How does stop-loss insurance work for a self-funded health plan?

General information, not legal or benefits advice. Have a licensed benefits professional and, for plan documents, ERISA counsel review any stop-loss contract before you rely on it. This page explains mechanics and gives no dollar attachment benchmarks. BRG is a broker and advisor paid by the provider selected in a competitive process; clients are not invoiced. That is a conflict worth knowing about when you read our view on funding arrangements.

What stop-loss insurance is

DOL Technical Release 2014-01 (6 November 2014; retrieved 15 Oct 2026) says stop-loss contracts “protect against claims that are catastrophic or unpredictable in nature,” with an attachment point that can apply to a single enrollee or to aggregate claims over a determined period. The NAIC white paper (2015; retrieved 15 Oct 2026) adds that a self-funded employer “remains ultimately responsible” if the stop-loss insurer fails to perform or denies a claim under the stop-loss contract, or if there are gaps or inconsistencies between the stop-loss policy and the employer’s obligations under the benefit plan. Stop-loss limits the employer’s exposure. It does not hand the plan to the insurer.

KFF’s 2025 survey (retrieved 15 Oct 2026) reports that 67% of covered workers are in self-funded plans: 27% at firms with 10 to 199 workers and 80% at larger firms. It describes level-funded arrangements as combining a relatively small self-funded component with stop-loss insurance. For how funding models compare, see self-funded or fully insured and level-funded health plans.

Specific and aggregate coverage

Per the NAIC white paper, a stop-loss policy usually has two components. The specific attachment point, also called a retention level, “protects against claim severity”: costs above it for a single covered person are covered by the policy. The aggregate attachment point “protects against claim frequency”: once the employer’s total claims payments, not counting any paid by the specific coverage, reach it, the policy covers remaining costs for the year, up to any policy limit. The paper notes that, except for very small employers, the aggregate attachment point will be well below the sum of the specific attachment points.

The NAIC model act defines an attachment point as “the claims amount incurred by an insured group beyond which the insurer incurs a liability for payment,” and expected claims as the claims projected for the group through its health plan in the absence of stop-loss or other insurance. The act’s legal floors, where a state adopts it, are a specific attachment point of at least $20,000 per individual; for groups of 50 or fewer, an aggregate at least the greater of $4,000 times group members, 120 percent of expected claims, or $20,000; and for groups of 51 or more, an aggregate at least 110 percent of expected claims (Section 3, July 2002 reprint). These are minimums in a model law, not what is typical. A state may differ, and DOL’s 2014 release says about ten states had adopted this approach.

Which claims count: incurred, paid, run-out and run-in

The NAIC white paper states that unless a policy provides extended coverage, it reimburses the employer “only for claims that were incurred and paid during the same policy year.” To limit gaps, a policy may include a run-out or extended reporting period, commonly called tail coverage, for claims incurred in the policy year but not reported or paid within it. It says a few states require insurers to provide tail coverage or at least offer it as an option. Insurers may also sell run-in or “nose” coverage for claims incurred in a prior policy year but paid during the current one. It adds that group health insurance covers on an occurrence basis, so nose coverage is not needed if the prior plan year was fully insured.

Run-out length varies. The paper says some policies it reviewed offered run-out periods as short as three months, that some claims take as long as 18 months to run out, including claims in the mandatory internal and external appeal process, and that if a claim resolves after the run-out window the employer may be solely responsible for it. The paper also says the contract should state whether there is any coverage for claims incurred before the policy period. Ask for the incurred and paid rules, the run-out length and the run-in treatment in plain words, not in a shorthand code.

Lasering

The NAIC white paper describes a “laser specific” as a higher attachment point for one or more individuals with preexisting high-cost conditions or other identified risk factors, offered either as a condition of coverage or in return for a lower premium. Its example is an employee whose condition is in remission, where the employer may accept the risk of relapse to avoid a larger premium increase, and it cautions that the employer should first have the cash to pay that claim. The model act text we read (NAIC #92) does not mention lasering. The white paper lists lasering among areas some states might choose to address through minimum standards, which is a regulator’s suggestion, not a statement that any state does.

