Rates —
Treasuries, swaps & the curve.

US Treasuries are debt securities issued by the US Department of the Treasury, backed by the full faith and credit of the United States government. They are the global benchmark risk-free asset and the most liquid securities market in the world.

The categories by maturity:

• Treasury Bills (T-Bills): Maturities of under one year — 4 weeks, 8 weeks, 13 weeks, 26 weeks, and 52 weeks. T-Bills do not pay a coupon; they are issued at a discount to face value and mature at par. The difference is the investor's return. T-Bills are traded by the money market desk.
• Floating Rate Notes (FRNs): 2-year maturity, floating rate linked to 3-month T-Bills with quarterly coupons. Introduced in 2014. Much less common than other Treasuries.
• Treasury Notes: Maturities of 2, 3, 5, 7, and 10 years. Fixed coupon, paid semi-annually. The 10-year Note is the global benchmark.
• Treasury Bonds: 30-year maturity only. Fixed coupon, paid semi-annually. The longest duration US government instrument.
• TIPS (Treasury Inflation-Protected Securities): Issued as 5-year Notes, 10-year Notes, or 30-year Bonds. The principal adjusts with CPI, so the coupon payment rises with inflation. Provides a guaranteed real return rather than a nominal return.

The UK equivalent is gilts — conventional gilts pay fixed semi-annual coupons and index-linked gilts adjust principal with RPI (Retail Price Index). For Germany the benchmark is Bunds (10yr/30yr), Bobls (5yr), and Schätze (2yr).

TIPS (Treasury Inflation-Protected Securities) are US government bonds where the principal adjusts daily with CPI. The coupon rate is fixed, but because it is applied to an inflation-adjusted principal, the actual dollar coupon payment rises with inflation. At maturity, the investor receives the greater of the inflation-adjusted principal or the original par value — providing downside protection in deflation scenarios.

TIPS yields are quoted as real yields — the return above inflation. If a 10-year TIPS yields 1.8% real and CPI averages 2.5% over that period, the nominal return is approximately 4.3%. If inflation surprises to the upside, TIPS outperform conventional Treasuries (which are fixed in nominal terms). If inflation comes in below expectations, conventional Treasuries outperform.

The difference between the conventional 10-year Treasury yield and the 10-year TIPS yield is the breakeven inflation rate — the market's implied expectation for average CPI over the next 10 years. This breakeven is one of the most closely watched inflation gauges in markets.

Investors buy TIPS when they are concerned about inflation eroding the real value of their fixed income portfolio, or when real yields are high enough to offer an attractive return above inflation in their own right. Pension funds and endowments use TIPS to match real liabilities.

Treasuries are sold to the market through a Dutch auction process run by the US Treasury. The mechanics:

At a specified time, all primary dealers — the ~25 large financial institutions (including Barclays, Goldman Sachs, JPMorgan, etc.) that are required to participate in all auctions — must submit bids specifying the yield they require and the quantity they want. Other institutional investors can submit non-competitive bids (accepting whatever yield the auction clears at).

Once the auction closes, bids are ranked from lowest yield (most aggressive) to highest yield (least aggressive). The Treasury accepts bids starting from the lowest yield until it has filled its target issuance size. The clearing yield — the highest yield needed to fill the auction — becomes the coupon rate for all bonds in that auction. Crucially, all successful bidders receive this same yield, even if they bid at a lower (more aggressive) yield.

Auction results move markets: a "tail" (auction clearing at a higher yield than the pre-auction market price) signals weak demand and pushes yields higher. A "stop-through" (clearing below the market) signals strong demand and rallies rates. The bid-to-cover ratio (total bids received divided by amount sold) is the primary gauge of demand strength.

Auction schedule: 2yr, 3yr, 5yr, 7yr notes are auctioned monthly. 10yr and 30yr securities auction every three months (quarterly). T-Bill auctions happen weekly.

The on-the-run Treasury is the most recently issued security at each maturity point — the current benchmark 2-year, 5-year, 10-year, and 30-year. When a new security is auctioned, it becomes the on-the-run and the previous on-the-run becomes "off-the-run."

On-the-run Treasuries are the most liquid — they trade on electronic exchanges, can be executed in large size with tight bid-ask spreads, and are the primary instrument used for hedging. This is where the bulk of rates trading volume occurs.

Off-the-run Treasuries are all previously issued securities of the same maturity. They trade over-the-counter with wider bid-ask spreads and lower volume. Their prices differ slightly from on-the-run securities with similar remaining maturities — they are typically cheaper (higher yield) because of the liquidity discount.

This liquidity premium creates a persistent spread between on-the-run and off-the-run Treasuries of similar duration — a classic relative value trade for rates funds. When markets are stressed, this spread widens as investors flee to the most liquid instrument; when markets are calm, it compresses.

By definition, there are 24 outstanding 2-year notes, 60 outstanding 5-year notes, 40 outstanding 10-year notes, and 120 outstanding 30-year bonds at any given time — each with a different coupon, different remaining maturity, and different liquidity profile. This complexity is one reason rates trading cannot be fully automated the way equity trading can.

STRIPS stands for Separate Trading of Registered Interest and Principal of Securities. They allow investors to hold and trade the individual coupon and principal components of a Treasury note or bond as separate zero-coupon securities.

A 10-year Treasury pays 20 semi-annual coupons plus a final principal repayment — 21 individual cash flows in total. Each of these can be stripped out and traded independently. The stripped coupon payments become zero-coupon bonds maturing on each coupon date; the stripped principal becomes a zero-coupon bond maturing in 10 years.

Zero-coupon bonds (STRIPS) have some useful properties: they have no reinvestment risk (all return comes at maturity), they have longer duration than coupon-bearing bonds of the same maturity (because all cash flows are at the end), and they trade at a deep discount to par (a 10-year zero-coupon bond might trade at 65–70 cents on the dollar, with the full return coming at maturity).

