Why Timing and Contingency Matter
An insurer may receive cash after a severe loss, but the receipt alone does not show that the insurer transferred risk. The payment may depend on insured losses and remain with the insurer, or it may be an advance that must be repaid. The first can reduce the insurer’s ultimate covered losses; the second can ease a near-term cash shortfall while leaving those losses substantially unchanged.
A useful analysis separates three questions: What event triggers payment? How much of the loss does the arrangement absorb? When does cash arrive, and must it later be repaid? Answering them distinguishes underwriting-risk transfer from financing and clarifies the effect on liquidity.
For a covered gross loss \(X\) and a contingent recovery \(R(X)\), the insurer’s retained covered loss is \(X-R(X)\), assuming the recovery is collectible and ignoring expenses and timing for this initial calculation. A fixed advance is different: it is not automatically a recovery \(R(X)\). If the advance is \(A\) and the insurer must repay \(B\), the cash flows include both the advance and repayment, while the underlying claim remains the insurer’s loss.
Measure the Loss Before Considering the Cash
A simple indemnity treaty illustrates the loss effect. Assume the stated loss outcomes and probabilities are exhaustive, the treaty responds only to those covered losses, and there is no counterparty default, premium, expense, or payment delay. All amounts in the example are in thousands of dollars.
Worked Example: A Capped, Loss-Contingent Recovery
A hypothetical insurer has a covered loss of either $100,000 with probability 0.60 or $500,000 with probability 0.40. A treaty pays 50% of the amount above a $100,000 attachment, subject to a maximum recovery of $150,000. Calculate the expected gross loss, expected recovery, and expected retained loss. Also compare the gross and retained variances.
Method. Calculate the recovery separately in each loss state, then probability-weight the gross losses, recoveries, and retained losses. This avoids applying an average loss to a nonlinear treaty.
At a gross loss of $100,000, the loss does not exceed the attachment, so the recovery is $0. At a gross loss of $500,000, the amount above the attachment is $500,000 − $100,000 = $400,000. Half of that is $200,000, but the cap limits the payment to $150,000. Thus the state-by-state results are:
The expected gross loss is \(0.60(\$100{,}000)+0.40(\$500{,}000)=\$60{,}000+\$200{,}000=\$260{,}000\). The expected recovery is \(0.60(\$0)+0.40(\$150{,}000)=\$60{,}000\). Therefore the expected retained loss is \(0.60(\$100{,}000)+0.40(\$350{,}000)=\$60{,}000+\$140{,}000=\$200{,}000\). The check is \(\$260{,}000-\$60{,}000=\$200{,}000\).
For the variance, measure losses in thousands of dollars. The gross mean is 260, so \(\operatorname{Var}(X)=0.60(100-260)^2+0.40(500-260)^2=0.60(25{,}600)+0.40(57{,}600)=15{,}360+23{,}040=38{,}400\) thousand-dollars squared. The retained mean is 200, so \(\operatorname{Var}(X-R)=0.60(100-200)^2+0.40(350-200)^2=0.60(10{,}000)+0.40(22{,}500)=6{,}000+9{,}000=15{,}000\) thousand-dollars squared. The treaty reduces the expected retained loss and, for these outcomes, its variance.
Common mistake avoided: Do not treat the treaty’s maximum recovery as the payment in every state. The $150,000 cap applies only when the loss reaches the high-loss state; in the low-loss state, recovery is zero.
Why the Retained-Loss Formula Holds
The state-by-state method above follows directly from the definition of retained loss. When a covered loss is \(X\) and the reinsurer pays \(R(X)\), the insurer pays the covered loss but receives the recovery against it. Its net covered amount is therefore \(X-R(X)\). Taking expectations gives the expected retained loss without any additional distributional assumption, provided the relevant expectations exist:
The corresponding variance identity is \(\operatorname{Var}(X-R)=\operatorname{Var}(X)+\operatorname{Var}(R)-2\operatorname{Cov}(X,R)\). It follows by expanding the variance of a difference. It explains why the effect on volatility depends on the relationship between loss and recovery, not only on the expected recovery. A recovery that rises with covered loss will generally be positively associated with loss, but a treaty’s attachment, cap, and covered-event definition still matter. A positive expected recovery alone does not establish that volatility falls in every possible model.
