
• Retention rose more than 500 basis points, and the share of new sales through higher-value channels was above 75%.⁴
• The Insurance segment's operating cost ratio fell 1.2 points in the second quarter to 7.1%.⁹
• CenterWell patients rose 27% in six months, and the company says it is offsetting V28 and expanding margin.⁹

Evidence it is narrowing

• Adjusted EPS fell from $26.09 in 2023 to $16.21 in 2024 and is guided to at least $9.00 for 2026.¹⁷¹⁸⁵
• Only 20% of members were in four-star plans for 2026, and the payment for 2027 still reflects those scores, improved by moving members to better contracts.¹⁰
• Prior-year claims reserves released $442M in the first half of 2026 against $638M a year earlier, so less of the profit is coming from old estimates.³
• Days in claims payable fell to 33.1 from 36.5 a year earlier, partly because pharmacy claims, which are paid fast, are a bigger share of costs. I can't tell from the filings whether reserves are cautious or thin.³
• The 2027 payment rate proposed in January's advance notice was 0.09%; it reached 2.48% only in the final announcement, and CMS excluded diagnoses from unlinked chart reviews, an effect of 1.53%.¹⁶

On balance I think the position is wide in integrated care, medium in Medicare Advantage scale, and unproven in quality scores. For the first I give credit. For the second I assume an ordinary margin in a good year and a poor one in a bad year, which is why my bear case is as low as it is. For the third I take the 2027 result as real but not permanent.

IIIThe competitive advantage

The honest comparison for Humana is the other large insurers, UnitedHealth, Elevance, Cigna and CVS Health through Aetna. Humana's difference is concentration. UnitedHealth and Cigna have large businesses outside Medicare Advantage, Elevance has Blue Cross plans, and CVS has a pharmacy benefit manager and drugstores. Humana has 70% of its premiums in one Medicare product. The table sets the five side by side on 2025 as filed, with operating income before any adjustment, and prices in Q4 2026.

Three things stand out. First, Humana earns a thinner operating margin than UnitedHealth, 2.1% against 4.2%, and the same order as CVS and Cigna C (calculated).²¹¹ Second, Humana and CVS are the two that trade at more than 40 times their latest GAAP earnings, 42 and 58 times, because their latest earnings are depressed C (calculated). Third, the read-across from this week is mixed. CVS reported third-quarter adjusted EPS of $2.09, raised its full-year guide to at least $8.20, and described cost trends as elevated with a cautious view for the rest of the year.¹⁹ Aetna is a direct Medicare Advantage rival, and its report is evidence that margin recovery is possible in 2026. It says nothing yet about 2027 pricing. The UnitedHealth page on this site notes that Humana's share of Medicare Advantage enrollment rose from 17% to 20% in a year while UnitedHealth's fell from 29% to 26%. That is the reverse of Humana's 2025.

Where Humana wins

• 70% of its premiums come from individual Medicare Advantage, so it has a deep base of data on seniors and the most to gain from better care of them.²
• It owns the clinics, pharmacy and home services its members use. 68% of individual Medicare Advantage members were in value-based arrangements at the end of 2025.²
• It needs little capital for its care businesses: capital spending was $546M in 2025, and is guided to about $650M for 2026.¹⁴³
• On its own measure its 2027 quality result is above the peer median by more than the 10% it set as a goal, while UnitedHealth and CVS/Aetna both lost share of members in four-star plans.⁸²¹

Where it loses

• Concentration. One product, one payer, and one large state: Florida is 14% of revenue.²
• Margin and consistency. Adjusted EPS fell 38% in 2024 and about 47% to the 2026 floor.¹⁸⁵
• Balance sheet. Debt to capital is 42.7%, with $2.3B of short-term debt and parent cash of $1.6B that the company says was lifted by the timing of a $1.05B Medicaid payment that settled just after quarter end.⁶³
• Scale in pharmacy. Its pharmacy benefit manager, unlike Optum Rx at UnitedHealth or Caremark at CVS, works for its own plans only, and the activity is eliminated in the accounts.⁴