Institutional investors — particularly pension funds matching long-dated liabilities — use STRIPS for duration management. The very long duration of principal STRIPS makes them efficient instruments for extending portfolio duration without adding credit risk.

Treasuries are quoted in price terms (as a percentage of face value) using a fractional convention based on 32nds — not decimals.

100-17 means 100 + 17/32 = 100.53125. A bond priced at "100-17" costs 100.53125 per £100 (or $100) of face value.

You'll also see a "+" notation: 100-17+ means 100 + 17.5/32 = 100.546875, where the "+" represents an additional half of a 32nd.

For Treasury futures, further precision is needed, so prices are quoted in 64ths — expressed as "100-170" meaning 100 + 17.0/32, or "100-172" meaning 100 + 17.25/32.

Why 32nds? It is a historical convention from when US government bonds first traded — fractional pricing in 32nds was the standard for US bond markets for over a century, and the market never fully converted to decimal pricing (unlike most other markets). Knowing this convention immediately signals you have actually engaged with Treasury markets, not just studied them from a textbook.

Yields, by contrast, are quoted as a decimal percentage — "4.32%" — or in basis points — "432bps." The day-to-day language on a rates desk mixes price (32nds) and yield (basis points) fluidly.

The major government bond markets beyond US Treasuries:

• UK Gilts: British government bonds, with conventional gilts (fixed coupon, semi-annual) and index-linked gilts (principal adjusts with RPI). Maturities run from 2 to 50+ years. The DMO (Debt Management Office) manages UK issuance. Gilts are the reference rate for sterling credit spreads. Currently one of the most watched markets globally given UK fiscal dynamics.
• German Bunds: The benchmark for European government bonds and the eurozone risk-free rate. Bunds (10yr and 30yr), Bobls (5yr), Schätze (2yr), and Bubills (bills). Germany's safe-haven status means Bunds often trade with negative yields in risk-off environments. The spread between peripheral European bonds (Italian BTPs, Spanish Bonos) and German Bunds is the key measure of eurozone credit stress.
• Japanese Government Bonds (JGBs): The largest government bond market by outstanding issuance. Maturities of 2, 5, 10, 15, 20, 30, and 40 years. JGBs have traded at extremely low yields for decades due to BoJ policy, making the current normalisation cycle one of the most watched stories in global rates.
• French OATs, Italian BTPs, Spanish Bonos: Eurozone peripheral government bonds that trade at a spread to German Bunds. The BTP-Bund spread is a key barometer of Italian sovereign risk and eurozone stability.

Duration is a measure of a bond's price sensitivity to changes in yield — specifically, the approximate percentage change in bond price for a 1% (100bps) change in yield. A bond with duration of 8 years will fall approximately 8% in price for a 1% rise in yield.

Duration is also the weighted average time to receive all cash flows from a bond. For a zero-coupon bond, duration equals its maturity exactly (all cash flow arrives at maturity). For a coupon-bearing bond, duration is less than maturity because some cash flows arrive earlier as coupons.

Three key rules: duration increases with maturity; duration decreases with coupon size (higher coupons bring more cash flow forward); duration decreases as yield rises (at higher yields, near-term cash flows are relatively more valuable).

Is the relationship linear? No — it is convex. Duration is only a linear approximation. For small yield changes, it works well. For larger yield changes, the actual price change differs from the duration-predicted change because of convexity — the curvature in the price-yield relationship. Positive convexity means bonds gain more in price when yields fall than they lose when yields rise by the same amount. This is a beneficial asymmetry for bondholders.

Modified duration is duration adjusted for yield, and gives the direct percentage price sensitivity. Macaulay duration is the weighted average time to cash flows in years.

DV01 (Dollar Value of a basis point, also called PV01 — Present Value of a basis point) measures how much the value of a bond or portfolio changes for a one basis point (0.01%) move in yield. It is the primary risk metric on a rates desk.

If a position has a DV01 of £10,000, every basis point move in yields generates a £10,000 gain or loss. A one basis point rise in yields costs £10,000; a one basis point fall gains £10,000.

Hedging using DV01: To make a position rate-neutral, you need to find an offsetting trade with an equal and opposite DV01. In practice, this is done using Treasury futures — the most liquid and cost-efficient hedging instrument.

Example: A corporate bond position has a DV01 of +£50,000 (it gains £50,000 per bp if yields fall). To hedge the rate risk, the trader shorts enough 10-year Treasury futures to generate a DV01 of −£50,000. The net position DV01 is zero — the position now only makes or loses money based on credit spread moves, not outright rate moves. The trader is "spread duration" only.

DV01 is additive: if you have ten bonds, you calculate each one's DV01 and sum them to get portfolio DV01. This makes portfolio risk management tractable even with hundreds of positions.

Convexity is the second-order measure of bond price sensitivity — it captures the curvature in the price-yield relationship that duration (a linear approximation) misses. Mathematically, convexity is the second derivative of the bond price with respect to yield.

For a conventional (non-callable) bond, convexity is positive, which means the bond gains more in price when yields fall than it loses when yields rise by the same amount. This is a beneficial asymmetry — bondholders benefit from yield volatility even if they have no directional view.

High convexity implies:

• Duration-based price estimates are increasingly inaccurate for large yield moves — the bond will outperform (positively) what duration alone predicts when yields fall significantly
• The bond is worth more to investors, all else equal, because of this asymmetric payoff — they will accept a slightly lower yield for a high-convexity bond
• Long-dated zero-coupon bonds have extremely high convexity (and therefore high price sensitivity to large rate moves)

Negative convexity is the dangerous case — callable bonds and MBS have negative convexity at certain price levels. When yields fall (prices rise), the embedded call option or prepayment option limits further price appreciation, while when yields rise, the bond falls in value just like any other bond. This is a one-sided disadvantage.

In swaps: DV01 is the sensitivity of the swap to rates (equivalent to duration); convexity is the sensitivity of DV01 to rates (the curvature term). High convexity means you benefit more from yield volatility.

