Self-Funded, Fully Insured, or Level-Funded: The Cost Comparison, With Numbers

Every other decision in a health plan sits downstream of one: who holds the claims risk. KFF's 2025 employer survey puts average annual premiums at $9,325 for single coverage and $26,993 for family coverage. For a 200-employee company, that is a seven-figure line item, and the funding model determines how much of it is fixed, how much is variable, and who keeps the difference when the year runs well. Most comparisons stop at definitions. This one is a decision framework: how each model works, who bears which risk, where the break-even sits in dollars, and when level-funding is the right first step rather than the final answer.

How Each Model Works

Three models, one spectrum. At one end the employer buys certainty by the month. At the other, the employer pays its own claims and buys insurance only for the outcomes it cannot absorb.

Level-funded is not a fourth kind of insurance. Legally it is a self-funded plan; financially it behaves like a fully insured one until the year-end settlement. That dual nature is exactly why it works as a transition vehicle, and why its contract terms deserve more scrutiny than its monthly bill suggests.

Who Bears the Risk

Claims risk never disappears. The employer either rents it out at a fixed price or holds it and caps it. Each choice has a bill.

In a fully insured plan, the carrier holds the risk and charges for the service. The ACA's medical loss ratio rules put a floor under what that costs: carriers must spend at least 80 cents of each premium dollar on claims and quality improvement in the small-group market, 85 cents in large group. The remainder funds administration, premium taxes, and margin. That spread is the permanent, structural price of certainty, and it is paid whether the year runs well or badly.

In a self-funded plan, the employer holds the working layer of risk and transfers the tail. The tail is not hypothetical: in Segal's SHAPE analysis of 2019–2023 claims, the top 1 percent of claimants drove roughly a third of medical spend, and the top 20 percent drove about 85 percent. Stop-loss exists to make that concentration survivable, and structuring it well is its own discipline. Our guide to specific vs. aggregate stop-loss and lasers covers the attachment points, corridors, and contract terms that decide whether the protection is really there.

In a level-funded plan, the carrier holds the risk inside the year: the monthly payment never moves, no matter what claims do. But the group's experience follows it to renewal. A heavy year comes back as a sharp increase, restrictive terms, or a nonrenewal, so the risk is deferred more than transferred.

The Real Cost Comparison: Fixed vs. Variable

A fully insured plan is one number. A self-funded plan is a build-up:

The savings in that example come from named places, not magic: retention dollars that never leave the company, state premium taxes of roughly 2 percent that self-funded plans largely avoid (the stop-loss premium is still taxed, and some states assess self-funded plans), mandated benefits the plan can decline to cover, and the employer keeping its own good experience instead of donating it to the carrier's pool. Stacked, they explain why published savings claims range so widely. Promotional material cites 15 to 30 percent. Read the build above instead: it lands at 8 percent in an expected year and 18 percent in a good one, and the same structure runs about 9 percent over the fully insured quote in the modeled worst case. Treat the savings as a favorable-years number, not an annuity: the structure that pays the employer in a good year charges it in a bad one.

Where the break-even sits depends mostly on size, because size is what tames volatility:

When Level-Funding Makes Sense

Level-funding earned its place in the market, and SHRM and others describe it as the steppingstone from fully insured to traditional self-funding, which is the right frame. The employer gets a fixed monthly bill it already knows how to budget, its first real claims data, and a refund check when the year runs well, all without taking on mid-year cash-flow swings.

The mechanics reward attention. The monthly payment is sized to the maximum claims liability plus fixed costs, which means it is set near the worst case, not the expected case. The value of the product therefore lives in the settlement terms, and those vary widely by contract.

One caveat belongs in every recommendation: level-funded products screen their risk. Underwriting favors groups with clean recent experience, which means the quote is least available exactly when the savings would help most. Price the fully insured fallback before proposing the switch, not after a declination. And treat level-funding as a bridge with a destination: once two or three years of data exist, rerun the three-way comparison and decide whether traditional self-funding is the next step.

What You Need to Self-Fund Confidently

Self-funding is an operating discipline, not a procurement event. The prerequisites are specific:

The macro backdrop raises the stakes in both directions. Surveys published in the summer of 2026 put 2027 medical trend between 9 and 11 percent before plan changes, with PwC at 9.0 percent for the group market and Segal at 9.9 percent for PPO plans, the highest readings in nearly twenty years. The fully insured premium for certainty is compounding fast, and so is the volatility a self-funded plan retains. A funding decision made once a decade is a decision made on stale numbers; the comparison deserves a fresh run at every renewal.

Before any switch, pressure-test readiness against the six prerequisites above, and model three scenarios for every option on the table: expected, adverse, and benign. A recommendation that only works in the expected case is not a recommendation.