In 2023, UnitedHealthcare denied one in every three claims it received. That same year, its parent company, UnitedHealth Group, reported $22.4 billion in net earnings, and its stock reached record highs. Those two facts are not in tension. They are the same fact.
The health insurance profits generated by America’s largest insurer did not come primarily from pricing risk accurately or managing care efficiently. A significant portion came from saying no and counting on most people not to fight back.
The mechanics are straightforward once you see them.
When an insurance company denies a claim, one of three things happens. The patient appeals and wins, which costs the company the claim plus administrative time. The patient appeals and loses, which costs the company administrative time but saves the claim. Or the patient does not appeal at all, which costs the company nothing.
Most people do not appeal. The process is confusing, the deadlines are short, and most sick people do not have the bandwidth to navigate a bureaucratic fight while also being sick. CMS data shows that approximately 80 percent of appealed UnitedHealthcare Medicare Advantage prior authorization denials were partially or fully overturned. Eight in ten. Which means the original denial was wrong eight times out of ten. And the overwhelming majority of those wrong denials were never challenged.
That gap between the number of denials issued and the number of appeals filed is where the money is.
Denying claims one at a time, with a doctor reviewing each case, is expensive. UnitedHealthcare found a more efficient solution.
The company acquired NaviHealth, a data analytics firm, and deployed its nH Predict algorithm to manage prior authorization decisions at scale, particularly for post-acute care, the rehabilitation and nursing facility care patients need after leaving a hospital. A US Senate Permanent Subcommittee on Investigations found that UnitedHealthcare’s post-acute care denial rate rose from 8.7 percent to 22.7 percent in the years directly following that algorithm’s deployment.
A separate federal lawsuit alleges the algorithm had a known 90 percent error rate. The company used it anyway.
What the algorithm produced was not better medical judgment. It produced faster denials at lower cost, applied consistently across hundreds of thousands of patients, most of whom would never appeal. The health insurance profits that followed were, in part, a direct output of that system.
In January 2025, a plastic surgeon in Austin, Texas, named Dr. Elisabeth Potter, was in the middle of a breast reconstruction surgery for a cancer patient when a note arrived in the operating room. UnitedHealthcare was on the phone. They needed to speak with her. It was urgent.
She stepped out. The patient was asleep on the table. The UnitedHealthcare representative wanted to know if the overnight stay was really necessary. Dr. Potter asked if they understood the patient was currently in surgery. The rep said that the information was in a different department.
She posted a video about it. 5.5 million views. UnitedHealthcare sent her a legal threat and told her to delete it. By late 2025, NBC News reported the two-year dispute may force her into bankruptcy.
That story is not an outlier. It is the system working exactly as designed, generating friction that most people absorb silently, surfaced publicly only because one doctor decided not to stay quiet.
Standard economic theory holds that competition drives efficiency and lowers prices. In most markets, that is broadly true. Health insurance is not most markets.
Switching insurance companies is difficult, expensive, and often not available to most people. Employer-sponsored plans limit choices, network restrictions tie patients to specific providers, and the administrative cost of changing plans mid-treatment is prohibitive for anyone managing a serious condition. The market power of large insurers in most regions is substantial. KFF analysis of ACA marketplace data found that UnitedHealth Group denied 33 percent of in-network claims in 2023, twice the national average, without losing market share in the states where it operated.
In a competitive market, a company that denied twice as many claims as its competitors would lose customers to those competitors. That did not happen. Because in most markets, the customers had nowhere else to go.
The result is a structure that economists call monopsony, concentrated buyer power that allows the dominant player to set terms that suppliers and customers have to accept. UnitedHealthcare is not the only insurer in America. But in enough local markets, across enough employer-sponsored plans, it is effectively the only option for millions of people.
The insurance industry’s argument for prior authorization is not entirely without merit, and presenting it honestly matters.
American healthcare spending is genuinely unsustainable. The US spends more per capita on healthcare than any peer nation while achieving worse outcomes on most population health indicators. Some of that spending is driven by unnecessary procedures, overdiagnosis, defensive medicine, and pharmaceutical overselling. Prior authorization, in its original design, was meant to catch that waste before it became a paid claim.
