Wrong-Way Risk in a Cross-Currency Swap with a Shipbuilder
A bank has a seven-year USD/KRW cross-currency swap with a Korean shipbuilder. Notional is USD 250m. The bank pays USD fixed 3.1% and receives KRW fixed 4.2%. The counterparty posts collateral under a CSA with a USD 25m threshold, no rating trigger, and a ten-business-day cure period. The shipbuilder earns USD revenue but reports debt in KRW; its rating is BBB− and its five-year CDS spread has widened from 90bp to 310bp in six months as newbuild orders fell. The bank’s CVA desk uses a Gaussian copula to link exposure to credit quality, with a single correlation parameter of 0.35 calibrated to sector equity returns. Front office argues the CSA makes CVA immaterial. Risk reports show the swap is USD 38m in-the-money to the bank on a replacement-cost basis before collateral. The bank must decide whether to hedge CVA, reset collateral terms, or reduce the position.
1. Identify the specific wrong-way risk in this trade and evaluate whether the 0.35 correlation parameter is adequate.
Read the worked answer
Specific wrong-way risk arises because the bank’s exposure and the counterparty’s credit quality are adversely linked. The bank receives KRW fixed and pays USD fixed, so its mark-to-market rises when KRW appreciates against USD. A Korean shipbuilder with USD revenue suffers when KRW appreciates, because its USD income converts into fewer KRW and its KRW debt service becomes harder. Thus the bank is most exposed exactly when the counterparty is most likely to default.
A single Gaussian copula correlation of 0.35 calibrated to sector equity returns is unlikely to capture this. It is a generic, average correlation, not a trade-specific linkage. It ignores tail dependence and the nonlinear effect of the CSA threshold: exposure can grow before a margin call is met. The parameter should be stress-tested, including correlation near 1 and a jump in CDS, and the CVA should be decomposed into specific wrong-way risk rather than treated as diversified. Front office’s view that the CSA makes CVA immaterial is incomplete because the threshold and cure period leave unsecured exposure.
2. Quantify the main collateral weaknesses and recommend mitigants, considering the USD 38m pre-collateral exposure.
Read the worked answer
The CSA leaves three weaknesses. First, the USD 25m threshold means the bank can be unsecured for up to USD 25m even after a call. Second, the ten-business-day cure period allows exposure to grow before the counterparty must post; with USD 38m pre-collateral, a default during the cure could leave a large unsecured claim. Third, there is no rating trigger, so collateral terms do not tighten as credit deteriorates.
A simple stress: if exposure rises to USD 50m and the counterparty defaults before posting, recovery on the unsecured amount might be 40%, giving a loss of about USD 30m, which is material against CVA. Mitigants include reducing the threshold to zero, adding a rating trigger and a shorter cure, requiring initial margin or a guarantee, and setting a wrong-way risk limit. Hedging with CDS is possible but introduces basis risk and may be illiquid. Reducing the swap notional or restructuring into a collateralised vehicle may be more effective. The choice should depend on the cost of each mitigant relative to the reduction in unsecured exposure.