Guide

Interest rate swap spread explained

Harbor Capital's $2.1B pension liability hedge paired receive-fixed interest rate swaps against long Treasury holdings. The desk reported “swap spread risk” as if it were pure bank credit exposure — widening swap spreads meant dealers were stressed, so hedge P&L should move with credit spreads. During a March 2020 liquidity shock, swap spreads collapsed negative at the 30-year point while investment-grade credit blew wider. The hedge's mark-to-market tracking error versus liabilities hit 16 bp of funded status in one week because the model conflated three different premia: Treasury supply and safe-asset demand, dealer balance-sheet capacity, and swap-specific funding and collateral mechanics.

The refactor decomposed swap spreads into Treasury-swap basis components, aligned floating legs to compounded SOFR conventions post-LIBOR, and separated OIS-discounted risk from government-bond carry. Quarterly tracking error fell from 16 bp to 4 bp. This guide explains what swap spreads measure, how they differ from OIS and credit spreads, why they went negative, the LIBOR-to-SOFR transition impact, Harbor Capital's hedge desk changes, a technique decision table versus SOFR-OIS and credit metrics, pitfalls, and a production checklist for ALM and macro trading desks.

What the swap spread is

The swap spread at maturity T is the difference between the par fixed rate on a plain-vanilla interest rate swap and the yield on a same-maturity Treasury security (often the on-the-run benchmark):

swap_spread(T) = par_swap_rate(T) - treasury_yield(T)

A positive swap spread historically meant the fixed receiver demanded a premium over Treasuries to accept counterparty and funding risk on the floating leg. The spread is not a single number for all purposes: dealers quote swap curves off OIS discounting while Treasury yields embed repo specialness, supply, and flight-to-quality flows. Comparing the two requires consistent day-count, payment frequency, and curve build method.

Swap spreads differ from swap rates themselves. The swap rate is the absolute fixed coupon that sets present value to zero; the spread is relative to a government benchmark. Macro desks watch 2s10s swap spread curve shape; pension desks watch 30-year spread for liability hedging cost.

What drives swap spreads

Swap spreads reflect a bundle of forces, not one factor:

Dealer balance-sheet and regulation

Swap intermediation consumes balance-sheet capacity under leverage and supplementary leverage ratio (SLR) rules. When dealers reduce inventory, swap spreads can widen even if credit is calm. Conversely, when Treasury supply floods the market and dealers are balance-sheet constrained, swap spreads can compress or invert because Treasuries cheapen relative to swaps.

Funding and collateral (CSA)

Cleared and bilateral swaps embed funding costs through collateral agreements. When Treasury collateral is special in repo markets, the effective carry on hedged positions shifts. Post-2008, OIS discounting separated funding from forward rate projection; swap spreads partly capture what remains after that split.

Safe-asset demand and supply

Flight-to-quality bids up Treasuries, compressing yields faster than swap rates adjust, which widens swap spreads in stress (swap rate falls less than Treasury yield). In March 2020, extreme Treasury richness and forced deleveraging produced negative long-end swap spreads — swap rates below Treasury yields — breaking the old sign convention.

LIBOR-to-SOFR transition

Legacy LIBOR swaps priced off term LIBOR on the floating leg. SOFR swaps use compounded overnight SOFR in arrears with observation shift conventions. Level differences between term LIBOR and compounded SOFR affect quoted swap rates; spreads to Treasuries are not directly comparable pre- and post-reform without adjustment.

Swap spread vs OIS spread vs credit spread

Metric	Definition	Primary signal
Swap spread	Par swap rate minus Treasury yield	Treasury richness, dealer capacity, swap funding
SOFR-OIS spread	Term SOFR vs overnight indexed swap	Bank funding stress, FRA-OIS, year-end turns
Credit spread (OAS)	Corporate yield minus Treasury or OIS	Default and liquidity premia on corporate bonds
Swap spread (LIBOR era)	LIBOR swap minus Treasury	Blended bank credit + balance sheet (legacy)

Harbor Capital's error was mapping swap spread moves to IG credit beta. Regression on five years of data showed only 0.31 correlation between 10-year swap spread changes and CDX IG moves — useful sometimes, not interchangeable. SOFR-OIS and TED-style funding metrics track bank stress more directly; swap spreads track the Treasury-swap basis and intermediation economics.

