What is correlation risk, and how does wrong-way risk amplify losses during market stress?

Correlation risk is the risk that the dependency structure between assets or risk factors changes unexpectedly. Wrong-way risk is a specific form where your exposure to a counterparty increases precisely when that counterparty's credit quality deteriorates.

Correlation Risk:

Portfolio risk models assume correlations are relatively stable. When they change dramatically, the portfolio's actual risk can be far different from modeled risk.

Three Ways Correlation Risk Materializes:

1. Correlation spike: Assets that were weakly correlated suddenly move together (equities during a crash)
2. Correlation breakdown: Assets that were positively correlated decouple (safe-haven breakdown when bonds sell off with equities)
3. Regime change: The correlation structure permanently shifts (post-QE changed equity-bond correlations)

Example — Ashford Multi-Strategy Fund:

• Held a 'correlation diversified' portfolio: long equities, short credit, long commodities
• Historical correlation: equities/commodities = 0.2, equities/credit = -0.3
• During March 2020: ALL positions moved against them simultaneously
• Realized correlations: equities/commodities = 0.85, equities/credit = 0.70
• One-day loss: 4x the 99% VaR

Wrong-Way Risk (WWR):

WWR occurs when credit exposure and counterparty default probability are positively correlated. Your exposure is highest precisely when your counterparty is most likely to default.

Types:

• General WWR: Macro factors affect both the exposure and the counterparty's creditworthiness (e.g., recession increases loan defaults AND reduces collateral values)
• Specific WWR: The exposure is directly linked to the counterparty (e.g., holding puts on a bank's stock as a hedge, with that same bank as counterparty)

Concrete Examples:

Measuring Correlation Risk:

• Stress test correlation assumptions: re-run VaR with correlations at +0.5, +0.8
• Use DCC-GARCH to model time-varying correlations
• Examine tail correlations separately from average correlations
• CVA (Credit Valuation Adjustment) models should incorporate WWR

FRM Key Points:

• Correlation risk is often called the 'silent killer' because it's hard to hedge and rarely visible in calm markets
• WWR can cause simultaneous mark-to-market loss AND counterparty default
• Basel CVA framework requires banks to assess WWR in derivative portfolios
• Diversification assumptions based on historical correlations are most unreliable during crises
• The 2008 AIG collapse was partly a massive wrong-way risk event (CDS exposure to housing)

Master correlation and wrong-way risk in our FRM Part II course.