Subscribe now

Molina beat Q1 2026 adjusted EPS at $2.35 vs. ~$2.17 consensus, with all three segment MCRs running ahead of full-year guidance, yet management deliberately refused to raise the $5.00 full-year EPS bar. The release materially de-risks the “2026 is the trough” thesis: Marketplace adjusted MCR landed at ~79.5% (right inside the 78-80% target), Medicaid acuity shift did not recur, Medicare integrated duals outperformed, and operating cash flow was $1.1B. But two pillars of the bull case showed cracks — open-market buybacks were zero in Q1, and Medicaid same-store (state) membership attrition was raised from -2% to -6%. Management punted the 2027-2029 earnings power discussion entirely to a May 8, 2026 Investor Day, where the path to mid-to-high $20s normalized EPS will either be validated or revised. The stock responded with a +10.8% move on April 23 and ~+14.8% by April 24, with Mizuho lifting its price target to $195.

The reported quarter was characterized by Zubretsky as

“solid under the circumstances, but that characterization is against the backdrop of current modest expectations.”

That single sentence is the lens through which the entire earnings release should be read: the underlying datapoints are positive, but the company is running a deliberate “time-tested” playbook of waiting for two quarters of confirmation before extending favorability into the back half. As Zubretsky put it on the call,

“we are merely reaffirming our full year 2026 adjusted earnings per share guidance of at least $5… bearing in mind in Medicaid with a 92% result in the first quarter, a 92.9% indication in our guidance for the full year, we can actually produce loss ratios north of 93% and still hit our guidance for the rest of the year.”

The headline numbers and where they landed vs. guidance

The GAAP Gap Molina’s earnings bridge was dominated by a $93M ($1.83/share) impairment. This is a clean-up move: the company is exiting the standalone MAPD product for 2027, leading to write-downs on licenses and trade names.

Underlying Fundamentals Beyond the one-time charges, the core looks solid:

• Reserve Development: Favorable development hit $253M, a massive jump from $186M last year. In plain English: Molina over-estimated its past medical costs and is now reaping the benefits. This "found money" acted as a significant tailwind for the quarter's profitability.

• DCP Timing: Days-in-claims-payable (DCP) dipped to 44 (vs. 46 YoY). CFO Mark Keim brushed off the drop as mere payment timing, noting that claims payable actually grew year-over-year.

Segment-level MCRs all ran below the full-year bar

Marketplace

The Marketplace reset is an important thing to understand. So let’s slow down and actually walk through it.

Recall what happened last year. Molina had grown its Marketplace book — these are the ACA individual insurance plans sold on the federal and state exchanges — to nearly 700,000 members. The MCR on that book exploded into the mid-90s. For context, the target is 78-80%. When your MCR is in the mid-90s on a book that’s supposed to run in the high-70s, you are losing serious money. That’s what happened in 2025. It was the single biggest contributor to the earnings collapse.

The fix was simple and brutal: raise prices 30%, exit unprofitable counties, and accept whatever membership you lose as a result. Molina went from 655,000 Marketplace members at the end of 2025 to 305,000 by the end of Q1 2026. That’s a 53% reduction in one quarter. They cut the book in half on purpose, kept the price discipline, and let the chips fall.

So the question for Q1 2026 is simple: did it work?

The reported Marketplace MCR was 84%. That sounds okay but not great — still above the 78-80% target. But that 84% number is misleading, and the CFO walked through exactly why on the call.

Here’s what’s sitting inside that 84%. Every year, after the ACA plan year closes, the federal government runs a settlement process across the entire insurance industry. It looks at every insurer’s member pool and asks: were your members sicker or healthier than the industry average? If your pool was sicker, the government sends you money. If your pool was healthier, you owe money. This is called risk adjustment, and it exists to stop insurers from gaming the system by only chasing healthy members.

The catch is the settlement takes time. So when it finally lands, it hits your current quarter’s books — even though it relates to a prior year. Old business, settling now.

On top of that, CMS — the federal agency that runs Medicare and Medicaid — has been running a major crackdown on fraudulent ACA enrollments. Over the past few years, shady brokers were signing people up for Marketplace plans without their knowledge to collect commissions. People were enrolled in plans they never asked for. The government has been systematically cleaning this up, and when it claws back the subsidy payments it made on those fraudulent enrollments, that hits the current quarter’s books too.

Together — the prior-year risk adjustment settlement and the CMS fraud clawback — those two items added approximately 450 basis points to Molina’s reported Q1 Marketplace MCR. In plain terms, they made the number look 4.5 percentage points worse than the actual 2026 business is performing.

Strip those out and the adjusted MCR is approximately 79.5%. Dead center of the 78-80% target.

That is the thesis working. In the first quarter with the reset book, the reset pricing, and the reset membership, the underlying Marketplace business is running exactly where it needs to be.

A few things make this even more encouraging. First, Q1 is seasonally the best Marketplace quarter — people who just enrolled in January haven’t fully utilized yet, so MCRs are naturally lower in Q1 and drift higher through the year. Management knows this, which is why the full-year guide is 85.5% rather than 79.5%. The cushion from Q1 will be needed. Second, the 70% of members who stayed are renewals — people who looked at the new higher prices and paid anyway. That’s a self-selecting group of people who value the coverage and use it consistently, not people who signed up because it was free and will disappear the moment it costs something. Third, Molina actually retained more members than it expected — 305,000 versus an internal expectation of 280,000 — suggesting the price increases didn’t chase away quite as many people as feared.

The next real checkpoint is June, when the industry-wide risk adjustment data drops. That will tell management whether the 2026 book’s risk profile is where they think it is. Until then, the 79.5% is the best available read — and it’s a good one.