IVManagement

Jim Rechtin joined Humana in January 2024 as president and chief operating officer and became chief executive on 1 July 2024. He ran Envision Healthcare from 2020 to 2023, was president of OptumCare in 2019, and held senior roles at DaVita Medical Group from 2014 to 2019.² Celeste Mellet became chief financial officer in January 2025 after being chief financial officer of Global Infrastructure Partners, Evercore and Fannie Mae, and she spent 18 years at Morgan Stanley.² The chair of the board, Kurt Hilzinger, has been a director since 2003.¹² Most of the senior team is new, and the Medicare Advantage and insurance leadership changed again in 2026.²²

The turnover cuts both ways. A new team can reset a business that had lost its way, and in 2025 it did things a settled team might not have done: it exited the employer-group medical business, left unprofitable plans and counties, and cut costs under what it calls value creation initiatives.² The other side is that the people promising a recovery by 2028 have run the company for about two years, and the people who priced 2024 are gone.

Pay, and what it pays them to do

Rechtin's reported pay for 2025 was $18.76M: a salary of $1.35M, stock awards of $13.87M, a cash bonus of $3.03M and $0.51M of other items.¹² The annual bonus has four parts. Pre-tax earnings count for 50%, the Star Ratings measure 20%, integrated health, which means members using two or more CenterWell services or growth in primary care patients, 15%, and clinical cost reductions 15%.¹² The board changed the earnings measure for 2025 from adjusted EPS to pre-tax earnings, saying EPS can be swung by things outside management's control.¹² Results came in at 168.6% of plan: pre-tax earnings of $2.841B against a target range of $2.63B to $2.777B, and the three other measures at the maximum. The board then used negative discretion, and the plan paid 112%.¹²

The long-term awards for 2025 were 60% performance units and 40% time-based units, with no options. The performance units pay on three years of total shareholder return against a health care provider index for 50%, Star Ratings performance for 30%, and productivity for 20%.¹² So follow the money. Management is paid for earnings before tax, for a quality score, for use of the group's own services, and for relative share price over three years. It is not paid for return on the capital it spends, for cash left after capital, or for the size of the balance sheet. The Stars measure in the awards was the right target for 2025 and 2026, and it has now been met for 2027, so the next set of targets matters more than this one. I'd like to see return on capital in the plan, because capital is exactly where the odds of a mistake are highest at an insurer that is growing 25% a year.

Their own money

Rechtin owned 44,251 shares at 15 January 2026, about $18.5M at $417.42, and he has bought shares with his own cash: 6,530 shares at $229.25 on 16 May 2025, about $1.5M.¹²¹³¹ Sanjay Shetty, the head of CenterWell, bought 810 shares at $185.21 on 25 February 2026.²⁵ All directors and executive officers as a group held 254,101 shares, about 0.2% of the company.¹² That is a stake, not a founder's.

The largest holders reported in the proxy are The Vanguard Group with 11.8%, Dodge & Cox with 8.5% and BlackRock with 7.7%.¹² Dodge & Cox's 13F filing shows 11,380,673 Humana shares at 31 March 2026 and 10,940,380 at 30 June 2026, a cut of 3.9%, worth $4.35B, 2.3% of its reported portfolio, or 9.1% of Humana's shares C (calculated).²⁶²⁷ That is a holding, not a stated view. I did not find anything Dodge & Cox has said about Humana in the documents I opened.

