In March 2021, one family office failed to meet its margin calls. Over the following weeks, several global banks disclosed combined losses of more than $10 billion. None of those losses came from a bad loan. Instead, they came from counterparty credit risk, the risk that a trading partner fails before it settles what it owes.
For risk analysts, this risk sits at an unusual crossroads. It blends credit analysis with market movements, legal contracts, and collateral operations. As a result, it needs a different toolkit from traditional lending.
This guide explains what counterparty credit risk is and how banks measure it. We also work through SA-CCR with a numerical example, cover wrong-way risk, and show how banks control the risk in practice.
Counterparty credit risk is the risk that the other party to a financial contract defaults before the final settlement of its cash flows. A loss occurs only if the contract has positive value to the surviving party at that moment. It arises mainly in OTC derivatives, repos, and securities financing transactions.
Unlike a loan, these contracts carry no fixed amount at risk. Their value moves every day with interest rates, exchange rates, and asset prices.
First, an interest rate swap between a bank and a corporate client can swing in value for either side over its life. Second, in a repo, a bank lends cash against government bonds. If the borrower defaults after bond prices have fallen, the collateral may not cover the cash. Third, securities lending creates similar exposure when a borrower fails to return the securities.
The scale is enormous. According to the Bank for International Settlements (BIS), the notional value of outstanding OTC derivatives reached $846 trillion at end-June 2025. However, notional is not exposure. Gross market value, the cost of replacing every contract at market prices, stood at $21.8 trillion. Even that overstates the credit risk, because netting and collateral shrink it further.
A term loan is simple from an exposure point of view. When a bank lends ₹100 crore, it knows the maximum amount at risk on day one. The borrower owes money, and the bank does not. Counterparty credit risk breaks all three of those assumptions.
The exposure is bilateral. In a derivative, either party can end up owing the other. Today, the bank may owe its client on a currency forward. Next month, after a sharp rupee move, the client may owe the bank. Therefore, each side carries credit risk to the other.
The future exposure is uncertain. A swap often starts with a value close to zero. Its future value depends on how markets move over several years. So the bank must estimate a distribution of possible exposures rather than a single number.
The exposure is market-driven. Because exposure depends on rates, currencies, and prices, credit officers cannot assess it in isolation. Instead, they need market risk models, simulation engines, and collateral data working together.
In practice, this creates a two-dimensional problem. Analysts must judge how likely the counterparty is to default and how large the exposure could be at that moment. Traditional lending largely answers the second question upfront. By contrast, counterparty risk leaves it open until the trade matures.
A related mistake is treating notional as exposure. A ₹500 crore swap rarely puts ₹500 crore at risk. Its exposure is usually a small fraction of notional, although it can grow quickly in volatile markets.
Banks use several related metrics, and each one answers a different question. Mixing them up is one of the most common mistakes in interviews and in model documentation.
Current exposure is the amount a bank would lose if its counterparty defaulted today. It equals the positive mark-to-market value of the contracts, net of any collateral held. If the value is negative, current exposure is zero, because the bank is the one that owes money. For example, if a client’s swaps are worth +₹15 crore to the bank and the bank holds ₹5 crore of collateral, current exposure is ₹10 crore.
Potential future exposure looks ahead. It estimates the highest exposure likely at a future date, at a chosen confidence level such as 95% or 99%. Think of PFE as a “bad but plausible” outcome. For example, a five-year swap might show a PFE of ₹40 crore at year two. Risk managers mainly use PFE to set and monitor counterparty limits.
Expected exposure (EE) is the average exposure at a future date across all simulated scenarios, with negative values set to zero. Whereas PFE focuses on the tail, EE captures the typical outcome.
Expected positive exposure (EPE) is the time-weighted average of EE over a period. Effective EPE (EEPE) then adds a conservative twist. It assumes exposure never falls during the first year, even if the portfolio would naturally run off. Regulators prefer this view because banks usually roll over maturing short-dated trades.
