Interest-Rate & Credit Derivatives

How to read this chapter

Derivatives transform risk through enforceable contracts. A derivative is a contract whose value depends on something else, such as an interest rate, yield curve, bond price, credit spread, reference entity, currency, commodity or index. That definition is correct, but it is not enough. The real banking question is: what risk is being transformed, who now carries it, how is it valued, how is it collateralised, how is it documented, how is it settled, and what can go wrong during the life of the trade?

Interest-rate derivatives help banks, corporates and investors manage exposure to changing rates. A borrower with floating-rate debt may want fixed-rate certainty. A bank with fixed-rate loans funded by rate-sensitive deposits may want to hedge structural interest-rate risk. A trading desk may quote swaps to clients and manage the resulting curve exposure. A treasury desk may use swaps to align funding cost with asset repricing. Credit derivatives help transfer or reference credit risk. A bank may buy protection on an issuer, sell protection to earn spread, trade a credit index, or use credit derivatives as part of portfolio management.

The danger is treating derivatives as only front-office trades. The trade ticket is only the start. A derivative can run for years and generate resets, coupons, valuations, collateral calls, margin disputes, novations, terminations, compression, exercises, credit events, settlements and reporting changes. The bank must keep legal terms, system economics, valuation models, risk sensitivities, collateral rules, accounting treatment and operational events aligned throughout the life of the contract. A small mismatch in day count, calendar, floating-rate option, payment frequency, clearing status, reference entity or collateral agreement can create real financial impact.

The BIS derivatives statistics track OTC and exchange-traded derivatives markets and show how large and operationally important these markets are for banks (BIS derivatives statistics). BIS data releases show interest-rate derivatives remain a dominant part of OTC derivatives activity, while notional amount should never be confused with actual risk, market value or exposure (BIS end-June 2025 derivatives release). For a learner, the scale matters because every contract must be booked, confirmed, valued, collateralised, risk-managed, reported and eventually closed.

Learning objectives

By the end of this chapter, you should be able to explain swaps, FRAs, futures, caps, floors, collars, swaptions, CDS, credit indices and structured credit derivatives in practical banking language. You should understand why derivatives are used for hedging, client risk management, ALM, trading, portfolio management and credit risk transfer. You should be able to connect derivatives to ISDA documentation, clearing, collateral, margin, valuation, counterparty credit risk, CVA, SA-CCR, market risk, accounting, product control, regulatory reporting and operations. You should also be able to design BA and testing scenarios for derivative trade capture, confirmation, reset, valuation, collateral, clearing, settlement, lifecycle events and exceptions.

1. The real purpose of derivatives

A derivative derives its value from an underlying reference, but the banking purpose is risk transformation. If a bank owns a fixed-rate bond, it is exposed to interest-rate movement. If it enters a swap that receives floating and pays fixed, it can change the risk profile. If a corporate has floating-rate debt and worries that rates will rise, it can use a swap or cap to create more certainty. If an investor wants exposure to credit risk without buying a bond, a credit derivative may provide a synthetic route. The derivative does not remove risk from the world. It moves, reshapes, prices or concentrates risk under a contract.

That distinction matters. A hedge may reduce one risk and introduce another. An interest-rate swap may reduce earnings sensitivity but create collateral calls when rates move. A credit default swap may hedge bond credit exposure but introduce counterparty risk, basis risk, documentation risk and settlement complexity. An option may protect against adverse movement but require premium and volatility valuation. A cleared derivative may reduce bilateral counterparty exposure but create daily variation margin and initial margin obligations. The hedge decision therefore needs both economic risk reduction and a funding plan for the resulting cash obligations.

Inside a bank, derivatives touch many teams. Front office quotes and books. Market risk measures sensitivities. Counterparty credit risk measures current and potential future exposure. Collateral teams calculate margin calls. Operations confirms trades and processes lifecycle events. Finance values trades and books P&L. Product control performs independent valuation checks. Legal maintains documentation. Compliance handles regulatory reporting and conduct obligations. Technology keeps product models, reference data, curves, calendars and interfaces working. A derivative implementation that only satisfies traders is incomplete.

The first professional habit is to ask why the derivative exists. Is it a client hedge, a bank hedge, a trading position, a balance-sheet management tool, a liquidity tool, a portfolio credit hedge or a structured product component? Purpose affects accounting, risk appetite, documentation, approval, reporting and conduct. A derivative booked without clear purpose is harder to govern.

2. Linear versus optional derivatives

Derivatives can be linear or optional. A linear derivative has value that moves broadly in a direct relationship with the underlying risk factor. A plain interest-rate swap has no option exercise decision, but its present value is not perfectly linear in rates: discounting, curve shape and convexity still matter. A forward rate agreement is linear. A futures contract is also broadly linear, although exchange margining creates daily cash settlement. Linear products can be complex, but their payoff is more direct than an option.

Options are different because they create asymmetry. The buyer has a right but not an obligation. A cap protects a floating-rate borrower against rates rising above a strike. A floor protects against rates falling below a strike. A swaption gives the right to enter a swap in the future. Option value depends not only on the current level of rates or spreads but also on volatility, time to expiry, moneyness, discounting and sometimes path features. This is why option risk uses Greeks such as delta, gamma, vega and theta.

The practical difference matters for users. A swap can become an asset or liability as rates move, and both parties are obligated. An option gives the buyer choice, but the buyer pays premium. A zero-premium-looking structure may embed obligations elsewhere. A borrower may prefer a cap because it protects against rising rates while allowing benefit if rates fall, but the premium can be expensive. A swap may be cheaper upfront but removes upside. A collar can reduce premium but gives away some benefit. Product suitability and explanation matter.

For systems, linear products and options need different data. A swap needs fixed rate, floating index, reset frequency, payment frequency, calendars and notional. An option needs strike, premium, expiry, exercise style, settlement method, underlying, volatility input and exercise workflow. A tester who treats an option like a swap will miss the most important events.

3. Interest-rate swaps

An interest-rate swap is a contract where parties exchange interest cash flows. The most common structure is fixed-for-floating: one party pays a fixed rate and receives a floating rate, while the other does the opposite. The notional amount is normally not exchanged in a vanilla interest-rate swap; it is used to calculate interest cash flows. If a bank pays fixed and receives floating, the swap gains value when floating-rate expectations rise relative to the fixed rate, all else equal. If rates fall, the value may move against the bank.

The swap has dates and conventions. It has trade date, effective date, termination date, notional, currency, fixed rate, fixed-leg day count, fixed-leg payment frequency, floating index, reset frequency, floating payment frequency, business-day convention, calendars and sometimes amortising or accreting notional schedules. A five-year quarterly floating leg versus annual fixed leg is not the same operational object as a ten-year semi-annual structure. Each date and convention affects cash flows.

