Every major employer health cost survey released this year lands within a few points of each other, and every one of them is worse than the last. What the surveys do not agree on is why five straight years of deductible hikes, narrow networks and vendor swaps have failed to bend the curve. Fetched primary data from an insurance broker, a regulator and a national benefits survey point to the same answer: the drivers behind 2027’s projected cost jump are not the ones health plans were built to manage.
The Compounding Math
Mercer’s own newsroom puts the baseline plainly. In its June 2026 employer survey, total health benefit costs were already projected to rise 6.7% in 2026, pushing the average cost per employee past $18,500, with prescription drug benefits alone climbing 9% for the year. That was before the 2027 numbers came in higher still, as Aon’s August 2026 health care cost report confirmed: a projected 9.5% jump in 2027, the fourth consecutive year of near double-digit growth, with average per-employee costs above $19,000.
“Employers have now experienced several consecutive years of health care cost increases that are approaching double digits,” said Mike Pasterick, North America Health Solutions Leader at Aon. Debbie Ashford, Aon’s North America Chief Actuary, put the response in structural terms: “The organizations best positioned for the future will be those that can proactively identify emerging risks and take targeted action before costs escalate.”
HRTech has covered the scale of this convergence before, and argued that the standard playbook of cost-shifting has already failed once. What the underlying data adds is a mechanism: four distinct forces are compounding at the same time, and none of them respond to the tools benefits teams have used for the last decade.
How the Five-Year Run Actually Adds Up
None of this is a one-year spike. Aon’s report frames 2027 as the fourth consecutive year of near double-digit growth, and the year-over-year detail shows why the compounding matters more than any single percentage. Employer-paid costs rose 8.8% in the most recent year Aon tracked, from $13,269 to $14,432 per employee. Total plan costs, employer and employee combined, rose 8.3%, from $16,212 to $17,562. Employee out-of-pocket costs rose faster than either figure, up 10.2%, from $1,966 to $2,167, even though employers still absorb the large majority of the total, 82.2% by Aon’s count.
KFF’s numbers, drawn from a broader employer sample, show the same shape over a longer window: average family premiums hit $26,993 in 2025, up 6% for the year and up 26% over five years, a pace that outstripped the 23.5% rise in general inflation over the same period though it trailed 28.6% wage growth. Average worker contribution toward family coverage now stands at $6,850 a year. The arithmetic that matters for a benefits budget is not the headline percentage. It is that five consecutive years of “unusually high” has quietly become the new baseline, and every plan built to absorb one bad year is now absorbing five.
Four Drivers, One Direction
GLP-1s Outran Every Budget Model
Weight-loss drug coverage is the clearest example of a cost employers did not model correctly. Mercer’s newsroom reports that 49% of large employers now cover GLP-1 medications for weight loss, but 6% dropped that coverage in 2026 and another 5% are planning to drop it or considering doing so in 2027. The reversal is not about the drugs failing. It is about utilization outrunning the actuarial assumptions plans were priced against.
The KFF 2025 Employer Health Benefits Survey shows how fast that adoption curve moved: among firms with 5,000 or more workers, GLP-1 coverage jumped from 28% in 2024 to 43% in 2025. Fifty-nine percent of large firms covering the drugs said utilization exceeded expectations, and 66% said the impact on prescription spending was significant. Mercer’s own pharmacy leader, Alysha Fluno, described the response taking shape: “The priority now is gaining more transparency, control, and confidence that every dollar spent is delivering maximum value,” she said, pointing to tightened utilization controls and renegotiated pharmacy benefit manager contracts as the emerging lever, not blanket coverage cuts.
A Regulatory Fix Became a Cost Driver
The least visible driver is the one with the clearest paper trail. The No Surprises Act’s independent dispute resolution process was designed to keep patients out of billing fights between insurers and out-of-network providers. Four years in, CMS’s own IDR reports page shows the mechanism has become a cost driver in its own right: 394,140 disputes were initiated in July 2026 alone, a 24% jump from June’s 318,421, bringing the cumulative total since the program’s April 2022 launch to more than 3.7 million disputes initiated and 3.3 million closed.
Every one of those disputes ends in a payment determination that flows back into premiums. Self-funded employers, who carry that cost directly rather than through an insurer’s risk pool, are the ones absorbing it first. It is a regulatory mechanism nobody projected onto a 2027 benefits budget three years ago, and it is now large enough to show up in one.
