Captives, Parametric, and the New Math of Holding Your Own Risk

Why is the property and casualty market splitting?

For roughly seven years, about 28 consecutive quarters, property was a hard market: pricing up, capacity tightening, carriers pulling out of certain geographies. Coming into 2026 that has finally cracked, and property has begun to ease. Easing means the direction has changed, not that pricing has snapped back to 2019 levels, so buyers are still in an elevated market. The difference is that property buyers have some leverage at renewal for the first time in years.

Liability is going the other way. The national liability index has climbed sharply since the start of the year, and this looks like a sustained hardening cycle with no clear end. For habitation owners buying both lines, the net effect is often close to flat or worse, because whatever they gain on property gets absorbed by casualty. Worse still, the liability coverage itself has shrunk. General liability towers have been carved up, assault and battery has been sub-limited, and AI carve-outs have been added, so buyers are paying more for a tower that pays out less on exactly the claims they are most likely to face. That squeeze is what is pushing operators toward alternative forms of risk transfer.

What is alternative risk transfer, and why now?

Traditional insurance is simple: you pay a premium and the carrier holds the risk. Alternative risk transfer is when you stop handing your risk to a carrier and start holding it yourself, more deliberately.

It helps to think in terms of risk retention. After a major loss, insurance does not pay immediately. Depending on how a coverage tower is structured, a large operator might carry $20 to $30 million of exposure before traditional insurance pays a dollar. That chunk is the operator's retention, effectively a very large deductible, and the question is what to do with it: sit on it passively and hope claims do not come, or hold it more intelligently. The urgency is new because, with assault and battery exclusions, AI carve-outs, and eroded sub-limits, a meaningful slice of liability that used to sit inside the tower has quietly fallen back onto operators as uninsured risk, a retention they never explicitly agreed to and often do not yet realize they hold. Insurance did not get cheaper, and it now covers less.

How do captives and parametric coverage actually work?

The two tools operators are reaching for solve different problems.

With a captive, instead of paying premiums to a third-party carrier, you pay them into your own insurance entity, and that entity pays your claims. When claims do not materialize, the underwriting profit stays with you rather than the carrier. The risk and the capital both stay in house.

Parametric coverage is different: it insures a loss trigger rather than a loss. In habitation it is most often used for weather. The operator and carrier agree up front that if, say, a hurricane of a defined category makes landfall within a set distance of the property, the policy pays a predefined amount, with no adjuster and no claims process. The trade-off is that there is no gray area. The trigger is either met or it is not, so a storm that lands just outside the defined geography pays nothing even if the property took real damage. That is why parametric does not replace the traditional property tower. It solves a timing problem. A major property claim can take 6 to 18 months to fully settle while the operator still has to repair buildings, manage tenants, and keep operating. A parametric policy pays out within about two weeks of the trigger, putting cash in hand immediately while the traditional claim works its way through.

What are the three types of captives?

The right captive structure depends on how large an operator is and what they are trying to fix. There are three main types:

• Single-parent captive:

  one operator owns the entire entity and has full control, but also carries the full cost of setting it up, capitalizing it, and managing it. Typically realistic only for larger operators.

• Group captive:

  several operators build the same structure together and share the cost and overhead. Each has less individual control, but it is accessible to operators who could not justify a single-parent captive alone.

• Cell captive (rent-a-captive):

  an operator buys a designated slot inside a captive program that already exists and is run by someone else. It has the lowest capital commitment and lowest barrier to entry, but the operator does not own or control the structure.

Two operators, two playbooks

The hosts walked through two operators facing the same problem, expensive traditional coverage with sub-limited assault and battery, but needing very different solutions.

Operator A is a mid-size owner with around 5,000 units across the Southeast, large enough to feel the market pressure but without the dedicated risk team or balance sheet of a national platform. A single-parent captive probably does not pencil out at that scale once capital, governance, and fronting costs are considered. The realistic move is a group captive or a rent-a-captive, where the cost of entry is shared or already absorbed. The captive then slots in behind the general liability tower to hold the sub-limited assault and battery layer. Instead of paying a carrier a premium baked with everyone else's claims history, the operator holds that layer itself, betting that its own loss experience is better than what the market would price. If the operator is right, it captures the underwriting profit; if it is wrong, it pays those claims out of the captive. That is exactly why the work has to start with data: the operator needs to study its assault and battery claims history over the past five years and understand what it is paying for sub-limited coverage before choosing any structure. This is a math problem the broker now has to help facilitate, not just a placement.

Operator B is a national REIT with more than 50,000 units across multiple states, heavy coastal exposure, and an in-house team that treats insurance as a capital-structure decision rather than an annual chore. At that scale, the REIT almost certainly stood up a single-parent captive during the property hard market around 2019 to hold coastal property risk, and that infrastructure can now be repurposed for casualty. With some restructuring, the captive expands to hold the casualty layer, picking up the assault and battery exposure a traditional carrier will not write at a sensible price, while parametric named-storm coverage is added on the property side for liquidity. The full stack ends up as a single-parent captive holding both the property retention and the casualty deductible layer, parametric named-storm coverage alongside it, and traditional property and general liability limits on top. For an operator with 50,000 units across hurricane states, the parametric speed matters enormously, keeping them liquid for the 90 to 180 days a large property claim can take to resolve.

The two look very different on the surface, but the common thread is the same: both are retaining risk on their own terms because the traditional market narrowed. The mechanism differs, the instinct does not.

Where to start: a data problem before a structure problem

The takeaway depends on who you are. On the property side, parametric is primarily a weather tool, not a liability one, so any operator with coastal exposure who has not talked to their broker about parametric options is overdue for that conversation, even in a softening property market, because that is how you work out whether the liquidity benefit is worth the premium. On captives, operators who already have a structure are asking whether they can extend it to hold their general liability retention, while those who do not are asking whether now is the time to build one or join one.

Underneath both questions sits the same prerequisite.

You cannot make an informed decision about retaining a risk you have never measured, which means data by peril and by coverage line is the starting point. Most operators are effectively flying blind, because that pricing data has historically been locked inside a few legacy brokerages. Without clean visibility into the carrier landscape and what comparable risks are actually paying, you cannot tell whether the market is mispricing your risk, whether a captive makes financial sense, or which carriers will write your peril at a sensible price. That visibility, the price benchmarking and insurance data analytics that public markets have had for decades, is what lets an operator treat insurance as a capital-structure decision rather than a compliance checkbox.