Banks use interest-rate swaps for several purposes. ALM may use swaps to manage structural interest-rate risk between assets and liabilities. Treasury may use swaps to change funding profile. Markets desks may intermediate client hedges. Investors may use swaps to express curve views. A bank with fixed-rate loans funded by variable-rate deposits may use swaps to reduce earnings sensitivity. A corporate with floating-rate debt may pay fixed through a swap to create budget certainty.

The swap does not eliminate all risk. It introduces valuation movement, counterparty exposure, collateral requirements, basis risk, model risk, documentation risk and operational lifecycle risk. If rates move sharply, the hedge may be economically useful but generate large collateral calls. If the floating index changes or benchmark fallback applies, systems must process the correct rate. If the trade is cleared, margin flows occur through the clearing house. If bilateral, the credit support annex defines collateral terms. A mature bank looks at economic hedge and cash collateral behaviour together.

From the borrower’s hedge to the bank’s cash flow

Consider a fictional corporate with USD 20 million floating-rate debt paying a reference rate plus 1.50 percent. It pays the bank 4.00 percent fixed and receives the same reference rate on a swap with matching notional and accrual periods. Ignoring fees, credit effects and timing differences, debt interest plus swap interest becomes 5.50 percent. The loan spread is still paid; the swap does not cancel it. The bank receives fixed and pays floating on this client trade and may hedge its resulting exposure with a separate market trade.

For one illustrative period, both swap legs use actual/360, the accrual period has 90 days, and the contract’s floating-rate calculation has produced an annualised coupon of 4.80 percent. These assumptions are deliberate: actual contracts may use different accrual fractions on the two legs.

Payment netting applies only to amounts meeting the agreement’s conditions, such as the same payment date and currency. It does not mean every obligation across currencies or legal entities may be netted. Net cash payment and net fair value are different: the USD 40,000 coupon says nothing by itself about the present value of remaining payments.

For an overnight-indexed swap, a period’s floating amount commonly compounds the daily overnight observations. A simplified no-spread, no-shift actual/360 example is N × [product(1 + r_i × d_i/360) − 1], where r_i is a decimal annual rate and d_i is the calendar-day weight for that observation. Friday’s rate may cover three calendar days. An observation shift, lookback, lockout, payment delay or spread method changes implementation and must follow the confirmation. A simple average of daily rates is not interchangeable with compounding. The New York Fed’s SOFR averages and index methodology illustrates compounding and day weighting; it does not determine every derivative’s contractual convention.

4. Forward rate agreements and futures

A forward rate agreement, or FRA, is an agreement on an interest rate for a future period. It allows a party to lock an interest-rate exposure without exchanging the underlying deposit. At settlement, the difference between agreed rate and market reference is paid, normally discounted according to market convention. FRAs are useful for understanding forward rates and short-end hedging.

Interest-rate futures are standardised exchange-traded contracts linked to interest-rate instruments or reference rates. They provide liquid risk management and transparent margining through an exchange and clearing house. Futures can be used to hedge short-term rate exposure, express views on policy expectations or manage duration quickly. Standardisation is both strength and limitation. The contract is liquid because everyone trades the same terms, but it may not perfectly match the exact exposure being hedged.

Exchange-traded futures are marked to market daily through variation margin. This means gains and losses are settled in cash as the market moves. That reduces build-up of counterparty exposure but creates daily liquidity needs. A hedge that is economically correct can still create cash strain if variation margin moves against the bank before the underlying exposure produces offsetting cash.

For BA and testing work, futures require product static, expiry, contract code, exchange, clearing broker, margin, position, trade allocation, daily settlement price, variation margin and expiry processing. FRAs require forward period, reference index, agreed rate, settlement date, fixing date and settlement calculation. Both require market data and calendar accuracy.

FRA and futures direction checks

For an illustrative FRA settled at the start of the interest period, the fixed-rate payer receives N × (L − K) × α / (1 + L × α), with simple money-market discounting. Here N is notional, L the fixing, K the agreed rate and α the period’s day-count fraction. With USD 10 million, L = 5.00%, K = 4.00% and α = 90/360, the receipt is 25,000 / 1.0125 = USD 24,691.36. A lower fixing reverses the payment direction. This formula is for the stated advance-settlement convention; do not apply it to a contract paying an undiscounted difference at period end.

Some short-rate futures are quoted as 100 − implied rate. Under that quotation a rise in the implied rate lowers the futures price. A borrower hedging against higher future rates may therefore sell those futures. Bond futures have different price and delivery mechanics. Hedge size needs the contract’s tick value and sensitivity, exposure tenor, basis and, where relevant, conversion factors. Equal notional is not proof of equal risk. Futures and forwards can also differ because daily margin cash flows interact with changing rates, creating a convexity adjustment in pricing comparisons.

5. Caps, floors, collars and swaptions

An interest-rate cap protects against a floating rate rising above a strike. It is economically a series of options called caplets. A borrower with floating-rate debt can buy a cap to limit maximum interest cost while still benefiting if rates remain low. The buyer pays premium. The seller receives premium and takes the obligation to pay if the reference rate exceeds the strike. A floor works in the opposite direction, protecting against rates falling below a strike. A collar combines a cap and floor, often reducing upfront premium by giving up some benefit.

Caps and floors are common in lending, structured products, treasury hedging and client risk management. They are easier to explain than many exotic products but still require careful lifecycle. Each reset period has a fixing, comparison to strike, calculation and payment if in the money. Premium can be upfront or structured. The product may be linked to a loan or standalone. Accounting and hedge designation depend on documentation and policy.

A swaption is an option to enter an interest-rate swap in the future. A payer swaption gives the right to enter a swap paying fixed and receiving floating. A receiver swaption gives the right to receive fixed and pay floating. Swaptions are used to hedge optionality in balance sheets, mortgage portfolios, callable debt, structured products and client exposures. They are also traded as volatility instruments.

Options introduce volatility risk. A cap’s value increases when expected volatility rises, because the chance of rates exceeding the strike becomes more valuable. Time to expiry matters. Moneyness matters. A cap far out of the money may have low delta but meaningful vega. A bank selling options must manage option risk, not only expected cash payouts.

Customer communication is important. A cap is protection, not free money. The premium may feel expensive when rates do not rise, but that is the cost of insurance. A collar may seem cheaper, but the customer gives up benefit if rates fall below the floor. Sales teams should explain outcomes under rising, flat and falling rates in plain language.