Subscribe now

Read more

Subscribe now

There’s a certain kind of stock that looks almost irresistibly cheap — beaten down, hated, and trading at valuations that seem to beg for a contrarian to swoop in and get paid. Novo Nordisk is exactly that kind of stock right now. And the question I keep turning over: is this a generational buying opportunity, or is it the kind of “cheap” that stays cheap forever?

Let’s actually dig into it.

How We Got Here: A $400 Billion Company Cut in Half

To understand where Novo is today, you need to remember what it was two years ago. In June 2024, Novo Nordisk was the most valuable company in Europe — trading at around $142 per share and sporting a P/E ratio of 35–40x. Wall Street had essentially declared it the Amazon of the pharmaceutical world. Ozempic and Wegovy weren’t just drugs; they were a cultural moment. The company couldn’t manufacture them fast enough.

Then came the unraveling.

It didn’t happen all at once. It was a series of gut punches, each one arriving just as investors thought they’d seen the worst:

February 2026 — The CagriSema Disaster. This was the biggest blow. CagriSema was supposed to be the “Lilly-killer” — the next-generation drug that would cement Novo’s dominance in obesity for the next decade. On February 23, 2026, NVO fell over 16% in a single day after CagriSema failed to demonstrate non-inferiority against Eli Lilly’s tirzepatide in the Phase 3 REDEFINE 4 trial. CagriSema achieved 23% weight loss after 84 weeks but did not meet the primary endpoint. Second place in a clinical trial is a losing position — you lose the label, and with it, the ability to command a premium price.

The Government Price Hammer. Under the Trump administration’s “Most Favored Nation” initiative, Novo was effectively cornered into slashing U.S. drug prices to align with what European countries pay. If Ozempic costs $150 in France, the U.S. government negotiated the right to pay close to that — not the $900+ Americans had been paying. Novo’s entire investment thesis for the past five years was built on charging Americans a massive premium over the rest of the world. That advantage is now legally constrained.

The 2026 Revenue Guidance. For the first time in its modern history, Novo guided for a sales decline — forecasting a drop of 5% to 13% for the full year 2026. For a company that had been growing at 30%+ annually, this was the moment the market fully repriced it from “growth stock” to “show me” stock.

The result? NVO is now down more than 56% from its 52-week high of $81.44, with a max drawdown of 56.46% recorded on March 30, 2026. From the all-time high, the damage is closer to 70%. A company that was briefly the most valuable in Europe is now trading like a mid-tier pharmaceutical also-ran.

The Financials

Here’s where the story gets genuinely interesting — and genuinely complicated.

Despite everything above, Novo Nordisk is not a company in distress. Not even close.

Subscribe now

Read more

Subscribe now

Introduction

As many readers know, I work in medical sales. When I read the annual letter and 10‑K for HIMS, I related to this company as an operator and would‑be owner more than I do with most businesses I look at. My goal, ever since I was young, has been to own a business. Early on, I was told that if you want to be a businessman, you need to understand sales. Ten years into a sales career, I can confirm that was excellent advice. At the right company, being in sales is effectively like running your own small franchise inside a larger organization.

Sales is, however, brutally competitive—especially on a residual‑based model. That’s why whenever I look at a business, the first thing I evaluate is the moat. The businesses I hate are the ones where customer retention ultimately comes down to which salesperson the staff likes more.

The Pathology Example and Moats

That dynamic is the foundation of many medical sales organizations. Take pathology, for example: once you set up a clinic, you need to move quickly to get them integrated with EMR systems, establish recurring pickups, and connect them with billing teams who can handle frustrated patients. Even all of that is a pretty thin moat.

What is very difficult to dislodge is a ten‑year relationship with a rep. That puts the chips in the representative’s hands and forces the pathology lab either to pay more to keep that rep or risk losing them—and the business—to a competitor lab.

There is a third option, and it’s the smartest one: make the lab so efficient and sticky to the clinic that neither the rep nor the clinic wants to leave. If the clinic feels like switching labs would create chaos, and the rep feels like no other lab can match the support and economics, then the lab finally owns the relationship.

This is a very commonsense way to think about business, but it’s often overlooked. Warren Buffett preaches the importance of a moat, but you don’t have to take it from him. It’s obvious if you’ve ever sold anything and taken the perspective of the owner signing your paycheck.

How HIMS Fits the Distributor Model

HIMS is unusual because it acts as both a distributor and a compounding pharmacy. There are thousands of compounding pharmacies and thousands of medical distributors in the U.S., which means HIMS has to differentiate itself in a crowded field.

The medical industry is built around the distributor model. At a high level, it is usually more attractive for a new product or drug to go through an existing distributor than to build an in‑house sales force. Founders of new therapies typically don’t have the skill set—or the desire—to also be sales leaders. Medical sales is complicated and relationship‑heavy: there are countless clinics, and most of them require an expensive lunch just to let you present. Sometimes even that isn’t enough.

Once you get into enough individual doctors’ offices, you then have to convince those doctors to request your product at the hospital. If you get enough of those requests, you might get approved at the hospital. Very few scientists have the time or temperament to create a new drug and then fight through bureaucracy and an ungodly number of personalities in the field.

So you get a web of distributors, some small and some huge. Maybe the new drug company signs a royalty deal with McKesson or Cardinal Health. Maybe a niche graft company starts locally and tries to cover Illinois with five independent distributors.

Those distributors then create webs of their own. It’s often just some guy or girl who connects a network of people. Say the graft company offers this person 20%, and then that person offers 50% of her cut to the people in her network, who then do the same thing downstream. That chain continues until the margin is so thin it’s no longer worth anyone’s time. Service quality deteriorates, and the model doesn’t scale. These arrangements tend to show up in short‑lived products where everyone wants a quick hit until the product is either commoditized or rolled up into one of the big distributors.

