Guaranteed Cost, Large Deductible or Captive

Expected annual cost

• Guaranteed cost: $3.9M, fixed. No retained claims, no collateral, no capital.
• $500k SIR: premium $3.9M x 0.92 = $3.59M, plus retained losses limited to $500k per claim. Limiting the ground-up history to $500k per occurrence removes $0.6M from each of the two $1.1M claims, so retained losses are about $6.2M - $1.2M = $5.0M over five years, roughly $1.0M a year. Total: about $4.6M, plus $1.0M collateral and full year-to-year lumpiness.
• Captive: retained losses about $1.0M, reinsurance $1.15M, fronting and overhead $0.42M = about $2.6M, plus $1.5M capital and $1.0M collateral.

Read the numbers sceptically

A captive that appears to save $1.3M a year is a warning, not a conclusion. That gap is somebody's money - it is either the insurer's margin and expense load (possible on a 34% loss ratio), or the feasibility study has under-quoted reinsurance and frictional cost, or the retained layer is under-reserved given that the two largest claims are still open and one has already developed upward by $400,000. Also add cost of capital (6% on $1.5M is $90,000) and check the reinsurance aggregate: if it has a $2M aggregate, a cluster of moderate claims leaves a hole the captive wears.

What each leaves unmanaged

• Guaranteed cost: hides the true cost of loss, transfers loss-control incentive to the insurer, and makes the company buy certainty at a price.
• SIR: frequency volatility and collateral, and it leaves claims handling with the insurer anyway unless negotiated.
• Captive: tail and aggregate risk, domicile regulation and capital adequacy, exit cost, fronting credit risk, and the loss of the insurer's safety and claims resources.

Recommendation: stay guaranteed cost for one year, and use the feasibility study as negotiating leverage.

Reasons:

1. The 8% credit fails arithmetic. Expected retained loss under a $500k SIR is about $1.0M against a credit of $0.31M. The company would pay roughly $0.7M more a year for the privilege of holding the volatility. Not a decision, just a worse price.
2. The 90% budget is the binding constraint. With a mean retained cost near $1.0M and annual incurred swinging between $0.7M and $1.9M, a 90th-percentile budget is plausibly $1.7M-$2.0M. Neither the SIR nor the captive fits that comfortably alongside $1.0M collateral, $1.5M capital, and an acquisition 14 months out that has first call on cash.
3. The captive would not solve the stated problem. Two large claims drive the volatility, and the captive retains only the first $500k of them. A captive helps frequency and frictional cost, not severity volatility. Selling it to a volatility-averse CEO as a volatility solution would be a misrepresentation.

Conditions to attach to the renewal

• Demand the credit be re-priced to at least expected transfer plus a margin (roughly 22-25%).
• Fix the data: obtain paid and incurred development on both open claims and on the claim that moved $400,000. That development is a reserving warning and it directly affects any future reinsurance quote.
• Negotiate loss-control services and a claims-handling protocol into the guaranteed-cost programme while it is still being paid for.

Triggers to reopen the captive

Loss ratio sustained below 35%, both large claims closed, 18 months of clean data post-acquisition, cash above $3M, and a firm reinsurance quote with an aggregate that matches a stochastic 1-in-100 estimate.

Under-priced items

• Capital adequacy. $1.5M of capital against a $500k retention plus an aggregate hole is thin. A stochastic model of the retained layer should set capital at a 1-in-100 outcome, which on this volatility is plausibly $4M-$5M, not $1.5M. Ask to see the actuarial capital study, not the feasibility summary.
• The reinsurance quote. $1.15M is priced off five years of data containing a claim that is still developing upward. Expect a correction at the first renewal, and expect the correction to arrive after the company has committed to the structure.
• Tail. $2M excess of $500k means coverage to $2.5M per occurrence. One $3M claim consumes the capital. The excess tower above $2.5M must be priced before any comparison is valid.
• Frictional and exit costs. Domicile fees, captive management, actuarial, audit, fronting fee, letters of credit, and the cost of unwinding the structure if it fails - typically one to two years of expense.
• Lost protections. Guaranty fund coverage, the insurer's claims and safety resources, and the discipline of an underwriter looking at the account annually.
• Tax and regulatory. Confirm the entity does not fall into the micro-captive scrutiny zone, and check whether the state requires qualifying self-insurer status for the underlying risk.

What would reverse the recommendation

• Both open claims closing materially below reserve, converting the loss ratio into a genuine and stable sub-30% number.
• The acquisition falling through, freeing the $2M and removing the timing conflict.
• A reinsurance quote with a credible aggregate and a capital study supporting a 1-in-100 figure inside the available cash.
• The incumbent refusing to re-price the SIR credit, which would signal a carrier with no appetite - at which point the captive becomes a defensive necessity rather than an optimisation.