Self-funding was a large-employer structure for two specific reasons, and both of them have changed. It is worth being precise about which, because the marketing around this has got well ahead of the substance.
The two barriers
A self-funded plan needs the employer to absorb month-to-month claims volatility, and it needs somebody to analyse the claims data well enough for the arrangement to be worth having. Small employers historically failed both tests: too few lives for claims to behave predictably, and no realistic access to actuarial analysis.
What captives did to the first one
A group medical captive is several employers jointly owning an insurance entity and sharing a layer of stop-loss risk. Each company’s claims fund stays its own. The shared layer smooths exactly the volatility that made solo self-funding uncomfortable below a few hundred lives.
The consequence that matters: pricing moves from community-rated toward experience-rated. A healthy group stops subsidising the pool it was placed in and starts being priced on itself. Unused premium in the shared layer typically returns to members.
What you give up is autonomy. Captives involve collective governance and usually a multi-year commitment, and leaving one is not a decision you make in a quarter.
What better analytics did to the second one
Four things are now routinely available to a group of sixty that used to require a benefits department:
Risk stratification
Reading claims history, demographics and utilisation to identify where the next large claim is likely to come from — so disease management and care coordination can be aimed rather than broadcast. This is the capability that genuinely did not exist at this scale without an actuarial team.
Plan design against real usage
Seeing which benefits are actually used, what drives cost, and how a design change would land — before the renewal rather than after it.
Member navigation
Tools that answer benefits questions at eleven at night and steer members toward higher-quality, lower-cost providers. For an employer with nobody dedicated to benefits, this is the difference between a plan people use well and one they use expensively.
Claims integrity
Anomalous billing, duplicates and outright errors, found at a scale nobody was reviewing by hand. Recovered dollars stay in the plan.
The honest caveat
None of this makes self-funding right for a group it is wrong for. A group with heavy known claims, unstable headcount or no cash cushion is still better off fully insured, and the analytics will tell you that too if you let them.
What has changed is the floor. The conversation now starts sensibly at twenty-five to fifty lives rather than at a hundred and fifty, and it starts with your own data rather than with a manual rate.
What to ask a vendor
- Do we get the underlying claims data, or only a dashboard?
- What exactly is the stop-loss structure — specific and aggregate attachment points, and who carries the risk?
- If this is a captive, what is the commitment period and what happens if we leave?
- Who owns the run-out liability if we move?
- Is compliance — 5500, ACA reporting, plan documents — handled, or is it now ours?
The last one gets skipped and matters. Self-funding moves obligations onto the plan sponsor that a fully insured carrier had been quietly absorbing.
How the mechanics work, or ask us to model it against your renewal.
General information for employers, not insurance advice.