Working with Interest Rate Swap Modeling

Introduction

You're pricing or hedging interest rate swaps and need a compact, accurate model that links market rates to cash flows so you can make trading and risk decisions fast. An interest rate swap is a bilateral contract where parties exchange cash flows-usually a fixed rate for a floating rate-over a defined notional; modeling matters because it yields present-value prices, sensitivities (DV01, basis risk), and scenario P&L used for pricing, regulatory capital, and hedge design. Typical users are treasury, asset managers, and derivatives desks running daily mark-to-market, hedge rebalances, and counterparty exposure reports. One-line purpose: map cash flows to rates so you can value and hedge swaps. For a concrete anchor, model a 5-year $10,000,000 notional swap to compute PV, DV01, and a 1-in-100 stressed loss-here's the quick math: project each leg's cash flows, discount using the market curve, net the legs to get PV; what this estimate hides: curve construction, convexity and credit adjustments require separate treatment, but this core mapping is what you'll use day one-defintely practical and necessary.

Key Takeaways

• Core purpose: map cash flows to market rates to value and hedge swaps (PV, DV01, stressed P&L).
• Conventions matter - payment frequency, day-counts, effective/expiry dates and notional conventions must be consistent.
• Use a multi-curve setup: OIS discounting + tenor-specific forward curves built by bootstrapping and sensible interpolation.
• Valuation basics: DCF fixed leg, forward-based floating leg, par swap rate solves PV equality; report DV01 and scenario losses.
• Implementation essentials: calibrate to market quotes, validate out-of-sample, and guard against bad data, stale quotes, or mis-specified accruals.

Swap mechanics and conventions

You're modeling swaps to price, hedge, or report exposure and need crisp rules so your numbers aren't wrong at settlement; the direct takeaway: get payment frequency, day-count, and effective dates right - they drive cash flows and accruals. One quick line: swaps are just mapped cash flows - fixed known, floating reset.

Fixed-for-floating structure, payment frequency, day-counts

In plain terms, a vanilla interest rate swap exchanges a fixed coupon stream for a floating coupon stream referenced to an index (historically LIBOR, increasingly Term SOFR or similar). The fixed leg gives you predictable cash flows; the floating leg resets each tenor and pays based on the realized rate for the prior accrual period.

Best practices and checks:

• Confirm currency-specific conventions
• Record fixed payment frequency precisely
• Record floating tenor and reset lag
• Document day-counts on trade ticket
• Apply business-day rules consistently

Common market practice you should expect: fixed leg typically pays semiannually, floating (3‑month tenor) pays quarterly; floating day-counts are usually ACT/360, while fixed legs vary by market (often 30/360 or ACT/365) - but always validate the trade confirmation. One clean line: mismatched day-counts break cash-flow math fast.

Notional exchange and effective/expiry dates

Most plain-vanilla swaps do not exchange notional; the notional is a bookkeeping reference used to compute coupon amounts. Exceptions exist (currency swaps, some structured deals) where principal is exchanged - double-check if notional moves or there are optional early-termination rights.

Key date rules to nail:

• Spot/settlement lag (e.g., T+2 typical)
• Effective date (trade vs. forward start)
• Maturity/expiry date alignment
• Business-day adjustment method
• Stub period handling

Practical steps: capture the trade date, compute the spot date (apply spot lag), then project payment dates using tenor and business-day convention (Modified Following or similar). One clean line: wrong effective date shifts every cash flow and valuation.

Numeric example - 5y pay-fixed 2.50% vs 3m LIBOR-like floating

Assume a plain example to map cash flows: notional $100,000,000, tenor 5 years, you pay fixed 2.50% (semiannual fixed), receive floating indexed to 3‑month LIBOR-like rate (quarterly resets), floating day-count ACT/360, fixed accrual 0.5 year per period.

Here's the quick math for fixed leg cash flows (no discounting, illustrative cash-flow amounts):

• Per semiannual payment = notional × fixed × accrual
• = $100,000,000 × 2.50% × 0.5
• = $1,250,000 per fixed payment

Floating example for one quarter (assume a forward for that quarter of 1.80% - illustrative only):

• Quarter accrual = 0.25 year
• Floating payment = $100,000,000 × 1.80% × 0.25
• = $450,000 for that quarter

What this estimate hides: no discount factors applied, ignores day-count edge cases and stub periods, and assumes resets match payment dates. Practical checklist:

• Confirm accrual fractions for every period
• Handle stub periods explicitly
• Apply business-day adjustments to each date
• Use actual market index (Term SOFR vs LIBOR) and any spread

One clean line: compute each cash flow as notional × rate × accrual, then discount - small mismatches cause material P&L drift, so be rigorous and defintely log every convention.