What they did with the money, over a full cycle

I take 2016 to 2025 as the cycle. It covers a failed merger with Aetna, a pandemic, a rise in profit to 2022, and the Medicare Advantage squeeze. In those ten years Humana generated $33.8B of operating cash. It spent $8.0B on capital, $6.0B on acquisitions, $3.3B on dividends and $12.2B on buybacks.¹⁴ Reported net income over the same ten years was $21.4B C (calculated).¹⁴

Three things stand out. First, the share count fell 20% in nine years, which is real value if the stock was bought at good prices. Second, it was not. In 2022, 2023 and 2024 Humana spent $4.5B on buybacks, in years when the average daily close was $469, $490 and $335, and the stock then fell to a low of $163.67 in March 2026 C (calculated).¹⁴¹ In the first half of 2026, with the stock near $180, it spent $103M, 0.6M shares at an average $182.13, a good price but a tiny amount.⁶ Third, the price is $417.42 against $167.42 in November 2016, up 149% in ten years, and adjusted EPS went from $9.57 to $17.14, up 79% C (calculated).¹²⁸⁴

The record is fair, not good. The company bought most of its stock in 2022 and 2023, when it was dear, and almost none in 2025, when it was cheap. The best call in recent years was the 2025 decision to give up members for margin, and the current one is to take members back. The 10-K says total net receivables rose with the changes to Part D, which is one reason cash flow lagged profit.² I call the capital allocation fair, not good, and note the dividend of $3.54 a share, a yield of 0.8% C (calculated), is a minor part of the return.

VFinancials

Humana's accounts are noisy, and four features make reported earnings a poor guide. First, the company reports an adjusted profit that leaves out the changing value of put and call options on its clinics, a transformation program that has charged $436M, $281M and $449M in the last three years, impairments, and amortization. Second, CMS payments are lumpy, so operating cash flow jumps around. Third, the profit depends on estimates of claims not yet paid. Fourth, 2026 is a year of 25% membership growth that costs profit before it earns any. I try to see through each of them.

Margins

The operating margin on reported figures ran between 2.1% and 7.9% and ended at 2.1% in 2025 C (calculated).¹⁴ Profit per dollar of revenue peaked in 2020 when people postponed care in the pandemic, and has fallen since as costs caught up with rates. The consolidated benefit ratio rose from 83.1% in 2020 to 90.2% in 2025, a worsening of 7.1 points that cost about $9B of pretax income a year at the 2025 level of premiums C (calculated).²⁹²

Earnings quality

Adjusted EPS and GAAP EPS tell different stories. Over the ten years the adjusted figures sum to $178.41 a share and the reported figures to $163.02, a gap of $15.39 C (calculated).^1428184 The 2017 GAAP figure includes the break-up fee from the terminated Aetna merger.³² For 2026 the gap is $2.48 a share: guidance is “at least $9.00” adjusted and “at least $6.52” GAAP, because of put and call charges of $1.47, value creation initiatives of $1.27, impairments of $0.17 and amortization of $0.30, less a tax effect of $0.73.⁵

A second point is the season. Humana's profit is made in the first half. Adjusted EPS was $7.61 in the second quarter and $17.91 for the six months; the third quarter is guided to about a $1.00 loss, which leaves a fourth-quarter loss of about $7.91 a share if the full year comes in at exactly $9.00 C (calculated).⁵⁹ The reason is cost. Members pass their deductibles, use more care, and take more drugs as the year goes on, and the new members of 2026 cost more than the old ones. The guide is a floor, “at least”, and the company calls its guidance approach appropriately conservative.⁴ If you annualize a good first half you will be wrong by a lot.

Reserve releases

An insurer books costs for claims it has not paid, and later adjusts the estimate. In the first half of 2026 favorable development of prior-period claims was $442M against $638M a year earlier, and in the second quarter $53M against $161M.³ The 10-K reports $1.0B of favorable development in 2025 and $701M in 2024, about 80 and 60 basis points of the benefit ratio.² So the release is a smaller part of profit this year. I cannot tell from the filings whether reserves are cautious or thin. Days in claims payable of 33.1 days at 30 June are down 3.4 days on a year earlier, which the company ascribes to fewer claims in process, timing of provider payments and more drug costs.³ A shortfall would show up as a charge in a later quarter, which is how 2024 went.

