Ridgeline Transport: Raising the SIR or Forming a Captive?
Ridgeline Transport runs 340 power units. Its auto liability program retains $250,000 per occurrence through a self-insured retention and buys $5M excess. Over three years: 62 claims, 7 penetrated the retention, total incurred $8.9M, and losses in the $250k–$500k band totalled $610k. Actuarial expected loss is $2.6M per year; the expiring premium is $3.1M. The CFO proposes raising the SIR to $500,000, which the broker prices at a $420,000 premium saving, and asks whether a small captive is better. The captive would require $1.4M of collateral costing 6%, plus $250k of annual administration and a TPA. Working capital is tight; the bank covenant requires $3M of liquidity headroom.
1. Quantify the economics of moving from a $250k to a $500k SIR. Is it a good trade?
Read the worked answer
Cost of the extra retention: losses in the $250k–$500k band were $610k over three years, or about $203k per year. That is the expected additional retained cost.
Benefit: $420k of premium saving.
Net before capital: $420k − $203k = $217k per year.
Liquidity check: the maximum single retained loss becomes $500k, about 17% of the $3M covenant headroom. Funding a $500k reserve is feasible but tight, and the SIR means Ridgeline pays claims as they come, not at a policy anniversary. Add the claims-handling burden: at a $250k retention the excess insurer's TPA typically handles everything; at $500k, Ridgeline owns more of the investigation and defence decisions and should budget for TPA fees and a letter of credit if the excess insurer requires collateral.
Verdict: yes, conditionally. The expected-value trade is favourable by roughly $217k a year, and the volatility is survivable if the company ring-fences a loss fund. Two conditions: a funded reserve so a $490k loss does not hit working capital in a single quarter, and a documented stop-loss - if two consecutive years exceed the expected band cost, revisit the retention.
2. How would you decide between the higher SIR and a captive?
Read the worked answer
A captive is a balance-sheet decision, not a discount. Its frictional cost here is $250k of administration plus 6% on $1.4M of collateral, about $84k - roughly $334k a year before it pays a single claim. That exceeds the entire expected cost of the $250k–$500k layer ($203k).
So a single-line captive writing only this layer cannot cover its own overhead. The question becomes whether the captive writes more: cargo, physical damage, or a broader liability program, or whether it exists to accumulate funds for shock losses rather than to save premium in a soft market.
Decision framework:
- Capital efficiency: does the captive free capital or lock it up? Collateral and regulatory capital both count.
- Volatility tolerance: a captive concentrates shock losses inside the company's own entity.
- Control: claims handling, data, and safety incentives may justify some frictional cost.
- Market conditions: the $420k saving reflects a soft market; a captive's cost base does not fall when rates do.
Recommendation: take the SIR increase now, keep the captive on the table as a multi-line feasibility study, and revisit in 24–36 months. A group captive is a reasonable middle path if the goal is risk-sharing rather than premium arbitrage.
3. What could make your recommendation wrong within 12 months?
Read the worked answer
Frequency shock. Rapid hiring during a driver shortage, new lanes, or a change in hiring screens can lift claim counts above the three-year norm - and the three-year data is the entire basis for the $203k estimate. A 30% frequency increase on a small sample is well within normal variation.
A single shock loss near $500k consumes two and a half years of expected band spending and lands directly on the covenant. The 7 claims that penetrated $250k show the retention is reachable regularly; the tail is thin data.
Reserve development. Some of those 7 claims may still be open. Adverse development on a $250k retention is capped, but at $500k it is not.
Market softening can erase the $420k saving at renewal, leaving the retention increase with no offsetting benefit. Conversely, a hard market makes the trade look better - so the decision should be re-tested, not locked in.
Operational and contractual: poor TPA performance, a change in excess insurer collateral terms, or a covenant renegotiation that shrinks headroom below $3M.
The mitigation is monitoring, not prediction: quarterly tracking of claim frequency against the actuarial expected loss, a funded loss reserve, and a pre-agreed review trigger.