Wrong Way Risk: When Exposure and Default Probability Rise Together

Key Takeaways

• Wrong way risk makes expected loss much larger than the product of average exposure and average default probability because the bad scenarios are correlated, high exposure and high default probability occur simultaneously.
• Specific wrong-way risk (SWWR) has a direct structural link (taking a bank's own equity as collateral); general wrong-way risk (GWWR) is a statistical correlation without a legal link.
• In the worked example, WWR-adjusted expected loss was 480 thousand versus 72 thousand under the naive calculation, seven times higher because a sovereign default triggers both the CDS exposure and the counterparty's failure at once.
• Standard CVA models assume independence of exposure and default; trades with identified wrong-way risk require explicit uplift or a separate stress framework.

What It Is

In plain terms, WWR means the thing that makes your counterparty default is also the thing that makes them owe you more. The Basel framework splits it into two categories.

Specific wrong-way risk (SWWR) is a direct legal or structural link. Taking equity of a bank as collateral for a trade with that same bank is the textbook example. General wrong-way risk (GWWR) is a statistical correlation without a legal link, for example an emerging market bank whose solvency depends on oil prices writing options on oil to commodity clients.

The Intuition

Counterparty credit models assume that exposure and default probability are independent, or nearly so. When they are positively correlated, the expected loss is larger than the product of average exposure and average default probability. The tail gets fatter because the bad outcomes line up.

The 2008 monoline crisis is the canonical case. Insurers like AMBAC and MBIA wrote credit protection on mortgage-backed securities. When the mortgage market collapsed, the insurers owed enormous sums at the same moment their own capital vanished, because their business depended on the very assets they had insured.

How It Works

SA-CCR embeds a WWR uplift in two ways. The alpha multiplier of 1.4 on effective EPE partly covers general correlation between exposure and default. For identified SWWR trades, the rules require modelling the trade as if default had already occurred and exposure had jumped to the stressed value, which typically doubles or triples the EAD.

Internal models take a more explicit route. You can introduce a correlation parameter between the counterparty's credit factor and the exposure driver.

Expected loss with WWR ~ E[EAD * 1{default}]  !=  E[EAD] * PD

The correction term grows with the correlation. A rough approximation used in practice is to scale the standalone expected exposure by a factor of (1 + rho x stress), where rho is the estimated correlation and stress measures how much exposure rises in the counterparty's default scenario.

Identification is where most of the real work happens. Banks maintain WWR registers that flag trades where the counterparty's revenue, collateral, or shareholder base is tied to the trade's underlying.

Worked Example

A US bank sells a 200 million notional credit default swap on Country X sovereign debt. The counterparty is a bank domiciled in Country X, with most of its assets in Country X government bonds.

Standalone counterparty credit risk analysis might show:

• 1-year PD of Country X bank: 1 percent
• Expected exposure on the CDS: 12 million
• LGD: 60 percent
• Expected loss: 0.01 x 12 x 0.60 = 72 thousand

That is the no-correlation answer. In reality, if Country X defaults, the CDS moves deep into the money (exposure perhaps 80 million) and the counterparty almost certainly fails at the same time.

Conditional expected loss | sovereign default = 80 x 0.60 = 48 million
Sovereign default probability = 1 percent
Wrong-way contribution to expected loss = 0.01 x 48 = 480 thousand

The WWR-adjusted number is roughly seven times the naive calculation. A desk limit set on the 72 thousand figure would materially understate the true risk.

Frequently Asked Questions

Q: What is wrong way risk in simple terms? Wrong way risk is when the two things you want to be independent, your exposure to a counterparty and the probability they default, actually move together. The more they owe you, the more likely they are to fail, which is the worst possible combination.

Q: How does wrong way risk affect investment decisions? Desks must flag trades where the counterparty's solvency is tied to the trade's underlying. A bank that buys CDS protection on a sovereign from a bank in that same country has specific wrong-way risk, the sovereign default that triggers the CDS payment would also likely destroy the counterparty writing the protection.

Q: What is a real-world example of wrong way risk? The 2008 monoline insurers (AMBAC, MBIA) wrote credit protection on mortgage-backed securities. When the mortgage market collapsed, they owed enormous protection payments at the exact moment their own capital was wiped out by the same exposures they had insured. Exposure and default materialized simultaneously.

Q: How can risk managers identify and control wrong way risk? Maintain a WWR register that maps counterparty revenues and collateral to trade underlyings. Use crisis-period correlations rather than five-year rolling correlations for calibration. Exclude counterparty-related instruments from eligible collateral under the CSA.

Q: How is wrong way risk different from general counterparty credit risk? General CCR models assume exposure and default probability are independent or nearly so. Wrong way risk is specifically the case where that independence assumption fails, the two are positively correlated. The difference can multiply expected loss by several times, as shown in the sovereign CDS example.

Common Mistakes

1. Screening only for legal links. SWWR is easier to see, but GWWR drives the bigger losses in practice. A systematic screen that maps counterparty revenue drivers to trade underlyings catches both.

2. Using historical correlations from calm periods. The correlation that matters is the one that shows up in stress. Five-year rolling correlations computed out of 2014 to 2019 are useless for sovereign or commodity WWR. Crisis-window calibration is required.

3. Trusting collateral that shares the risk. Posting the counterparty's own bonds, or the currency of a weak sovereign they are exposed to, does not actually reduce your loss. Collateral eligibility rules should exclude issuer and related-entity instruments.

4. Forgetting CCP concentration. Clearing houses can carry WWR against their own members. If a large clearing member fails in a market move that also stresses collateral posted at the CCP, the default fund can be insufficient.

5. Pricing with a CVA model blind to WWR. Standard CVA assumes independence of exposure and default. Trades with flagged WWR should be priced with an uplift or, for material SWWR, moved to a specific stress framework rather than the general book.

Sources

1. Basel Committee on Banking Supervision. "The standardised approach for measuring counterparty credit risk exposures." BCBS 279. https://www.bis.org/publ/bcbs279.htm
2. Basel Committee on Banking Supervision. "Sound practices for backtesting counterparty credit risk models." BCBS 185. https://www.bis.org/publ/bcbs185.htm
3. ISDA. "Counterparty Credit Risk and Wrong-Way Risk." https://www.isda.org/a/OEiDE/wwr-paper-final.pdf
4. GARP. "Risk Intelligence on Wrong-Way Risk." https://www.garp.org/risk-intelligence

Disclaimer

This article is educational content only and is not financial advice. Nothing here is a recommendation to buy, sell, or hold any security. Consult a licensed advisor before making investment decisions.