Rising MLRs: Why Health Plans Are Rethinking Lines of Business

This is Part 2 of our MLR series. Part 1, the consumer side, covered the rebate mechanics, the $759 million number, and who gets the checks. This post is the strategic sequel: what the same data means inside the plan that writes the checks.

Last month, Elevance Health (the country’s second-largest health insurer) told investors it was walking away from Washington, D.C.’s Medicaid managed care program. It also signaled that more state exits would follow over the next 18 months. The official language cited “elevated medical costs” and “ongoing profitability challenges” in the post-redetermination Medicaid population. What none of the earnings materials said explicitly is that the MLR in that business had stopped being a regulatory compliance number and had become a capital-allocation veto.

The MLR was designed as a consumer floor. The Affordable Care Act’s 80/20 rule (rebranded from a wonky actuarial metric into a populist guardrail) said: if you do not spend at least 80% of premium dollars on medical care (85% for large groups), you write a check back to your members. It was transparent, enforceable, and politically durable. Fifteen years later, it has quietly become a ceiling, not on what insurers must spend, but on what they can earn. And the plans that live under that ceiling every quarter are making decisions that have nothing to do with rebate checks and everything to do with whether a line of business survives at all.

The MLR stopped being a consumer number. It is now a portfolio signal.

When Humana announced in February 2023 that it would exit the Employer Group Commercial Medical Products business (fully insured, self-funded, and Federal Employee Health Benefit plans), the press release described the business as “no longer positioned to sustainably meet the needs of commercial members over the long term.” A cleaner translation: the employer group MLR runs persistently near or above the 85% threshold, and the administrative cost load on a fully insured commercial book leaves almost no room for margin without hitting the rebate trigger. Humana’s CEO, Bruce Broussard, was explicit about the destination: the company would “focus our health plan offerings primarily on government-funded programs (Medicare, Medicaid, and Military) and Specialty businesses.”

That pivot, from commercial employer to government, is not a one-off. It is a structural reallocation of capital happening across the industry, and the unifying signal is the MLR.

The 80/20 rule was designed as a consumer floor. But the math of a floor looks the same as the math of a ceiling when you are the one standing under it. A plan that runs an 87% MLR in the individual market (80% threshold) has a 13% corridor for administrative costs, taxes, and profit. That sounds adequate until you subtract the 8-10% most insurers spend on admin and the 2-3% that goes to premium taxes, and you are left with a 0-3% pre-tax margin on a book that carries utilization risk. In the large-group market, where the threshold is 85%, an 88% MLR leaves 12% for everything else. That is a margin so thin that any adverse claims year pushes you straight into rebate territory.

When a plan’s MLR in a given line of business runs persistently above the rebate threshold, the math tips from “thin margin” to “capital destruction.” The rebate machinery, originally built to protect consumers from insurer profiteering, now acts as a hard profitability cap, and plans are responding by exiting the markets where the cap bites hardest.

The three-year average masks the real-time decision.

Here is the structural tension that makes the MLR a uniquely dangerous management metric: the rebate calculation uses a rolling three-year average, but the internal decision to stay or exit a line of business uses real-time data. (For an explainer of that rebate mechanism, see Part 1 of this series).

A plan can be losing money right now while the three-year MLR still looks adequate, and sometimes even adequate enough to avoid rebates. That creates a window where the external number says “fine” and the internal P&L says “get out”. By the time the three-year average catches up to the deterioration, the plan has already decided. The rebates that eventually arrive are not the cause of the exit; they are the trailing confirmation of a decision that was made months or years earlier.

The individual market is the clearest case study. In 2025, the simple loss ratio (total claims divided by total premiums, the closest real-time proxy for the regulatory MLR) averaged 93% across individual-market plans, according to KFF’s analysis of insurer filings. At 93%, a plan has seven cents of every premium dollar left for everything: administration, quality improvement, taxes, risk margin, and profit. The ACA MLR threshold for individual plans is 80%. So the ratio looks generous by that standard, but the practical margin is razor-thin, because individual-market plans carry disproportionately high administrative costs (broker commissions, risk-adjustment administration, exchange user fees) relative to group coverage.