Caplet payout and exercise control

For a simple term-rate caplet paid at the end of a 90-day actual/360 period, USD 10 million notional, 4.00 percent strike and 5.20 percent fixing produce 10,000,000 × max(0.0520 − 0.0400, 0) × 90/360 = USD 30,000 payable to the cap buyer. A 3.50 percent fixing produces zero, not a payment by the buyer. The premium is a separate obligation and must remain in the customer’s total-cost analysis. Overnight compounded cap products have their own observation and payout conventions.

An exercise notice is a distinct operational event from an in-the-money valuation. For a physically settled swaption, valid exercise creates the underlying swap and its future cash-flow schedule; cash settlement creates a cash obligation calculated under the agreed method. The bank should control notice deadline, time zone, authorised recipient and evidence, then reconcile the option’s closure with the new swap or cash payment. A system that books both a replacement swap and a cash payout for one exercise can double the obligation.

6. Credit default swaps

A credit default swap, or CDS, is a contract where the protection buyer pays periodic premium to the protection seller in exchange for compensation if a defined credit event occurs for a reference entity or obligation. The reference entity may be a corporate, sovereign or other entity. The contract specifies credit events, obligations, maturity, coupon, notional, settlement terms and documentation. CDS can be used to hedge credit exposure, express a credit view or trade relative value.

The buyer of protection is economically short credit risk. If the reference entity deteriorates or defaults, protection becomes more valuable. The seller of protection is economically long credit risk, receiving premium but exposed to loss if a credit event occurs. A CDS spread is the market price of credit protection. Wider spreads indicate higher perceived credit risk or weaker liquidity. Narrower spreads indicate stronger credit perception or stronger risk appetite.

CDS is documentation-sensitive. What is the reference entity? Which obligations are covered? What counts as bankruptcy, failure to pay or restructuring? How is settlement determined? Is the contract subject to auction settlement? Which definitions apply? ISDA’s 2014 Credit Derivatives Definitions and related protocol materials are an important industry source for credit derivative documentation changes (ISDA 2014 credit definitions protocol). The practical lesson is that credit derivatives are not only spread trades. They are legal definitions with market consequences.

A CDS can hedge a bond, but hedge effectiveness is not automatic. The bond and CDS may reference the same issuer but differ in maturity, seniority, deliverable obligations, restructuring treatment, liquidity and basis. CDS-bond basis can move. A hedge can reduce credit spread exposure while creating basis and counterparty risk. Risk teams should not accept the word hedge without reviewing the actual terms.

Credit-event payment versus credit-spread valuation

Suppose an illustrative CDS has USD 10 million protection notional and an applicable auction final price of 35 percent of par. The protection payment is 10,000,000 × (1 − 0.35) = USD 6,500,000 from seller to buyer, subject to the contract’s settlement terms. Accrued premium, any upfront payment and earlier premiums are separate from this calculation. The reference entity’s default does not mean that the protection seller has defaulted; these are different legal events.

A quoted market spread is also not necessarily the coupon booked into a standard contract. A CDS can trade with a fixed running coupon plus an upfront amount, whose sign depends on agreed coupon versus the market valuation. Operations must reconcile both cash components. A trader’s mark-to-market gain from buying protection before spreads widen is a valuation movement and does not by itself trigger a default payout.

7. Credit indices, baskets and tranches

Credit indices reference a basket of credit default swaps. They allow market participants to trade broad credit exposure more efficiently than trading many single-name CDS contracts. Index CDS can be used for macro credit hedging, trading, portfolio overlay and liquidity management. Because indices are more liquid than many single names, they are often used to quickly adjust credit exposure.

Index composition, roll dates, series and version matter. A trade on one series is not the same as a trade on another. Constituents can change when a new series is launched. Liquidity can migrate to the current on-the-run series. Systems must capture index family, series, version, maturity, coupon, notional and clearing status. A wrong series can create wrong risk.

Tranches reference slices of portfolio loss. They are more complex because the investor takes exposure to a specific layer of loss, such as equity, mezzanine or senior tranche. Tranches depend on default correlation, recovery assumptions, spread levels and portfolio behaviour. They are not beginner products and require strong product approval, model validation and risk control. A small change in correlation assumption can change valuation materially.

For learners, the key point is that credit derivatives can move from simple single-name protection into portfolio and structured risk. As structure increases, so does model risk, documentation risk, valuation uncertainty and conduct risk. Banks should match product complexity with control strength.

What a tranche attachment point means

For a simplified portfolio tranche attaching at 3 percent and detaching at 7 percent, the tranche width is 4 percent of portfolio notional. If cumulative portfolio loss is 5 percent, the loss allocated to that tranche is min(max(5% − 3%, 0), 4%) = 2% of portfolio notional, or 50 percent of tranche notional. At 2 percent portfolio loss it has not yet absorbed loss; at 8 percent it is fully exhausted. This waterfall illustrates loss allocation, not a tranche pricing model. Actual contracts also govern default timing, recoveries, premium notional reductions and settlement. Correlation assumptions affect the distribution of portfolio loss, so a hedge that matches a spread sensitivity may still fail under clustered defaults.

8. ISDA documentation and legal architecture

OTC derivatives live through legal documentation. The ISDA Master Agreement provides a standard framework for many privately negotiated derivatives. Confirmations define trade-specific economic terms. Credit Support Annexes or other collateral documents define collateral rights and obligations. Definitions booklets define product terms. ISDA’s 2021 Interest Rate Derivatives Definitions provide a framework used in confirmations for many privately negotiated interest-rate and currency derivative transactions (ISDA 2021 interest-rate definitions).

Documentation matters most when conditions are stressed. In calm markets, counterparties may resolve small issues through relationship. In default or dispute, the document decides. Netting, termination events, events of default, close-out amount, collateral eligibility, thresholds, minimum transfer amounts, dispute rights and governing law all matter. A bank should not book products that legal documentation cannot support.

The economics in the system must match the confirmation. If the confirmation says quarterly reset and the system books semi-annual reset, one side is wrong. If the day-count convention differs, cash flows differ. If the floating-rate option is wrong, coupons differ. If business-day convention is wrong, payment dates differ. If the reference entity is wrong, a credit derivative may not hedge the intended risk. Confirmation matching is therefore a financial control.

Business analysts should treat legal terms as data requirements. They should capture agreement type, governing law, netting set, collateral agreement, clearing status, product definitions, trade confirmation status and lifecycle authority. Legal documentation is not outside systems. It is one of the sources of truth systems must reflect.