Cardinal Health, McKesson, and Cencora (formerly AmerisourceBergen) dominate U.S. drug distribution. They control more than 90% of the market, generate hundreds of billions in revenue, and live on razor‑thin margins—roughly mid‑single‑digit gross margins and around 1% operating margins. HIMS, on the distribution side of its business, is playing on that same thin‑margin field.

The distribution model can be extremely lucrative for an individual rep with a tight, loyal panel of prescribers. If I have a close‑knit group of 10 productive doctors and enough products to sell them, I can plausibly earn seven figures a year. The moment I try to turn that into a big web where other people do the work for me, the economics start to erode. That’s when I stop liking the business.

The Real Engine: Compounding

HIMS understands that distribution is low margin. The real engine is compounding. They buy raw ingredients for pennies, mix them, and sell finished products for cash prices.

In 2025, HIMS reported about $2.31 billion of online revenue—which is where the compounding and direct‑to‑consumer business sits—and about $36 million of wholesale revenue. Total revenue grew to roughly $2.35 billion in 2025 from $1.48 billion in 2024 and $872 million in 2023.

Management has framed the 2025 weight‑loss GLP‑1 business at around $725 million of revenue. At the same time, gross margin fell from about 79% in 2024 to roughly 74% in 2025. Wholesale revenue ticked down from roughly $39 million to $36 million. Depreciation and amortization jumped from about $17 million in 2024 to about $55 million in 2025 as new compounding facilities came online, and the company incurred higher transportation and cold‑chain costs to ship GLP‑1s.

If you take management’s GLP‑1 revenue target and the change in blended margin, you can back into an implied GLP‑1 gross margin of roughly the high‑50s percent, contributing on the order of $400‑plus million in gross profit. That’s a very healthy margin for a fast‑growing segment.

But the GLP‑1 compounding trade was never going to be a durable business. It was an arbitrage. Under federal law (Section 503A and 503B of the Food, Drug, and Cosmetic Act), pharmacies can only mass‑produce “copycat” versions of a drug when that drug is on the FDA’s official shortage list. By early 2025, the FDA began declaring the semaglutide shortages “resolved.” Once a drug is off the shortage list, it is legally “commercially available,” and mass compounding near‑copies becomes patent infringement, not clever pharmacy work.

2026: The Unraveling

In 2026, the HIMS story broke.

On February 7, 2026, the FDA came down on HIMS for an “illegal copycat” product. HIMS had tried to launch a compounded weight‑loss pill—essentially an oral Wegovy—at $49 for the first month to keep its weight‑loss growth engine humming. Regulators quickly labeled it an illegal copy of a commercially available drug and forced HIMS to pull it within 48 hours.

Two days later, on February 9, Novo Nordisk sued HIMS for mass patent infringement. At that point, the debate about whether HIMS was a telehealth platform or a quasi‑pharmacy ended. In the eyes of Novo and regulators, it looked like an unlicensed drug manufacturer.

The stock then hit what technicians call a “death cross,” with the 50‑day moving average falling below the 200‑day. Short interest spiked well north of 30%. Shares fell more than 70% from their highs.

The Novo Nordisk Deal and the Rebound

Fast‑forward to the recent rally. HIMS shares jumped about 50% on news that Novo Nordisk would distribute Wegovy and Ozempic—including the oral Wegovy pill—through the HIMS platform. In exchange, Novo agreed to drop its lawsuit, and HIMS agreed to stop promoting compounded GLP‑1s.

The primary thing this deal accomplished was to remove the existential legal overhang. Bankruptcy risk from a massive settlement moved off the table. That alone justifies some multiple expansion.

The Street quickly rebuilt a growth narrative. If HIMS can offer the oral pill, it can expand its weight‑loss funnel, because many patients prefer a pill over injections. A second bull argument emerged that weight loss could become a huge top‑of‑funnel product. If HIMS can use GLP‑1 demand to bring a lot of people onto the platform, maybe it can cross‑sell them higher‑margin compounded products for hair loss, sexual health, dermatology, and so on. In other words, GLP‑1s might become the loss leader that drives a more profitable bundle.

On paper, that’s a nice story. I just don’t buy it.

Subscribe now

Read more

Subscribe now

From Honor System to Data Moat

In the early 1990s, commercial insurance ran on a virtual honor system. Insurers had no real-time way to know whether a surgeon’s bill reflected the local going rate or something wildly inflated, so they relied on paper claims and fuzzy “Usual, Customary, and Reasonable” (UCR) benchmarks built from internal folders. In that “Wild West” environment, two doctors in the same neighborhood could bill $1,500 and $3,000 for the same procedure, and the insurer had little empirical basis to push back.

UnitedHealth’s Ingenix entity (the predecessor to OptumInsight) turned that opacity into a data advantage. By aggregating claims from nearly 100 insurers into a 1.3‑billion‑record database, Ingenix could benchmark prices by CPT code, specialty, and zip code. That allowed payers to move from guesswork to percentile‑based pricing tables: if local charges clustered around $1,200 at the 80th percentile, a $2,000 claim suddenly looked objectively “unreasonable.” United didn’t just use those tables internally; it sold the benchmarks as a product to competitors like Aetna and Cigna, effectively turning its own cost‑control infrastructure into a high‑margin software business. Once most of the market used the same United‑owned yardstick, physicians’ leverage to negotiate higher-paying contracts shrank dramatically.

At the same time, Ingenix bought coding and compliance shops and built “claims scrubbers” that scanned incoming bills for upcoding and billing anomalies. Algorithms compared each claim to expected patterns by specialty, diagnosis, and geography, flagging instances where a provider billed for a complex visit when documentation suggested a simple one, or where frequency of a code looked out of line with local peers. Over time, utilization analytics evolved from price policing to behavioral policing: if most cardiologists ordered two EKGs per patient per year but one billed for six, that pattern could be automatically flagged as potential “waste, fraud, and abuse” and routed for denial before payment.