From profit to cash

Operating cash flow was $0.9B in 2025 against net income of $1.2B, and $3.0B in 2024.² The 2025 drop was timing: receivables rose with the Part D changes. In the first half of 2026 it was $3.2B against $1.6B, helped by a $2.9B rise in benefits payable and a $1.05B Medicaid payment that came shortly after quarter end.⁶³ Guidance for the full year is $2.5B to $2.9B, with capital spending of about $650M.³ Premiums received before claims are paid are real cash, but they are the insurer's float and sit in regulated subsidiaries, so I do not count them as owner cash.

What an owner could take out

Here is the number I care most about. If Humana kept everything in good order, how much could it pay its owners each year? I start from adjusted income and take off the things it leaves out that a shareholder still pays for. Stock pay is already in adjusted income, so I do not add it back. I deduct the average charge for the transformation program, $389M a year before tax for 2023 to 2025, because it has come every year, and I deduct the capital the insurance subsidiaries must keep as premiums grow. The company disclosed more than $3B of added capital needs between the end of 2024 and 2026, on about $45B of added premiums, which is 7 cents on the dollar.⁴³

The answer for 2028 is about $3.0B, or $25.12 a share, 80% of my base adjusted EPS of $31.28 E (estimate). For 2026 the same arithmetic gives a negative number, because a company that grows its premiums by $32B in a year has to tie up about $2.3B of capital while it earns $1.1B. That is not a defect in the company, it is what the first year of a growth swing looks like, and it is why I will not pay for 2026 adjusted EPS. I can't pin the capital rate more tightly, because Humana does not tell us how much of the $15.5B of statutory capital in its subsidiaries is required and how much is spare. A rate of 5% instead of 7% would add about $1 a share to 2028 owner cash, and 10% would take off about $1.60.⁶

The balance sheet

At 30 June 2026 Humana held $6.9B of cash and $17.6B of investments, mostly in regulated subsidiaries; the parent held $1.6B of cash and short-term investments, helped by the timing of the $1.05B Medicaid payment.⁶³ Debt was $14.2B, of which $2.3B was short term, mostly commercial paper of $1.3B and notes due within a year. Equity was $19.2B, and goodwill $10.5B, up $0.8B with MaxHealth. Benefits payable were $12.9B, up from $10.0B at the end of 2025.⁶ Debt to total capital was 42.7%, above the company's long-term target of about 40%. The ratings were BBB at S&P and Baa2 at Moody's.⁶ In March 2026 it issued $1.0B of junior subordinated notes at 6.625% due 2056, which the company says gets partial equity credit from the rating agencies, and in May 2026 it set up $1.5B of contingent capital facilities.⁶³⁸⁹ These are the actions of a company that financed growth by stretching its balance sheet. I do not subtract net debt in my valuation, because the owner cash I value is already after the interest paid; but the balance sheet matters in a bad year, and it is why my bear case is as low as it is.

VIValuation

Start with what would go wrong

Before any numbers, I ask what would most likely make this a bad investment from here, and work back from that. For Humana the problem is not a lack of customers; membership is growing 25%. It is that the price paid for each customer is set by the government, while the cost of caring for them is not. Here are the ways it goes badly wrong, in the order I think most likely.