Financing Can Improve Cash Availability Without Transferring Loss
A loan or other repayable advance can help an insurer meet claims before premium or investment cash is available. To measure its effect, keep the claim payment separate from the borrowing and repayment. Assume the insurer pays the claim immediately, receives an advance immediately, and repays a fixed amount one year later. The interest rate is an effective annual rate of 8%; there are no fees, and all amounts are nominal dollars unless stated otherwise.
Worked Example: A Fixed Advance With a Fixed Repayment
A hypothetical insurer’s immediate claim is $180,000 with probability 0.70 or $360,000 with probability 0.30. A lender advances $200,000 immediately. The insurer must repay $216,000 in one year, regardless of the claim amount. Calculate expected claim cost, the net nominal cash outflow over the full year in each state, and the present value of the repayment at 8%.
Method. Treat the claim as the underwriting loss, and then add the borrowing cash flows separately. The repayment is fixed, so it does not vary with the claim.
The expected claim is \(0.70(\$180{,}000)+0.30(\$360{,}000)=\$126{,}000+\$108{,}000=\$234{,}000\). At the claim date, the advance reduces the insurer’s net cash outflow by $200,000. Across both dates, the nominal net outflow is claim plus repayment minus advance.
For the $180,000 claim, the full-period net outflow is \(\$180{,}000+\$216{,}000-\$200{,}000=\$196{,}000\). For the $360,000 claim, it is \(\$360{,}000+\$216{,}000-\$200{,}000=\$376{,}000\). The expected full-period nominal outflow is \(0.70(\$196{,}000)+0.30(\$376{,}000)=\$137{,}200+\$112{,}800=\$250{,}000\). This agrees with the expected claim plus the net nominal financing cost: \(\$234{,}000+(\$216{,}000-\$200{,}000)=\$250{,}000\).
At an effective annual rate of 8%, the discount factor is \(v=1/1.08\). The present value of the repayment is \(\$216{,}000/1.08=\$200{,}000\). Thus the advance and repayment have equal present values at that rate, before fees or other costs. The advance still supplies $200,000 of immediate liquidity, but the claims remain $180,000 or $360,000. Because repayment is fixed regardless of losses, this arrangement finances the claims rather than transferring their underwriting risk.
Common mistake avoided: Do not subtract the advance from the insurer’s ultimate loss and call the result a recovery. The advance is accompanied by a repayment obligation. It changes the timing of available cash; it does not make the insurer’s claim amount depend on the lender.
Timing Changes Liquidity and Present Value
Even when a payment is genuinely contingent on loss, a delay can leave the insurer needing to fund the claim first. Ultimate retained loss and immediate cash need are therefore different measures. In the next example, assume losses are paid immediately and a contingent recovery arrives six months later. The effective annual interest rate is 8%; for a half-year period, use an accumulation factor of \((1.08)^{0.5}\). There are no premiums, expenses, or credit losses.
Worked Example: A Delayed Recovery After a Covered Loss
A hypothetical insurer faces a covered loss of $80,000 with probability 0.50 or $320,000 with probability 0.50. A treaty pays 50% of the amount above an $80,000 attachment, capped at $100,000. Any recovery is paid six months after the insurer pays the claim. Calculate the expected nominal recovery, expected ultimate retained loss, and expected present value of recoveries.
Method. First apply the treaty to each loss state. Then discount only the payments that are actually made, using the stated effective annual interest rate and payment delay.
At $80,000, the loss is exactly at the attachment, so recovery is $0. At $320,000, the amount above the attachment is $320,000 − $80,000 = $240,000. Half is $120,000, which exceeds the $100,000 cap, so the recovery is $100,000. The ultimate retained losses are therefore $80,000 and \(\$320{,}000-\$100{,}000=\$220{,}000\).
