When an employer self-funds its health plan, it stops buying insurance and starts being the insurer. There is no carrier standing between the company and its employees' medical bills. When an employee has surgery, the money to pay for it comes out of the employer's own funds. The company hires administrators to run the plan and buys insurance to cap its worst-case losses, but the fundamental relationship has inverted: instead of paying a premium to transfer risk away, the employer holds the risk and pays claims as they come.
This sounds like something only a large corporation would do, and for a long time it mostly was. But self-funding has crept steadily down-market, and the reason is a single federal statute passed in 1974 that gives self-funded plans a set of advantages no fully insured plan can touch. To understand why a company would choose to become its own insurer, you have to understand that law.
ERISA: the reason self-funding is powerful
The Employee Retirement Income Security Act of 1974 — ERISA — was written mainly to protect workers' pensions after a series of collapses, most notoriously the failure of Studebaker's pension plan, left employees with nothing. But ERISA also governs employee welfare benefit plans, health insurance among them, and buried in its structure is a provision that reshaped the entire group health market: preemption.
ERISA preempts state laws that "relate to" employee benefit plans. For self-funded health plans, this preemption is transformative, because it means a self-funded plan is generally exempt from state insurance regulation. State benefit mandates — the laws that require insured plans to cover, say, in-vitro fertilization or acupuncture or a particular cancer screening — do not bind a self-funded plan. Neither do state premium taxes, which can add a few percent to the cost of a fully insured policy. A self-funded plan answers to federal law and to the plan document the employer writes, not to the fifty-state patchwork of insurance mandates.
There is a crucial subtlety here, often called the "deemer clause." ERISA says states may not treat a self-funded plan as if it were an insurance company for purposes of state regulation. But states retain authority to regulate actual insurance — including the stop-loss insurance that self-funded plans buy. So a self-funded plan itself is beyond state insurance mandates, while the stop-loss policy protecting it is not, and states have used stop-loss regulation (setting minimum attachment points, for instance) as one of the few levers they have over self-funding, particularly to keep very small employers from using thinly-disguised stop-loss to escape the small-group rules.
Preemption is the whole reason self-funding is attractive, and it is also why the arrangement is contested. A multi-state employer can run one uniform national plan under one set of federal rules rather than complying with fifty different state mandate regimes — an enormous administrative and cost advantage. Critics point out that the same preemption lets self-funded plans opt out of consumer protections that state legislatures enacted deliberately. Both observations are true.
What the employer still owes: the ACA and federal floor
Escaping state insurance regulation is not escaping regulation. Self-funded plans remain subject to a substantial body of federal law, and the Affordable Care Act reaches them in specific, important ways.
A non-grandfathered self-funded plan must observe the ACA's ban on annual and lifetime dollar limits for essential health benefits, must cover recommended preventive services without cost-sharing, must extend dependent coverage to age 26, and must honor the annual out-of-pocket maximum — which for the 2026 plan year is capped at $10,600 for self-only coverage and $21,200 for family coverage. Self-funded plans also owe the Patient-Centered Outcomes Research Institute fee (the PCORI fee, reported on IRS Form 720), and they carry their own ACA reporting burden: a self-funded employer files the coverage information the government uses to administer the individual and employer mandates, and a large self-funded employer combines that reporting on Forms 1094-C and 1095-C.
What the ACA does not impose on a self-funded plan is the small-group machinery. A self-funded plan is not required to offer the full slate of essential health benefits as a benefit mandate the way a fully insured small-group plan is (though the out-of-pocket cap applies to whatever essential benefits it does cover), it is not bound by adjusted community rating, and it is not part of the single risk pool. That is precisely the freedom that makes self-funding — and its packaged cousin, level funding — appealing to employers who believe community rating is charging them more than their own risk warrants.
Beyond the ACA, self-funded plans must comply with COBRA continuation rules, HIPAA privacy and portability requirements, the Mental Health Parity and Addiction Equity Act, and ERISA's own fiduciary duties — because when an employer becomes the plan sponsor and holds plan assets, it takes on fiduciary responsibility for administering the plan prudently and in participants' interest. That fiduciary role is not a formality; it is a real legal exposure, and it is one of the reasons self-funding demands more sophistication than writing a premium check.