Term structure and curve construction

You're building curves to value and hedge swaps; the short takeaway: build a clean OIS discount curve and separate tenor-specific forward curves, then bootstrap with stable interpolation to avoid pricing noise. Here's a practical, step-by-step guide you can run against market feeds and backtest within a week.

Define zero (spot) curve, forward curve, and discount curve

You need three related but distinct objects.

Zero (spot) curve - gives the continuously-compounded or discrete zero rates r(0,T) for every maturity T; from it you compute discount factors DF(0,T)=exp(-r(0,T)T) for continuous compounding or DF=1/(1+rT) for simple money-market conventions.

Forward curve - shows expected short rates for future periods, f(t1,t2). For a given tenor (e.g., 3m LIBOR-like), forwards are derived from zero rates: f = (DF(t1)/DF(t2)-1)/(t2-t1) under simple compounding, or via log differences for continuous compounding.

Discount curve - the set of discount factors used to PV cash flows; after 2008 the discount curve is typically the OIS discount curve for collateralized trades, not the tenor forward curve.

Steps:

• Collect clean deposit, futures, FRA, swap quotes.
• Convert quotes to a consistent day-count and compounding.
• Derive initial short DFs for very-short maturities directly from deposit/futures.

One-liner: get DFs right first - forwards follow from DFs.

Outline market inputs OIS for discounting, IRS/par swaps for forwards, FRAs, futures

Use market instruments according to the trade's collateral and tenor exposure.

OIS (overnight indexed swaps) - use OIS rates to build the discount curve for collateralized, cleared, or CSAs that pay overnight rates (USD: Fed Funds OIS; EUR: ESTR OIS). OIS is the anchor for PV discounting.

Interest rate swaps (IRS) / par swaps - use par swap rates across maturities to solve for forward rates and zero rates beyond short tenors. Par swaps provide long-maturity points where deposit/FRA data lack coverage.

FRAs and futures - fill the short- to mid-curve forward points: FRAs give single-period forward rates; Eurodollar/Fed Funds futures give market-implied forward rates for short dated tenors (adjust convexity for futures).

Best practices:

• Prefer OIS for discounting when trade is collateralized.
• Use basis swaps (e.g., 3m vs 6m) to map tenor spreads and build tenor-specific forward curves.
• Filter quotes by liquidity and timestamp; drop stale or wide-bid/ask levels.

One-liner: discount with OIS; build forwards from tenor market quotes and basis swaps.

Explain bootstrapping steps and interpolation choices linear log-linear cubic

Bootstrapping turns discrete market quotes into a continuous curve of DFs or zeros. Keep the steps deterministic and numerically stable.

Concrete bootstrap steps:

• Step 1 - input short rates: deposits and futures → compute DFs for first few maturities.
• Step 2 - use FRAs to solve single-period forwards and corresponding DFs where possible.
• Step 3 - for each par swap tenor, solve for the missing discount factor so that PVfixed = PVfloating (i.e., the model-implied par rate equals market par rate).
• Step 4 - once all instrument tenors are solved, construct zero rates r(0,T)= -ln(DF)/T or via appropriate discrete compounding.
• Step 5 - validate: re-price inputs and ensure residuals < tolerance (e.g., < 1e-8 absolute) and no negative DFs.

Numeric example (illustrative): given a 1y deposit rate 2.00% (ACT/360 money-market simple), DF(1y)=1/(1+0.021)=0.98039. Given a 2y par swap at 2.50% annual, solve DF(2y) so fixed leg PV equals floating leg PV using existing DF(1y).

Interpolation choices and trade-offs:

• Linear on zero rates - simple, monotone in rate space, but can produce kinks in DFs.
• Log-linear (linear on log(DF)) - preserves positive DFs and produces smooth discount behavior; often preferred for discount curve.
• Cubic splines on zero rates or on log-DFs - give smooth first/second derivatives (good for Greeks), but can oscillate and create arbitrage if unconstrained.

