Counterparty

Plain English

What is a counterparty, really?

Strip away all the finance jargon for a second.

A counterparty is simply the other side of a deal.

If you buy, someone else has to sell. If you lend, someone else has to borrow. If you insure something, someone else has to be the one taking on that risk for you. Every deal has two sides — you can't clap with one hand, and you can't do a deal all by yourself.

Whoever is standing on "the other side" of your transaction is your counterparty.

One important thing: "counterparty" does not mean "customer"

A counterparty can be:

• a customer buying a product from you;
• a bank you have no retail relationship with at all — you just exchange messages or trade with them;
• on a centrally cleared exchange, your legal counterparty may not even be the trader on the other side of the screen at all — more on why in the technical section below.

So every customer is, in a loose sense, a counterparty. But not every counterparty is a customer.

That gap matters a great deal in regulated finance, because the rules that apply — and the protections a firm owes someone — depend heavily on which side of that gap the counterparty falls on. More on that in the Professional view section below.

---

Simple example

Picture a coffee shop. You hand over £3.50, and get a coffee back. For those five seconds, you and the coffee shop are counterparties to each other — the barista is just the person serving the coffee shop's side of the deal, not the counterparty themselves.

Now scale that up. Two banks agree that Bank A will pay Bank B $50 million next Tuesday, and Bank B will pay Bank A £40 million on the very same day. Same basic shape as the coffee purchase — two sides, each expecting the other to deliver — just with a lot more zeros, a lot more paperwork, and a lot more at stake if one side doesn't show up.

Simple way to remember it

Every deal has two sides.
Whichever side isn't you, is your counterparty. 🤝

Why this word carries so much weight in finance

In the coffee shop, both sides hand things over at the same moment — coffee for cash, right there. In big financial deals, the two sides often don't hand things over at the same moment. A wire transfer can take hours. A bond can settle days later. A derivative might not finish for years.

During that gap, one question sits underneath almost everything:

"What if my counterparty doesn't hold up their end?"

That single question is the seed of a huge amount of financial regulation, risk management and — relevant to this lexicon — client documentation and communication. Who a firm is dealing with, how well it knows them, and how it is required to treat them, all trace back to this one word.

---

In practice

Important: the word "counterparty" on its own does not tell you how sophisticated the other side is, how risky the relationship is, or which rules apply. That is decided separately, through classification — covered in the Professional view section.

A simple walk-through

Imagine a UK asset manager wants to hedge currency risk on a US investment.

1. The manager enters into a forward FX contract with a bank. (FX contract: a contract that sets out which currencies will be exchanged, how much, at what rate, and when.)
2. The bank is now the manager's counterparty — the manager is relying on the bank to deliver dollars at the agreed rate on the agreed date.
3. If that contract is instead routed through a clearing venue, a central counterparty may step in between the two original firms. From that point on, both firms deal with the CCP, not with each other.

Why this matters for communications

This is where the word stops being abstract and starts touching the practical, document-facing side of the business: what a firm has to disclose to a counterparty, how it is required to classify them, and what evidence it must keep, all depend on exactly what kind of counterparty it is dealing with — a retail client, a professional client, an eligible counterparty, a correspondent bank, or a central counterparty. Getting the classification wrong is not just a labelling slip; it can mean sending the wrong protections to someone, or withholding protections someone was legally entitled to. [3]

---

Professional view

Classifying counterparties under UK conduct rules (FCA COBS 3)

In UK regulated financial services, a counterparty is often also a client. Where COBS 3 applies, the FCA's Conduct of Business Sourcebook requires firms to categorise that client into one of three categories:

• Retail client — a client who is neither a professional client nor an eligible counterparty. This is the default category, and it comes with the highest level of regulatory protection. [1]
• Professional client — either a per se professional client (an entity that automatically qualifies, such as certain regulated financial institutions, large undertakings and institutional investors) or an elective professional client (someone who has opted up from retail status after meeting qualitative and quantitative tests). [1]
• Eligible counterparty (ECP) — the most sophisticated category, carrying the least built-in protection. [1]

Important nuance: a client is only an eligible counterparty in relation to specific eligible counterparty business — broadly, dealing on own account, execution or reception/transmission of orders on the client's behalf, and closely related ancillary services. [2] Outside that limited scope, the same client has to be categorised as a professional or retail client instead. So "eligible counterparty" is not a permanent label stamped on an entity; it's a status that attaches to specific services or transactions, and a firm can be dealing with the same legal entity as an eligible counterparty for one activity and a professional client for another.