Treasury futures are standardised, exchange-traded contracts that allow market participants to take views on future interest rate movements or hedge existing rate exposure — without needing to hold the underlying Treasury securities on-balance-sheet. They trade on the CME (Chicago Mercantile Exchange).

The six US Treasury futures contracts:

• TU — 2-year Treasury Note futures
• FV — 5-year Treasury Note futures
• TY — 10-year Treasury Note futures
• TN — 10-year Ultra Treasury Note futures (longer duration than standard TY)
• US — 30-year Treasury Bond futures
• AUL — 30-year Ultra Bond futures (longest duration contract)

Futures expire quarterly: March (H), June (M), September (U), December (Z). A contract named TYZ26 means the 10-year Treasury Note futures contract expiring in December 2026.

At expiry, the seller of a futures contract must deliver an actual Treasury security (not another futures contract) that qualifies as a "deliverable bond" under the contract specifications. The seller will choose the cheapest-to-deliver (CTD) bond — the eligible Treasury that is most economical to deliver after accounting for the futures price and a conversion factor.

Futures are preferred for hedging because: they are highly liquid, off-balance-sheet (saving capital), standardised, and can be rolled forward to maintain a hedge as contracts expire.

When a Treasury futures contract expires, the seller must deliver an actual Treasury security from a specified "basket" of eligible bonds. Because the bonds in the basket have different coupons and maturities, they have different prices. The futures contract uses conversion factors to standardise these differences, but the conversion factor system never perfectly neutralises the price differences across bonds.

The seller of the futures contract has the right to choose which eligible bond to deliver. Rationally, they will deliver the bond that is cheapest for them to purchase in the market and deliver against the futures contract — this is the cheapest-to-deliver (CTD) bond.

The CTD bond effectively determines what the futures contract is tracking. Because futures prices are anchored to the CTD bond's price (via arbitrage), a shift in which bond is CTD can cause the futures price to jump — a process called "CTD switching."

In practice, the CTD is typically a bond with a coupon close to the futures' notional coupon (6% in the US) and a maturity near the middle of the deliverable basket. When yields are far from 6%, the CTD shifts toward shorter- or longer-duration bonds in the basket, depending on the rate environment.

Understanding CTD is important because it determines the duration of the futures contract — which changes as the CTD switches — and affects the precision of DV01 hedges using futures.

The yield curve's shape reflects the market's expectations about future rates, inflation, and economic growth — embedded in the difference between short-term and long-term yields.

Steep yield curve (long yields much higher than short yields): Signals that the market expects rates to rise over time — either because economic growth is expected to accelerate, inflation expectations are elevated, or the central bank is expected to tighten policy in the future. A steep curve is typically associated with early expansion phases of an economic cycle. Investors demand a significant term premium for locking up capital long-term.

Flat yield curve (short and long yields converging): Signals uncertainty about the future direction of the economy or rates. Often occurs during transitions — for example, when the central bank has raised short rates aggressively but the market believes growth will slow enough to force future rate cuts.

Inverted yield curve (short yields higher than long yields): The classic recession predictor. It occurs when the market believes the central bank has tightened too far and rates will need to fall significantly in the future. The Fed's aggressive hiking cycle in 2022–23 produced the deepest 2s10s inversion on record. Every US recession in the modern era has been preceded by a yield curve inversion.

Current state (2026): The curve is in the process of disinversion — the 2s10s spread turned from deeply negative back toward positive as the Fed began cutting short-term rates. This bull steepening (short rates falling faster than long rates) is the textbook early-easing-cycle curve move.

A 2s5s flattener expresses the view that the spread between 2-year and 5-year Treasury yields will narrow — i.e. the curve between those two points will become flatter.

How to put it on: You sell (short) 2-year Treasury futures and buy 5-year Treasury futures. The trade must be duration-weighted — the DV01 on each leg must be equal and opposite so that a parallel shift in rates (all yields moving up or down together) does not generate P&L. You only profit if the spread between 2s and 5s changes.

Profit scenarios:

• 2-year yields rise more than 5-year yields (curve flattens from the front end)
• 5-year yields fall more than 2-year yields (curve flattens from the back of that segment)
• 2-year yields rise while 5-year yields fall (curve inverts between 2s and 5s)

When to put this on: Classic flattener rationale — the Fed is hiking aggressively (pushing up short-end yields) while growth expectations are softening (capping long-end yields). Pre-2022, many rates traders were running flatteners as the Fed hiked, expecting the hiking cycle to eventually slow growth. The trade worked as the curve went from steep to flat to inverted.

The opposite is a 2s5s steepener: buy 2-year futures, sell 5-year futures. Puts on in early easing cycles when the Fed is expected to cut (pulling down the front end) while growth expectations keep long-end yields sticky.

A 5s10s steepener expresses the view that the spread between 5-year and 10-year Treasury yields will widen — the curve between those points becomes steeper.

Positions: Buy (long) 5-year Treasury futures and sell (short) 10-year Treasury futures, duration-weighted so the DV01 on each leg is equal.

Profit scenarios:

• 10-year yields rise more than 5-year yields (the back end sells off more)
• 5-year yields fall more than 10-year yields (the front of this segment rallies more)
• 5-year yields fall while 10-year yields rise (the segment steepens sharply)

When it makes sense: A 5s10s steepener makes sense when you expect term premium to rise at the long end — due to fiscal concerns, increased Treasury supply, or inflation uncertainty at longer horizons — while the 5-year is more anchored by near-term rate expectations. In 2024–2026, rising term premium (driven by fiscal deficits and QT reducing demand) has been a rationale for steepener positions in the 5s10s and 10s30s segments.

Bear steepeners (both 5s and 10s selling off, with 10s selling off more) vs bull steepeners (both rallying, with 5s rallying more) have very different macro implications — bear steepeners signal inflation/fiscal concern; bull steepeners signal easing/recession.