Practical recommendations:

• Use log-linear interpolation on discount factors for stability and positivity.
• Use monotone-cubic or tension splines for zero-rate smoothing where you need convexity stability, and check for local arbitrage.
• Set solver tolerances to 1e-10 for bootstrapped points; cap extrapolation slopes beyond the longest market quote to avoid wild forward rates.
• Backtest daily: reprice all instruments and track P&L drift; defintely flag any swap reprice > 1bp.

One-liner: bootstrap deterministically, interpolate on log-DF, and validate by re-pricing inputs every run.

Quant/Finance: build a 13-week OIS test curve, bootstrap a 5y par swap, and report any pricing mismatch > 1bp by Friday.

Working with Interest Rate Swap Valuation and Pricing Models

You want a clear, usable method to value swaps: use cash‑flow DCF with OIS discounting and tenor‑specific forward curves, solve for the par fixed rate, then map risks to the right curves.

Here's the quick takeaway: get discount factors right, use forward rates for float, and solve par = (PV floating)/(annuity).

Cash-flow DCF approach

Fixed leg present value (PV) equals the sum of each fixed coupon times its accrual fraction times the discount factor: PV_fixed = sum_i (Notional × fixed_rate × tau_i × DF(t_i)).

Floating leg PV is computed by projecting future floating coupons using forward rates, then discounting: PV_float = sum_i (Notional × fwd_i × tau_i × DF(t_i)) + Notional × DF(0) - Notional × DF(T) which simplifies, for standard par swap pricing, to Notional × (1 - DF(T)).

Practical steps and a numeric example (assumptions shown):

• Assume Notional = 100,000,000
• Fixed = 2.50% paid annually, 5y maturity
• Discount factors (OIS) at yearly payment dates: DF1=0.975, DF2=0.940, DF3=0.900, DF4=0.860, DF5=0.820

Compute fixed leg PV: coupon = 100,000,000 × 0.025 = 2,500,000. Sum DF = 0.975+0.940+0.900+0.860+0.820 = 4.495. So PV_fixed = 2,500,000 × 4.495 = 11,237,500.

Compute floating leg PV (par simplification): PV_float = Notional × (1 - DF5) = 100,000,000 × (1 - 0.820) = 18,000,000.

What this estimate hides: day‑count conventions, accrual stubs, first coupon reset timing, and any upfront fees; these change tau_i and the DF schedule, so implement exact accrual math in code.

One liner: always compute fixed cash flows with exact accruals and discount with the OIS DF schedule.

Par swap rate and solving for the fixed leg

Define par swap rate (S) as the fixed rate that sets PV_fixed = PV_float. Algebraically: S = (PV_float / Notional) / sum_i (tau_i × DF(t_i)).

Using the example above: numerator PV_float/Notional = 18,000,000 / 100,000,000 = 0.18. Denominator sum tau×DF = 4.495 (annual taus = 1). So S = 0.18 / 4.495 = 0.040044 → 4.00% (rounded).

Practical steps for implementation:

• Compute exact accrual fractions tau_i per market day‑count (ACT/365, 30/360, etc.).
• Use the same DFs used for pricing to compute the annuity (sum tau_i×DF_i).
• If irregular schedule or non‑annual payments, sum each cash flow exactly; avoid approximations.
• When S must be found numerically (nonlinear adjustments), use a stable root finder (bracketing + Brent) with tolerance ~1e‑9 on rates or ~1e‑2 dollars on PV.

Validation checklist: reproduce market par swap quotes to within basis‑point tolerances; if you miss >1bp, check day‑counts, DFs, or whether you discounted with OIS vs libor/tenor curve.

One liner: par rate = (1 - DF_T) / annuity - implement exactly with market day counts and you'll match swap desks.

Multi-curve setup: OIS discounting and tenor-specific forward curves

Post‑2008 practice separates the discount curve (usually OIS) from forward curves (3m, 6m, etc.). Discounting uses OIS DFs; forward projection uses tenor-specific curves calibrated to IRS, FRAs, futures, or cap/floor markets.

Practical calibration steps:

• Build OIS discount curve from OIS deposits and OIS swaps; bootstrap using market quoted OIS rates.
• Build tenor forward curve (e.g., 3m) by bootstrapping instruments that reference that tenor: FRAs, futures, and IRS that pay the tenor float.
• Use basis swaps (3m vs OIS) to fit spread between forward curve and OIS discount curve where needed.
• Choose interpolation on zero rates or log(DF); prefer monotone cubic or log-linear for stability across maturities.