The strategic twist was that United monetized these tools broadly across the industry. It positioned itself simultaneously as a payer using the rules and as the vendor selling the rulebook.

The Risk‑Score Engine

The first act of Ingenix was cost control: cap unit prices, reduce inappropriate utilization, and scrub claims. The second act was revenue optimization inside government programs, particularly Medicare Advantage. CMS pays more for “sicker” members via risk adjustment, and diagnostic codes largely define that sickness. If you can systematically find more codes - or more severe codes - without materially changing care, you can grow revenue without adding much cost.

United deployed chart‑review and data‑mining programs to surface missed diagnoses in historical records. Coders and algorithms would comb old notes for conditions such as chronic heart failure or malnutrition that were never formally captured as Hierarchical Condition Categories (HCCs), then pull those diagnoses into the current year’s risk profile. The member didn’t necessarily receive new treatment, but the plan’s payment ticked up. A recent Senate Judiciary staff report, based on 50,000 pages of internal documents, concluded that United had turned risk adjustment into a “major profit centered strategy,” using aggressive coding tactics and its vertical integration to stay ahead of CMS efforts to curb coding intensity.

Behavioral health became a major extension of this engine. CMS’s risk model assigns substantial additional payment when a member is coded with certain chronic mental health conditions, such as major depressive disorder or severe anxiety. Internal analytics identified these as historically under‑coded categories, particularly given the prevalence of symptoms in primary and specialty care. If a single high‑value behavioral code can raise a member’s annual risk‑adjusted payment by thousands of dollars, the incentive to “harvest” those codes through better documentation, and more aggressive interpretation of symptoms, is obvious.

United uses two distinct but reinforcing channels to capture that value. On the Optum‑owned side, physicians practice inside a vertically integrated environment where internal dashboards and workflows emphasize closing “care gaps,” including capturing any possible diagnoses hinted at in the encounter. On the non‑owned side, United offers natural‑language‑powered coding tools that read clinicians’ notes in real time and “suggest” specific ICD codes that would increase the risk profile. A phrase like “struggling with sleep and mood” can trigger a recommendation to assign a formal depressive disorder diagnosis that the physician had not yet coded.

Get Full Access

Read more

When investors talk about UnitedHealth Group (UNH), the story usually splits into two engines: the insurance arm (UnitedHealthcare) and Optum. Optum itself breaks into three divisions:

• Optum Health: employs clinicians and runs a sprawling care delivery network.

• Optum Rx: manages pharmacy benefits and drug distribution.

• Optum Insight: the data, analytics, and tech stack stitching the ecosystem together.

This article is about that second engine - Optum Rx - because it’s where Wall Street sees a fortress moat and where I see a throne with a sword hanging over it.

Subscribe now

Read more

This report will replace my usual Tuesday write-up, so there won’t be another article until Thursday.

Subscribe now

Read more

Subscribe now

Medicare Advantage didn’t drop out of the sky. Before you can understand why it looks the way it does, and why it now costs what it does, you need a basic grip on three things: how Original Medicare (Parts A and B) was built in the 1960s, what PPOs and HMOs actually are, and how Part C was designed to mimic HMO‑style managed care onto that old A/B framework. If you already live in this world and don’t need the history or the HMO/PPO refresher, feel free to skip ahead to the second half of the article, where we get into the money: benchmarks, bids, rebates, and why MedPAC thinks Medicare is now overpaying for MA.

Short history: Parts A and B

How Medicare’s A and B buckets came together

Medicare’s story starts in 1965, when President Lyndon B. Johnson signed the Social Security Amendments that created Parts A and B in one stroke. It was a Great Society project aimed at a very specific failure: private insurance simply did not work for older Americans. Roughly half of seniors had no coverage at all, and those who did often paid punishing premiums for thin policies that could be canceled or capped just when they needed care most. After watching earlier efforts stall under Truman and Eisenhower, Johnson used his post‑1964 landslide to push through a federal guarantee: once you turned 65, the government, not your former employer or a skittish insurer, became the payer of last resort.

Medicare’s two original pieces were a political compromise that still shapes everything built on top of it. Part A was designed as automatic hospital insurance: if you worked and paid Medicare payroll taxes, you “earned” Part A and don’t pay a monthly premium for it in retirement. Part B was set up very differently. It was voluntary “medical insurance” for doctor and outpatient services, and seniors had to actively enroll and pay a monthly premium, with the federal government subsidizing most of the cost behind the scenes. Put simply: Part A is automatic and premium‑free for most people because it’s funded through payroll taxes during your working years; Part B is optional and comes with a monthly bill in retirement. Together, those two pieces are what people now call Original Medicare.

The easiest way to grasp the A/B split is to follow a single episode of care. Imagine a 70‑year‑old who comes to the hospital with chest pain, is admitted for a few days, then discharged to a skilled nursing facility for short‑term rehab, and finally follows up with a cardiologist in clinic. Everything that looks like room and board in an institution, the hospital stay itself and the days in the nursing facility, is Part A territory. That’s the “hospital insurance” bucket paying the facility bills, subject to its own deductible and day‑based cost‑sharing. Everything wrapped around that stay on the professional side - the ER doctor’s time, the cardiologist’s work, the echocardiogram, the follow‑up office visit, the outpatient cardiac rehab sessions - runs through Part B, the “medical insurance” bucket, with its separate annual deductible and roughly 20 percent cost‑sharing for the patient.

This two‑bucket design solved the immediate coverage crisis: seniors were no longer actuarial orphans that private insurers could avoid or punish at will, because the federal government had committed to paying hospital and doctor bills in old age. But it also locked in a particular kind of system - fee‑for‑service, volume‑driven, split between facility and professional charges - that would later invite experiments with managed care. Medicare Advantage is best understood as one of those experiments: a late‑arriving attempt to take this 1960s hospital‑plus‑doctor construct and hand it to private plans to manage differently.