Together, EE and PFE form exposure profiles that analysts plot across the life of a portfolio. A plain interest rate swap usually shows a hump. Its exposure rises as rates drift, peaks part-way through its life, and then falls as fewer payments remain.
CVA is the market price of counterparty credit risk. It adjusts the risk-free value of a derivative for the chance that the counterparty defaults. In simple form:
CVA ≈ LGD × Σ EE(tᵢ) × PD(tᵢ₋₁, tᵢ) × DF(tᵢ)
Here, LGD is loss given default, PD is the marginal default probability for each period, and DF is the discount factor. The PD inputs usually come from CDS spreads or internal ratings, so sound PD estimation matters here too.
Banks also calculate debit valuation adjustment (DVA), the mirror image that reflects their own default risk. CVA also moves when a counterparty’s credit spread widens, even without a default. The Basel Committee noted that roughly two-thirds of counterparty-related losses during the global financial crisis came from CVA losses, not actual defaults.
For capital purposes, these measures feed into one number: exposure at default. EAD is the amount the bank expects to have at risk when a counterparty defaults. Regulators then apply a risk weight to it. Basel III offers two main routes for derivatives.
The standardised approach for counterparty credit risk (SA-CCR) replaced both the Current Exposure Method (CEM) and the Standardised Method in the Basel framework. The Basel Committee published the standard in March 2014, with an effective date of 1 January 2017. Its core formula is short:
EAD = α × (RC + PFE), where α = 1.4
Here, RC is replacement cost, which captures today’s exposure. For an unmargined netting set, RC = max(V − C, 0). V is the net market value of the trades, and C is the net collateral held after haircuts.
PFE captures possible future increases in exposure. It equals a multiplier times the aggregate add-on. The add-on depends on asset class, notional, maturity, delta, and a supervisory factor. Meanwhile, the multiplier reduces PFE when a netting set is out of the money or over-collateralised. However, it never cuts PFE below 5% of the add-on.
Finally, alpha is a conservative scaling factor borrowed from the internal model framework. There, it captures what average exposure alone misses, such as low portfolio granularity, correlated exposures across counterparties, and wrong-way risk.
Large banks with supervisory approval can use the Internal Model Method instead. Under IMM, the bank simulates thousands of market scenarios and revalues every trade along each path. EAD then equals alpha × Effective EPE. Alpha defaults to 1.4, although supervisors may allow a bank’s own estimate with a floor of 1.2. Basel III also requires stressed calibration when it produces a higher charge. For these reasons, IMM approval usually stays with large banks that have strong data, backtesting, and validation teams.
Suppose a bank has one unmargined five-year interest rate swap with a corporate client. The notional is ₹500 crore, and the swap is currently worth +₹12 crore to the bank. The client has posted no collateral.
Step 1 – Replacement cost: RC = max(12 − 0, 0) = ₹12 crore.
Step 2 – Adjusted notional: SA-CCR scales interest rate notionals by a supervisory duration. For a five-year swap starting today, SD = (1 − e^(−0.05 × 5)) ÷ 0.05 ≈ 4.42. So the adjusted notional is about ₹2,212 crore.
Step 3 – Add-on: The supervisory factor for interest rates is 0.5%. With a delta of 1 and a maturity factor of 1, the add-on is 0.005 × 2,212 ≈ ₹11.06 crore.
Step 4 – PFE: The netting set has positive value and no excess collateral, so the multiplier equals 1. Therefore, PFE = ₹11.06 crore.
Step 5 – EAD: EAD = 1.4 × (12 + 11.06) ≈ ₹32.28 crore.
Now assume the client had posted a one-off ₹10 crore cash independent amount at inception. There is still no variation margin agreement, so the netting set stays unmargined. Cash carries no haircut, so C = ₹10 crore. RC would fall to ₹2 crore, and EAD would drop to 1.4 × (2 + 11.06) ≈ ₹18.28 crore. In other words, collateral cut the exposure by more than 40%.