The events that move rates markets, roughly in order of impact:

• FOMC meetings (8x per year): The rate decision, statement, dot plot (FOMC members' rate projections), and the Chair's press conference. The dot plot and "higher for longer" vs "easing" language often matters more than the rate decision itself.
• Non-Farm Payrolls (first Friday of every month, 8:30am EST): The headline jobs number, unemployment rate, and average hourly earnings. The most market-moving regular release because employment directly informs Fed policy timing.
• CPI/Core PCE releases (monthly): Inflation data. Core PCE is what the Fed officially targets. A surprise print can move the entire rates curve within seconds of release.
• Treasury auction results: Particularly 10yr and 30yr auctions. A "tailed" long bond auction can spike long-end yields 5–10bps within minutes as it signals weak demand for US fiscal deficits.
• Fed Chair speeches and Jackson Hole symposium: The August Jackson Hole speech is often the venue for signalling major policy shifts — Powell's 2022 Jackson Hole speech confirmed the aggressive hiking path.
• Quarter-end/year-end rebalancing: Institutional portfolio rebalancing creates predictable but sometimes large flows.

The single most important in the current environment: NFP, because the speed of Fed easing — and therefore the entire rates market direction — is primarily data-dependent on the labour market.

An interest rate swap (IRS) is an agreement between two parties to exchange interest rate cash flows over a specified period. The most common structure is the fixed-for-floating swap:

• Fixed-rate payer / floating-rate receiver: Pays a fixed interest rate and receives a floating rate (linked to SOFR in the US, SONIA in the UK). This party profits if floating rates rise above the fixed rate they are paying — they receive more than they pay.
• Fixed-rate receiver / floating-rate payer: Receives the fixed rate and pays floating. This party profits if floating rates fall below the fixed rate they receive.

Key terms of a swap: settlement date, term (duration — swaps can be 1 to 30 years), payment frequency (quarterly or semi-annual), notional amount (the principal on which interest is calculated — importantly, the notional is not exchanged, only the net interest payments are), fixed rate, and floating reference rate.

On day one the swap has zero net value — both legs are priced to be equivalent at initiation. Over time, as rates move, one side will be "in the money" and the other "out of the money," creating mark-to-market gains and losses.

Practical use: A company with floating-rate debt (e.g. a leveraged loan at SOFR + 300bps) worried about rising rates can enter a swap as the fixed-rate payer — converting its floating borrowing cost into a fixed rate, eliminating interest rate risk. Conversely, a pension fund receiving fixed pension income can receive fixed in a swap to hedge duration.

On day one, an interest rate swap has zero net present value. Both legs — the fixed and the floating — are priced so that their discounted cash flows are exactly equal at initiation. This is a fair trade: neither party is giving up value to the other at inception.

The fixed rate in the swap (the "par swap rate") is set at the level that makes the fixed leg worth exactly as much as the floating leg at current market rates. If you are paying 4.5% fixed and receiving SOFR (currently ~4.5%), the two legs net to approximately zero at initiation.

How value changes: After initiation, the swap's value fluctuates as interest rates move. If rates rise, the floating leg becomes more valuable (the floating-rate receiver will receive higher payments in future) while the fixed leg is less valuable. The fixed-rate receiver's position falls in value; the fixed-rate payer's position gains.

This mark-to-market change creates counterparty credit risk — if a party's position is deeply in the money, they face the risk that the counterparty defaults and cannot pay. This is why swaps are governed by ISDA agreements and are increasingly centrally cleared through CCPs (Central Counterparties like LCH) which require daily margin payments to cover mark-to-market changes.

Beyond the standard fixed-for-floating IRS, several variants exist:

• Basis swaps: Both legs are floating, but linked to different reference rates. For example, swapping USD SOFR against GBP SONIA — used to take a view on relative changes between two floating rate curves, or to manage cross-currency floating rate exposure.
• Cross-currency swaps: Exchange interest payments and principal in two different currencies. A UK company that has issued USD bonds but operates in sterling can use a cross-currency swap to convert its USD fixed obligation into a sterling obligation, eliminating the FX risk.
• Forward-starting swaps: The swap starts at a future date rather than immediately. Useful for hedging future anticipated debt issuances or locking in today's rates for a swap that won't begin until next year.
• Amortising swaps: The notional amount decreases over time (amortises), matching the paydown profile of an underlying amortising loan or mortgage portfolio.
• Callable/puttable swaps: The swap contains an embedded option allowing one counterparty to terminate the swap early under certain conditions.
• Overnight index swaps (OIS): Fixed for floating, but the floating leg is the daily compounded overnight rate (SOFR in USD, SONIA in GBP). OIS rates are considered near-risk-free and are used as the discounting rate for derivative cash flows.

The two key risk measures for interest rate swaps mirror the concepts used for bonds:

DV01 (PV01): The sensitivity of the swap's value to a one basis point parallel shift in interest rates. A £10,000 DV01 means each basis point move in rates changes the swap's value by £10,000. For a fixed-rate receiver, DV01 is positive (they benefit from falling rates); for a fixed-rate payer, DV01 is negative (they benefit from rising rates). Swaps can be sized precisely by DV01 to hedge a specific rate exposure.

Convexity: The sensitivity of DV01 to changes in rates. Positive convexity means the swap's DV01 increases as rates fall (the position becomes more sensitive to further rate moves). High convexity is beneficial because it means you gain more than duration alone predicts in large rate moves.

For swaps specifically, convexity is also important in the context of convexity adjustments in certain products — for example, Eurodollar futures (now SOFR futures) require convexity adjustments when converting from futures prices to swap rates because futures settlement is linear while swap cash flows are non-linear.

Together, DV01 and convexity give a complete first- and second-order picture of a swap's rate risk — analogous to delta and gamma in options pricing.

LIBOR (London InterBank Offered Rate) was the world's most important interest rate benchmark — the rate at which large banks claimed they could borrow from each other in the London interbank market. It was the floating rate reference for hundreds of trillions of dollars of loans, bonds, derivatives, and mortgages worldwide.