Two basic flavors of insurance: PPO and HMO

To understand what Part C was trying to do, you need a quick detour into the two dominant flavors of commercial health insurance: PPOs and HMOs. PPO stands for “preferred provider organization.” In a PPO, the insurer contracts with a big network of doctors, hospitals, and third‑party services (including diagnostics labs) and labels them “preferred.” If you stay in that network, the plan pays a larger share of the bill; if you leave it, the plan might still pay something, but on worse terms. That flexibility is what people like about PPOs, and it’s why they tend to rely heavily on cost‑sharing: higher premiums, higher deductibles, and plenty of other knobs the insurer can turn to protect its underwriting margin. The more “PPO‑ish” a product is, the more it feels like classic insurance with a big network and lots of ways to split the bill with you.

HMOs, or “health maintenance organizations,” are tighter by design. The plan builds a smaller, more controlled network, often requires you to pick a primary‑care doctor as a gatekeeper, and leans hard on referrals and prior authorizations. If a lab like mine isn’t in the HMO’s network, it generally doesn’t get used or paid, except in very specific circumstances. The trade‑off is that you usually get lower premiums and lower deductibles with an HMO, but you have to live inside that tighter network and accept more “no” along the way. At the plan level, HMOs are typically on a fixed per‑member budget for your care; under the hood they might still pay individual doctors per visit, but the overall business model is “here’s a pot of money for this population, manage inside it.”

In between these poles sit a bunch of hybrids and niche products, plans that look a little like PPOs on some dimensions and a little like HMOs on others, because there’s always demand for something that threads a different needle for a particular employer, geography, or risk pool. But conceptually, almost everything in the commercial market is just a variant of those two basic ideas: pay more for freedom, or accept more constraints in exchange for cheaper, more predictable coverage.

Why Part C was built to look like an HMO

Medicare Part C was created in 1997, under the Clinton administration’s Balanced Budget Act, originally branded “Medicare+Choice” before being renamed Medicare Advantage a few years later. At its core, Part C exists to do one thing: turn what Medicare would have spent on you in Parts A and B into a fixed, risk‑adjusted, per‑member‑per‑month payment to a private plan. The plan then has to live inside that budget. To enroll, you need to qualify for Part A, be enrolled in Part B, and live in a county where a plan is offered. You keep paying your Part B premium, and Medicare sends a fixed monthly payment to the plan to take over your A and B benefits. On paper, every Part C plan has to cover at least what Original Medicare would cover under Parts A and B - hospital, skilled nursing, hospice, doctor visits, outpatient care, tests, equipment, and so on.

Once you see the PPO/HMO contrast, this structure makes more sense: Congress built Part C to save money and improve efficiency by moving Medicare beneficiaries out of the old, open‑tab fee‑for‑service system and into HMO‑style plans. PPOs tend to be more cost‑sharing in nature, and doctors are usually paid per service rather than out of a flat budget. When a doctor is paid per unit of work instead of from a fixed pot, the incentives change.

Think about it this way. If you’re on a work trip with an unlimited expense budget, you’ll probably loosen up: grab a Starbucks coffee in the morning, order room service, pick a nice place for lunch, and have a few beers at night after you’ve done whatever you were there to do. If you’re on vacation with your wife and kids and paying out of pocket, you go downstairs, look at Starbucks, decide the free lobby coffee is fine, grab the complimentary cereal box, and maybe hit McDonald’s for lunch.

Doctors see the same thing. When care is effectively on an open tab and each extra visit or test generates another payment, it’s very easy for the system to tilt toward “one more thing,” especially in the gray areas where multiple options are defensible. When they’re working inside a flat budget for a patient population, they think harder about what is medically necessary and what is not.

Take a simple wound. In a loose, PPO‑style setup, that acute wound is more likely to end up with a $600 PCR panel if it’s covered and available. In a tighter, HMO‑style setup running on a fixed budget, the same wound is more likely to get a $75 culture - or even a “clean it, wrap it, stay off it, and come back next week” - unless there is a clear clinical reason to escalate. None of this is about “good” or “bad” doctors. It’s basic human nature, the same way people behave differently on a company card than on their own.

Once you hand a flat budget to a commercial insurer, they decide what kind of product to wrap around it. In Medicare Advantage that usually means an HMO‑style plan and, less often, a PPO‑style plan. MA leans heavily toward HMO designs because the economics and the customer base push it there: older, sicker members worry about big, late‑life bills and prefer lower premiums and predictable copays over wide‑open networks, while plans need tight networks and strong utilization control to make the math work. HMO‑ish products are the natural compromise: less freedom, but fewer nasty surprises.

From the member’s perspective, Part C feels like a Medicare HMO. You’re in a defined network, you deal with referrals and prior‑auth rules, and the plan, not the government, is the front door for your care. In return, everything is bundled into one product - one card, one set of copays, and an annual out‑of‑pocket maximum that Original Medicare doesn’t offer on its own.

Read more

Subscribe now

Abusing the System

Abuse is a feature of any system with rules and money. Give people a formula and enough time, and someone will eventually optimize it past what the designers intended.

Matt Levine’s Spotify story is a perfect illustration. In his “Fake Songs Made Real Money” column, he explains how Spotify allocates royalties. You pay your monthly fee, Spotify pools those funds, takes roughly a quarter, and divides the rest among artists based on total streams. If you barely listen, your share of the pool goes to whatever everyone else is streaming.

If money flows to whatever gets played the most, you can get rich by creating the songs and doing all the listening yourself. Michael Smith allegedly used AI to mass-produce low-effort tracks under fake artist names, then used bot armies and scripted accounts to stream those songs on repeat. Each bot either paid a subscription or generated ad revenue, meaning almost all that listening revenue flowed back to him. Over time, it added up to more than ten million dollars. He wasn’t hacking Spotify; he was simply optimizing the payout rule to its logical extreme. The same rule that fairly compensates Taylor Swift for millions of real streams also allowed a guy with fake fans to cash out millions from fake music.