A note on simplification: this example uses a single trade to show the mechanics. Real portfolios aggregate add-ons across hedging sets and asset classes. Margined netting sets also use different replacement cost and maturity factor formulas.
Wrong-way risk occurs when exposure to a counterparty rises just as its credit quality falls. It turns an ordinary risk into a dangerous one, because the two drivers reinforce each other. Simple models often assume exposure and default are independent, so they understate losses when that assumption breaks.
General wrong-way risk comes from broad market factors. For example, imagine a bank with FX forwards against an emerging-market company. The company pays US dollars and receives local currency, yet it earns local-currency revenue. If the local currency collapses, the bank’s exposure jumps. At the same time, the company’s dollar obligations become harder to meet.
Specific wrong-way risk comes from a direct legal or economic link. A classic case is a counterparty posting its own shares as collateral. Similarly, a bank might buy credit protection on a company from that company’s own subsidiary. If the reference entity fails, the protection seller will almost certainly fail too.
Basel standards require banks to identify specific wrong-way risk and apply more conservative exposure treatment to affected trades. The opposite pattern, right-way risk, arises when exposure tends to fall as the counterparty weakens.
Measurement alone does not protect a bank. The real defence comes from contracts, collateral, market infrastructure, and discipline.
The ISDA Master Agreement lets two parties treat all their trades as a single contract. If one side defaults, the survivor nets positive and negative values into one close-out amount. Without netting, a bank could owe money on some trades yet still queue as an unsecured creditor on others.
The effect is substantial. According to ISDA’s analysis of BIS data, close-out netting reduced mark-to-market exposure by 85.3% at end-2025.
Enforceability matters just as much as the contract. Banks rely on legal opinions confirming that close-out netting works in each counterparty’s jurisdiction. In India, the Bilateral Netting of Qualified Financial Contracts Act, 2020 gave close-out netting clear statutory backing.
A Credit Support Annex (CSA) sits alongside the ISDA agreement. It sets the collateral rules, including eligible assets, haircuts, thresholds, and minimum transfer amounts. Thresholds deserve close attention. If a client’s threshold is ₹25 crore, it posts nothing until exposure crosses that level, so the bank carries that slice unsecured.
Variation margin (VM) covers changes in the current value of trades, usually daily. Initial margin (IM), by contrast, covers potential losses while the survivor closes out a defaulted portfolio. For uncleared derivatives, global margin rules typically calibrate IM to a ten-day horizon. Together, VM removes most current exposure, while IM absorbs the gap risk. In India, the RBI applies these rules through its 2024 directions on margining for non-centrally cleared OTC derivatives.
After the 2008 crisis, G20 leaders committed to clearing standardised OTC derivatives through central counterparties. A CCP steps between buyer and seller and becomes the counterparty to each side. As a result, many bilateral exposures turn into exposures to one well-capitalised entity.
CCPs protect themselves with initial margin, default funds, and a loss waterfall. When a clearing member defaults, the CCP first uses that member’s margin and default fund contribution. Next, it uses its own capital, and only then the default fund contributions of surviving members. However, clearing concentrates risk rather than eliminating it, so banks must still monitor their CCP exposures. In India, the Clearing Corporation of India Ltd (CCIL) plays this role in government securities, repo, and foreign exchange markets.
Banks set counterparty limits, usually based on PFE, and monitor them daily. Good practice also includes stress testing, concentration limits, and clear escalation when a counterparty breaches its limit. In addition, margin terms should tighten as a client’s positions grow, rather than stay fixed from onboarding.
Finally, many banks run a dedicated CVA desk. It prices counterparty risk into new trades and hedges CVA sensitivity with credit default swaps and market hedges. This reduces earnings volatility when counterparty spreads widen.
Archegos Capital Management was a family office run by Bill Hwang. It built large, concentrated equity positions, mainly through total return swaps with several prime brokers. The swaps gave Archegos economic exposure without direct share ownership, and they kept its large stakes out of public ownership filings. Because it spread these positions across several banks, no single lender could see its full leverage.