Best practices and checks:

• Price instruments with the same conventions used in the market (same discounting and forward tenors).
• Reconcile that par swaps priced with OIS discount + 3m forward reproduce market par swap quotes for that tenor.
• Compute curve sensitivities per curve (PV01 per curve bucket) so hedges attach to the right market instruments.
• Watch basis risk: using the wrong discount curve or single‑curve pricing induces systematic P&L and hedge mismatch.

Common pitfall to avoid: using one curve for discounting and forwarding will defintely misprice and mis-hedge; always specify which curve each cash flow uses.

One liner: always discount with OIS and project forwards with the tenor curve that matches the floating leg - no shortcuts.

Working with Interest Rate Swap Risk Metrics and Sensitivities

You're managing swap exposure and need clear, actionable risk numbers so you can price, hedge, and report accurately. The quick takeaway: measure DV01/PV01 precisely, add convexity and vega checks for nonlinear risks, and run a set of scenario and stress tests that include curve twists and counterparty spread moves.

DV01 and PV01 explained

DV01 (dollar value of a one basis point move) or PV01 (present value of a one basis point move) is the primary linear sensitivity for swaps. It tells you how much the swap NPV moves per 1 basis point = 0.0001 change in rates. Use it for P&L attribution and initial hedge sizing.

Practical steps to compute DV01

• Coupon method: sum accrual times discount factor.
• Par-swap simplification: DV01 ≈ notional × sum(alpha_i × DF_i) × 0.0001.
• Check with central bump: NPV(rate +1bp) and NPV(rate -1bp); DV01 = (NPV_up - NPV_down)/2.

Quick numeric example, actionable so you can reproduce:

Assume a 5-year annual-pay par swap, notional = $100,000,000, flat discount curve at 3.00%, year fractions = 1. Sum of discount factors t=1..5 ≈ 4.5797. Then DV01 ≈ 100,000,000 × 4.5797 × 0.0001 = $45,797 per 1bp. Per $1m notional that's ≈ $458 per 1bp.

What this hides: for multi-tenor swaps use accrual alpha per period (e.g., 0.25 for 3m) and tenor-specific forward curves; always compute DV01 using the exact cash-flow schedule not a duration proxy.

Convexity, vega, and basis risk between curves

Convexity is the second-order sensitivity; it captures curvature in price with rate moves. Vega is sensitivity to implied volatility for caps/floors. Basis risk is exposure from mismatched curves or tenors, e.g., OIS discount vs LIBOR-like forwarding. All three are common sources of hedge slippage.

How to measure and act on convexity

• Compute central second difference: Convexity ≈ (NPV_up + NPV_down - 2×NPV_base)/(0.0001)^2.
• Report convexity per 1bp^2 and convert to dollar impact for expected move ranges.
• Hedge: use options or treasury futures for convexity if DV01 hedges leave residual curvature.

How to measure and act on vega for caps/floors

• Price caps with Black formula using forward rates and discount factors.
• Compute vega by bumping implied vol by 1 vol point and recalculating PV (or use analytic vegas where available).
• Hedge: buy/sell options or use strip hedges; track vol surface term-structure and liquid expiries.

How to measure and manage basis risk

• Quantify basis exposure as sensitivity to the basis spread: bump tenor basis by 1bp and compute NPV change.
• Use basis swaps to hedge tenor mismatches; size using basis DV01.
• Watch collateral/CSA terms: if discounting switches to OIS, previous hedges tied to a different discounting base will create basis P&L.

Best practices: central-difference for convexity, maintain a cap/floor vol surface calibration, and keep a mapping table of which curve (OIS, 3m, 6m) funds which cash flows so you can immediately identify basis sources. Also, defintely tag trades with collateral terms to avoid surprise basis.

Scenario and stress testing: rate shifts, curve twists, and counterparty credit spread moves

Scenarios give you forward-looking P&L and capital implications beyond linear sensitivities. Build a menu that covers parallel shifts, steepeners/twists, butterfly moves, and counterparty spread widening. Run both point shocks and path-dependent sequences for funding and collateral changes.

Concrete scenario set to run weekly

• Parallel shifts: ±100bp and ±25bp.
• Twists: short-end +50bp / long-end -50bp and the opposite.
• Butterfly: front and back move ±50bp with mid unchanged.
• Credit: counterparty spread widen by +50bp and +200bp for stressed counterparties.
• Liquidity/funding: OIS vs unsecured funding shock of +50bp with collateral rehypothecation changes.