You see this pattern everywhere. Compounding pharmacies were a textbook example. For decades, compounding was a legitimate niche, custom medications for patients with allergies or unique needs. But when Medicare Part D and TRICARE began paying based on ingredient cost plus a dispensing fee, and when no one was watching, the system exploded. Pharmacies realized they could bill staggering amounts by inflating ingredient prices, creating convoluted formulas, and aggressively marketing expensive topical creams for pain, scars, and other questionable uses.

The numbers defied reason. TRICARE spent $5 million on compounding in 2004. A decade later, that was $514 million. In 2015 alone, they paid $1.75 billion, including $545 million in a single month. Medicare compounded topicals grew 2,353% from 2010 to 2016, from $13 million to $323 million, and the average cost per cream jumped from $54 to $422. Marketers targeted vulnerable Medicare seniors, funneled them to telemedicine doctors who never examined them, and churned out prescriptions for overpriced creams mixing a few cheap ingredients billed at thousands per tube. Pharmacies kicked back a share to doctors and marketers. Federal investigators later found that only about 5% of the billed cost reflected actual compounding expense, margins were around 90%.

Then the hammer dropped. In May 2015, TRICARE implemented strict ingredient screenings. Spending fell from $75 million in the first eleven days of May to under $5 million for the rest of the month. Enforcement came fast - DOJ, FBI, and the Medicare Fraud Strike Force brought charges against dozens of pharmacies, doctors, and marketers. TRICARE spending plunged from $1.75 billion in 2015 to under $100 million annually by 2017. A loose system turned into tight oversight overnight, killing the fraudulent business model entirely.

The same movie played again in wound care and skin substitutes. Medicare spending shot from $256 million in 2019 to more than $10 billion in 2024, a 39-fold increase in five years. The schemes were identical: murky reimbursement formulas, minimal oversight, and massive economic incentives to overuse expensive products.

So once again, regulators shut the door. As of January 1, 2026, CMS implemented a flat reimbursement rate of $127.14 per square centimeter for all skin substitutes - down from as much as $2,000 to $3,000 previously. The change effectively slashed Medicare spending by 90%, saving an estimated $19.6 billion annually and instantly vaporizing an entire cottage industry.

A colleague of mine who was earning $700,000 a year on wound grafts as a side business got caught in that reset. They weren’t uniquely unethical, they were simply optimizing the rules as written. They danced while the music played. But now the music is over. They’re no longer dancing, though still a millionaire from the ride.

My own company came out of the compounding boom. When that went bust, they pivoted to PCR testing, thinking they’d found the next gold rush. They didn’t realize that the same pattern would repeat. CMS began reviewing reimbursement for molecular diagnostics and found the same structure: loose rules, high margins, and massive abuse potential. Labs billed $10,000 for genome sequencing on wound samples when a basic culture cost $75. Patient balances were written off, Medicare was milked, and when the crackdown came, reimbursements fell by 90%. My company, once flying high, suddenly faced clawbacks, audits, and medical-necessity denials.

But we’re not really here to talk about PCR tests, compounding creams, or Spotify bots. Those are just chapters in a recurring story. The same playbook now defines what’s happening in Medicare Advantage.

Read more

Cracking Open Molina

Molina Healthcare is the cleanest way to study what happens when a pure‑play gov…

Read more

Molina (MOH): down 50%
Centene (CNC): down 35%
UnitedHealth (UNH): down 45%

How Medicaid Insurers Actually Make Money

Medicaid insurers make almost all of their money through a thin underwriting margin, not by investing a big float the way property‑and‑casualty insurers in the Berkshire universe do. Margins are low, claims get paid quickly, and the whole game is keeping medical costs below a fixed per‑member rate the state pays.

This makes the unit economics incredibly simple: premiums minus expenses equals profit or loss. Admin costs are there, but they’re tiny compared to medical claims. In a traditional fee‑for‑service world, the insurer basically reimburses doctors and hospitals every time a patient touches the system—every visit, every test, every procedure. Medicaid managed care flips that on its head: the plan gets a fixed “capitated” payment per member per month, whether that person never sees a doctor or racks up a seven‑figure ICU stay. Fee‑for‑service is pay‑per‑use; capitation is a flat check up front and a promise to eat whatever shows up in the claims bucket.

For the rest of this discussion, when we talk about Medicaid insurers—managed care organizations, or MCOs—we’re talking about that capitated book. That’s where roughly three out of four Medicaid enrollees sit today and where essentially all of the underwriting risk lives.

One final example, because I want this model to be crystal clear. Imagine an MCO in Florida getting paid $400 per member per month by the state. That’s the capitation. The plan then cuts deals with doctors—maybe a flat $25 per member per month to a primary care group—to take care of that patient. The doctor can order more or fewer services, but the MCO is still only getting $400 from the state, so everything that gets spent between $0 and $400 is what determines whether the plan earns an underwriting profit or takes a loss on that member.

What Went Wrong: Unwinding and Acuity

So what actually happened to these companies that led to the massive drawdowns? As you can probably guess, claims ran hotter than expected and earnings got crushed. But the “why” behind that matters.

First, this isn’t a ‘one bad apple’ story. The Medicaid-heavy managed care names have been hit across the board. When you see multiple players in the same niche down 30–50% at the same time, you’re not looking at one incompetent management team—you’re looking at an industry in distress.

Second, a lot of this starts with what’s called the Medicaid “unwinding” after COVID. During the pandemic, Congress basically told states: “We’ll give you extra federal money, but you’re not allowed to kick people off Medicaid.” Redeterminations—the regular process of checking whether people are still eligible—were paused for about three years, nobody was getting removed for the usual reasons—missed forms, small income changes, moving, whatever. Enrollment just kept ratcheting up quarter after quarter until you had 90‑plus million people on Medicaid and CHIP, up from the mid‑70‑millions before COVID, an all‑time high.