1. Costs outrun the rate for another two years. The government raised the 2027 payment 2.48%, 4.98% with its estimate of rising risk scores, while Humana plans for medical and drug costs to rise at a high single-digit rate.¹⁶⁹ Each point of benefit ratio on about $155B of premiums is $1.55B of pretax income, or $9.54 a share E (estimate). A swing of three points, about the 2.9 points between 2023 and 2025, would take about $29 off a share, nearly a third of my 2028 base.³¹²
2. The rules on risk adjustment turn against it. CMS excluded diagnoses from unlinked chart reviews starting in 2027, a 1.53 point effect on payments, and has told Humana it will use its final audit rule without the fee-for-service adjustment the company relies on; a court vacated that rule in September 2025 and the government has appealed.¹⁶⁶ The Justice Department's Civil Division has also sent Humana information requests about its risk adjustment practices, and a securities class action over 2022 to 2024 statements is pending.⁶ I can't put honest odds on these, so I don't try. I show what a lost margin point does: $8.0 a share in 2028 E (estimate).
3. A growth bet meets a stretched balance sheet. The company grew individual members 23% in a year, found the capital through debt, junior notes and reinsurance, and sits above its own leverage target, with $1.6B at the parent.⁶³⁸ If the 2027 bids are wrong or the third quarter surprises, it has less room than UnitedHealth to absorb it, and before the new joint venture, calling the first two groups of clinics in 2026 would have cost $1.0B to $1.6B.⁶¹⁵

A fourth way is simply the price. The shares have risen 155% since March, so a good outcome is already partly in the price, and a small miss on a base this low can move the stock by a third. I can show what each of these does to the value; I can't say which will happen.

What the price implies

Over ten years the price went from $167.42 to $417.42, up 149%, while adjusted EPS went from $9.57 to $17.14 in 2025 and a guided $9.00 floor for 2026. At its 2022 high of $563 it was 22 times that year's adjusted EPS of $25.88; now it is 46 times a floor, 24 times last year, and 16 times the 2023 peak C (calculated).¹³⁷¹⁷⁴ So on the company's own history, the price is not low. The multiple of 2026 is not meaningful, because 2026 is the trough of a swing in membership that the company chose.

Reverse DCF: what the price needs

Held at my base on everything else, the price needs the individual Medicare Advantage margin to reach about 2.1% in 2028 and the base growth after it, which gives 2028 adjusted EPS of about $32 C (calculated). The company commits to “at least 3%”.⁹⁴⁰ So the price is about 0.9 point below management's target and above my own 2.0%. Two simpler tests show how much of the price is the recovery. If Humana earned the $17.14 of 2025 for ever, growing 2.5%, my model gives $194 a share, and the price is 2.2 times that. If it earned the $26.09 of 2023, its peak, for ever, it gives $325, still 22% below the price; and if it earned the guided $9.00 floor, $74 E (estimate). So the price is not describing the business as it is. It is paying for adjusted earnings of about $32 a share to arrive by 2028 and then grow. That is a different claim from “cheap” or “dear”, and it is the claim the market is making.

There are two other ways to read the same price. One is that it equals my base cash flows at a discount rate of about 9.3%, which is 4.4 points above the Treasury and 0.2 points under the 9.5% I require. The other is the three cases below, in which the same price returns 4.8%, 9.3% and 12.7% a year E (estimate). The bear case returns less than the 4.90% Treasury.³⁹

From segments to EPS

I build 2027 and 2028 from the segments, because that is where the company's own targets sit. I back-tested the method. In 2025 the two segments earned $1.66B and $1.34B, and the proxy gives adjusted pre-tax earnings of $2.841B, which leaves about −$0.16B for corporate items, interest and other income; at an adjusted tax rate near 27% that is $17.14 a share C (calculated).²¹²⁴ For 2026 the guide has the Insurance segment at about breakeven and CenterWell at $1.3B to $1.8B on a GAAP basis, a 25.5% tax rate and 121M shares; $9.00 needs about $1.46B of pre-tax profit, so the middle of the CenterWell range covers it almost exactly C (calculated).³ That is the first thing to see in the table: in 2026 the whole profit is CenterWell.

Each point of individual Medicare Advantage margin in 2028 is about $1.3B of pretax profit, or $8.0 a share E (estimate). That is why the range is so wide. Management's target of 3% against my 2.0% is $8 a share in 2028, and my bear case of 0.5% is another $12 below. A reader who disagrees with me about one number can start with this one.

