Reinsurance Credit Risk

Study guides: Exam 9, Exam CP-341

Reinsurance Credit Risk is the risk that a reinsurer fails to pay, in full or on time, the recoverables it owes under its reinsurance contracts — through insolvency, dispute or slow payment — while the cedant remains fully liable to its own policyholders. It turns part of the insurance risk a cedant has shed back into credit risk against its reinsurers.

$E[\text{Credit loss}] = \sum_j PD_j \times LGD_j \times EAD_j$

$EAD_j = \max\bigl(R_j - \text{Collateral}_j,\; 0\bigr)$

$R_j$ is everything due from reinsurer $j$ — ceded paid losses not yet collected, ceded case and IBNR reserves, and ceded unearned premium — and collateral is funds withheld, trust accounts or letters of credit the cedant can draw on. $PD_j$ is the reinsurer's default probability over the whole collection period, not one year.
What makes it distinctive.
- Long horizon. Recoverables on long-tailed and excess layers are collected years after the cession, because ceded losses develop later than gross; the reinsurer's credit quality can migrate a long way in that time.
- Wrong-way risk. The largest recoverables arise from catastrophes that also strike reinsurers' own books, so a reinsurer is most likely to fail exactly when it owes the most.
- Concentration. Programmes are placed with a handful of reinsurers (Concentration Risk).
- Disputes. Coverage disagreements and slow payment cause losses without any insolvency.
Exposure depends on the contract. Across the types of reinsurance, a quota share generates a steady stream of recoverables on every claim, while excess and catastrophe covers concentrate them in a few large, late or event-driven recoveries — see Reinsurance Recovery.
Regulatory recognition. In the U.S., Schedule F of the statutory annual statement computes a provision for reinsurance, a liability charged against surplus, for recoverables from reinsurers that are not authorized in the cedant's state and not adequately secured, and for recoverables that are overdue or in dispute; Risk-Based Capital adds a credit charge on recoverables. In Canada the MCT applies credit risk factors to recoverables, and cessions to unregistered reinsurers earn capital credit only against collateral (Registered Reinsurance).
Managing it. Select counterparties by rating and financial strength; set limits by reinsurer; require collateral; include downgrade (special termination) clauses that let the cedant cancel or demand security if a reinsurer is downgraded; settle long-dated recoverables early by commutation; and use fully collateralised capacity such as CAT Bonds, which removes counterparty risk at the cost of basis risk.

Expected Credit Loss on a Reinsurance Panel A cedant's recoverables ($$$$$M), with default probabilities over the expected collection period:

Reinsurer A (AA): recoverable $40$ , no collateral, $PD = 0.5\%$ , $LGD = 50\%$

Reinsurer B (BBB): recoverable $25$ , no collateral, $PD = 3\%$ , $LGD = 60\%$

Reinsurer C (unrated captive): recoverable $20$ , trust fund $15$ , $PD = 10\%$ , $LGD = 80\%$

Compute the expected credit loss.
Answer Reinsurer C's exposure is $\max(20 - 15, 0) = 5$ .
 $\begin{align*} EL_A &= 0.005 \times 0.50 \times 40 \\ &= 0.10 \\ EL_B &= 0.03 \times 0.60 \times 25 \\ &= 0.45 \\ EL_C &= 0.10 \times 0.80 \times 5 \\ &= 0.40 \end{align*}$ 
Total expected credit loss is $\$0.95$ M on $\$85$ M of recoverables, about $1.1\%$ . Without its trust fund, C alone would cost $0.10 \times 0.80 \times 20 = \$1.6$ M — collateral is doing most of the work on the weakest counterparty, while the unsecured BBB reinsurer is now the largest source of expected loss.

Wrong-Way Risk on a Catastrophe Cover A catastrophe treaty would pay $\$200$ M in a 1-in-100-year event. The reinsurer's unconditional annual default probability is $1\%$ , but its default probability given that event is $10\%$ . $LGD = 50\%$ . Compare the expected credit loss using each default probability.
Answer The recoverable exists only if the event happens, so the relevant default probability is the conditional one:
 $\begin{align*} EL_{\text{naive}} &= 0.01 \times 0.01 \times 0.50 \times 200 \\ &= \$0.01\text{M} \\ EL_{\text{conditional}} &= 0.01 \times 0.10 \times 0.50 \times 200 \\ &= \$0.10\text{M} \end{align*}$ 
Treating the default as independent of the catastrophe understates the expected credit loss tenfold. Given the event, the expected shortfall is $0.10 \times 0.50 \times 200 = \$10$ M — arriving exactly when the cedant is weakest, which is why panels are spread across reinsurers with different catastrophe exposures and why collateralised capacity is valued above its expected-loss arithmetic.