UnitedHealth (UNH): The Cost Ratio Turned, Is That Enough?

UnitedHealth trades around $376.90 as I write this, down about 40% from the $630.73 high it set in November 2024. That gap did not open on one bad headline. It opened because a single ratio moved by roughly three percentage points and refused to move back for the better part of two years, until this summer.

The ratio is the medical care ratio: the share of every premium dollar that goes back out the door as medical claims. UnitedHealth prices its Medicare Advantage plans a year ahead of the utilization trend it expects, and when members use more care than that pricing assumed, the company eats the gap until the next renewal cycle. For years before 2024, the ratio ran in the low-to-mid 80s, 82% to 83% in 2019, 2021, 2022 and 2023 by the company’s own quarterly filings, which is part of why the jump into the high 80s rattled the market so hard. Guidance was cut more than once, and the company changed chief executives in the middle of it.

Shares now sit at 24.2 times trailing earnings, close to the five-year average of 25.4, which tells me the market has stopped pricing UnitedHealth as broken and started pricing it as merely uncertain. That is progress. It is not the same thing as proof the margin problem is behind the company for good.

The medical care ratio actually improved

Start with the number that caused the crash in the first place. UnitedHealth’s medical care ratio came in at 86.7% for the second quarter of 2026, down from 89.4% in the same quarter of last year, according to the company’s second-quarter earnings release filed with the SEC. Management credited benefit design, pricing discipline, a shift in member mix, and medical cost management programs, in that order.

Two and a half points sounds small, but applied to more than $112.0 billion of quarterly revenue it is a large number. Earnings from operations came in at $8.0 billion for the quarter, versus $5.2 billion a year earlier, a jump of more than 50%. About $860 million of that quarter’s result came from favorable development on prior-period medical reserves, meaning claims booked in earlier periods ended up costing less than the company had set aside. Favorable development is real money. It is also not the kind of item that repeats on command, and a reader who backs it out still finds a meaningfully better underlying ratio.

Optum still runs on its own clock

UnitedHealthcare, the insurance side, and Optum, the services side, share a ticker and not much else operationally. Optum is paid to manage total cost of care rather than collect a premium and pay claims after the fact, through a pharmacy benefit manager, an employed physician network of its own, and a data and analytics business that prices risk for other payers. When the medical care ratio spikes, that is a UnitedHealthcare problem first. Optum’s margin runs on separate contracts and a separate pricing cycle, and the two segments deserve to be judged apart, the way I found it useful to split Disney’s parks, streaming, and ESPN businesses into three separate bets rather than one blended story.

I do not have a clean, current per-unit margin for Optum Health, Optum Rx and Optum Insight to quote here, and I would rather say that than manufacture one. What I can say is that a services business paid on cost-of-care outcomes does not fix a mispriced insurance book by itself; it can offset it partially, on a lag, if the risk contracts are structured that way. Investors who bought UnitedHealth for Optum alone, and were surprised when the stock fell anyway, misjudged how much of the total earnings base still runs through the insurance side.

The multiple already assumes a real comeback

Here is where I get less comfortable. Wall Street expects $19.97 of earnings per share over the next year, implied growth of 28% off the trailing $15.56. That puts the forward P/E at 18.9, a real discount to the trailing multiple of 24.2 and to the five-year average of 25.4. A forward multiple that low, next to growth that fast, is the market betting the recovery holds and then accelerates. I made a similar point about Microsoft trading at a premium to its own history: the gap between a trailing and a forward multiple is a specific bet on a specific number showing up on schedule, not free money.

Revenue was never the problem. Full-year 2025 revenue came to $447.6 billion, up 12% from $400.3 billion in 2024. Net margin, though, is a thin 3%, well below where UnitedHealth ran before the utilization spike, and operating margin sits at 4%. Getting margin back, not growing revenue, is the entire thesis. The reverse-DCF math I laid out separately applies here too: a market multiple only holds up if the earnings path it is pricing actually shows up on schedule, and a 28% growth assumption leaves very little room for a third bad quarter.

Metric	UnitedHealth now	Five-year average
P/E (trailing)	24.2	25.4
P/E (forward)	18.9	n/a
Price/sales	0.8	1.2
Price/book	3.5	5.0

What the targets and the quant score say now

Twenty-one analysts cover the stock. Eighty-one percent rate it a buy.

The average price target is $487, which is 29% above where shares sit today, and even the low target still implies double-digit upside. That spread is unusually narrow for a stock that fell by 40%, and I read it as a sign the sell side has mostly moved past the crisis phase and into a debate about the size of the recovery, not whether one is happening. Short interest sits at just 1.7% of the float, which is not the positioning of a stock the market expects to keep falling.

Our own quant score moved the other way, from a B before the reset to a C now, a reminder that a cheaper multiple and a lower quality score can be true at the same time. The stock got less expensive because the business got less predictable, not because the price fell for no reason.

The counter-case: one good quarter is not a trend

Here is where I could be wrong, and it is a specific risk rather than a general one. Medicare Advantage plans are repriced annually, and the 2027 bid cycle will show whether UnitedHealth priced this year’s plans for the utilization trend it is now seeing, or for a friendlier trend it hopes returns. If the medical care ratio moves back above 88% in an early-2027 quarter, that would tell me the improvement leaned more on favorable reserve development and pricing catch-up than on a durable fix, and I would treat that as a reason to size down rather than add.

The dividend, at $8.95 a year and a yield of 2.37%, is not under threat either way; it is a small piece of the return here compared to the multiple question.

Worth naming the second risk too, since it rarely gets mentioned alongside the first. Washington sets Medicare Advantage reimbursement rates and risk-adjustment rules, and a policy change unrelated to UnitedHealth’s own execution could move the medical care ratio in either direction regardless of how well the company prices its plans. I have no specific rate change to point to here, and I would treat any forecast of one as a guess rather than a fact, but it belongs on the list of things that could break the recovery thesis even if UnitedHealth does everything right on its own end.

What the reserve number is actually telling you

Favorable prior-period development, the $860 million I mentioned above, deserves a closer look than most coverage gives it. It shows up when a company sets aside money for medical claims it expects to pay out, then later finds the real bills came in lower than reserved. Run that in one direction long enough and it looks like a genuine improvement in cost control. Run it the other way, unfavorable development, and it looks exactly like what triggered the 2024 crisis in the first place: reserves that turned out to be too thin, forcing the company to book the miss all at once rather than smooth it in gradually.

The reason this matters for UnitedHealth specifically is that reserving is partly a judgment call, not a mechanical output. A company under pressure to show margin recovery has every incentive to lean toward optimistic reserve assumptions, and the market will not find out which way that judgment leaned until claims actually get paid a few quarters later. None of this means the $860 million was manufactured. It means a single quarter of favorable development is weaker evidence than a full year of it, and I would rather see two or three quarters in a row before treating the medical care ratio’s improvement as settled rather than as a data point still building its case.

What the December quarter has to confirm

I would want to see the medical care ratio hold under 87% without a repeat of the $860 million reserve benefit doing the heavy lifting, and I would want operating margin move off 4% and toward the high single digits that defined the business before 2024. Until then, $376.90 buys a company the market believes is recovering, priced at a forward multiple that leaves room to be right and very little room to be wrong twice in the same year.

Analysis and opinion only, not investment advice. Figures come from UnitedHealth Group’s second-quarter 2026 results on SEC EDGAR and its investor site; valuation multiples and price targets are approximate and were checked on September 18, 2026.