The discount rate

I discount at 9.5%. It is the 10-year Treasury at 4.90% plus about four and a half points.³⁹ I use a higher rate than the 8.5% I used for UnitedHealth and the 9.0% for CVS Health, for four reasons. Humana is smaller, with $50.1B of market value against $367B at UnitedHealth C (calculated).¹¹ It has 70% of premiums in one government product. Its adjusted EPS has fallen 66% from the 2023 peak to the 2026 floor, and the company says 2028 will include a benefit it calls one-time.⁸ And its debt to capital is above its own target with a BBB rating. 9.5% is the least I ask for a business whose profit can fall by half in a year.

Bear, base and bull

My base case is Humana recovering to a 2.0% margin on its largest product and no further. Adjusted EPS goes from $9.00 to $19.31 in 2027 and $31.28 in 2028, which is 20% above the 2023 peak of $26.09 and 1.8 times 2025, then grows 6% a year, fading to 2.5%. It is the pretax profit of about $5.1B on revenue I estimate at $183B, a margin of 2.8% against 4.3% in 2023 E (estimate).¹⁷ The revenue is 75% larger than in 2023, which is why a thinner margin still gives a higher EPS. The bull case is the company's own: a 3.0% margin in 2028, which gives about $43. The company has not given an EPS target that I could check. Trade summaries of its June 2025 investor conference put the 2028 range at $36 to $44 and $38 to $42, and the conference deck was not open to my access, so I use the margin it states and build the EPS myself. The bear case has a margin of 0.5%, about $14, less than 2025.

What I can't judge

Some inputs I can't know, and I would rather say so than pretend. I don't know how much of the 2028 Stars payment will be given back in benefits. The company says it will manage its plans as if it were at the Top Quartile mark, and that some results above that mark in 2028 are one-time and will go to investments and shareholders, and it will say more when it gives 2028 guidance on the fourth-quarter 2027 call.⁸ TD Cowen's Ryan Langston estimated the recovery could be worth $3B or more of extra revenue in 2028.²¹ I don't know how medical costs will compare with the bids Humana filed for 2027. I can't forecast the Fifth Circuit rulings or what CMS will do with audits. I don't know whether reserves are cautious. I don't know what BlackRock's return protections will cost. I take the whole range to be wider than my three values suggest, and I would not argue with someone who put the bear at $80 or the bull at $750.

Multiply it out

My base needs several things at once. The 2027 bids hold, so that the individual margin reaches about 0.8% in 2027: I would give that 65%. The Star Ratings payment arrives in 2028 on the 2027 ratings, with no change to the rules: 90%. Medical costs run no higher than the bids for two years: 55%. No ruling or audit policy takes more than $1B a year from risk-adjusted payments: 75%. CenterWell reaches about $2.0B of income through the BlackRock transition: 70%. Multiplied, that is 17%, about one chance in six that all five hold. Management's 3% needs each of them to go better than my base, not just hold. Those odds are guesses and I do not use them in the values. They are why my base is only the middle, and not the likely outcome.

Quality and price are two questions

How good is the business? Mixed. The care model and the senior base are real, the membership is the largest it has been, and the quality score is back. But it has one product, one payer, and a record of $26.09 falling to $16.21 in a year, and 2026 is guided to $9.00. How good is the price? Not low. It sits 3% above my base value, 205% above my bear, and returns 9.3% on my base against the 9.5% I require. A good business at an uncertain price is one thing; an uncertain business at a price that already assumes the recovery is another, and this looks like the second.

It helps to compare with the next best use of the same dollar. A ten-year Treasury pays 4.90% with no risk to the dollars promised.³⁹ Humana returns about 9.3% a year on my base, 4.4 points more, and 4.8% on my bear, which is less than the Treasury. On this site, UnitedHealth was valued at a base of $376 against a price of $400.44 on 2 November, a price that equals my base cash flows at about 8.2%, and CVS Health at a base of $111 against $88.51, about 10.7% on its base. Those pages use discount rates of 8.5% and 9.0%, and their prices moved on 4 November. On those pages' numbers, CVS offers the highest return and the largest gap to its base, UnitedHealth sits 7% above its base as Humana sits 3% above, and Humana has the widest range of the three: its bull is 4.6 times its bear, against 3.3 for UnitedHealth and 2.3 for CVS. At a similar expected return, a narrower range is worth more to me.