Steps to implement scenario runs

• Define shock matrix and affected curves (discount, each forward tenor).
• Reprice swaps and related hedges using re-calibrated forward curves for each scenario.
• Compute NPV change, DV01 repricing error, and convexity residuals; flag trades where model delta differs from finite-difference DV01 by > 5%.
• For counterparty moves, rerun CVA/DVA with bumped credit spreads and update collateral thresholds if scenario triggers margin changes.

Operational best practices and guardrails

• Use central differences and multiple bump sizes to catch nonlinearity.
• Store scenario results in a time series for backtesting vs realized P&L.
• Automate overnight runs; human-review material moves (> $250k per desk or > 5% of position DV01).

What to watch: stale market data will understate real stress exposures; mismatched discounting assumptions blow up scenario results; and failing to reprice optionality under new forward curves yields wrong vega/CVA outcomes.

Next step: Quant/Finance to run the weekly scenario matrix, produce DV01/convexity/vega tables, and price a 5-year par swap using the 13-week test curve by Friday; attach discrepancies > 5% to the report.

Model implementation, calibration, and common pitfalls

You're building or validating a swap-pricing stack and need repeatable, auditable curves that match market quotes and hold up in stress - here's the focused playbook to implement, calibrate, and avoid the usual traps. Direct takeaway: pick explicit numeric conventions, calibrate to liquid quotes with tight tolerances, and automate data hygiene checks.

Recommend numeric choices: day-count handling, interpolation, stable solver tolerances

Start by locking the mechanical conventions - day-count, business-day roll, and accrual formulas - at the code-entry point and never let ad-hoc overrides slip in downstream. If the instrument says ACT/360, code ACT/360; if it says 30/360 ISDA, code 30/360 ISDA; store the convention with the instrument record.

One-liner: enforce convention at ingestion and treat any mismatch as a hard error.

• Implement day-count functions in double precision and unit-test them against ISDA examples
• Apply business-day adjustments using exchange calendars (NY, LON, TARGET) per-leg
• Calculate accrual factor once per coupon and cache it - repeated recompute invites tiny binary-diff drift

For interpolation, pick methods that fit the financial quantity you're bootstrapping:

• Interpolate discount factors in log-space (log-linear) for monotone discount curves
• Use monotone cubic (Fritsch-Carlson) for forward rates when smooth curvature matters
• Use linear interpolation on short ends (first 30-90 days) where instruments are dense

Solver and numeric settings to start with: root tolerance 1e-10, relative tolerance 1e-12, max iterations 100, and Jacobian check frequency once every 10 calls. These keep calibration stable without spinning forever. What this estimate hides: tighter tolerances cost CPU and increase sensitivity to noisy input - tune to production SLA.

Calibration: fit model to market quotes (swaps, FRAs, caps) and validate out-of-sample

Choose a minimal liquid instrument set that pins the curve: short OIS/Fed funds/FRA instruments to anchor discounting, money-market futures/FRA for the short forward curve, and IRS/par swaps across tenors to define mid-to-long forward points. Add caps/floors or swaption vols if you calibrate volatility surfaces.

One-liner: calibrate to liquid quotes, validate on the rest.

• Build an objective function on par-rate errors or PV errors; weight by liquidity (higher weight for on-the-run swaps)
• Use constrained nonlinear least squares (Levenberg-Marquardt) or trust-region methods; initialize from prior curve to speed convergence
• Regularize to enforce monotonic discount behaviour (penalize negative forward rates or large oscillations)

Target calibration thresholds: par-rate residuals 0.1 bps and PV error $500 per $100m notional as a practical starting rule - tighten if your desk requires sub-bp accuracy. Validate out-of-sample against:

• Nearby tenors not used in fit (e.g., 18m if you fit 1y and 2y)
• Caps/floors, futures, or client mid-curve swaps
• Historic intraday reprice tests (stability test)

Operational step: record calibration diagnostics (residuals, Jacobian condition number, iterations) to a daily audit table and alert if residuals exceed thresholds.

Warn about data quality, stale quotes, incorrect discount curve, and mis-specified accruals

Data is the frequent root cause of bad marks. Treat every quote with metadata: timestamp, venue, bid/ask, and liquidity flag. If a quote lacks timestamp or source, mark it suspect and exclude by default.

One-liner: bad input beats the best model every time.