9. Trade capture and derivative data model

Derivative trade capture must record the complete economic contract. A swap needs trade date, effective date, maturity date, notional, currency, fixed leg, floating leg, reset dates, payment dates, calendars, day-count conventions, business-day adjustment, clearing status, counterparty, book, trader, legal entity, confirmation status, broker, UTI or USI where applicable and downstream reporting flags. A CDS needs reference entity, seniority, restructuring clause, credit events, coupon, maturity, settlement method and documentation version.

A derivative is not safely captured when only headline fields are present. Many breaks come from the fields people assume are minor: holiday calendar, payment lag, fixing source, compounding convention, observation shift, day count, reset frequency, stub period, clearing account, collateral agreement, netting set and legal entity. These details decide cash flows, valuation, collateral and reporting.

The data model should keep trade version history. Amendments, cancellations, terminations, novations, allocations, compression and exercises should not overwrite history without audit. Users should be able to see what changed, when, why, who approved it and which downstream systems received the change. Derivatives live for years, so history is a control.

A useful requirement is simple: the trade captured in the system must be explainable from the legal confirmation, reproducible in valuation, visible in risk, usable in collateral, payable in operations, reportable to regulators and reconcilable to finance. If any of those fail, the trade capture is incomplete.

10. Derivative operating algorithm

Before execution, product approval and counterparty onboarding establish permitted products, legal agreements, credit capacity and reporting responsibility. After execution, the booked event feeds several processes in parallel: confirmation, clearing submission where applicable, risk, valuation, accounting and regulatory reporting. Reporting deadlines do not necessarily wait for confirmation or cash settlement. Margin and payment events then recur throughout the trade’s life. Each receiving system must know the trade version and return a processing acknowledgement or an actionable exception.

This operating sequence matters because derivative failures often come from weak handoff. A trader may book a swap correctly in front office, but confirmation may reveal a mismatch. Risk may receive the trade but use the wrong curve. Collateral may miss the netting set. Settlement may use the wrong payment date. Regulatory reporting may reject the UTI. Finance may book valuation before product control completes review. A derivative is controlled only when every handoff works.

The algorithm also creates accountability. If confirmation is unmatched, operations owns follow-up with front office and counterparty. If valuation is outside tolerance, product control owns challenge. If collateral is disputed, collateral operations and counterparty risk investigate. If a report is rejected, regulatory reporting owns correction. If market data is missing, market data operations and risk own resolution. A workflow without ownership becomes email noise.

For BA and developers, the operating algorithm should become workflow design. Do not only build screens for booking. Build statuses, exception queues, owner fields, timestamps, evidence, approval, downstream acknowledgements and ageing. A derivatives platform without exception management is not a bank platform.

One trade, several independently proved states

The diagram is an illustrative operating model. Execution creates economic exposure even if the confirmation is still unmatched. A clearing submission remains pending until the clearing service accepts it under its rules; rejection must trigger the agreed bilateral or unwind procedure, with limit and client impact checked. “Booked,” “confirmed,” “cleared,” “reported” and “cash settled” are separate states. A repository acceptance acknowledges a data submission, not the payment of a coupon.

Back office reconciles the scheduled coupon to the confirmed calculation and approved settlement instructions, the generated payment to the payment system’s outcome, and the outcome to the actual bank-account entry. Collateral movements have a separate reconciliation to custodian or cash balances. A failed cash payment leaves an unsettled obligation and potentially an overdue exposure; it must not silently close the trade. A duplicate lifecycle event needs idempotent processing using a stable event key and version, not a second payment.

11. Clearing and central counterparties

Many standardised derivatives are centrally cleared. In central clearing, a central counterparty stands between buyer and seller, becoming the counterparty to both sides under the clearing rulebook. Clearing can reduce bilateral counterparty risk and improve netting, but it introduces clearing-house margin, default fund, operational, liquidity and membership requirements. Cleared derivatives are not risk-free. They transform counterparty risk into clearing and liquidity processes.

Cleared trades are subject to the applicable margin rules. Initial margin covers potential exposure during close-out after default, within the model’s assumptions. Variation margin follows current mark-to-market exposure and may be called intraday as well as daily. Its legal character can be collateral or settlement, depending on the clearing service and contract. Either model creates cash-flow volatility. A hedge may be economically beneficial but create variation margin outflows before the hedged item produces offsetting cash. Treasury must include margin liquidity in stress.

Bilateral OTC derivatives may be subject to margin requirements for non-centrally cleared derivatives. The BCBS-IOSCO framework for margin requirements for non-centrally cleared derivatives increased the operational importance of collateral, eligible assets, thresholds, segregation, dispute resolution and margin call timing (BCBS-IOSCO margin framework). For banks, margin is not only a risk calculation. It is a daily operational cash and collateral process.

For systems, clearing status changes the workflow. A cleared trade may flow to a clearing broker or CCP, receive clearing confirmation, generate margin calls and settle through clearing accounts. A bilateral trade depends on counterparty confirmation and CSA terms. A platform must not treat cleared and bilateral trades the same just because both are swaps.

Collateralised-to-market and settled-to-market

Under a collateralised-to-market arrangement, VM secures the exposure while the derivative and collateral rights remain separately recognised under the relevant accounting policy. Under a legally settled-to-market arrangement, VM extinguishes the covered current exposure; the accounting and prudential analysis must support that treatment. Do not infer settlement treatment solely from the fact that a trade is cleared. The US interagency guidance published by the OCC specifically requires legal and accounting analysis for the covered contracts. Initial margin remains distinct from settlement of daily gains and losses.

A client clearing through a broker also faces client-to-broker arrangements, account segregation and portability conditions. It should not assume the same rights as a direct clearing member. A CCP reduces some bilateral exposures but concentrates dependency on its margin methodology, settlement banks, default management, recovery arrangements and the clearing broker’s operational performance.

12. Collateral, margin and UMR

Collateral reduces counterparty exposure by requiring one party to post cash or securities when the derivative has positive value to the other party. The details are defined in collateral agreements or clearing rules. Eligible collateral, currency, haircut, threshold, minimum transfer amount, independent amount, initial margin, variation margin, timing, dispute process and segregation all matter. A collateral call is both a risk event and an operational event.

Uncleared margin rules, often called UMR, made margin processes more important for non-centrally cleared derivatives. Firms in scope must exchange variation margin and, where applicable, initial margin under prescribed standards and documentation. The practical effect is that collateral operations, legal documentation, custody, eligible collateral, dispute workflow, margin model governance and settlement timing become core parts of derivative readiness.

Margin operations begin with exposure calculation. The bank values trades, applies netting and collateral terms, calculates required margin, sends or receives calls, agrees or disputes, settles collateral and reconciles balances. A dispute may arise because counterparties use different prices, curves, models, trade populations or collateral terms. If disputes are not controlled, credit exposure can grow.