Managing 70 million members without any utilization review would, the industry argues, push premiums to levels that price even more Americans out of coverage. The administrative costs of the prior authorization system are real, but so are the costs of unlimited open approval.
These arguments have merit. The problem is that a system where 80 percent of appeals succeed is not catching waste. It is denying necessary care at scale and relying on patient exhaustion to make most of those denials stick. That is not utilization management. That is a revenue strategy.
On December 4, 2024, UnitedHealth Group CEO Brian Thompson was shot and killed outside a hotel in Manhattan. The messages found on the shooter’s ammunition, “deny,” “defend,” “depose,” generated more public discussion about health insurance claim denials than years of policy advocacy had managed to produce.
Within twelve months, UnitedHealthcare’s denial rate dropped from 33 percent to 20 percent. The largest single-year improvement of any major national insurer.
The rate did not fall because the company fixed its algorithm or changed its financial incentives. It fell because the reputational and regulatory pressure became impossible to manage at the previous level. Which tells you something important: the prior behavior was a choice, not a necessity. A company that can reduce its denial rate by 13 percentage points in one year when it needs to was never constrained by actuarial requirements. It was constrained only by the absence of consequences.
The health insurance profits continue. UnitedHealth Group’s revenue exceeded $400 billion in 2024. The structural incentives that produced the 33 percent denial rate are still in place. The algorithm still runs. The appeals window is still short. As examined in the analysis of the recession risk that household debt is building into the broader economy, the medical debt that is denied claim generates is now a macroeconomic problem, not just a personal one.
The political response to the Thompson assassination produced real but limited change. HHS held a press conference. Insurance companies made voluntary commitments to reduce prior authorization burdens. Congress accelerated hearings on the issue. Some states passed legislation strengthening appeal rights.
None of it touched the underlying incentive structure. Insurance companies are still compensated based on the gap between premiums collected and claims paid. The wider the gap, the better the financial results. Every claim denied is a contribution to that gap. Every patient who does not appeal is a contribution to shareholder returns.
The pattern of how institutional capital extracts value from services people cannot avoid is consistent across American healthcare. The specific mechanism differs, prior authorization here, private equity ownership there, but the logic is identical. Find the inelastic demand. Control the access. Extract while the regulatory friction is low enough to permit it.
Health insurance is the most inelastic demand in the economy. You cannot choose not to need medical care.
Fixing this requires changing what insurance companies are rewarded for, not just regulating individual practices.
Countries that have managed to provide universal or near-universal healthcare coverage at lower cost, such as Germany, France, the Netherlands, and Japan, have done so through a combination of regulated pricing, mandatory coverage requirements, and nonprofit or heavily regulated insurer structures that remove the profit motive from the denial decision. None of them has eliminated private insurance. All of them have removed the specific financial incentive that makes denial a revenue strategy.
The US political environment makes those structural changes difficult. The insurance industry spent hundreds of millions on federal lobbying in the five years through 2025. The legislative proposals that would most directly address the incentive structure of a public option that competes with private insurers, mandatory loss ratio requirements that cap how much of premium revenue can go to non-claims expenses, and binding arbitration for denial disputes have stalled repeatedly under that pressure.
As examined in analysis of how economic anxiety translates into political outcomes, the gap between what voters say they want from healthcare and what the legislative system produces is one of the most persistent features of American political economy. Polling consistently shows overwhelming support for prior authorization reform. The legislative calendar consistently fails to deliver it.
UnitedHealthcare denied 33 percent of claims in 2023. Its stock hit record highs. In 2024, under pressure it could no longer ignore, the denial rate dropped to 20 percent.
The stock fell.
UnitedHealth Group’s share price declined significantly through 2025 as the combination of regulatory scrutiny, reputational damage, and a Department of Justice investigation into its Medicare billing practices compressed its earnings multiple. The market, in its own way, confirmed what the data had been saying all along: the health insurance profits were not built on efficiency. They were built on denial.
The system is still running. The incentives are still in place. One hundred million Americans still carry medical debt. And somewhere right now, a claim is being denied by a wrong algorithm 90 percent of the time, for a patient who will probably not appeal, generating a fraction of a cent of profit for a company whose CEO was replaced after his predecessor was shot on a Manhattan sidewalk.
That is not a broken system. That is the system working exactly as it was built to work.