The per se eligible counterparty list

Certain entity types are automatically treated as eligible counterparties (unless categorised differently for a specific purpose), including: [2]

• investment firms;
• credit institutions;
• insurance companies;
• UCITS funds or their management companies;
• pension funds or their management companies;
• other authorised or regulated financial institutions;
• national governments and bodies that manage public debt;
• central banks;
• supranational organisations (such as the World Bank, the IMF, the ECB or the EIB).

Professional shorthand

Retail client: Does this person or entity get the full set of conduct-of-business protections by default?

Professional client: Has this entity earned, or automatically holds, the right to be treated as sophisticated enough to need less protection?

Eligible counterparty: Is this wholesale-to-wholesale business between two sophisticated market participants, where most retail-style protections simply don't apply?

Classification is not necessarily permanent. A firm must allow a professional client or an eligible counterparty to ask to be re-categorised into a category with more protection, and — separately — a firm may, under certain conditions, treat a client as an elective eligible counterparty at their own request. [1]

Classifying counterparties for derivatives: UK EMIR's FC / NFC split

A different classification system runs alongside COBS, built specifically for OTC derivatives: UK EMIR (the onshored, UK version of the EU's European Market Infrastructure Regulation).

Under UK EMIR, every counterparty to a derivative is either:

• a Financial Counterparty (FC) — broadly, regulated financial institutions such as credit institutions, investment firms, insurers and reinsurers, UCITS funds and their managers, and pension funds; or
• a Non-Financial Counterparty (NFC) — essentially, any other undertaking that isn't a financial counterparty or a central counterparty. [4][5]

Clearing thresholds are set per asset class. As of mid-2026, the FCA's published UK EMIR thresholds are: [4]

• credit and equity derivatives: €1 billion;
• interest rate and FX derivatives: €3 billion;
• commodity derivatives: €6 billion (raised from €3 billion in May 2026, after industry feedback that rising commodity prices had effectively eroded the old threshold).

How the mechanics actually work:

• The calculation is a two-step, group-wide test. Step 1: derivatives that are objectively measurable as reducing risk directly related to the NFC's commercial activity or treasury financing — genuine hedging — are excluded from the calculation entirely. Step 2: the gross notional value of everything left over is compared, per asset class, against the thresholds above. [4]
• An NFC may run that calculation itself, using its aggregate month-end average position over the previous 12 months. If it chooses not to, it must clear all of its OTC derivatives subject to the clearing obligation, regardless of size.
• If the calculation shows it exceeds the threshold in a given asset class, the NFC must notify the FCA. It only becomes subject to the clearing obligation for that asset class 4 months after notifying — not the instant the threshold is crossed. [4]
• Even an NFC that stays below every threshold isn't free of obligations: it still has to meet risk-mitigation requirements for uncleared trades — timely trade confirmation, portfolio reconciliation, dispute resolution and, where applicable, portfolio compression (a duty that only bites once an entity holds a large enough number of non-centrally-cleared OTC contracts with the same counterparty). What it does not pick up while below the threshold is the bilateral margin requirement — that only applies once an NFC crosses into NFC+ territory, putting it in the same scope as financial counterparties. [12]

Financial counterparties get a parallel, but not identical, carve-out. UK EMIR REFIT created the category of small financial counterparty (SFC) — an FC whose derivatives activity sits below the same asset-class thresholds used for NFCs. An SFC is exempt from the clearing obligation, but — unlike a small NFC — it stays in scope for margin as well as the other risk-mitigation obligations, because FCA margin rules apply to financial counterparties generally, not just those above a threshold. Confirming SFC status follows the same 12-month rolling calculation, notification and 4-month lead time as the NFC route. [4][12]

A note on EU EMIR: the EU's own regime has been diverging from UK EMIR. Under "EMIR 3," the EU is moving to a calculation methodology based mainly on uncleared OTC positions for NFCs (with FCs calculating both uncleared and aggregate cleared-plus-uncleared positions), alongside recalibrated thresholds. Following ESMA's February 2026 final technical standards, the European Commission formally adopted the delegated regulation implementing this on 14 July 2026. As a delegated act, it's still subject to a European Parliament and Council scrutiny period before it can be published in the Official Journal, after which it enters into force 20 days later. Counterparties can apply the new calculation in the same annual cycle they already use, or sooner if they want the benefit earlier — so for many, this may mean their next annual calculation cycle in 2027, though the regulation itself doesn't fix a single date for everyone. UK and EU thresholds and mechanics should therefore be checked separately, not assumed to match. [5]

Why two classification systems exist side by side

COBS classification answers one question: "How much protection does this client get when we serve them, and for which activities?"