Why it was replaced: LIBOR was a survey-based rate — banks submitted what they claimed they could borrow at, not what actual transactions occurred at. This left it vulnerable to manipulation. In 2012, a major scandal emerged: traders at Barclays, UBS, Deutsche Bank, and others had colluded to submit artificially high or low LIBOR rates to benefit their own derivatives positions. Fines exceeded $10 billion collectively. The credibility of LIBOR was permanently damaged and regulators decided to replace it with rates based on actual transaction data.

Replacements:

• SOFR (Secured Overnight Financing Rate): The US replacement. Based on actual overnight repo transactions secured by US Treasuries — a deep, liquid market with genuine transaction data. Now the standard for new USD derivatives, loans, and bonds.
• SONIA (Sterling Overnight Index Average): The UK replacement. Based on actual overnight unsecured sterling transactions. Now the standard for new GBP interest rate swaps and loans.
• €STR (Euro Short-Term Rate): The eurozone replacement for EURIBOR in the overnight space.

The transition was completed by June 2023 for most LIBOR currencies. Legacy contracts referencing LIBOR have been transitioned using fallback provisions or synthetic LIBOR rates published by the FCA for a limited period.

The most important practical difference is that LIBOR was a term rate — you could look up 3-month LIBOR today and know in advance exactly what interest you would pay for the next 3 months. SOFR is fundamentally an overnight rate — compounded daily over a period — which means you only know the actual rate paid at the end of the interest period.

This creates a cash flow uncertainty that LIBOR did not have. For commercial loans, this was addressed by developing "Term SOFR" — a forward-looking rate derived from SOFR futures that mimics the advance-knowledge feature of LIBOR. Term SOFR (1-month, 3-month, 6-month) is now widely used in floating-rate corporate loans.

A second difference: LIBOR included a credit risk component (it reflected interbank borrowing costs, which include some bank credit risk above the risk-free rate). SOFR is based on secured repo transactions and is essentially risk-free. This means SOFR rates are slightly lower than LIBOR rates for the same tenor under normal conditions. To compensate for this, the ARRC (Alternative Reference Rates Committee) established standard credit spread adjustments when converting legacy LIBOR contracts to SOFR.

Operationally, the transition required enormous effort from legal, technology, and risk management teams at every bank — updating hundreds of thousands of contracts, systems, and risk models to handle the different compounding mechanics of SOFR.

Value-at-Risk (VaR) is a statistical risk measure that estimates the maximum potential loss on a portfolio over a given time horizon at a specified confidence level.

A 99% 1-day VaR of £5 million means: "We are 99% confident that the portfolio will not lose more than £5 million in a single day." Equivalently, there is a 1% probability of losing more than £5 million on any given day.

VaR is used on rates desks to:

• Set daily trading limits — traders cannot take positions that breach the desk's VaR limit
• Aggregate risk across a portfolio of positions in different instruments and maturities
• Report risk to senior management and risk committees in a single standardised number
• Satisfy regulatory capital requirements (Basel III uses VaR-based measures for market risk capital)

Limitations of VaR: VaR assumes a normal distribution of returns, which underestimates the probability of extreme losses (fat tails). It also says nothing about the size of losses beyond the threshold — a 99% VaR of £5m does not tell you whether the 1% scenario involves losing £6m or £60m. This is why banks also calculate Expected Shortfall (ES, also called CVaR) — the average loss in the worst 1% of scenarios — which is now the regulatory standard under Basel 3.1.

Basel III introduced two key liquidity ratios that directly affect how banks manage their rates portfolios and balance sheets:

LCR — Liquidity Coverage Ratio: Requires banks to hold sufficient high-quality liquid assets (HQLA) to cover net cash outflows over a 30-day stress period. The formula: LCR = HQLA / Net Cash Outflows over 30 days ≥ 100%.

HQLA are assets that can be converted to cash quickly in a stress — primarily government bonds (Treasuries, gilts), central bank reserves, and AAA-rated securities. This requirement means banks must hold large amounts of government bonds, creating a structural demand for Treasuries and gilts that supports their prices (and depresses their yields) relative to what pure supply/demand would dictate.

NSFR — Net Stable Funding Ratio: Ensures banks are not funding long-term assets with short-term liabilities — the "maturity transformation" risk that caused the 2008 liquidity crisis. The formula: NSFR = Available Stable Funding / Required Stable Funding ≥ 100%.

Available Stable Funding includes long-term liabilities and stable deposits. Required Stable Funding reflects the long-term funding need of assets. The NSFR penalises banks that fund long-dated assets (like 30-year mortgages or corporate bonds) with overnight or short-term wholesale funding.

Both ratios must exceed 100% at all times and are reported to regulators. Breaches require immediate remediation — typically selling assets, extending funding maturities, or raising capital.

Mark-to-market (MTM) accounting requires that securities in a bank's trading book are revalued at their current market price each day. If the market value of a bond falls, the bank recognises that loss immediately on its P&L — regardless of whether it has sold the bond or intends to. This creates real-time transparency about the value of trading positions.

MTM accounting can be destabilising in stressed markets. During the 2008 financial crisis, mortgage-backed securities fell sharply in price — even for bonds that were ultimately performing — forcing banks to take large mark-to-market losses that reduced their regulatory capital. This triggered margin calls, forced selling, and a feedback loop of falling prices and further losses.

Mark-to-model applies to illiquid securities where no reliable market price exists — for example, complex structured products, private credit loans, or bespoke derivatives that trade infrequently. Banks use proprietary pricing models to estimate fair value. Regulators scrutinise these models heavily because they can be used to overstate asset values (as happened with some CDO tranches in 2007–08).

The distinction is important for rates desks: benchmark Treasuries and liquid swaps are always mark-to-market (prices are transparent). Complex structured products, bespoke cross-currency swaps, or illiquid credit instruments may use mark-to-model components. The less transparent the price, the more subjectivity — and potential for valuation disputes — exists.