Consolidation and AI-Priced Claims
Provider consolidation and AI-enabled billing infrastructure round out the driver list. Aon’s report names both directly: growing prevalence of chronic conditions and high-cost claims alongside “provider adoption of AI and enhanced clinical documentation” as factors pushing per-claim costs higher. The pattern is consistent across insurers and brokers this cycle: AI is not lowering health system costs yet. It is making billing more precise, and more precise billing captures more revenue per claim.
Specialty drug spending outside the GLP-1 category compounds the same problem. Mercer reports 27% of large employers tightened, or plan to tighten, utilization controls on specialty medications generally in 2026 or 2027, a category that includes cancer treatment, a cost driver Aon lists separately from GLP-1s. The two drug categories do not compete with each other for budget attention. They add.
The Counter-Argument, and Why It Does Not Fully Hold
The obvious pushback is that this is cyclical: health costs have spiked before, employers redesigned plans, and growth moderated for a few years before the next cycle. That pattern held through the 2000s and 2010s. What is different now is that three of the four drivers above are structural rather than cyclical. GLP-1 utilization is not going to plateau back to pre-2024 levels once patients are established on the drugs. IDR dispute volume has grown every reporting period CMS has published, not fluctuated. And AI-enabled billing infrastructure is a permanent capability providers and payers are both investing in, not a temporary market condition. A plan redesign can blunt the impact of a cyclical spike. It cannot make a structural driver cyclical again.
There is a milder version of the counter-argument worth taking seriously: employers have absorbed structural cost growth before, through decades of medical inflation running ahead of general inflation, without health benefits becoming unaffordable at scale. That is true, but it assumes the absorption capacity is unlimited. KFF’s five-year premium data shows wage growth still narrowly outpacing premium growth nationally. If GLP-1 utilization, dispute volume and AI-priced claims keep compounding at their current rate while premium growth stays ahead of wages for even two more cycles, that assumption stops holding for a meaningful share of employers, not just the ones already furthest behind on plan design.
What HR Leaders Are Actually Doing About It
The response so far is mostly defensive. Mercer’s survey found 48% of large US employers expect to raise deductibles or copays in 2027, and 27% have tightened or plan to tighten utilization controls on specialty drugs. Non-traditional medical plans, built around narrower, pre-selected provider networks, are offered or planned by 31% of large employers, with another 38% considering the shift. Mercer’s press release also flags employers reevaluating pharmacy benefit manager contracting models outright, rather than accepting renewal terms, a sign that some of the cost pressure is being redirected at the vendor relationship instead of the employee.
KFF’s data shows where that cost-shifting lands: the average worker contribution for family coverage already sits at $6,850 a year, and one in five large firms report “high” employee concern about cost-sharing affordability. Camaj’s framing is the one worth sitting with: “Employers are using different levers to manage costs, both traditional cost-sharing tactics and strategies that guide their people to higher-value care.” The traditional lever, raising what employees pay, is doing most of the work. The higher-value-care lever, the one that would actually change the trajectory, is the one fewer employers have built the infrastructure to pull.
That gap between the two levers is where retention risk creeps in. Benefits are still one of the few remaining differentiators in total compensation that a competitor cannot match with a signing bonus. An employer that solves cost purely through cost-shifting is asking employees to fund the fix themselves, in a year when wage growth on KFF’s own numbers is already outpacing premium growth by a shrinking margin. That is a harder conversation at open enrollment than a percentage increase on a slide.
What It Means for the HR Leader
Four things follow from reading these data sets together rather than one survey at a time. First, GLP-1 coverage decisions need pharmacy-specific utilization data, not a plan-wide yes or no; the employers dropping coverage and the employers tightening controls are solving the same problem two different ways, and the tightened-control group is the one keeping access while managing cost. Second, self-funded employers should ask their broker directly how much of this year’s claims spend traced to IDR payment determinations rather than negotiated network rates; that number was close to zero five years ago and is not anymore. Third, none of the 2027 numbers assume a recession or a benefit design overhaul. They assume employers keep doing roughly what they did in 2026, only more expensively. Fourth, treat the PBM contract as an active negotiation this renewal cycle rather than a rollover, since transparency and control, in Fluno’s words, is where employers are finding the only real leverage left. A flat renewal is no longer on the table for most plans; the choice is between a planned redesign now or an unplanned one during open enrollment.
Source: Mercer