The margin of safety

My margin-of-safety price is $263, 35% below my base value. At $263 my bear case returns 6.2% a year, below the 9.5% I require but above the Treasury, and the base returns 12.9% C (calculated). A 35% cushion is what I use when the bull is more than four times the bear: the wider the range of outcomes, the more I want to be paid for it, and 35% is the same as I used for UnitedHealth. The price is 59% above it, and the stock last closed below $263 on 7 May 2026, at $247.12, with a low of $163.67 on 12 March 2026.¹

VIIThe case for the price

The price is 3% above my base value, so the question here is not whether the bull case can reach it. It can: my bull is worth $635. The question is what has to be true for the price to be right, what the market is leaning on, and what each side has to show.

What most people miss

First, the whole 2026 profit is CenterWell. The guide has the Insurance segment at about breakeven and CenterWell at $1.3B to $1.8B, and that covers the $9.00 almost to the cent C (calculated).³ The insurer, 96% of outside revenue C (calculated)², has not yet earned the money the stock is priced on. Second, the 95% Star Ratings figure pays in 2028, not 2027. The 2027 payment still rests on the 2026 ratings, when preliminary data had 20% of members in plans of four stars or more; the company moved members into better contracts, which it said would lift the share in 2027 by a number it did not give.¹⁰ Third, growth was financed by balance-sheet work. The company says it avoided about $3.2B of capital contributions for 2026 through reinsurance and changes to legal entities, issued $1.0B of junior subordinated notes and set up $1.5B of contingent capital.³⁸⁹

The bull case

The bull case, as I understand it: Humana deliberately gave up members in 2025 to fix margin, then won them back in 2026, over 70% of the new sales being switchers from rival plans⁴; those members stay longer, cost less over time, and now sit in plans the government rates highly. From 2028 the quality bonus returns, the rate and the cost trend stop fighting each other, and a 3% margin on a $135B book is $4.1B of pretax profit, about $25 a share on its own and about $43 of EPS with the rest of the group in my bull.

Management's version has dates. On 29 July 2026 the company said its approach to 2027 bids would drive meaningful progress toward a sustainable pre-tax margin of at least 3% in 2028, and CEO Jim Rechtin said the first half was where it said it would be at its 2025 investor day.⁹⁵ It reaffirmed the guide on 9 October and said it had exceeded its Top Quartile goal for the 2028 payment year; it will give a mark to market against its framework on 10 December 2026.⁸ Among analysts, TD Cowen's Ryan Langston estimated on 9 October 2026 that the Stars recovery could net Humana $3 billion or more of extra revenue in 2028.²¹ Revenue of that size with little added cost would be about $18 a share after tax on 121M shares if all of it reached profit C (calculated). My bull adds $12 to my base, because the company says it will spend part of the benefit on members and investments.

The one holder I can check is Dodge & Cox, which held 10,940,380 shares at 30 June 2026, down from 11,380,673 at 31 March.²⁷²⁶ That is a holding, not a thesis, and it was trimmed in a quarter when the stock went from about $173 to about $397 at its quarter-end values. I found no one else with a current, public, attributable view that I could verify.

Against reality

Much of the bull case is already true. Membership is up 23%, retention is better, the quality score is back, the operating cost ratio is down 1.2 points, and the adjusted guide of at least $9.00 has not changed since February.⁹⁸⁷ I tested what happens if the company does what it says. A 3.0% margin on about $135B of individual Medicare Advantage revenue in 2028, with CenterWell at $2.2B, gives adjusted EPS of about $43. If Humana does that and keeps growing 9% a year after it, my bull is worth $635, 52% above the price. So yes, if they hit their targets, the price is possible.