For BA work, margin requirements should include netting set, CSA identifier, collateral currency, eligible collateral, haircut, threshold, minimum transfer amount, call frequency, valuation time, dispute reason, call status, settlement status, custody account and ageing. A derivative platform that ignores collateral cannot explain real counterparty exposure.

Margin is a separate cash and securities workflow

Assume an illustrative bilateral collateral agreement has zero threshold, no minimum-transfer constraint for this call and one netting set. The bank’s net positive derivative value rises to USD 1.20 million, while USD 0.90 million eligible cash collateral has already settled. Ignoring collateral interest and other adjustments, the bank calls USD 0.30 million additional VM. A call agreed today is not received collateral until settlement evidence exists. If the counterparty disputes USD 0.05 million, the undisputed USD 0.25 million should follow the agreement’s settlement process while the disputed part is investigated and exposure remains monitored.

If eligible securities are offered instead, a 5 percent agreed haircut means USD 1.00 million required collateral value needs about 1,000,000 / 0.95 = USD 1,052,631.58 market value, before any currency haircut or concentration constraint. Securities move to the agreed custody or collateral account; cash VM moves through the specified cash account. Eligibility, settlement, substitution and return controls must keep the collateral inventory distinct from the derivative notional.

The BCBS-IOSCO non-cleared margin framework distinguishes IM from VM, including treatment of IM segregation. Scope, phase-in history, thresholds, exemptions and eligible assets depend on the implementing jurisdiction and covered entities. UMR is not a rule that every derivative customer posts IM, and segregated IM is not unrestricted funding cash for the recipient.

13. Valuation, curves and market data

Derivative valuation depends on expected cash flows and market inputs. A swap valuation needs discount curves, projection curves, fixing history, payment dates and collateral discounting assumptions. An option needs volatility as well as rates. A CDS needs credit spread curves, recovery assumptions, discounting and credit event assumptions. A credit tranche needs correlation and portfolio assumptions. The more complex the product, the more important model validation becomes.

Modern interest-rate valuation often uses multi-curve approaches because discounting and forward projection may use different curves depending on collateral, currency and benchmark. A beginner does not need to build a curve model, but must know that index, discounting and collateral terms affect valuation. A swap referencing one index cannot automatically be valued with a generic curve.

Market data governance is crucial. Missing rates, stale curves, wrong calendars, incorrect fixing, bad volatility input or wrong credit spread can create valuation errors. Independent price verification checks front-office marks against independent sources or models. Product control investigates differences. Finance uses fair value. Collateral teams may use valuations for margin. If valuation numbers do not reconcile, disputes and reporting issues follow.

Testing should include normal valuation, missing market data, stale market data, curve shift, volatility shift, credit spread widening, rate fixing correction, benchmark fallback, day-count error and independent valuation difference. Valuation is not only math. It is controlled data plus approved model plus explainable result.

Valuing the remaining swap payments

For a receive-fixed/pay-floating swap, a simplified clean value from the bank’s perspective is sum[N_i × K × α_i × D_i] − sum[N_j × F_j × β_j × D_j]. Fixed coupons use fixed-leg accrual fractions α_i; projected floating coupons use their own fractions β_j; D discounts each payment to the valuation date. Already fixed coupons use the contractual fixing rather than a newly projected rate. A pay-fixed swap has the opposite sign. Collateral terms, index projection, multiple curves, stubs and valuation adjustments must be handled consistently.

For one remaining illustrative period, notional is USD 10 million, both accrual fractions are 0.25, fixed coupon is 4 percent, projected floating coupon is 5 percent and the discount factor is 0.98. Receive-fixed value is (100,000 − 125,000) × 0.98 = −USD 24,500. Pay-fixed value is +USD 24,500 on the same clean assumptions. This is a future-cash-flow valuation, not today’s settled coupon. A front-office mark, accounting fair value and contractual margin valuation can differ legitimately; reconciliation should identify methodology and adjustment differences rather than force them to match by manual override.

14. Benchmark reform and fallback discipline

Interest-rate derivatives historically referenced benchmarks such as LIBOR. Benchmark reform moved many markets toward overnight risk-free or near risk-free rates such as SOFR, SONIA and €STR, depending on currency and product. This changed documentation, systems, valuation, reset processing, confirmations, fallback language and client communication. The lesson is broader than LIBOR: any benchmark-linked product needs robust source, calculation and fallback logic.

A derivative confirmation should define the floating-rate option, reset dates, observation method, payment dates and fallback. Systems must know where to obtain the rate, what happens if the rate is missing, whether compounding applies, whether there is a lookback or observation shift and how holidays affect dates. A missing fixing should not become a manual guess without governance.

Fallback events can be operationally difficult. If a benchmark is discontinued, changed or unavailable, legal terms may define replacement rates and spread adjustments. Trades may need remediation. Clients may need communication. Valuation models may need changes. Hedge accounting may be affected. Reporting may need updates. A benchmark event is therefore legal, operational, technology, risk and client work together.

For BA requirements, benchmark fields should include rate source, administrator, publication calendar, fallback hierarchy, compounding method, observation convention, lookback, lockout where applicable, payment delay and manual override approval. Benchmark governance is now a permanent derivatives skill.

15. Lifecycle events

Derivative lifecycle events are actions that occur after initial trade booking. For interest-rate swaps, lifecycle includes rate fixing, reset, coupon calculation, payment generation, valuation, collateral calls, amendments, novations, compression, termination and maturity. For options, lifecycle includes premium, expiry, exercise, settlement and barrier events where relevant. For credit derivatives, lifecycle includes premium payments, credit events, auction settlement, physical or cash settlement, succession events and restructuring treatment depending on documentation.

A derivative can be booked correctly and still fail later. A wrong reset calendar may produce wrong coupon years after trade date. A missed novation may leave the wrong counterparty in systems. A termination may be agreed commercially but not reflected in risk. A compression event may reduce notional but fail in downstream reporting. A credit event may require fast operational interpretation. Lifecycle processing is therefore not routine administration. It is ongoing risk control.

Events must be authorised. Amendments should record what changed, why, who approved and what downstream systems were updated. Novations require agreement among transferor, transferee and remaining party. Terminations require valuation and settlement amount. Compression reduces gross notional but must preserve net risk correctly. Exercises require notice and timing. Credit events require documentation and market process.

Testing lifecycle is essential. Many derivative defects are not found on new trade booking. They appear at first reset, first payment, first margin call, first amendment, first novation, first exercise or first termination. A strong test plan follows trades through time.