UK EMIR classification answers a different question: "What derivatives-specific obligations — clearing, reporting, risk mitigation — does this entity have?"

The same legal entity can sit in different boxes under each regime at the same time — for example, a large corporate could be a per se professional client under COBS for one service, while separately being a non-financial counterparty below the clearing threshold under UK EMIR. Neither system overrides the other; a firm typically has to track both, independently.

---

Understanding counterparty risk

This is where the word "counterparty" moves from a classification label to a quantitative risk problem — the layer that keeps risk teams, quants and prudential regulators busy.

What is counterparty credit risk (CCR)?

Counterparty credit risk is the risk that the other party to a contract defaults before the contract is complete, and therefore cannot make all the payments it owes.

CCR is considered a hybrid risk: part credit risk (will the counterparty default?) and part market risk (how much would that default cost, given how the market has moved?). It applies mainly to OTC derivatives, securities financing transactions, and certain cash transactions in securities, FX and commodities. [6]

A key building block here is the netting set — a group of transactions with a single counterparty that are covered by a legally enforceable bilateral netting arrangement, so that exposure is calculated on the net position rather than transaction-by-transaction. [6]

Central counterparties (CCPs) and novation

The Bank for International Settlements defines a central counterparty as a clearing house that sits between the parties to contracts traded on one or more financial markets, effectively becoming the buyer to every seller and the seller to every buyer, so that the performance of open contracts is assured even if one original party defaults. [7]

The legal mechanism behind this is called novation: the original single contract between two parties is replaced with two new contracts — one between each original party and the CCP. From the moment of novation, both sides look only to the CCP for performance, not to each other. [8]

CCPs manage the risk they take on through initial margin, variation margin, default fund contributions and, ultimately, a "waterfall" of financial resources that absorbs losses if a member defaults.

Settlement risk (Herstatt risk)

Even outside derivatives, counterparty risk shows up in plain payments — this is called settlement risk, and it's often nicknamed Herstatt risk.

The name comes from the 1974 collapse of Bankhaus Herstatt in Germany. German regulators closed the bank in the afternoon of 26 June 1974 — but several of Herstatt's counterparties had already irrevocably paid Deutsche marks that day, expecting to receive US dollars later the same day in New York. Herstatt's New York correspondent then suspended the outgoing dollar payments, leaving those counterparties exposed for the full value of what they'd already paid. [9]

That single episode is a big part of why international central bank cooperation on payment systems — and eventually the Basel Committee itself — took the shape it did, and it's still the reference point for why payment-versus-payment (PvP) settlement matters specifically for FX: a PvP mechanism only lets one leg of a currency trade settle if the other leg settles at the same time, removing exactly the timing gap that caused the Herstatt losses. The clearest large-scale example is CLS (Continuous Linked Settlement), a specialist settlement system set up in 2002 in direct response to the G10 central banks' post-Herstatt strategy, which today settles a large share of global FX turnover on a PvP basis. Even so, BIS data shows a meaningful share of FX turnover still settles outside PvP protection, so Herstatt-style settlement risk hasn't disappeared — it has shrunk. [9]

Credit Valuation Adjustment (CVA) and the Basel III capital charge

CVA is an adjustment to the value of a portfolio of transactions with a counterparty, to account for the risk that the counterparty might default.

Basel III introduced a dedicated capital charge for CVA risk — the risk of a mark-to-market loss caused purely by a deterioration in a counterparty's creditworthiness, separate from an actual default. Three broad approaches exist under the Basel framework: the standardised approach (SA-CVA, which needs supervisory approval), the simpler basic approach (BA-CVA), and — for banks with limited derivatives activity below a materiality threshold — the option to set the CVA charge equal to their counterparty credit risk capital requirement. [6]

A detail that trips people up: direct transactions with a qualifying central counterparty are exempt from the CVA capital calculation — one of several places in the framework where clearing through a CCP is deliberately given a lighter capital treatment than an equivalent bilateral trade. [6]

CVA is also the best-known member of a wider family of "XVA" valuation adjustments that grew out of the same underlying problem — accounting for the fact that a derivative isn't really risk-free just because it's been priced as if it were.

Wrong-way risk

Wrong-way risk occurs when a firm's exposure to a counterparty is negatively correlated with that counterparty's own credit quality — in plain terms, the more money the counterparty would owe you if it defaulted right now, the more likely it actually is to default.

A textbook example: a company sells a put option on its own shares. If the share price falls, the company's own credit quality tends to fall with it — at exactly the moment its liability under the option is growing. Regulators distinguish this "specific" wrong-way risk (tied to how a trade is structured with a particular counterparty) from "general" wrong-way risk, which arises from broader macroeconomic factors affecting many counterparties at once, such as a shock to interest rates.