Contrast this with the large-group market, where the 2024 simple loss ratio was approximately 88%. The threshold is higher, 85% instead of 80%, but the absolute corridor is wider both because the denominator (premiums) is larger per member and because administrative costs run structurally lower in employer-group business. A plan running 88% in large group has the same percentage margin as one running 83% in individual, but the real economics favor the group plan because its admin costs do not eat the entire spread.

The three-year averaging mechanism obscures all of this. The MLR data that reaches the public, and the rebate checks that reach consumers, reflects 2022-2024 experience, not the 2025 experience that is driving current management decisions. The median state-level individual market MLR filing from Mark Farrah Associates data can show a 2024 three-year average of 84% while the contemporaneous simple loss ratio for 2025 is 93%. That nine-point gap is the space where capital-allocation decisions get made before the public data catches up.

Which lines of business are under the most pressure right now.

Not all markets are under equal strain. The MLR pressure lands differently across the four major lines of business, and the response from plans is different in each.

Individual market (on-exchange): the margin trap

The individual market is the most structurally difficult line of business for MLR management, and it shows. Premiums are rising approximately 20% for the 2026 plan year. That sounds like a pricing win for insurers, but the MLR creates a catch-22: if premiums rise faster than claims, the MLR falls, and plans owe rebates; if claims rise with premiums (which they tend to do in a market where the risk pool has aged and grown sicker), the MLR stays high and margins stay thin. The plans that survive in the individual market are the ones that can thread that gap year after year, which typically means they have scale advantages, tight provider network management, and sophisticated risk-adjustment operations.

The dynamics also create a forward-looking squeeze. Elevated 2026 premiums, if claims do not rise proportionally, will generate elevated margins in 2026, which translates to rebate exposure in 2028-2029 when those years enter the three-year average. But the plans that stay must survive two years of razor-thin margins first. The dilemma: price high to survive and risk future rebates, or price low to maintain market share and risk near-term losses. Several smaller carriers have chosen a third option: exit.

Medicare Advantage: the compression from above and below

Medicare Advantage has historically been the most profitable line of business in health insurance. In 2024, per-enrollee gross margins averaged $1,655, nearly double the fully insured group market ($846) and more than two-and-a-half times Medicaid managed care ($608). The headline MLR in MA was approximately 90% in 2024, comfortably above the 85% regulatory minimum.

So why is MA under pressure? Because the MLR itself is not the problem; it is everything compressing the effective margin around it.

The V28 risk-adjustment model phase-down is reducing revenue per enrollee by changing how diagnoses translate into payments (see our CMS-HCC V28 deep dive). Star Ratings revenue is at risk for plans that slipped in the 2026 ratings release (see Star Ratings Cut Points), and unlike the MLR, Star Ratings bonus revenue is not subject to rebate. RADV (Risk Adjustment Data Validation) audit exposure (see Prospective vs. Retrospective Risk Adjustment) creates a contingent liability that narrows the effective margin even when the reported margin looks healthy. Together, these forces are compressing the spread between premiums and claims from both sides: revenue is softening while utilization, especially in outpatient and specialty drug categories, is rising.

The result is that MA gross margins fell 17% from 2023 to 2024. That is not a crisis yet, $1,655 per enrollee is still a healthy number, but the trajectory matters. Plans are responding not by exiting MA (almost no one is doing that), but by narrowing their footprints to the counties and contracts where their Star Ratings, provider relationships, and risk scores give them an advantage. The MLR is the quiet denominator in all of those calculations.

Medicaid managed care: the predictability premium

Medicaid managed care had the highest simple loss ratio of any market in 2024 at 91%, up from 88% in 2023. The three-point jump is the largest single-year increase in any market, and it reflects the post-redeterminations reality: as states disenrolled roughly 15 million people between April 2023 and December 2024, the remaining population skewed sicker and more expensive, while the denominator (premium revenue) shrank unpredictably.

