Level-Funded vs. Fully Insured vs. Captives: Which Funding Model Actually Fits a 50 to 1,000 Life Employer

Last updated: September 8, 2026. By Rich Westermayer, Founder, BeneSmart.

The Advice You Get Depends on Who’s Giving It

Direct answer: The funding advice you get is shaped by how the person giving it is paid and by which model they place most often. Work out whose economics sit behind the recommendation before you weigh it. Then judge the three models on surplus ownership, volatility, and data access rather than on the confidence of the pitch.

Ask three brokers which funding model fits your company and you may get three confident, incompatible answers. That is not because the question is unanswerable. It is because a broker who places one model well recommends it, and a broker paid a percentage of premium prefers the arrangement where premium is largest.

We laid out how that compensation works in Exposing the Hidden Costs of Health Insurance Broker Commissions. Here it is a filter: before you weigh the models, understand why the person explaining them prefers one.

That filter applies to us, so here is our answer to it. BeneSmart is paid a disclosed, carrier-paid per-employee-per-month fee based on enrollment. It does not move with your premium. The rate is set by your strategy and funding structure, and it increases only if scope increases.

So this post does not pick a winner. It gives you the comparison, the readiness tests, and the exit question, so the decision belongs to you and is defensible in a board meeting.

Level Funded vs. Fully Insured vs. Captive: What Each One Is

Direct answer: Fully insured means you pay a fixed premium and the carrier keeps any surplus and absorbs any excess. Level-funded means you self-fund claims through a packaged arrangement with fixed monthly payments and built-in stop-loss, with a potential surplus return. A captive means you self-fund and pool the volatility layer with other employers, and the real difference is not price but who owns the surplus, who absorbs volatility, and how much claims data you can see.

Fully insured

You pay a premium. The carrier pays claims. If your group runs well, the carrier keeps the difference. If it runs badly, the carrier absorbs it and then re-prices you at renewal.

The trade is clean: maximum budget certainty, minimum upside, and the least visibility into your own claims. That last point is the one employers underestimate. Without claims data you cannot diagnose your cost drivers, so every renewal conversation runs on the carrier’s information advantage.

Level-funded

A packaged form of self-funding. You pay a fixed monthly amount covering a claims fund, administrative fees, and stop-loss premium. If claims come in under the funded amount, some or all of the surplus may return to you on the terms your contract defines. If they run over, stop-loss is the backstop.

It behaves like fully insured on your cash flow calendar and like self-funding on your economics. It is the most common first step out of fully insured for groups in the 50 to 1,000 life range, and the model where the fine print does the most work.

Captive

You self-fund and join a group of employers who collectively finance a layer of risk that would otherwise be sold to a carrier. Pooled, community, private, and single-parent structures each behave differently on governance, capital, and how losses are shared.

A captive is a decision to become a long-term owner of a piece of your own risk, with the governance obligations that implies. It suits employers with stable populations, a multi-year horizon, and the appetite to participate in decisions rather than receive a renewal.

If terms like attachment point, aggregate corridor, or run-out are unfamiliar, the High Performance Benefits Glossary covers them in plain English.

The Six-Dimension Comparison

Direct answer: Compare the three models across six dimensions: cash flow predictability, who keeps the surplus, claims data access, volatility exposure, administrative and governance burden, and reversibility. Read across the rows. The point is what each model costs you on the dimensions it does not advertise.

Two rows carry most of the decision. Claims data access decides whether you can manage cost rather than absorb it. Reversibility decides what a bad year costs you in optionality, and it is the row employers read last, if at all.

What renewal actually looks like under each model

The table is static. Renewal is where the models separate in practice, because renewal decides how much of your own experience you get to see and argue with.

Under a fully insured plan, renewal arrives as a number. The carrier prices your group on experience you cannot inspect, blended with pooled assumptions you cannot see, and your only counter is to market the group to carriers who will underwrite it the same way. A good year does not come back to you.

Under a level-funded plan, renewal is two conversations. First the fixed costs: administrative fees and stop-loss premium, where your claims drive the quote and prior large claimants can be lasered or priced in. Then the surplus settlement, on its own timetable, often long after the plan year closes and sometimes conditioned on renewing. A fee increase paired with a surplus you never receive is not a good renewal.

Under a captive, renewal is a governance event. Your funding rate, the pooled layer’s loss experience, and any assessment or capital call run through the captive’s process, and other members’ claims can move your cost. In exchange you sit where the assumptions are visible.

Why “Which Is Cheapest” Is the Wrong Question

Direct answer: No funding model is inherently cheaper, because none of them changes what your population costs to care for. What they change is who bears variance, who keeps favorable results, and how much information you have to act on. The right question is not which model costs less, but which one converts good years into retained surplus and gives you the data to create more of them.

Your claims are your claims. A funding model is a container for them.

What changes across containers is the answer to three questions. In a good year, who keeps the money? In a bad year, who absorbs it, and how long does it follow you? And throughout, can you see enough to intervene?

That reframing kills the most common failure mode, which is not choosing wrong. It is choosing right and then being unprepared for the volatility that the choice deliberately handed you.

Which is why the stop-loss contract underneath a level-funded or captive arrangement deserves more scrutiny than the funding decision itself. That contract is where retained volatility is priced. See Traps in Stop-Loss Coverage and, at renewal, No New Lasers vs. Rate Caps.