16. Portfolio compression, novation and tear-up

Portfolio compression is a process where counterparties reduce gross notional by terminating or replacing economically offsetting trades while preserving materially similar net risk. It can reduce operational burden, gross notional, counterparty exposure and capital or margin costs depending on rules and portfolio. Compression is common in mature derivatives markets because old trades accumulate over time.

Compression is not simply deleting trades. The bank must verify that replacement or tear-up preserves intended risk, finance treatment, collateral exposure, regulatory reporting and accounting. Operations must terminate old trades, book replacement trades if any, update cash settlements, update risk, update regulatory reports and preserve audit trail. If compression is processed incorrectly, the bank may lose hedge protection or misstate exposure.

Novation transfers a trade from one counterparty to another with required consent. It changes legal counterparty and therefore credit exposure, collateral, netting, confirmation and reporting. A novation may be commercially simple but operationally sensitive. Systems must close exposure to the transferor and open exposure to the transferee correctly.

For testing, include full compression, partial compression, cash settlement, replacement trade booking, regulatory reporting update, risk preservation, accounting update and failed participant scenario. For novation, test three-party approval, counterparty change, collateral agreement mapping, UTI treatment, reporting and finance impact.

17. Hedge accounting and effectiveness

Derivatives are often used for hedging, but accounting hedge treatment is not automatic. Hedge accounting depends on applicable accounting standards, formal designation, documentation, risk management objective, hedge relationship, effectiveness assessment and ongoing compliance. This chapter is not accounting guidance, but banking learners should understand the practical point: a trade may be an economic hedge but still fail hedge accounting requirements if documentation or effectiveness rules are not met.

Hedge relationships can include fair value hedges, cash-flow hedges or net investment hedges depending on exposure and accounting framework. Under IFRS, the IFRS 9 standard overview notes the permitted policy choice to continue IAS 39 hedge accounting. US GAAP uses its own requirements under ASC 815; the design cannot assume IFRS qualification or effectiveness rules apply unchanged. A fair value hedge may hedge changes in fair value of a fixed-rate asset or liability. A cash-flow hedge may hedge variability in future cash flows, such as floating-rate interest payments. Hedge accounting affects where changes in derivative value and hedged item value appear in financial statements.

For systems, hedge designation must be controlled. The system should link derivative and hedged item, store hedge documentation reference, effectiveness testing results, start date, end date, designation type and de-designation events. If the hedge is amended, terminated or fails effectiveness, accounting treatment may change. Finance, ALM, risk and front office must stay aligned.

The practical risk is promising management that a derivative solves an accounting volatility problem when documentation is incomplete. Hedge accounting is a disciplined process, not a label added after booking. BA teams should involve finance early when derivatives are used for accounting hedges.

Valuation entries do not prove cash settlement

Consider an illustrative derivative measured at fair value through profit or loss, without hedge accounting, tax, netting presentation or other valuation adjustments. If it increases from zero to a USD 100,000 asset, the simplified entry is debit derivative asset USD 100,000 and credit valuation gain USD 100,000. No cash has moved. If it later falls to USD 70,000, debit valuation loss USD 30,000 and credit derivative asset USD 30,000. Each entry balances.

If the counterparty separately posts USD 100,000 cash VM under a collateral arrangement that finance treats as refundable collateral, receipt is debit cash USD 100,000 and credit collateral payable USD 100,000. It is not another USD 100,000 trading gain. A settled-to-market contract requires different entries consistent with extinguishing current exposure. The actual journal design depends on legal terms, framework, accrued coupons, netting and the bank’s chart of accounts.

In qualifying hedge accounting, finance may present effective cash-flow hedge movements in other comprehensive income and recognise ineffectiveness or reclassification as the applicable standard requires. A fair value hedge generally also recognises the hedged-risk adjustment of the hedged item. Neither treatment follows automatically from the trader calling the swap a hedge. The system needs designation and valuation lineage that finance can independently reproduce.

18. Counterparty credit risk, SA-CCR and CVA

Derivatives create counterparty credit risk because the contract can have positive value to the bank. If the counterparty defaults when the derivative is in the bank’s favour, the bank may lose replacement value after netting and collateral. Counterparty risk depends on current exposure, potential future exposure, maturity, volatility, netting agreement, collateral, wrong-way risk and counterparty credit quality.

SA-CCR, the standardised approach for measuring counterparty credit risk exposures, is the Basel framework for measuring exposure associated with OTC derivatives, exchange-traded derivatives and long-settlement transactions (Basel SA-CCR standard). A learner does not need to calculate every supervisory formula at this stage, but should understand why trade data, netting sets, collateral, maturity, asset class and margin status matter for counterparty credit capital.

Credit valuation adjustment, or CVA, reflects the market value of counterparty credit risk in derivative valuation. CVA can move because the counterparty’s credit spread changes and because market factors change expected exposure. Debit valuation adjustment, funding valuation adjustment and other XVA components may also be considered depending on bank policy and accounting. The practical point is simple: derivative fair value is not only clean discounted cash flow.

Wrong-way risk is especially important. It occurs when exposure to the counterparty increases at the same time the counterparty’s credit quality worsens. Controls include counterparty limits, netting agreements, collateral terms, exposure simulations, stress testing, wrong-way risk review, margin monitoring and dispute management. A derivative trade should not be approved only because its market risk is hedged. Counterparty risk may remain.

SA-CCR output is exposure, not the capital charge

The Basel CRE52 SA-CCR chapter calculates exposure at default per recognised netting set as EAD = 1.4 × (RC + PFE), with separate margined and unmargined rules. As an arithmetic example only, RC of USD 0.20 million and PFE of USD 0.80 million produce EAD of USD 1.40 million. The example assumes RC and PFE were already calculated correctly; it is not a shortcut from notional to exposure. EAD then feeds the applicable counterparty capital treatment. CVA capital addresses a different risk and must not be confused with this exposure number or with an accounting CVA reserve.

Legal enforceability matters before netting benefits are recognised. Two trades with the same commercial customer but different counterparties, agreements or legal entities cannot simply be offset. A negative clean mark does not establish zero future exposure. Initial margin, VM, settlement exposure and default-fund contributions may have different prudential treatments.

19. Credit derivatives and default process

Credit derivatives require a clear default process. When a reference entity experiences a possible credit event, the bank must determine whether the event is covered by documentation. Market processes may involve determinations committees and auction settlement depending on product and documentation. Operations, legal, risk, trading and finance must work together because timing and interpretation matter.

A credit event can trigger settlement obligations. In physical settlement, the protection buyer may deliver an eligible obligation and receive par. In cash settlement, payment is based on market-determined value or auction result. Deliverable obligations, maturity, ranking, currency, restructuring and documentation terms matter. A protection buyer expecting coverage may discover that a particular bond is not deliverable. A protection seller may face a large payment.