Expected gross loss is \(0.50(\$80{,}000)+0.50(\$320{,}000)=\$40{,}000+\$160{,}000=\$200{,}000\). Expected nominal recovery is \(0.50(\$0)+0.50(\$100{,}000)=\$50{,}000\). Expected ultimate retained loss is \(0.50(\$80{,}000)+0.50(\$220{,}000)=\$40{,}000+\$110{,}000=\$150{,}000\), which checks as \(\$200{,}000-\$50{,}000=\$150{,}000\).
The present value of the recovery in the high-loss state is \(\$100{,}000/(1.08)^{0.5}=\$96{,}225.04\), rounded to the nearest cent. There is no recovery in the low-loss state. Expected present value is therefore \(0.50(\$0)+0.50(\$96{,}225.04)=\$48{,}112.52\), rounded to the nearest cent. The expected present value of the insurer’s net loss, with claims paid immediately, is \(\$200{,}000-\$48{,}112.52=\$151{,}887.48\). The insurer must still fund the full claim when it is paid; the later recovery reduces its ultimate covered cost and provides cash only after six months.
Common mistake avoided: Do not discount the recovery from the claim date as if it arrived immediately. Also do not treat the $100,000 recovery as certain: it occurs only in the high-loss state.
Exam Traps and Practical Interpretation
A reliable comparison makes the cash-flow dates explicit and keeps underwriting losses separate from funding cash flows. For each possible outcome, identify the claim, any contingent payment, any advance, and any repayment. Then calculate expected values only after the state-specific amounts are correct.
- Probability-weighting error. If a $100,000 payment occurs with probability 0.40, its expected contribution is \(0.40(\$100{,}000)=\$40{,}000\), not $100,000. If a $400,000 payment occurs in the other state with probability 0.60, the expected payment is \(\$40{,}000+\$240{,}000=\$280{,}000\).
- Calling every receipt a recovery. A payment received today can be a loan or advance. Check whether it must be repaid, offset, or returned under the contract.
- Ignoring contractual limits. Attachments, caps, exclusions, and covered-event definitions determine the recovery in each state. Apply them before taking expectations.
- Confusing liquidity with ultimate cost. A delayed recovery may reduce ultimate loss but not the initial cash required to pay claims. A fixed advance may improve that initial position without reducing ultimate underwriting loss.
- Assuming transfer means zero risk. A treaty may cover only part of a loss, have a limit, or respond only to specified events. The insurer retains losses outside the contract and may also face collection and timing risks.
This distinction matters when comparing risk-management alternatives. A contingent recovery can change the insurer’s retained loss distribution; financing primarily changes the dates and sources of cash. Evaluating both dimensions gives a clearer view of loss protection and liquidity. These distinctions are aligned with the CAS Exam 9 (Risk Management for Actuaries) syllabus.
Check Your Understanding
For each problem, distinguish the effect on retained loss from the effect on cash timing. Treat each stated model as complete, and ignore expenses and counterparty default unless specified.
- A hypothetical treaty pays 60% of covered loss above $100,000, subject to a $90,000 cap. Loss is $50,000 with probability 0.70 and $250,000 with probability 0.30. Find the expected gross loss, expected recovery, and expected retained loss. (Answer: expected gross loss $110,000; expected recovery $27,000; expected retained loss $83,000.)
- An insurer receives a fixed $75,000 advance today and must repay $81,000 in one year, regardless of claims. The effective annual interest rate is 8%, and there are no fees. What is the present value of repayment, and does the advance transfer underwriting risk? (Answer: present value $75,000; no underwriting risk is transferred because the repayment is fixed and independent of claims.)
- Loss is $40,000 with probability 0.25 and $240,000 with probability 0.75. A treaty pays 25% of covered loss above $40,000, capped at $40,000. Calculate expected recovery and expected retained loss. (Answer: expected recovery $30,000; expected retained loss $160,000.)
- A recovery of $60,000 is paid one year later only if a covered event occurs; the event probability is 0.20. The effective annual interest rate is 5%. Find the expected nominal recovery and the expected present value of the recovery. (Answer: expected nominal recovery $12,000; expected present value approximately $11,428.57.)