Claims-reimbursement timing and the employer’s cash

The employer pays claims first and is reimbursed afterwards. The white paper gives the example that if a catastrophic claim occurs in January the employer must fund the whole specific retention up front, and if the aggregate attachment point is reached in September it has paid the full aggregate retention in the first nine months. Some policies include a “monthly aggregate accommodation” or advance-funding provision that allows a temporary negative claims balance. The paper says repayment terms, penalties and interest should be specified, that usually the full amount of advances must be repaid immediately if the contract terminates, and that advances funded by an outside lender fall outside the insurance contract.

The paper also describes, as provisions found in some policies it reviewed and not in every policy: proof-of-loss reporting required within 30 days of the employer becoming aware of facts suggesting a claim may reach a stated share of the specific deductible; a rule that employer payments are not credited toward the retention if not made within 30 days of adequate proof of loss; reporting of newly identified conditions; and re-underwriting or rate triggers if the employee census changes by a stated percentage. Missing a reporting date can cost coverage, so read these clauses with your administrator.

Renewal and terminal liability

Per the NAIC white paper, stop-loss policies are written with one-year terms and are newly underwritten each year, so terms and price can change at renewal. An insurer may decline to renew or, in some cases, cancel mid-term, and where the group has significant claimants it may quote a higher premium, higher limits or lasers. The paper says the typical limits on termination are the state-law rules for commercial or casualty lines, such as timely notice of cancellation or non-renewal. It identifies, in the sample it reviewed, some policies that allow termination by the insurer without cause on 30 days’ notice, which is a reason to read the cancellation clause.

The white paper notes that a Vermont regulation requires disclosure of whether a “terminal liability” option is available and a clear description of it, and that a Utah regulation’s disclosure form addresses terminal liability funding. We did not find a regulator definition of the term, so we do not define it. In practice, ask your advisor what happens to claims incurred before termination but not yet paid, who pays them, and what the policy offers for that case.

Why stop-loss is not health insurance for the employee

The NAIC white paper calls stop-loss a “third-party” line of coverage: the person who had the medical event is not insured under it. It says the insurer “has no direct contractual obligations to the plan participants” and that participants rely on the employer, not the stop-loss insurer, for benefit payments. The model act says nothing in it shall be construed “as treating any stop-loss policy as a direct policy of health insurance” and bars a policy that “provides direct coverage of health care expenses of an individual.” DOL’s release says stop-loss is generally not treated as health insurance under state law and does not usually guarantee payment of benefits to plan participants, and that a state law limiting the attachment points insurers may offer would not, in the Department’s view, be preempted by ERISA. See ERISA and Form 5500 for employers for the plan’s own duties.

Where this fits in your review

Stop-loss terms belong beside the claims-funding and renewal questions already in benefits benchmarking and PEO renewal review, because a funding arrangement is only as clear as the stop-loss contract behind it. If a PEO or administrator offers a self-funded or level-funded medical plan, ask to see the stop-loss terms and ask who the policyholder is.

What we could not confirm

We did not find a regulator or other neutral source for typical attachment-point dollar levels, so this page gives none beyond the model act’s legal floors. The model act is a 2002 reprint of a 1995 act as amended in 1999, and the NAIC white paper is dated 2015; DOL’s release is from 2014. We did not check what has changed since, or read any state’s stop-loss statute, so which rules apply in your state is a question for counsel or your state insurance department. The white paper’s policy provisions come from a limited sample of policies and its authors say they are not in every policy, so none is a statement about what your contract contains. We did not find a regulator definition of “terminal liability,” “laser” beyond the white paper’s description, or the “12/12”-style contract-period shorthand used in industry material, so we do not use them. We did not read any carrier’s policy, and nothing here applies to a contract until a licensed benefits professional has read it.