The young-healthy premium gap

How the rules move cost onto the young

The ACA lets individual and small-group insurers vary price by only four things: age, tobacco, geography and family size. Everything else that predicts cost is pooled. Four rules do the work.

Community rating on health status

A 26-year-old who runs marathons pays the same as a 26-year-old with Type 1 diabetes. The largest transfer — bigger than age.

The 3:1 age band

A 64-year-old can be charged at most 3× a 21-year-old. Pre-ACA the spread was ~4.8×, roughly the cost spread. The SOA found ages 21–29 are charged ~27% more than their cost; age 21 alone ~50%.

No gender rating

Pre-ACA, young women paid up to 45% more than young men. Now pooled, so young men subsidise young women’s higher utilisation (largely maternity).

Guaranteed issue + loss-ratio floors

Anyone can buy at any health status in open enrollment, so the pool skews sicker. Insurers keep up to 20% (80% MLR floor); the individual market ran ~85% in 2024.

The design is intentional. The law needs young, healthy enrollees to overpay to fund older ones; KFF describes the system as depending on young people enrolling in enough numbers to generate that surplus. The problem is that the person funding it never sees the bill itemised.

The fair price, not the $400

Comparing a ~$5–6k premium with ~$400 of care ignores what insurance is for: the 1-in-200 cancer diagnosis, the ICU stay. The thesis survives that objection, and is sharper once it’s answered. So we priced a healthy young adult’s year as a deliberately tail-heavy distribution.

claims curve · illustrative · tail-heavy

Scenario (non-preventive)ProbabilityAllowed cost

Routine — a sick visit or generic Rx60%$150

Minor acute — urgent care, labs, imaging26%$1,000

Moderate — ER with CT, outpatient procedure11%$5,000

$5,000 deductible — the plan pays $0 above this line

Major — inpatient stay or surgery2.5%$35,000

Catastrophic — cancer, major trauma, ICU0.5%$150,000

Plus $250 a year of preventive care every plan covers at $0. Expected total care: $2,775 a year — MEPS puts the bottom half of US spenders at $374 (2019), so this overstates a healthy 26-year-old’s cost and understates the gap.

The key is what the deductible does. Under a $5,000 deductible the plan pays $0 in three of five scenarios. The member self-insures the first $5,000 and pays full premium. The insurer’s expected payout comes almost entirely from the bottom two rows: ~$1,340 of the ~$1,590 it expects to pay.

Fair premium= E[claims the plan pays]loss ratio = $1,5900.85 ≈$1,870

So the honest framing isn’t “premium vs. $400.” It’s premium vs. the fair price of catastrophe cover — about $1,900, against a $6,800 bill. Same conclusion, defensible math.

Where the premium dollar goes

Less than a third of a healthy young adult’s premium pays for their own expected claims. Health-status pooling, not age, is the largest leak.

Your own expected claims Age-band subsidy to older members Health-status / pool subsidy Insurer admin, broker, taxes, margin

Aged-off freelancer$6,827

$1,590 · 23%

$1,234 · 18%

$2,979 · 44%

$1,024 · 15%

yours · 23%

Big tech employee$10,735

$2,052 · 19%

$7,395 · 69%¹

$1,288 · 12%

pool subsidy · 69%

Lean minimizer$5,032

$1,575 · 31%

$909 · 18%

$1,793 · 36%

$755 · 15%

¹ Group plans don’t age-rate per employee, so age and health subsidies can’t be separated; it’s one transfer to the rest of the workforce. Method: the claims budget is premium × loss ratio; the age share uses the SOA finding that ages 21–29 are charged ~1.27× their cost; health pooling is the residual. The age-vs-health line is directional; the total transfer is not.

Why 2026 made it worse — and 2027 will again

The target customer took three hits in one year, and none of the relief went to them.

Benchmark premiums

+26%

Age-40 benchmark $497 → $625 a month. Insurers filed a median ~15% increase for 2027.

The 400% FPL cliff

$62,600

Enhanced credits expired 31 Dec 2025. Above this single income, the subsidy is $0 — a cliff, not a phase-out.

Even below the cliff

9.96%

Required contribution at 400% FPL is $6,235 — already above a 26-year-old’s ~$6,010 benchmark. Age runs out the subsidy before income does.

Max out-of-pocket

$12,000

$9,200 (2025) → $10,600 (2026) → $12,000 (2027). The worst year gets worse as premiums rise.

Older and lower-income enrollees are partly shielded by subsidies that scale with premium. The healthy, single, above-cliff young adult absorbs the full sticker increase.

Why the young can’t just opt out

The overpayment is sticky because every exit has a real cost. That stickiness is part of the problem, not a footnote.

The tail is real. ~3% of years in the model involve a $35k+ event. Going bare turns a $5k-a-year overpayment into an uncapped five- or six-figure exposure.
Mandates. California, Massachusetts, New Jersey, Rhode Island and DC penalise going uninsured — in California, at least $950 per adult.
Enrollment lock-in. Guaranteed issue only applies in open enrollment or after a qualifying event. You can’t buy when you get sick, so you buy before — at the pooled price.
Employer plans are opt-out in name only. Declining doesn’t return the employer’s ~$9k contribution as cash; it’s forfeited.
The lean options are still pooled. Bronze and catastrophic plans are cheaper because they pay less, not because they price your risk. The lean minimizer’s gap stays at ~63%.

The result: a customer who is rationally overcharged, knows it vaguely, and has no clean way to stop paying it.

Open questions we’re testing next

Calibrate the claims curve

Replace the illustrative distribution with real claims data for healthy 21–30 enrollees (HCCI, MEPS microdata, or a carrier). The tail probability drives the fair premium more than anything else.

Size the population

How many unsubsidised 26–34-year-olds sit above 400% FPL in the individual market, and how many more in employer plans?

Persistence of “healthy”

MEPS shows high spenders mostly stay high and young adults mostly stay low. How much of a 26-year-old’s healthy status survives to 35?

Who would object hardest

Every dollar recaptured is a dollar the pool loses. Insurers, regulators and older enrollees all have standing — worth mapping before any solution work.

Tax treatment

Self-employed premiums are deductible and employer premiums are pre-tax, which shrinks the after-tax gap for the first two personas. Worth modelling after-tax.