Credit derivatives also have premium accrual and settlement mechanics. The protection buyer pays periodic premium until maturity or credit event. If a credit event occurs, accrued premium and settlement are handled according to terms. Systems must process event dates, notices, auction results, settlement amounts and termination. A manual process under stress is risky.

For testing, include reference entity downgrade, credit event notice, disputed event, auction settlement, physical settlement, deliverable obligation issue, premium stop, recovery assumption update, index event and reporting. Credit derivatives are documentation-heavy because credit events are messy in real life.

20. Product control and independent valuation

Product control is the independent function that helps ensure trading P&L, valuation and balance-sheet reporting are accurate. In derivatives, product control is vital because models, curves, volatility, credit inputs and collateral terms can all influence valuation. A trader may have a commercial view, but finance needs defensible fair value. Independent price verification, valuation adjustments and P&L explanation are key controls.

P&L explain breaks down daily profit and loss into market moves, new trades, time decay, carry, valuation adjustments, fees, reserves and unexplained items. A swap desk should explain P&L through curve moves, basis, carry and trades. An options desk should explain delta, gamma, vega, theta and other effects. A credit desk should explain spread moves, default risk, carry and index basis. Unexplained P&L is a warning sign.

Valuation adjustments may include bid-offer, model uncertainty, close-out cost, funding, credit, debit or liquidity adjustments depending on product and bank policy. These are not arbitrary buffers. They recognise that model value may not equal executable exit value or accounting fair value under uncertainty.

Business analysts should include product-control requirements in derivatives platforms: independent marks, market data source, IPV tolerance, P&L explain feed, valuation adjustment, approval workflow, stale data flags and audit. Without product control, derivatives reporting can become front-office self-marking.

21. Regulatory reporting and trade repositories

Derivatives are subject to regulatory reporting in many jurisdictions. Regimes such as EMIR in Europe, Dodd-Frank reporting in the United States and other local rules require reporting of trade details, lifecycle events, counterparties, collateral and valuation information to trade repositories or regulators. Requirements differ by jurisdiction and change over time, so a bank must use current local rules. This chapter does not list all rule fields because implementation must be jurisdiction-specific.

The practical point is that derivatives reporting is data-heavy. Product type, UTI, counterparty identifiers, notional, price, maturity, clearing status, collateralisation, valuation, lifecycle event, execution venue and timestamps can all matter. A wrong field can create regulatory breaks. Reconciliation between counterparties can reveal mismatches. Reporting must update after amendments, terminations, novations and clearing changes.

Regulatory reporting is not an afterthought. The trade capture design must include reporting fields early. If a field is not captured, downstream reporting teams may rely on manual enrichment. Manual enrichment creates operational risk. A strong derivatives architecture treats regulatory reporting as a downstream consumer from day one.

Testing should include new trade report, valuation update, collateral update, amendment, termination, novation, correction, rejection from repository, delegated reporting and reconciliation mismatch. BA teams should involve regulatory reporting specialists because field interpretation can be subtle.

A reporting obligation matrix and a rejected event

For each jurisdiction, the implementation should record regulated entity, product scope, responsible reporter, delegation, repository, current schema and validation rules, event type, deadline, identifiers and correction procedure. US swaps and security-based swaps do not share one universal regulator or reporting rule set. Cross-border entities may have multiple obligations. Delegation changes who submits the report; it does not justify dropping the firm’s own completeness and oversight controls.

For an EU example, ESMA’s EMIR reporting page states that the revised RTS and ITS applied from 29 April 2024 and provides ISO 20022 schemas and validation material. EMIR 3 active-account reporting under Article 7b is a separate obligation for entities in scope; its RTS entered into force on 26 February 2026. An XML file passing schema validation is only the first check: business validation, repository acceptance, inter-repository reconciliation and source-to-report completeness still matter.

Suppose a partial termination reduces notional from USD 20 million to USD 12 million. The trade store, cash-flow engine, risk, collateral and reporting consumers must receive the same authorised version. If the repository rejects the event, operations should retain the rejection code, correct the reporting data under the applicable action/event rules, resubmit and prove acceptance. Risk must still reflect the economically effective termination; it should not wait for the repository to accept a message. A daily control reconciles outstanding reported trades with the source population and investigates missing, duplicated or stale events.

22. Systems architecture for derivatives

A derivatives architecture normally includes trade capture, product control, market data, valuation engine, risk engine, confirmation platform, collateral management, clearing connectivity, settlement, accounting, regulatory reporting, document management, counterparty data, static data and data warehouse. Some banks have one integrated platform. Others have many systems connected by feeds. The architecture must support accuracy, timeliness, reconciliation and audit.

The trade capture system records the deal. The valuation engine calculates fair value. The risk engine calculates sensitivities and stress. The confirmation platform matches legal terms. The collateral system calculates margin. The clearing system manages cleared workflows. The settlement system generates payments. Accounting posts entries. Regulatory reporting sends messages to repositories. Data warehouses support MI, finance, risk and audit. If trade identifiers do not align, reconciliation becomes painful.

Reference data is central. Counterparty, legal entity, netting set, CSA, clearing broker, product taxonomy, calendars, day counts, floating-rate options, credit reference entities, curves and settlement accounts must be correct. A derivative may fail because a calendar is wrong, not because pricing theory is wrong.

Architecture should support event-driven lifecycle. A reset event should update cash-flow calculation, payment, valuation and reporting. A termination should update risk, finance, settlement, collateral and regulatory reporting. A novation should update counterparty, legal documentation and exposure. A collateral dispute should be visible. Good architecture does not let lifecycle events disappear into emails.

Reconcile by purpose rather than by headline total

Store an immutable event history, but do not assume every consumer supports the same schema. An integration may use a vendor API, financial products markup, an internal event or a network message appropriate to the destination. A message standard transports agreed data; it does not itself establish contract enforceability, clearing acceptance or settlement finality.

23. BA requirements for derivatives

A strong BA does not write “process swap trade.” That is too vague. A useful requirement states product type, economic terms, lifecycle, validations, data sources, downstream feeds, controls and exceptions. For a swap, specify notional schedule, currency, fixed leg, floating leg, calendars, reset dates, payment dates, clearing status, counterparty, book, confirmation status, valuation and reporting. For a CDS, specify reference entity, seniority, restructuring clause, credit events, maturity, coupon, settlement method, clearing status and documentation version.

Acceptance criteria should be measurable. The system should reject invalid value dates. It should calculate cash flows using correct day count. It should create reset events on fixing dates. It should feed valuation to risk and finance. It should generate payment instructions on payment dates. It should update collateral exposure daily. It should report lifecycle events. It should maintain audit trail for amendments. These criteria are testable.

