What Is Counterparty Risk?

Michael Whitman · Trading since 2017, full-time since 2020 · YouTube creator, 137K subscribers · Published 2026-09-14 · Updated 2026-10-27

Being right about a trade is only half of getting paid. The other half is whoever owes you the money still being able to hand it over, and that is a separate risk with its own rules.

How it works

Every contract has two sides. If the agreement is directly between you and another party, your gain depends entirely on their ability and willingness to pay it.

This is invisible most of the time. Counterparties pay, contracts settle, and the risk looks like an abstraction — right up until one does not.

Exchange-traded instruments largely solve it. A clearing house becomes the buyer to every seller and the seller to every buyer, so you face the clearing house rather than a stranger.

Over-the-counter contracts do not. A forward, a swap, a bespoke option — these are private agreements, and the person on the other end is the guarantee.

Netting shrank dealers’ derivatives gross market value of $22.80 trillion to credit exposure of $3.35 trillion at the end of 2025, as counterparty credit risk reports.

A shorter cycle leaves less time for a trade to go wrong between execution and delivery; US equities reached T+1 settlement on 28 May 2024.

The exposure grows when you are winning

A contract worth nothing to you carries no counterparty risk. Nobody owes you anything, so nobody can fail to pay.

As the trade moves your way, the amount owed grows — and so does the loss if they default. The exposure is at its largest exactly when the position is at its most valuable.

Collateral is the standard answer. The losing side posts assets as the trade moves against them, so the winner’s exposure is covered. That reduces the risk substantially and never to zero — collateral is posted with a lag, valued at a price that can move, and disputed in exactly the conditions where it matters.

A worked example

Take a hypothetical contract whose collateral is set for a move of 1% a day, the size its margin model treats as ordinary, and then a day that moves 5%.

Collateral arrangements are sized on typical movement. A margin call issued against yesterday’s close assumes today looks roughly like an ordinary day.

On the 5% day, it does not. The move is five times what the model allowed for, so the collateral posted covers about a fifth of the new exposure and the gap is uncovered until the next call — which arrives after the damage.

And volatility clusters, so large bars arrive in runs rather than singly. Several consecutive under-collateralized days is how a manageable exposure becomes an unmanageable one.

Where you already have it

Your broker. Cash and securities held with them are a claim on them. Segregation rules and investor compensation schemes exist precisely because that claim can fail, and both have limits.

Your bank. A deposit is a loan to the bank. Insurance covers a stated amount; above it you are an unsecured creditor.

Any structured product. A note issued by a bank pays out because the bank does. The underlying index performing as promised is irrelevant if the issuer has gone.

And any physically-unbacked exposure. If the instrument tracks something rather than holding it, somebody is promising to deliver the difference.

The original data

Counted on S&P 500 closes from 1950 to 2026, a slide of at least 5% from the prior record began 68 times, a little under once a year. The worst of them, from October 2007 to March 2009, reached 56.8% — the same stretch in which counterparty failures stopped being an abstraction.

The distance between a routine pullback and that extreme is the collateral problem in one comparison. Any arrangement calibrated on typical movement is under-provisioned for the extreme, and the extreme is what produces defaults.

How institutions actually manage it

Netting is the first tool. Two parties with many contracts between them agree that only the net amount is owed, so a hundred offsetting positions collapse into one exposure. That single change removes an enormous amount of gross risk without either side altering a position.

Initial margin covers the gap collateral leaves. Variation margin settles today’s move; initial margin is posted up front to cover the move that happens between a default and the position being closed out. It exists precisely because collateral arrives with a lag.

Credit limits cap how much exposure any one counterparty may accumulate. A desk that would happily do a trade will refuse it once the limit is reached, regardless of how attractive the trade looks — which is a risk control that overrides the profit motive by design.

And credit default swaps let the exposure be insured. You can buy protection against a counterparty failing, from a third party — who is then a counterparty themselves, which is the recursion at the heart of why this risk is so hard to remove rather than relocate.

Currency forwards show how much of this exposure sits between dealers: in April 2025, 39.25% of forward market currency turnover was one dealer trading with another.

How this connects

Counterparty credit risk is the formal version of this idea, built for derivatives whose exposure has to be estimated across a contract’s life rather than read off today’s value.

An interest rate swap is the textbook case. It starts worth nothing to either side and becomes a claim on whichever party the rates moved against.

A central counterparty is the institutional fix, and settlement risk is what remains in the short window before a trade completes, when one side may have delivered and the other not.

On a crypto exchange the platform is the counterparty. Coins left there are a claim on the exchange, which is why self custody in a cold wallet removes that exposure rather than insuring it.

Leverage can make you the risky side. A leveraged account that falls into deficit is the broker’s exposure to you, and the forced close-out is how the broker manages it.

Cash above protection limits is often moved into a treasury bill, swapping a bank’s promise for a claim on the government; a money market fund holds similar short paper on your behalf.

And concentration risk magnifies it. A single broker, bank or exchange holding everything turns one institution’s failure into the whole of your loss.

When it fails

The characteristic failure is assuming a hedge removes risk rather than relocating it. A position is hedged with an over-the-counter contract, the hedge performs exactly as designed, and the counterparty fails at the same moment — because whatever caused the market move is also what damaged them. The hedge was correct and worthless, and the correlation between “my hedge pays out” and “my counterparty cannot pay” is not accidental. It is the same event.

A second failure is treating a clearing house as absolutely safe. It concentrates the exposure rather than deleting it, which systemic risk covers.

A third is ignoring it on long-dated contracts. A twenty-year swap depends on a counterparty existing in twenty years, which is a much stronger claim than it existing today.

A fourth is holding balances above compensation limits without treating the excess as a position.

The question: The trade was right. Did they pay?

And a fifth is assuming a credit rating settles it. A rating is an opinion about likelihood, produced by an agency the issuer pays, and it moves slower than conditions do.

Related

Systemic risk covers what happens when these failures propagate. Risk management covers sizing against an exposure that grows as you win. And corporate bond covers the instrument where this risk is priced explicitly as a spread.

This page is educational, not financial advice. Test every idea on your own charts before risking money.