The machinery: TPAs, ASO, and stop-loss
A self-funded employer rarely processes its own claims or builds its own provider network. Instead it hires a third-party administrator, or contracts on an administrative-services-only basis with a carrier that rents out its network and claims platform without taking the risk. Under an ASO arrangement, a familiar carrier's name may still be on the ID card and members may not notice any difference, but the carrier is being paid a fee to administer a plan whose claims the employer is funding — a fundamentally different deal from fully insured coverage where the same carrier would be bearing the risk.
Stop-loss insurance is what keeps self-funding from being a gamble. As with level funding, it comes in two forms: specific stop-loss caps the plan's liability for any single individual's claims above a chosen attachment point, and aggregate stop-loss caps the plan's total annual claims. The attachment points are the dials the employer turns to set its risk appetite — a lower specific attachment means more protection and a higher stop-loss premium; a higher attachment means the employer keeps more risk in exchange for a cheaper policy. Setting those points well is a real actuarial exercise, and getting them wrong in either direction — over-insuring and giving away the savings, or under-insuring and exposing the budget — is a common and costly mistake.
The employer's cash flow under self-funding is genuinely different from anything on the fully insured side. Claims arrive unevenly. A quiet month costs little; a month with a serious hospitalization costs a great deal, and the money comes from the company's own account before any stop-loss reimbursement catches up. This is why self-funding rewards employers with the reserves and the stomach to absorb month-to-month volatility, and why it can strain a business that cannot.
The MEWA question and pooling for smaller employers
Self-funding's cost advantages have always tempted smaller employers who cannot self-fund alone to pool together and self-fund as a group — several businesses combining to share risk and administration. These arrangements are called multiple employer welfare arrangements, or MEWAs, and their history is a cautionary one. In the years after ERISA passed, some operators used the preemption argument to run MEWAs while claiming exemption from state oversight, and a number collapsed into insolvency or turned out to be outright fraud, leaving members with unpaid claims. Congress responded in 1983 by amending ERISA to explicitly permit states to regulate MEWAs — even self-funded ones — closing the loophole that let them escape both state and meaningful federal supervision. A small employer looking at a self-funded pooling arrangement today should treat the MEWA label as a signal to examine the arrangement's solvency and regulatory standing carefully, because the structure's past includes real failures.
Who self-funding fits
The classic self-funded employer is large enough that its claims experience is statistically credible — a bigger group's costs are more predictable, which makes self-funding's risk more manageable — and financially stable enough to weather a bad month without distress. For that employer, self-funding offers the fullest control: it keeps the dollars that a fully insured carrier's risk charge and margin would have consumed, it can design the plan exactly as it wishes rather than accepting a carrier's off-the-shelf options, it earns interest on reserves it holds rather than remits as premium, and it gains detailed visibility into its own claims data that a fully insured plan would keep behind the carrier's wall.
The trade is complexity and exposure. The employer becomes a fiduciary, takes on cash-flow volatility, must manage stop-loss and administration and compliance, and shoulders the possibility that a catastrophic year — even with stop-loss — is more stressful to run than simply paying a premium. Level funding exists precisely to give smaller and more risk-averse employers a taste of these advantages with the sharp edges sanded off, which is why many businesses arrive at true self-funding only after outgrowing a level-funded plan.
Self-funding is not a small-employer default and should not be treated as one, but the line for who can reasonably self-fund has moved down-market as stop-loss products and administrators have matured, and plenty of mid-sized businesses that assume the option is closed to them would benefit from actually modeling it. That modeling — projected claims, attachment points, worst-case exposure, the fiduciary and administrative load — is the work, and it is worth doing with someone who can build the numbers rather than trusting a rule of thumb about company size.
See how this plays for your business
Every structure in this series looks different once your census, your state, and your renewal are on the table. We’re an independent, carrier-neutral brokerage — we model fully insured, level-funded, self-funded, and the HRA routes side by side and tell you what we’d do in your seat. Start with the group coverage guide, or jump straight to the guide for 2–49 employees or 50+ employees. It’s free and there’s no obligation.
This is general educational information for employers, not legal, tax, or benefits advice. Figures reflect the 2026 plan year and are indexed annually; statutory citations are current as of publication. Rules, thresholds, and IRS-set figures change — confirm the current specifics for your state and situation with a licensed advisor. Written by Matthew T. Giberti (NPN 20698856). Effective date: July 8, 2026.