At any given time, there are:

• 24 outstanding 2-year Treasury Notes (monthly auctions × 2 years)
• 36 outstanding 3-year Notes
• 60 outstanding 5-year Notes
• 84 outstanding 7-year Notes
• 40 outstanding 10-year Notes (quarterly)
• 120 outstanding 30-year Bonds (quarterly)

That is approximately 364 individual securities — each with a different coupon, a different remaining maturity measured in exact days, and therefore a different price and yield. Unlike equities where every share of Apple is identical, bonds issued six months apart at different coupons have meaningfully different prices even at the same maturity point.

This heterogeneity is why rates trading is structurally more complex than equity trading. You cannot simply program a uniform pricing algorithm — each security requires individual pricing based on its specific cash flows. The complexity also means that a rates trader's book contains hundreds of positions across the maturity spectrum simultaneously, making risk management substantially more involved than a simpler asset class.

This is also why full automation of rates trading has been slower to develop than in cash equities — the on-the-run securities can be traded electronically, but the long tail of off-the-run securities still requires human judgment and voice trading for large blocks.

A basis swap is an interest rate swap where both legs are floating — rather than one fixed and one floating. The two floating rates are linked to different reference rates, indices, or currencies.

Common examples:

• 3-month SOFR vs 6-month SOFR: Used when a position has exposure to different tenor floating rates that don't move in perfect lockstep
• USD SOFR vs GBP SONIA: A cross-currency basis swap — exchanges USD floating-rate cash flows for GBP floating-rate cash flows, used by banks to manage multi-currency funding costs
• SOFR vs Fed Funds: Captures the spread between secured (repo-based) and unsecured overnight rates

When clients use them: A European bank that funds itself cheaply in euros but needs US dollar liquidity can enter a EUR/USD cross-currency basis swap — paying EURIBOR and receiving SOFR — to convert its euro funding into dollar funding at a known spread. This is common in cross-border bank treasury management. Similarly, a company that issues a floating-rate bond in one currency but operates in another can use basis swaps to manage the mismatch.

The cross-currency basis — the spread in basis swaps between different currency floating rates — is itself a closely watched market signal, reflecting relative funding demand and supply across currencies.

Interest rate swaps are governed by ISDA (International Swaps and Derivatives Association) documentation. The ISDA framework has three layers:

• ISDA Master Agreement: The overarching legal contract between two counterparties that governs all OTC derivatives transactions between them. It sets out default provisions, termination events, and netting rights (which allow offsetting gains and losses across all trades in the event of a default). Two parties typically have one master agreement that covers thousands of individual trades.
• Schedule: Annexes to the master agreement that customise the standard terms for the specific relationship — for example, specifying which currencies are covered, which entity is the primary obligor, and which governing law applies.
• Confirmation (Trade Confirmation): The document specific to each individual trade — setting out the exact terms: notional, fixed rate, floating rate reference, payment dates, maturity, and settlement details. Confirmations reference the master agreement rather than restating all its terms.

Post-GFC, the Dodd-Frank Act (US) and EMIR (Europe) mandated that standardised swaps be cleared through Central Counterparties (CCPs) like LCH SwapClear. Cleared swaps replace bilateral ISDA documentation with CCP rulebooks and margining requirements, reducing counterparty credit risk.

Bilateral (uncleared) swaps still exist for bespoke structures but require additional documentation under ISDA's 2016 VM CSA (Credit Support Annex for variation margin).

Interest rate swaps can be structured for virtually any maturity, but liquidity concentrates at standard "vanilla" tenors. The most actively traded maturities are 1, 2, 3, 5, 7, 10, 12, 15, 20, 25, and 30 years.

The most liquid points — where bid-ask spreads are tightest and execution is easiest — are 5-year and 10-year in most currencies. These correspond to the benchmark Treasury/gilt maturities and attract the most client hedging and speculative activity. The 30-year point is also liquid but thinner.

For currency-specific context:

• USD: The entire curve from 2yr to 30yr is liquid. The 10yr is the benchmark.
• GBP (Sterling SONIA swaps): Liquid from 2yr to 50yr, with longer tenors important for UK pension fund liability matching
• EUR: Liquid from 2yr to 30yr, 10yr is the benchmark
• JPY: Liquid from 2yr to 40yr given the BoJ's history of ultra-long-end buying

For bespoke maturities — say a 13-year swap — the desk will construct the price from interpolation between the 10-year and 15-year par swap rates, adding a small liquidity premium for the non-standard tenor.

Use case 1 — Corporate hedging of floating-rate debt: A company has a £500m leveraged loan at SONIA + 350bps. If rates rise, its interest payments increase. The company enters a pay-fixed swap: it pays a fixed rate (say 4.5%) and receives SONIA. The floating SONIA on the swap offsets the floating SONIA on the loan, converting its total borrowing cost to a fixed 4.5% + 3.5% = 8%. Rate risk is eliminated — the company now has certainty over its interest expense regardless of rate moves.

Use case 2 — Pension fund duration extension: A UK pension fund has long-dated liabilities (pension payments 20–40 years into the future) but holds mostly short-duration assets. It is exposed to rates falling, which increases the present value of its liabilities without corresponding gains on assets. By entering receive-fixed swaps at 20–30 year tenors, it extends its asset duration to match liabilities — eliminating the mismatch without needing to buy expensive long-dated gilts in size.

Use case 3 — Macro directional trade: A macro hedge fund believes the Federal Reserve will cut rates more aggressively than markets are pricing. It enters a receive-fixed, pay-floating USD swap at the 5-year tenor. If rates fall as the fund expects, the fixed rate it receives becomes more valuable than the declining floating rate it pays — the swap gains in mark-to-market value, which the fund can realise by terminating the swap or entering an offsetting position.

The tail in a Treasury auction is the difference between the yield at which the auction clears (the stop-out rate) and the yield at which the "when-issued" (WI) market was trading just before the auction closed.