The BA should map downstream impact. A trade change may affect risk, P&L, collateral, settlement, finance, regulatory reporting and client statements. If the BA captures only front-office booking, delivery will fail later. Derivatives reward end-to-end analysis.

A BA should also capture exception ownership. Who handles unmatched confirmation? Who approves manual rate fixing? Who resolves valuation dispute? Who authorises termination? Who owns rejected regulatory report? Who signs off stale market data? Good requirements name not only data but accountability.

24. Testing scenarios

Testing derivatives requires time travel. A trade booked today may reset next month, pay next quarter, mature in five years and be amended in between. Test environments must simulate dates, market data, fixings, collateral calls and lifecycle events. Happy-path booking is only the first test.

For interest-rate swaps, test new trade, effective date, first reset, coupon calculation, payment generation, rate fixing missing, holiday adjustment, amortising notional, clearing acceptance, collateral call, valuation movement, amendment, novation, compression, partial termination, full termination and maturity. For caps and floors, test premium, reset, in-the-money payment, out-of-the-money period, volatility change, expiry and cancellation. For swaptions, test premium, expiry, exercise and underlying swap creation. For CDS, test premium payment, spread revaluation, credit event, auction settlement and termination.

Operational exception tests should include unmatched confirmation, wrong counterparty, wrong calendar, wrong day count, missing curve, stale fixing, failed payment, collateral dispute, rejected regulatory report, clearing rejection, duplicate trade message and manual amendment. Risk tests should include limit breach, stress shock, sensitivity mismatch, hedge ineffectiveness and wrong book classification.

The best derivatives test pack follows the trade from front office to legal, risk, collateral, settlement, finance and reporting. A derivative is only safely implemented when every downstream team receives the right version of the truth.

25. Practitioner casebook

A bank hedges fixed-rate mortgages with interest-rate swaps. Rates rise. The swaps help offset the mortgage risk economically, but margin movements create cash needs. The lesson is that hedge economics and collateral liquidity must be analysed together. ALM cannot stop at duration.

A corporate buys an interest-rate cap on floating debt. Rates never rise above the strike, so the cap pays nothing. The customer complains that the product was useless. The bank must explain that the premium bought protection, like insurance, and protection can be valuable even if not used. The lesson is that pre-trade explanation prevents post-trade disputes.

A swap confirmation is unmatched because the bank booked quarterly fixed payments while the counterparty has semi-annual fixed payments. The trade may look similar in notional and maturity but cash flows differ. The lesson is that confirmation breaks are not paperwork. They are wrong-risk warnings.

A bank buys CDS protection on a bond issuer, but the bond held is subordinated while the CDS references senior unsecured obligations. Spread movement is partly hedged, but credit-event settlement may not match. The lesson is that reference obligation, seniority and deliverability matter. A hedge label is not enough.

A clearing house increases initial margin after market volatility. The bank’s positions are hedged, but cash and collateral needs rise. Treasury must fund margin. The lesson is that derivatives create liquidity needs through margin even when market risk is controlled.

A benchmark fallback event occurs. Some trades move to a replacement rate correctly. Others fail because old confirmations or system fields are incomplete. The lesson is that benchmark governance is legal, system, operations and client work together.

A portfolio compression run removes thousands of old offsetting trades. Risk looks unchanged, but one legal entity mapping fails and regulatory reporting rejects part of the update. The lesson is that compression is a controlled lifecycle event, not a housekeeping exercise.

A collateral dispute grows for several days because both parties use different curve sets. Counterparty exposure increases while the dispute is open. Product control, collateral operations and market risk investigate. The lesson is that valuation differences become credit risk when collateral is not agreed.

26. Controls and governance

Derivatives governance begins with product approval. The bank should understand product purpose, target clients, legal documentation, valuation model, market data, risk measures, collateral treatment, clearing requirements, accounting, regulatory reporting, settlement process, operational capacity and conduct risks before launch. A product should not be approved merely because another bank trades it. The bank must be able to support it.

Limits should cover market risk, counterparty risk, credit risk, wrong-way risk, tenor, product type, notional, sensitivity, stress loss, collateral exposure, settlement exposure and operational exceptions. Limit breaches must have clear escalation. A derivatives desk with strong pricing but weak limit discipline can create losses quickly.

Model governance is essential. Pricing models should be approved, documented and validated. Model limitations should be known. Inputs should be controlled. Changes should be tested. Complex products require stronger model review. A model that nobody understands should not be trusted because it produces a number.

Operational governance includes confirmation timeliness, settlement breaks, collateral disputes, regulatory reporting rejections, market data exceptions and manual amendments. Aged exceptions should be visible to management. Derivatives risk often hides in old breaks.

27. Source references used for this chapter

• ISDA - 2021 Interest Rate Derivatives Definitions - ISDA source for the 2021 framework used in confirmations of privately negotiated interest-rate and currency derivative transactions.
• ISDA - 2021 Interest Rate Derivatives Definitions InfoHub - ISDA implementation and information hub for the 2021 definitions.
• ISDA - 2014 Credit Derivatives Definitions Protocol - ISDA source explaining the protocol to incorporate the 2014 Credit Derivatives Definitions.
• BIS - OTC derivatives statistics overview - BIS source explaining OTC and exchange-traded derivatives statistics.
• BIS - OTC derivatives statistics at end-June 2025 - BIS statistical release on outstanding OTC derivatives amounts and market values.
• BCBS-IOSCO - Margin requirements for non-centrally cleared derivatives - Official Basel Committee and IOSCO framework for margin requirements on non-cleared derivatives.
• Basel Committee - Standardised approach for counterparty credit risk - Official Basel Committee source for SA-CCR exposure measurement for derivatives and long-settlement transactions.

Final takeaways

Interest-rate and credit derivatives are not mysterious when read correctly. They are contracts that transform risk. The difficulty is that the transformation must be legally documented, economically valued, operationally processed, collateralised, reported and controlled for the full life of the trade. A swap is not only fixed versus floating. It is cash flows, dates, curves, collateral, counterparty exposure, risk sensitivity, accounting and lifecycle. A CDS is not only a spread. It is reference entity, credit event definition, settlement process, legal terms, counterparty exposure and basis risk.

The learner should leave this chapter with a practical instinct. When you see a derivative, ask what risk is being transformed, who receives that risk, how cash flows are calculated, what documentation governs the trade, how valuation is produced, what margin can be called, how lifecycle events are processed, how the trade is reported, what can break and who owns the break. That is derivatives inside a bank.