The pattern we see most often

The employers who move well and the employers who move badly rarely differ on the model they chose. They differ on what happened in month four. A group leaves fully insured for the surplus, retains variance for the first time, and meets a bad quarter before a good one arrives. Treated as a multi-year strategy, that quarter is a data point. Treated as a price decision, it becomes a referendum, and the group stampedes back, having paid the entry cost of self-funding and collected none of the benefit. The deciding factor is almost never the contract. It is whether anyone set the expectation before the first bad month.

This describes a recurring pattern across engagements, not a specific client, and contains no client figures.

The Four Readiness Tests

Direct answer: Four tests predict whether a move to level-funded or a captive will work: cash flow tolerance for month-to-month variance, population stability, willingness to actually use claims data, and a multi-year horizon. Failing one is a reason to wait, not a reason to never move.

1. Cash flow tolerance. Not “can you afford a bad year,” but can you afford a bad quarter without the decision becoming a referendum? Self-funding converts a smooth premium line into a variable one. Model your worst plausible quarter before you sign.
2. Population stability. Rapid headcount swings interact badly with attachment points, enrollment-based guarantee triggers, and captive underwriting. If you are mid-acquisition or scaling fast, that is a timing problem, not a permanent disqualification.
3. Willingness to use the data. The single best predictor of success. Claims data has no value unless someone reviews it on a cadence and acts on it. If nobody will own that hour each month, the biggest advantage of leaving fully insured never arrives.
4. Multi-year horizon. These models compound, and one year is not enough for a good year to pay for a bad one. If leadership will re-litigate the decision every renewal, the volatility wins before the strategy gets a chance.

Three yeses and a fixable no is a plan. Two noes is a reason to strengthen the fully insured position this year and revisit, starting with what your broker is paid to place it.

The Exit Question Nobody Asks

Direct answer: Before entering any self-funded arrangement, ask what leaving costs. The answer lives in run-out liability, terminal liability provisions, surplus settlement timing, and, for captives, capital return and exit notice terms. An arrangement cheap to enter and expensive to leave is a multi-year commitment sold as an annual one.

Ask these four, in writing, before you sign anything:

1. “If we terminate at the end of the plan year, what claims remain our responsibility, and for how long?” Run-out is the surprise that arrives after everyone has moved on.
2. “Is terminal liability coverage available, what does it cost, and what exactly does it cover?”
3. “When and how is surplus calculated and paid, and what can reduce or eliminate it?” A surplus that only settles if you renew is a retention device, not a return.
4. “What are the exit provisions: notice periods, capital return timing, and any obligation that continues after we leave?” Critical for captives, where exit terms are negotiated rather than standard.

What makes those questions concrete is the calendar behind them. Run-out is typically claims incurred in the plan year and paid within a set window after termination, commonly twelve months, and without terminal liability coverage they come out of next year’s budget for a plan you no longer have. Surplus settlement runs on a separate clock, often six to twelve months past plan year end, so a group that terminates December 31 can be writing run-out checks in March and still waiting on a surplus number in September. For a captive, add notice measured in months and capital returned in stages after final loss development. Ask for the dates and the dollar mechanics.

An employer who can answer those four is making a decision. An employer who can’t is accepting one.

You have more leverage here than the complexity suggests, because complexity is the only reason it feels one-sided. For our clients the funding decision is not a standalone choice made at renewal under time pressure. It is one output of a defined sequence: a Self-Funded Blueprint™ or an Alternative Funding Blueprint™, depending on how you’re funded today. That is the direct path and the one most employers should take. If you would rather see evidence before you commit, a Strategic Diagnostic puts the proof in front of you first. Ready to act, Blueprint. Want evidence first, Diagnostic. Either way the decision gets made inside a process, which is why it holds up when the first bad quarter arrives.

Frequently Asked Questions

What is the difference between level funded and fully insured?

Fully insured means you pay a fixed premium and the carrier keeps any surplus and absorbs any excess claims. Level-funded means you self-fund claims through a packaged arrangement with a fixed monthly payment and built-in stop-loss, where favorable claims experience may return to you as surplus under your contract’s terms. The core difference is who keeps the money in a good year.

Is a captive better than level funded for a small employer?

Neither is universally better. Level-funded is generally the more accessible first step for a 50 to 1,000 life employer, while a captive suits employers with a stable population, a multi-year horizon, and the willingness to take on governance participation and capital obligations. The deciding factors are commitment horizon and appetite for ownership, not size alone.

Can you switch back from level funded to fully insured?

Usually yes, but the cost of switching back is not zero. Run-out claims, terminal liability, and surplus settlement timing all follow you past the termination date, so the exit terms should be understood before entry rather than discovered at renewal.

Which funding model is cheapest?

None of them is inherently cheapest, because the funding model does not change what your population costs to care for. What changes is who keeps favorable results, who absorbs volatility, and how much claims data you can see and act on. Cost differences show up over multiple years, through management, not at signing.

What claims data should an employer get under a level funded plan?

Access varies sharply by vendor, so it should be confirmed in writing before signing. At minimum, ask for regular claims reporting with enough detail to identify cost drivers, high-cost claimant tracking in whatever form your vendor is able to release, and pharmacy detail, which is the data you need to manage the spend you have just agreed to retain.