A tailing auction means the auction cleared at a higher yield than the WI market expected — in other words, the Treasury had to offer a higher yield to attract enough buyers to fill the auction. This signals weak demand. Tails of even 1–2 basis points can move the broader market — if the 30-year auction tails by 3bps, long-end yields often spike 5–8bps in the immediate aftermath as the market concludes that fiscal demand for Treasuries is less robust than expected.

A "stop-through" is the opposite — the auction cleared below the WI yield, meaning demand was so strong that bidders accepted lower yields. This sends rates lower.

Bid-to-cover ratio is total bids submitted (by notional value) divided by the amount the Treasury was selling. A bid-to-cover of 2.4x means $2.40 of bids were received for every $1.00 of Treasuries sold. Historical averages are around 2.3–2.8x depending on tenor. Below 2.0x is considered weak; above 3.0x is strong.

Both metrics are released simultaneously with the auction results and markets react within seconds. Large institutional investors and rates traders have entire workflows dedicated to pre-auction positioning and post-auction analysis.

Futures and forwards both fix a price today for a transaction occurring at a future date, but differ in several important ways:

Futures:

• Standardised contracts traded on organised exchanges (CME for Treasuries, ICE for gilts)
• Centrally cleared — variation margin paid daily to a CCP, eliminating counterparty risk
• Standard contract sizes, maturities, and delivery specifications — no customisation
• Highly liquid — tight bid-ask spreads, large daily volumes
• Off-balance-sheet (only margin posted as collateral, not the full notional)

Forwards:

• Customised bilateral contracts negotiated directly between two parties (OTC)
• No daily margin settlement — the profit or loss settles at expiry (though ISDA agreements increasingly require variation margin even for forwards)
• Fully customisable maturity, notional, underlying — can be tailored to exact client needs
• Carry counterparty credit risk (the party in-the-money is exposed to the counterparty defaulting)
• Less liquid — harder to unwind before maturity without negotiating with the original counterparty

When to use each: Futures for liquid, standardised hedges and speculative positions where exchange liquidity and margin efficiency matter. Forwards for bespoke situations — for example, a corporate hedging an exact dollar amount of exposure at an exact future date that doesn't align with a standard futures expiry.

The overnight interest rate is the primary tool — in the US, the Federal Funds Rate; in the UK, the Bank Rate set by the MPC. By setting this rate, the central bank influences the entire yield curve and credit conditions across the economy. Rate changes are transmitted through banks to mortgages, corporate loans, and deposits.

Additional monetary policy tools:

• Reserve requirements: The minimum proportion of deposits banks must hold as reserves at the central bank. Increasing reserve requirements reduces the amount banks can lend, tightening credit. This tool is less commonly used in developed markets today.
• Open market operations (OMO): The central bank buys or sells government securities in the secondary market to manage the level of reserves in the banking system and keep the overnight rate near its target. Routine OMOs are the operational mechanism for maintaining the rate target.
• Quantitative easing (QE) / tightening (QT): Large-scale asset purchases (QE) or balance sheet rundown (QT) that go beyond overnight rate control to directly influence longer-dated bond yields and broader financial conditions. The Fed purchased over $4 trillion of Treasuries and MBS during COVID. QT (allowing the balance sheet to shrink by not reinvesting maturing securities) has been ongoing since 2022.
• Forward guidance: Communication about the future path of rates. Increasingly important as a tool in itself — "higher for longer" guidance can affect long-end yields even without any actual rate change.
• Special facilities: Emergency lending programs like the Bank Term Funding Program (BTFP) deployed in the 2023 US regional banking stress, or the Fed's CPFF (Commercial Paper Funding Facility) during COVID.

Basel III requires banks to hold minimum capital against their risk-weighted assets (RWA). For trading desks, the key requirements:

Market risk capital: Banks must hold capital against potential mark-to-market losses on their trading books. Under the Fundamental Review of the Trading Book (FRTB — the latest Basel standard), this uses Expected Shortfall (ES) rather than VaR as the primary risk measure, and applies more stringent requirements for illiquid positions.

Minimum 8% capital ratio: Banks must maintain Tier 1 capital of at least 6% and total capital (Tier 1 + Tier 2) of at least 8% of RWA. For Global Systemically Important Banks (GSIBs — which includes Barclays, HSBC, Deutsche Bank, etc.), an additional buffer of 1–3.5% applies.

Practical impact on rates desks: Capital requirements make holding risky or illiquid positions more expensive. A trader who holds a large position in off-the-run Treasuries or illiquid structured products needs to set aside more capital than one holding on-the-run Treasuries. This capital cost is reflected in bid-ask spreads — desks need to earn enough spread to justify the capital deployed. The higher the regulatory capital requirement, the wider the spread the desk needs to charge clients to maintain an acceptable Return on Equity (ROE).

This capital constraint is one reason banks have pulled back from market-making in illiquid products since the GFC — the capital cost can make certain activities uneconomic, reducing market liquidity in those segments.

Mark-to-model is used when a reliable market price cannot be observed — i.e., for illiquid securities where transaction data is sparse or non-existent. Instead of using a market price, the bank uses a proprietary pricing model to estimate fair value.

Securities commonly valued by model:

• Complex structured products (bespoke CDO tranches, synthetic CLO equity)
• Private credit loans without secondary market trading
• Exotic or long-dated derivatives with no observable market
• Real assets on bank balance sheets

The danger: models require assumptions — about default rates, correlation, interest rate paths — that can be wrong or manipulated. During the 2007–08 crisis, banks were marking AAA CDO tranches at or near par using models, even as the underlying mortgages were defaulting, because the market had effectively seized and there were no transaction prices. When these positions were ultimately liquidated, losses far exceeded the model-derived valuations.

Regulators now require extensive model validation processes, independent model risk management functions, and conservative overlays for mark-to-model positions. Under FRTB, positions that cannot be marked to market are subject to higher capital requirements — creating a financial incentive for banks to trade in liquid instruments where market prices are observable.