Then that deal ended in 2023. States had to go back and re‑check essentially everyone: “Are you still eligible? Did your income go up? Did you move? Did you send back the right form?” That’s the redetermination process. When you run that at scale, a lot of people fall off—not always because they’re truly ineligible, but because they didn’t return a form, the packet went to an old address, the online portal didn’t work, or they missed a deadline while juggling two jobs and childcare. Those are “procedural disenrollments.” On top of that, you’ve got the marginal folks who actually do have other lanes they can slide into: a spouse’s employer plan, an ACA exchange plan once their income rises a bit, or they’re healthy enough that when the Medicaid letter shows up, it just goes in the trash.

Who fights to stay? Not the person who hasn’t seen a doctor in two years. It’s the diabetic on insulin, the dialysis patient, the kid with complex medical needs, and the families and social workers around them. Put that together and the unwind doesn’t trim everyone proportionally. It strips out a lot of the healthier, lower‑cost, more functional enrollees first and leaves a smaller pool that is poorer and sicker on average. The capitation rates, however, were built off the blended COVID‑era pool that still included all of those cheaper lives.

Third, the pricing was never built for that shift. States set capitation rates off historical data, not off a live feed of who just got kicked off and who stayed in. Acuity—the average severity and costliness of the people left in the pool—jumped faster than the rates did. That’s how you blow up a medical loss ratio, that’s how you end up cutting guidance two or three times in a year, and that’s how you get charts for MOH, CNC, and UNH that look like somebody took them out back and shot them.

Pharmacy: The GLP‑1 Shock

Pharmacy is its own little wrecking ball in this story.

On paper, Medicaid should be “stable and boring.” In reality, you’ve got high‑cost specialty drugs, GLP‑1s, and a pipeline of shiny new therapies hitting the system with price tags that look more like small cars than prescriptions. The problem is simple: states priced a lot of these capitation contracts in a world where those drugs either didn’t exist yet or barely mattered. Now they matter a lot, and every extra dollar is landing directly in the claims bucket.

GLP‑1s are a good example of why this feels different. These are chronic drugs, not one‑time shots—you’re supposed to stay on them, and if you stop, a lot of the benefit, including weight loss, tends to fade. And unlike a rare‑disease biologic, the eligible pool here is massive: diabetes alone is a big population, and once you add obesity indications, you’re talking about tens of millions of adults who meet BMI and comorbidity criteria, not a tiny sliver of patients. Medicaid coverage for GLP‑1s purely for weight loss is still patchy and varies by state, but coverage is broad for diabetes and is slowly expanding for obesity and related conditions. In other words, this isn’t a niche edge case creeping into the pharmacy line—it’s a growing, recurring spend item.

Commercial plans at least have a few levers. They can jack up cost sharing, design tighter formularies, or make people jump through prior‑auth hoops until they give up. Medicaid MCOs don’t have that same freedom. The benefit design is heavily regulated, cost sharing is minimal, and politically it’s very hard to be the plan seen as “denying” a breakthrough drug to a poor or disabled patient. So when these high‑ticket meds show up, there’s nowhere to hide: the plan pays, and it flows straight into the medical loss ratio (MLR).

Put differently: the pharmacy line item has quietly shifted from “annoying but manageable” to “this alone can blow up your year” for Medicaid-focused insurers. When you combine that with already-squeezed capitation rates and a sicker post‑unwinding pool, it’s not hard to see how a few bad quarters of drug trend can take a thin underwriting margin and just erase it.

Policy Overhang: One Big Beautiful Bill

On top of the unwinding mess and medical cost trend, you’ve now got “One Big Beautiful Bill” hanging over the whole sector. The short version: Washington just made it harder to get on Medicaid, harder to stay on Medicaid, and harder for states to finance Medicaid. None of those are friendly to MCOs.

Start with work requirements and more frequent checks. Medicaid is now tied, for a big chunk of adults, to logging at least 80 hours a month of work or “qualifying activities,” and proving it at application and then at least every six months. If you miss the paperwork or your hours bounce around, you can lose coverage even if you’re exactly the kind of low‑income person the program was built for. The official scorekeepers expect this alone to knock millions of people off the rolls over time and deliver a big chunk of the law’s Medicaid “savings.” From an MCO’s perspective, that’s a smaller membership base with the same fixed overhead and a lot more churn.

Then layer on the financing side. Medicaid is co‑funded: states put in a share, and the federal government usually picks up about 50–75% of the tab, depending on the state. One of the main ways states have come up with their share is by taxing hospitals and health plans. Imagine a hospital with $1 billion of patient revenue. At a 6% provider tax, the state can collect $60 million from that hospital and treat it as Medicaid funding. If the federal government roughly doubles that with its match, the state now has on the order of $120–150 million it can use to support Medicaid rates and payments. It’s a legal loophole that turns provider taxes into extra federal dollars. One Big Beautiful Bill shrinks that lever. Over time, that same hospital can only be taxed at 3.5%, so the state might collect $35 million instead of $60 million. Less tax means less match, which means less total Medicaid money sloshing around. If states still want to raise capitation rates to MCOs, they now have to do it with real general‑fund dollars instead of provider‑tax magic. That’s a much harder political ask, and it makes big, generous rate increases a lot less likely.