How far is today from that? The Insurance segment earned $2.26B in the first half and the guide implies about zero for the year, so the second half loses about $2.3B C (calculated).³ The individual margin is guided to slightly below breakeven for 2026 and has to rise by about 3.5 points in two years.⁴ The third quarter is guided to about a $1.00 loss and the benefit ratio to slightly above 94%.⁹ The 2027 payment rate is 2.48% against a cost trend in the high single digits.¹⁶⁹ And the stock has already risen 155% from March on the recovery. That is why my base is a 2.0% margin and not 3%. The case for the market is the same facts read the other way: if 2026 is the bottom of a swing that the company chose, the price of 13 times my 2028 base is a fair price for a recovery that still has two years of proof to go.

What a 10x would take. Ten times $417.42 is $4,174 a share, about $501B of market value on today's share count, 1.4 times the $367B of UnitedHealth C (calculated).¹¹ At 15 times earnings that needs about $33B of net income, 8.8 times my 2028 base. At a 3% net margin it needs about $1.1T of revenue, 9 times the 2025 figure and 2.5 times what UnitedHealth, the largest US insurer, took in last year C (calculated).¹¹ A tenfold move in ten years is 26% a year. Humana itself is up 2.5 times in ten years. For a company this size a tenfold rise in a decade has almost never happened, and I see no route for it here. A rise of about half is possible: my bull is 1.5 times the price.

Where I come down

What moved me: the 2026 profit is all CenterWell, so the insurer's recovery is still a promise, and the 2027 Star Ratings are real, with 95% of members in good plans, though they pay only from 2028. What did not move me: a price of 46 times a floor and 13 times my 2028 base, the company's 3% target, which I can't test until the bids it has filed for 2027 are paid, and a stock up 155% since March. What would raise my values: a 2027 guide at or above my base of about $19, a 10 December update that shows how it gets to 3%, and a 2028 CMS rate of at least 2.5% before risk scores.

VIIIRisks, and what would change my view

The ways this goes badly wrong are at the start of Valuation. These are the rest.

• Policy and rates. The 2028 advance notice comes in early 2027. CMS kept the 2024 risk model for 2027 and excluded unlinked chart review diagnoses, a 1.53 point effect, and it can do either again.¹⁶
• Audits and courts. CMS has selected Humana contracts for risk adjustment audits for payment years 2018 to 2021 and plans a rolling schedule. A court vacated the audit rule on 25 September 2025 and the government's appeal is pending at the Fifth Circuit; Humana's own challenge to the 2025 Star Ratings was rejected on 14 October 2025 and is also on appeal.⁶
• Cost trend. The company plans for a blended high single-digit trend, with medical at the higher end of mid-single digits and pharmacy in the low double digits. Pharmacy is a growing share of costs.⁴³
• Concentration. Florida is 14% of revenue and 83% of revenue is federal.²
• Reserves. Days in claims payable fell to 33.1 and favorable prior-period development fell. A reserve shortfall would arrive as a charge in a later quarter.³
• Capital and clinics. The balance sheet is above its leverage target. The new BlackRock venture removes the put on the clinics but gives BlackRock return protections and a right to force a sale after 2033, and Humana may buy clinics from 2028.¹⁵
• Medicaid. The Texas and Georgia contracts face procurement challenges, and the company points to national Medicaid enrollment headwinds.⁹⁴
• Timing. Third-quarter results arrive on 6 November 2026, two days after the price on this page. They could move the stock more than anything written here.⁴¹

What would raise my base value: two quarters in which the Insurance benefit ratio comes in at or below the guide with the third-quarter loss no larger than the $1.00 guided. What would lower it: a 2027 guide near $15, a court or audit result that removes the fee-for-service adjustment, or a fourth-quarter loss much larger than the roughly $7.91 the guide implies.⁹⁵