The Fed has deployed a range of special lending facilities during crises, beyond its standard open market operations:

• Commercial Paper Funding Facility (CPFF): Deployed in 2008 and 2020. Allowed the Fed to directly purchase commercial paper from issuers, backstopping the short-term corporate funding market when it seized up. Without this, companies unable to roll maturing commercial paper would have faced sudden liquidity crises.
• Term Asset-Backed Securities Loan Facility (TALF): 2008 and 2020. Provided loans to investors to finance purchases of ABS backed by auto loans, credit card loans, student loans, and small business loans — supporting the consumer credit market.
• Primary Dealer Credit Facility (PDCF): Allowed primary dealers (the banks required to participate in Treasury auctions) to borrow directly from the Fed against a broad range of collateral. Used in 2008 and again in March 2020 during the COVID liquidity shock.
• Bank Term Funding Program (BTFP): Introduced in March 2023 after the Silicon Valley Bank failure. Allowed banks to borrow against their Treasury and agency securities at par — crucial because many banks were sitting on large unrealised losses on their bond portfolios due to the 2022 rate hikes. The BTFP was wound down in 2024.

The common thread: these facilities backstop specific market segments that are critical to the functioning of credit markets, preventing liquidity crises from becoming solvency crises.

A repurchase agreement (repo) is a short-term collateralised borrowing transaction. The borrower sells a security (typically a Treasury) to the lender and simultaneously agrees to repurchase it at a slightly higher price on a specified future date. The price difference represents the interest — the "repo rate."

Economically, repo is equivalent to a secured overnight loan: the borrower receives cash, the lender receives the security as collateral. If the borrower defaults, the lender keeps the security. The collateral overcollateralisation (the repo market's "haircut") provides an additional buffer.

Why it is systemically important:

• The repo market is the primary funding mechanism for the US financial system — primary dealers fund their Treasury inventories overnight in repo, using the Treasuries they hold as collateral to borrow cash
• Repo rates are closely linked to the Fed Funds rate — the repo market is one of the transmission mechanisms through which Fed policy affects the broader economy
• Disruptions in repo can have immediate systemic effects: in September 2019, unexpected stress in the repo market caused overnight rates to spike from 2.2% to 10%+ in a single day, forcing the Fed to intervene with billions in temporary repos
• The reverse repo facility (RRP) — where banks and money market funds park cash with the Fed overnight in exchange for Treasuries — is a key tool for draining excess liquidity from the system during QT

A reverse repo is the mirror image of a repo — from the perspective of the party providing cash (lending it). The cash lender buys a security and agrees to sell it back at a higher price, effectively earning the repo rate on their cash. The lender is running a "reverse" repo; the borrower is running a "repo."

The Fed's Overnight Reverse Repo Facility (ON RRP) allows eligible counterparties (primarily money market funds and primary dealers) to deposit cash with the Fed overnight, receiving Treasuries as collateral and earning the ON RRP rate. It functions as a floor for short-term rates — institutions with excess cash will lend to the Fed rather than accepting lower rates elsewhere.

Why it became important post-COVID: The Fed's massive QE purchases injected enormous amounts of cash into the financial system. Money market funds and banks flooded with cash had nowhere to safely invest it at competitive rates. ON RRP balances surged from near-zero to over $2.5 trillion at the peak in 2023, effectively sterilising excess liquidity that the financial system couldn't usefully deploy.

As the Fed raised rates and began QT, ON RRP balances declined as money market funds found better yields in short-term Treasuries and elsewhere. Monitoring ON RRP balances has become a key indicator of excess liquidity in the financial system and the Fed's success in draining reserves without destabilising money markets.

The term premium is the extra compensation investors demand for holding a long-dated bond rather than rolling over a series of short-term bonds. It reflects uncertainty about future inflation, rates, and the economy — the longer you lock up capital, the more compensation you need for that uncertainty.

The term premium is not directly observable — it must be estimated using models (the most widely cited is the Adrian, Crump, and Moench ACM model from the New York Fed). When estimated, it can be negative (investors accept less yield on long bonds than short-term rates would imply) or positive.

Why it rose in 2024–2026:

• Fiscal concerns: US fiscal deficits have been running at 6–7% of GDP — unusual outside of wartime or recession. The Treasury must issue unprecedented volumes of debt. The market increasingly demands higher yields to absorb this supply.
• QT reducing demand: The Fed is no longer buying Treasuries (QT ended the QE-era suppression of term premium). With the Fed stepping back, private investors must absorb all new supply at market prices.
• Inflation uncertainty: Even after CPI normalised, uncertainty about whether inflation could re-accelerate adds a risk premium to long-dated bonds.
• Global demand reduction: China and Japan (historically large buyers of US Treasuries) have reduced their holdings. Foreign demand is less reliable than during the 2000–2020 era.

Rising term premium is one reason long-end yields have stayed elevated even as the Fed has cut short-term rates — it is a structural headwind for a big bond rally.

A spot-starting swap begins immediately (or on the standard 2-business-day settlement date). Both parties start exchanging fixed and floating cash flows from the settlement date. This is the standard structure for most IRS.

A forward-starting swap begins at a future specified date — for example, a "1-year into 5-year" swap starts in one year's time and runs for five years from that point. The fixed rate is agreed today, but cash flow exchange does not begin until the start date.

Why clients use forward-starting swaps:

• A company planning to issue a bond or take out a loan in six months can enter a forward-starting swap today to lock in the current swap rate for the hedge that will begin at issuance — eliminating the risk that rates move adversely before the loan starts
• A pension fund anticipating a large liability cashflow beginning in 3 years can begin managing duration exposure for that future liability today
• A rates trader who believes rates will be higher in one year's time (making it advantageous to pay fixed at today's rate for a swap starting then) can express this view efficiently with a forward-starting swap

Forward-starting swaps create some FX risk in cross-currency contexts — because one party is committing to an FX exchange at a future date, some residual FX exposure exists and may need hedging separately.