You can see why the market jumps straight to “shrinking pie + sicker members = permanently lower margins.” You’ve already got a post‑unwinding pool that’s poorer and sicker on average, and now federal policy is explicitly designed to push enrollment and financing down over time. The structural part here is real: more frequent redeterminations, work requirements, and tighter financing aren’t going away next quarter. Where the market may be over‑discounting things is in assuming that every dollar of lost enrollment is pure margin destruction. States still need plans to show up, networks to function, and access standards to be met. If rates don’t eventually reflect the higher acuity and tighter rules, you don’t just get lower insurer profits—you get MCOs walking away from contracts and patients losing access to care. That’s a political problem governors and legislators can’t ignore forever.

Is This Still a Good Business?

Yes—just not a pretty one right now.

Structurally, Medicaid managed care is still an oligopoly business with high switching costs. States don’t want a long list of tiny plans; they want a handful of scaled players who can handle compliance, networks, IT, and the political fallout when something goes wrong. Getting into this game is hard—RFPs (state contract bids), capital requirements, regulatory scrutiny, star ratings, years of operational track record. Once you’re in and you’ve got a big Medicaid book in a state, you’re not easily replaced. Contracts are multi‑year, renewals are lumpy, and ripping out an incumbent is painful for everyone involved.

The underlying demand story hasn’t changed either. Medicaid isn’t going away. The population it serves—low‑income families, disabled adults, duals—isn’t shrinking, and healthcare costs aren’t getting cheaper. States can trim eligibility at the margin and make people jump through more hoops, but when the dust settles, they still need someone to actually run the program. They don’t have the in‑house infrastructure to do it themselves at scale, and trying to rebuild that would be more expensive and risky than paying MCOs to take the risk and the blame.

So yes, margins are getting squeezed, rates are lagging, and policy is pushing in the wrong direction. But the core ingredients of a decent business are still there: recurring revenue from government contracts, a concentrated set of players, high operational barriers to entry, and a customer (the state) that can’t just walk away. The question isn’t “does this model survive?” It’s “at what level of margin and multiple does it settle out once the current shock and policy overhang are digested?”

How to Think About Investing in This Mess

When an entire niche blows up at once, the temptation is either to run away or to blindly buy the dip. Neither is a process. The way to approach Medicaid insurers right now is to treat it like any other distressed but potentially durable business: start with what “normal” looks like, figure out what has to go right to get back there, and then ask whether the market is over‑ or under-pricing that path.

First, define a sane end state. For these names, that means a medical loss ratio that lives in a sustainable band (say mid‑80s to very low‑90s, depending on mix), a mid‑single‑digit to high‑single‑digit underwriting margin, and a mid‑teens ROE. That’s roughly what these businesses looked like in the “boring good” years before the unwinding and GLP‑1 shock. If you assume they can get back to something in that neighborhood, you can back into what an earnings power number and a reasonable multiple might look like a few years out.

Second, build a simple checklist for separating value from value trap:

Are rates catching up? Look for evidence in state filings and management commentary that capitation renewals are actually reflecting higher acuity and drug trend, not just hand‑waving.

Is the MLR stabilizing? One bad year is a shock; three years of elevated MLR with no sign of flattening is a broken pricing loop. You want to see the MLR stop climbing and start oscillating in a tighter range, even if it’s off the old lows.

How exposed is the book? Break out Medicaid vs Medicare Advantage vs ACA. A plan that’s hugely concentrated in Medicaid expansion adults in the most aggressive cost‑cutting states is very different from one with a more balanced mix and better states. (I’m not going to unpack every state and product line in this writeup, but mix matters; that’s something I’ll dig into name by name in the follow‑up.)

What’s management actually doing? Are they exiting bad contracts, walking away from underpriced counties, and being disciplined in RFPs, or are they “chasing volume” to keep the top line growing? In a capped world, growth for growth’s sake is how you die.

Balance sheet and cushion. Thin margins plus policy risk means you don’t want a levered balance sheet or a management team that’s spending every spare dollar on buybacks at the wrong time. You want enough capital and conservatism to survive a few more bad rate cycles.

Finally, sanity‑check valuation against that framework. If you assume a reasonable normalized margin and the current price already bakes in “this business never recovers and earnings stay permanently impaired,” you might have the makings of a value investment. If, on the other hand, you have to assume heroic rate increases, perfect GLP‑1 management, and a friendly political backdrop just to get to today’s price, you’re probably staring at a value trap. The goal here isn’t to be optimistic; it’s to be roughly right about what a dull, steady‑state version of this industry looks like and only pay for that, not for a fantasy.

Where Does That Leave Us?

We’ve got an industry that was already running on thin underwriting margins, hit by a three‑part shock: a sicker, poorer post‑unwinding membership pool that the old rates weren’t built for; a pharmacy trend led by GLP‑1s and other high‑cost drugs that went from footnote to line‑item; and a policy turn (One Big Beautiful Bill) that shrinks enrollment at the margin and makes it harder for states to fund generous capitation increases. None of that is great news if you own the stocks at 20–25x peak earnings.

But it’s exactly the kind of setup that can create a mispriced opportunity. The core business model is still there: states outsourcing a critical function to a small group of scaled MCOs, on multi‑year contracts, serving populations that aren’t going away. The question is not whether Medicaid managed care survives; it’s what “normal” margins and multiples look like once the risk pool, rates, pharmacy trend, and new rules find an uneasy truce. If the market is pricing in a permanent collapse in economics where what we really have is a nasty but ultimately cyclical reset, that’s where the value may be.

This writeup was about understanding the storm at the industry level—how the model works, what broke, and why the headlines look as ugly as they do. The next step is where it actually gets interesting: going company by company, walking through rate adequacy, medical loss ratio trends, product and state mix, balance sheets, and management behavior to see which of these names, if any, deserve to come out of the “messy Medicaid” bucket and into the portfolio.

Subscribe now

Disclaimer

This analysis is educational commentary based on publicly available SEC filings, earnings calls, and published reports. It is not investment advice, and readers should conduct their own due diligence. The author has no position in Shift4 Payments. The author may hold positions in other securities discussed. Past performance is not indicative of future results.