In late March 2021, ViacomCBS shares fell sharply. Archegos could not meet its margin calls and defaulted on 25 March 2021. Those that moved early largely contained their losses, while slower banks absorbed heavy ones.
Total losses across banks exceeded $10 billion, according to widely reported figures. Credit Suisse lost about $5.5 billion, and Nomura reported a loss of roughly $2.87 billion.
An independent review by law firm Paul, Weiss for Credit Suisse’s board found a fundamental failure of management and controls. It highlighted unaddressed limit breaches and a failure to prioritise dynamic margining.
The lessons map directly onto this article. First, static margin eroded as positions grew, so collateral fell behind the exposure it was meant to cover. Second, concentrated single-stock positions created severe wrong-way risk. Third, limit breaches meant little without escalation. Finally, banks need a full view of a client’s leverage, not just their own slice. The Basel Committee’s December 2024 guidelines on counterparty credit risk management draw on lessons from episodes like this one.
Under Basel III, banks hold capital for counterparty credit risk in two layers. First, default risk capital applies a risk weight to EAD under SA-CCR or IMM. Second, a separate CVA risk capital charge covers mark-to-market losses from widening counterparty spreads. The final Basel III reforms overhauled this CVA framework, introducing a basic approach (BA-CVA) and a standardised approach (SA-CVA). Banks also hold capital against CCP exposures, although trade exposures to qualifying CCPs attract a low 2% risk weight.
In India, the picture is still evolving. The RBI issued SA-CCR guidelines in 2016 but later deferred them, so banks still compute exposure under the Current Exposure Method. The RBI updated the CEM add-on factors in March 2026. Then, in June 2026, it released draft SA-CCR amendment directions, proposed to take effect from 1 April 2027. In August 2026, it also issued draft directions replacing its 2011 CVA framework with BA-CVA. For Indian practitioners, the practical lesson is to learn both CEM and SA-CCR, because transition projects will need both skills.
Counterparty credit risk rewards analysts who can think across disciplines. It asks you to combine default probabilities with simulated market paths, legal netting, and collateral mechanics.
The core ideas become manageable once you see how they connect. Current exposure tells you what you could lose today. PFE and EE describe tomorrow. SA-CCR and IMM translate them into exposure at default for capital, while netting, margin, and clearing shrink that number in practice.
For banking professionals, these skills sit at the centre of treasury risk, model validation, and regulatory capital roles. If you want structured, hands-on practice with PD, LGD, and EAD models, a guided programme will shorten the learning curve.
Explore our Credit Risk Modeling Certification Training to master Risk Analytics.
Both involve a borrower or counterparty failing to pay. In lending, the exposure is known upfront. In counterparty credit risk, exposure moves with markets, can switch between the two parties, and must be estimated as a distribution over the life of the trade.
Under SA-CCR, EAD = 1.4 × (RC + PFE). Replacement cost (RC) is today’s net value after collateral, floored at zero. PFE is a supervisory add-on for possible future exposure, scaled by a multiplier that gives credit for excess collateral or a negative mark-to-market value.
Wrong-way risk arises when exposure to a counterparty increases as its credit quality worsens. General wrong-way risk comes from market-wide factors such as currency moves. Specific wrong-way risk comes from direct links, such as a counterparty posting its own shares as collateral.
Explore Dexlab Analytics’ Credit Risk Modeling certification program to build PD, LGD, and EAD models from scratch, work through IFRS 9 ECL frameworks, and learn model validation techniques used by practicing risk teams.
.
Credit and Market Risk, Credit Risk, credit risk analysis, Credit Risk Analytics And Modeling, credit risk analytics training, Credit risk certification India, Credit risk modeling course India, Credit Risk Modelling, Credit Risk Modelling Using SAS, Credit scorecard modeling, Expected credit loss ECL, IFRS 9 modeling course India, Logistic regression scorecard, PD LGD EAD modeling, Python credit risk modeling
Comments are closed here.