SA-CCR

For the exposure measurement that feeds into a bank's capital requirements, Basel III's post-crisis reforms introduced the standardised approach for counterparty credit risk (SA-CCR) — a non-modelled methodology for measuring counterparty credit risk exposure arising from derivatives, replacing the older, simpler approaches that regulators considered insufficiently risk-sensitive. [6]

---

Professional judgement

What should experienced compliance and risk professionals watch for?

Classification is a snapshot, not a permanent fact

An entity's COBS category or UK EMIR status reflects a point-in-time assessment. Both regimes explicitly allow — or require — reclassification when circumstances change: a professional client can ask to move to retail protection, an elective eligible counterparty arrangement can be requested or ended, and an NFC- can become an NFC+ — not simply because its "derivatives book grew" in a general sense, but specifically because its relevant, non-hedging, group-wide OTC exposure crossed an asset-class threshold. [1][4]

The practical implication:

A counterparty classification made two years ago is a claim to be re-verified, not a fact to be assumed.

The word "counterparty" alone tells you almost nothing about risk

"Counterparty" describes a role in a transaction, not a risk level. A sovereign central bank and a thinly-capitalised trading firm are both, technically, counterparties. The risk sits in the classification, the credit quality, the netting arrangements and the collateral — never in the bare word itself.

Two classification systems, one entity, two different answers

Because COBS and UK EMIR classify the same entity for different purposes, it's entirely possible — and normal — for a single counterparty to be, say, a professional client for conduct purposes and a small financial counterparty (SFC) for clearing purposes. Treating these as interchangeable, or assuming one implies the other, is a common source of avoidable errors.

A sophisticated classification doesn't decide your AML, sanctions or KYC position on its own

Being classified as an eligible counterparty under COBS, or as exempt from clearing under UK EMIR, is a conduct-of-business or derivatives answer — not a financial-crime one. Those obligations have to be assessed separately, under whichever AML, sanctions and due-diligence regime actually applies to the relationship, and the required depth of that assessment genuinely does vary by counterparty type. Wolfsberg's own correspondent banking guidance, for instance, allows a lighter financial-crime due-diligence standard for reporting-only, non-customer relationships than for transactional ones — while still expecting sanctions exposure to be considered in every case. [10][11]

Ownership matters

Dormant or stale counterparty classifications tend to accumulate quietly unless someone is explicitly accountable for reviewing them. That's a general control-design point worth applying to COBS and EMIR classifications too, in the same spirit as the ongoing-review discipline our SWIFT RMA entry describes for correspondent relationships — though it's a practical recommendation here, not something COBS or EMIR mandate by name.

---

Key takeaways

A counterparty is simply the other side of a financial transaction.

Not every counterparty is a customer, and not every customer is (only) a counterparty.

Under UK conduct rules (FCA COBS 3), every client is categorised as a retail client, professional client or eligible counterparty — and eligible counterparty status only applies to specific "eligible counterparty business," not as a blanket label.

A separate system, UK EMIR, splits derivatives counterparties into Financial Counterparties and Non-Financial Counterparties, with asset-class-specific clearing thresholds and a 12-month rolling calculation — crossing a threshold triggers notification, not an instant reclassification.

Even a non-financial counterparty below the clearing threshold has risk-mitigation duties (confirmation, reconciliation, dispute resolution) — but margin requirements specifically only kick in for financial counterparties and for non-financial counterparties above the threshold.

The EU's own regime has diverged from UK EMIR under EMIR 3 — the two shouldn't be assumed to match.

The same entity can carry different classifications under different regimes — and even different classifications under the same regime for different activities — at the same time.

Counterparty credit risk (CCR) is the risk that a counterparty defaults before a contract is complete — a hybrid of credit risk and market risk.

Central counterparties (CCPs) reduce bilateral counterparty risk through novation, margining and default-fund waterfalls.

Settlement risk, or Herstatt risk, is what happens when payment timing gaps let one side lose the full value of what it already paid.

Basel III introduced a dedicated capital charge for CVA risk, alongside the existing capital charge for counterparty default risk.

Wrong-way risk is when exposure to a counterparty rises exactly as that counterparty's credit quality falls.

A sophisticated classification doesn't settle a counterparty's AML, sanctions or KYC position — that has to be assessed separately, and the required depth genuinely varies by counterparty type.

---

Official sources

🛡 Always confirm against current official standards and your organisation's own policies.