Social Security and Medicare take 15.3 % of covered wages and everybody knows it. Health insurance takes more than that from the same pot, is invisible to the worker, and grew four and a half times over while wages grew a fifth. This is the rest of the deduction.
Two different measures, both here. BLS reports employer cost per hour worked, averaged across all workers — including part-timers and the roughly one in three private-sector workers whose employer covers nobody. KFF reports the premium for workers who actually have family coverage. The first tells you what the labour market costs on average; the second tells you what it costs if it is you. Neither is wrong; they answer different questions, and mixing them is how this subject gets argued badly.
The share that is not wages
30.1%
Of what a private-sector employer spends on an employee, 30.1 % is not wages. It is $14.01 of every $46.60 per hour worked.
March 2026. Health insurance alone exceeds the entire Social Security and Medicare payroll tax.
Source: BLS, Employer Costs for Employee Compensation, March 2026, table 1, private industry column.
Sources: BLS ECEC March 2026 for the hourly figures; KFF 2025 survey for the family-coverage employer contribution; payroll tax computed at the 2026 national average wage of $74,046.
The arc
The average family premium was $5,809 in 1999 and is $26,993 in 2025. Adjust both to 2026 dollars and it still 2.4×. The national average wage over the same period grew 1.22× in real terms.
The gap between the lines is inflation. The lower line is what the premium costs in constant 2026 dollars — and it still more than doubles.
Source: KFF/Kaiser-HRET Employer Health Benefits Surveys 1999–2025, family coverage, exported from KFF's published series. Deflated with the CPI series underlying the 2026 Trustees Report.
Same money, same starting point. By 2025 the premium is at 240 and the average wage is at 122.
Sources: KFF as above; average wage from SSA's national Average Wage Index.
At the kitchen table
Your share is not spared. The worker's own contribution to a family plan went from $3,101 to $7,008 in constant 2026 dollars — it 2.26×d in real terms while your real wage grew 22 %.
And that is before the deductible, which is a separate and also-growing bill.
In the committee room
The premium went from 19.1 % of the national average wage in 1999 to 37.5 % in 2025. Every dollar of it is compensation that could have been wages.
This is the largest untaxed, unindexed, unlegislated transfer in the compensation system, and no vote was ever held on its size.
The triggering moment
This is the part that gets argued wrong in both directions. Premiums did not decouple from wages steadily across twenty-five years. They exploded in one six-year window and have converged on wage growth in every period since — and over the last five years they have actually grown slightly slower than wages.
Percentage points per year by which the real premium outgrew the real wage. Above the line, premiums are pulling away; at or below it, they are not.
Computed from the KFF and SSA series above, both deflated to 2026 dollars, as compound annual real growth over each period.
At the kitchen table
If your premium feels ruinous, that is a level problem, not a current sprint. The bill you carry today was set by an explosion that finished twenty years ago and never came back down.
Anyone telling you premiums are currently running away from wages is describing 2003, not 2026.
In the committee room
The structural cause is older than the series. Wartime wage controls in 1942 pushed employers to compete with benefits instead of pay; a 1943 ruling and the 1954 tax code made those benefits permanently untaxed. That is the subsidy that built the system.
What the data can show is that the modern escalation is visible from the very first year KFF measured, so the inflection sits at or before 1999 — the series cannot see further back, and no honest reading of it can either.
Who gets the subsidy
A state or local government employer spends 2.2× what a private employer spends per hour on health insurance. But the genuinely enormous gap is retirement, at 5.6× — and on payroll tax, where the statute is identical for everyone, the two are indistinguishable.
Source: BLS ECEC March 2026, table 1. Note these are averages per hour worked across all workers in each sector, so the health gap reflects both plan generosity and the share of workers offered coverage at all — government coverage rates are far higher. It is not a like-for-like comparison of two identical plans.
At the kitchen table
The intuition that public employees get a better deal is correct, but the money is mostly in the pension, not the health plan. Retirement is where a government employer spends five and a half times what a private one does.
In the committee room
Federal is the surprise. The FEHB government contribution is statutorily the lesser of 72 % of the programme-wide weighted average or 75 % of the chosen plan — and the private-sector employer already pays 74.6 % of the average family premium.
The 2026 federal maximum for family coverage is $20,229 a year against a private-sector average of $20,143. The federal advantage is not the size of the subsidy. It is that the share is written into law and cannot be cut at renewal.
The same three mechanisms
The programmes on the other pages grew by a rate ratchet, a cap removal, and a frozen threshold. Every one of those recurs here.
$7,000 in 1983 is $22,713 in 2026 dollars — the base has lost 69 % of its real value. Sources: IRS Topic 759 and CRS R44527 for FUTA; SSA's contribution-and-benefit-base page for the Medicare ceiling; California EDD for SDI.
At the kitchen table
A Californian earning $300,000 saw their SDI bill go from about $1,378 to $3,900 without the rate ever explaining it. Removing the ceiling did that.
In the committee room
State paid family and medical leave programmes are the next cycle, standing up since 2019 at rates small enough that nobody objects. Social Security also began at 2 %.