More than 1,800 American employers were asked what they expect to spend on health benefits next year. The answer, published on Wednesday, is that the cost per employee will rise 8.2 percent in 2027 — the steepest increase since 2003, and up from 6.7 percent this year.
That is the number that will be quoted. It is not the number that describes what is happening to the price of care.
The 8.2 is the answer after the homework
The same survey reports that without the changes employers plan to make, the increase would be about 11 percent. The gap between the two figures — call it just under three percentage points — is not a forecast of restraint in the health system. It is a forecast of what employers will do to their own plans in response.
Fifty-nine percent say they will make such changes. The survey is specific about the most common one: roughly two-thirds of employers with 500 or more staff expect to raise the share of the premium the employee pays.
So the widely reported figure is already net of a transfer. The 8.2 percent is the increase in what the employer carries. The eleven is closer to the increase in what the care costs. The difference between them lands on a payroll deduction.
What the employee's number actually is
For a household, the arithmetic runs the other way. The employee absorbs their share of the underlying increase and then absorbs the shift on top of it. A worker at a large employer raising contributions does not experience 8.2 percent. They experience something between 8.2 and eleven, depending on how much of the mitigation their employer took out of their paycheque rather than out of the plan.
That is worth stating plainly because the two numbers get used interchangeably in open-enrollment coverage every autumn, and they are not the same number pointing at the same person.
The other lever is plan design. Twelve percent of large employers say they will offer variable copay plans in 2027, rising to 18 percent among those with 20,000 or more staff, and more than a third say they will offer some non-traditional medical plan. Fifty-eight percent say they will steer members toward higher-quality providers — which is the one mitigation on the list that could reduce the cost of the care rather than move it, and also the slowest to show up in a premium.
Three drivers, and one of them is new
The survey names the causes. Two are familiar. Advances in diagnostics and therapeutics cost money, and GLP-1 medicines alone account for about one percentage point of the growth — a single drug class carrying roughly an eighth of the increase. Health system consolidation is the second, and it is the one this desk has written about as a question of who owns the practice: a consolidated provider negotiates prices from a stronger position, and the price is what the employer pays.
The third is newer. The survey cites AI-enabled billing software raising the volume of claims.
That deserves to be read carefully, because it is not a claim about fraud. Coding software that is better at identifying billable elements of an encounter produces more billable elements per encounter. Every one of them may be defensible. The aggregate is still a larger bill for the same visit, arriving faster and in greater volume than the payer's review capacity was sized for.
It is the first time the AI build-out has shown up in this paper's pages as a line in somebody's health premium rather than as capital expenditure, and it will not be the last. The technology was sold as an efficiency. On the provider's side of the ledger it is one. Efficiency at generating claims is a cost to whoever receives them.
Why employers keep reaching for the same lever
Cost-shifting is the most reachable mitigation because it is the only one an employer controls unilaterally and can implement in a single plan year. Steering to higher-quality providers requires data, contracts and a change in employee behaviour. Renegotiating with a consolidated hospital system requires leverage the employer may not have. Raising the contribution requires a decision.
This is the same pressure that has pushed employers toward subscription primary care and back into running their own clinics — attempts to buy care directly because buying insurance against care keeps getting more expensive. Those experiments are small. The contribution increase is universal.
Employers now also have hospital price data that is finally usable, which in principle makes the steering strategy work. In practice, knowing that one hospital charges three times another does not help an employer whose staff live nearer the expensive one.
The number to watch
Not 8.2 percent. Watch the employee contribution share — the percentage of the total premium deducted from pay — and watch it against wage growth in the same year.
If contributions rise faster than wages, the reported 8.2 percent understates what happened to households by exactly the amount employers succeeded in moving. If they rise more slowly, employers absorbed the increase, and the mitigation in this survey came from plan design rather than from payroll.
One of those is a cost-control story. The other is a wage story wearing a benefits uniform.
The 8.2 percent projection for 2027, the 11 percent figure before plan changes, the 6.7 percent figure for 2026, the sample of more than 1,800 US employers, the roughly one percentage point attributed to GLP-1 medicines, the 59 percent of employers planning changes, the share of large employers expecting to raise premium contributions, the 12 and 18 percent variable copay figures and the named cost drivers are from the employer health survey published by Marsh McLennan on 2 September 2026 and reported the same day by The Hill, the Washington Post, the New York Times and Fierce Healthcare; several outlets attribute the survey to Mercer, its benefits business. The analysis is our own.
Topics healthemployersbenefits