• Detect stale quotes: reject quotes older than your market-tick threshold (e.g., rejects > 5 minutes for liquid IRS in high-vol times)
• Compare cross-venue prices and discard outliers beyond a robust z-score (e.g., > 5σ)
• Verify discount curve used matches collateralization - using the wrong discount curve (OIS vs unsecured) creates systematic PV shifts
• Test accruals: run unit tests that compare present-day accrual from trade terms to reference calculators

Common failure patterns and quick checks:

• Mis-specified day-count: small per-coupon errors compound over many coupons - run a parity test versus market par rates
• Wrong discounting: reprice a 5y par swap with both OIS and unsecured discount; check PV difference and flag > 0.5% of notional
• Stale liquidity: if bootstrapped nodal DFs jump > 10% intraday, isolate the quote causing it

Operational fix list: automate quote freshness checks, require dual-feeds for critical tenors, log the quote chain back to the trader, and add a daily reconciliation that reprices a canonical instrument set. Also, defintely include human sign-off when a curve change moves P&L materially.

Next step: Quant/Finance to build a 13-week test curve, price a 5y par swap, and upload calibration diagnostics to the audit table by Friday; owner: Quant/Finance.

Working with Interest Rate Swap Modeling - Conclusion

Restate core takeaway: accurate curves + consistent conventions = reliable swap valuation

You're validating swap models and need a crisp rule to judge results. The core takeaway: accurate curves + consistent conventions = reliable swap valuation.

Keep three pillars front and center:

• Build an OIS discount curve and tenor-specific forward curves (e.g., 3m SOFR forwards)

• Fix conventions: day-count, business-day shifts, stub rules, payment frequency, and interpolation

• Calibrate to live market instruments (deposits, OIS, swaps, FRAs, futures, caps) with clear tolerance targets

One-liner: get the curves and conventions right, and your pricing, hedges, and risk will track sensibly.

Immediate next step: build a 13-week test curve and price a 5y par swap to validate models

You should run a short, repeatable validation using market-close data as of 28-Nov-2025. This is a focused functional test - fast to run, high signal.

Step-by-step test recipe:

• Pull market quotes: overnight OIS/deposit rates, 13-week instruments (deposits or futures), 3m SOFR futures, and par swap rates across 1y-5y from Bloomberg/Refinitiv/CME by timestamp

• Bootstrap a 13-week discount curve using standard formulas: DF(t) = 1 / (1 + r × daycount_fraction) for short deposits, then OIS swap instruments for longer points; use ACT/360 for money-market legs and OIS conventions for discounting

• Interpolate between nodes with a clear rule (recommend: log-linear on discount factors for stability; document choice)

• Generate a 3m forward curve from the bootstrapped discount factors (forward = DF(t1)/DF(t2) - 1 adjusted by accrual)

• Price a 5y pay-fixed par swap: compute fixed-leg PV = fixed × sum(accrual × DF(payment_date)); compute floating-leg PV = DF(0) - DF(5y) (or sum of forward cashflows); solve for par fixed rate where PV_fixed = PV_floating

• Record outputs: par rate, PV (USD), PV01 (price sensitivity per 1bp), and calibration residuals for each input instrument

Quick math example (illustrative only): if annuity = 4.50 and DF(5y) = 0.78 then par ≈ (1 - 0.78)/4.50 = 4.89%. This example is to show the algebra; use live DFs in your run - it's defintely worth the effort.

One-liner: run the 13-week curve then price a 5y par swap - if both match market within tolerance, your pipeline is OK.

Owner: Quant/Finance to run the test and report discrepancies by Friday

Assign clear owners, deliverables, and escalation rules for a fast iteration. Deadline: 05-Dec-2025 (Friday).

• Quant: build the 13-week OIS discount curve, construct tenor forward curves, produce par swap price, PV01, and convexity

• Finance: independently price the same 5y par swap using the production library and provide side-by-side numbers

• Market data/Operations: certify source and timestamp for all input quotes and attach raw snapshots

• Risk: run scenario shifts (+/- 25bp parallel, +/- steepener/flatteners) and report P&L and basis exposures

• Deliverable packet: par rate, PV (USD), PV01, calibration residuals, interpolation method, and raw market snapshots

Escalation rules: if par-rate divergence > 1bp or PV01 difference > 5%, escalate to Head of Quant and Head of Trading with the packet.

One-liner: Quant/Finance run the test and send results to Risk and Trading by 05-Dec-2025.