The Elevance D.C. exit is only the most recent and most prominent example. Several carriers have exited state Medicaid contracts in 2024-2025, and the pattern is consistent: the states where exits happen are the ones where (a) the MLR requirement is aggressive (some states go above the federal 85% floor), (b) the redeterminations-driven membership churn makes premium forecasting unreliable, and (c) the state’s rate-setting process lags the actual acuity shift by 12-18 months. In those conditions, the MLR becomes not just a profitability metric but a cash-flow metric: plans are paying claims at 2024-2025 acuity levels on 2023-2024 premium rates.

A majority of states now require Medicaid plans to pay remittances if they fail to meet MLR thresholds, and unlike the commercial market, there is no standard three-year averaging provision; state rules vary. The compliance burden and financial exposure are both higher, and the margin to absorb errors is lower. For plans evaluating which state Medicaid contracts to retain and which to shed, the state-level MLR filing is the single most important piece of data on the table.

What the data tells plans that plans will not say publicly.

There is a quiet signal in the MLR data that is invisible in quarterly earnings calls but visible to anyone who knows where to look: a sustained simple loss ratio above 95% in a state-level market is the leading indicator of an exit, typically appearing 12-18 months before the public announcement.

This signal works because state-level MLR data, the same Mark Farrah Associates filings that KFF uses for its rebate analysis, reveals line-of-business profitability at a granularity that public financial reporting does not. An earnings call can report “consolidated Medicare Advantage MLR of 89.5%” while obscuring that one state contract is running at 97% and another at 84%. The states at 97% are the ones where exits happen: not immediately, but after two or three quarters of confirming the trend is not noise.

The corollary is equally important: the plans that stay in a line of business when competitors exit are the ones that can move the numerator, not just the denominator. Premium pricing (the denominator) has a ceiling, set by competitors, regulators, and the MLR rebate threshold itself. Medical spending (the numerator) can be managed downward through quality improvement spend, care management, accurate HCC capture, and provider network optimization. A plan that reduces its claims cost by 2% through better care management accomplishes the same margin improvement as a 3% premium increase, but without the rebate exposure that the premium increase creates.

This is where the internal link to Quality Health’s platform becomes concrete. The plans that are thriving in high-MLR markets are not the ones with the highest premiums; they are the ones with the most effective care-gap closure programs, the most accurate risk-adjustment documentation, and the tightest quality-improvement-to-claims ratio. These are not insurance functions; they are healthcare operations functions. And they are the only sustainable lever for margin improvement when the MLR has already consumed the pricing runway.

The MLR as a strategic dashboard

The July 14 post explained what the MLR rebate is and who gets it. That is the public-facing story. The strategic story, the one that matters inside the plan, is that the MLR data, sliced by state and line of business, is the closest thing the industry has to a real-time capital-allocation dashboard. It tells you which markets are generating enough margin to reinvest, which ones are treading water, and which ones are consuming capital that could be deployed elsewhere.

The plans that read the dashboard well, that see the 95% loss ratio in a state Medicaid contract and act before the three-year average flags it, that recognize the individual-market margin compression for what it is rather than waiting for the rebate cycle to confirm it, that invest in the numerator-side levers (quality, care management, risk adjustment) before the denominator-side levers (premium pricing) are exhausted, are the ones that will still be in those markets five years from now.

The plans that do not will be writing exit announcements that cite “strategic repositioning.” The data will already have told the story.

Sources

• KFF, “2026 Medical Loss Ratio Rebates,” July 13, 2026. kff.org
• KFF, “Health Insurer Financial Performance in 2024,” February 23, 2026. kff.org
• Humana, “Humana to Exit Employer Group Commercial Medical Products,” February 23, 2023. humana.com
• STAT News, “Elevance Health to Shrink Medicaid Portfolio Amid High Costs,” July 15, 2026. statnews.com
• CMS, Medical Loss Ratio (45 CFR Part 158) reporting data. cms.gov
• Mark Farrah Associates, Health Coverage Portal.

See also: Part 1, ACA MLR Rebates in 2026: $759M and the Cycle Behind the Number, for the consumer-side explainer and interactive rebate tool.