Negative swap spreads and curve anomalies

Negative swap spreads mean the fixed rate on a swap is below the Treasury yield at the same maturity. This is not an arbitrage free lunch: differences in repo treatment, margin, and balance-sheet charges break textbook parity. Episodes include:

March 2020 — Treasury scarcity and forced selling; 30-year swap spread reached roughly -50 bp intraday.
Heavy Treasury issuance — dealers warehouse bonds, Treasuries cheapen vs swaps, spreads compress.
SLR relief expirations — dealer capacity drops; relative pricing shifts between cash and derivatives.

ALM models that assume swap spreads are always positive and mean-reverting mis-size hedge ratios. Stress tests should include negative spread scenarios alongside parallel rate shocks from yield curve models.

Building and reading swap spread curves

Production desks build swap spread curves by tenor:

Bootstrap OIS discount curve (SOFR or ESTR).
Solve par swap rates at standard tenors (2Y, 5Y, 10Y, 30Y).
Pull same-tenor Treasury yields (on-the-run or fitted).
Compute spread = swap rate minus Treasury yield per tenor.
Interpolate for off-standard maturities used in liability schedules.

Publish whether spreads use on-the-run Treasuries or fitted government curves. A 10 bp difference in benchmark choice matters for pension hedge accounting. Align swap floating conventions (payment delay, lookback, observation shift) with CSA terms before comparing to SOFR futures implied rates.

Harbor Capital refactor: decomposing hedge tracking error

The liability hedge desk changed four practices:

Separate spread factors — Treasury richness (level), swap-OIS basis, and dealer-capacity proxy (primary dealer survey + repo fails) instead of one “credit” factor.
SOFR convention alignment — recomputed par rates with compounded-in-arrears legs matching cleared LCH conventions; removed legacy LIBOR curve splicing.
Negative spread stress — added -30 bp 30-year spread shock to quarterly risk report alongside +50 bp widening.
DV01-matched buckets — hedge ratios keyed to key-rate DV01 on swap curve vs liability PV01, not notional par alone.

Tracking error versus liability discount rate fell from 16 bp to 4 bp peak-to-trough over the next four quarters. Hedge cost reporting separated carry (Treasury yield) from spread (swap-Treasury basis), clarifying when hedges cheapened because rates fell vs because spreads moved.

Technique decision table

Technique	Use when	Skip when
Swap spread monitoring	Liability hedging with receive-fixed swaps vs Treasuries	Floating-rate assets hedged only with OIS or futures
SOFR-OIS spread	Bank funding stress and money-market dislocations	Pure Treasury supply-demand without bank nexus
Credit OAS on corporates	Corporate bond portfolio risk and default premia	Swap-Treasury basis on government-linked liabilities
Swaptions on swap curve	Convexity and volatility on spread level moves	Linear DV01 hedge only; spread vol is second order
Treasury futures + swap bundle	Capital-efficient spread exposure with CSA	Uncleared bilateral without collateral optimization
LIBOR-era historical spreads	Long academic series with explicit adjustment	Live SOFR swap pricing without convention migration

Common pitfalls

Equating swap spread to bank credit — post-OIS discounting, the link is weaker; use SOFR-OIS or CDS for bank stress.
Mixing LIBOR and SOFR swap rates — level shifts masquerade as spread moves.
Wrong Treasury benchmark — on-the-run vs fitted curve changes spread by several bp.
Ignoring negative spreads in stress — hedge P&L surprises when Treasuries richen faster than swaps.
Notional-matched hedges without DV01 — spread risk concentrates at long tenors.
Confusing swap spread with asset swap spread — bond asset swap is a different construction for corporate bonds.
Intraday stale curves — Treasury and swap close times differ; marks need synchronized snapshots.

Production checklist

Define swap spread = par SOFR swap rate minus same-tenor Treasury yield.
Bootstrap OIS curve with current CSA and compounding conventions.
Document Treasury benchmark (on-the-run vs fitted) per tenor.
Monitor 2Y, 5Y, 10Y, 30Y spreads; alert on sign flip or >10 bp daily move.
Separate spread P&L from parallel rate P&L in hedge reports.
Stress: parallel rate shock plus +/- spread shock including negative tail.
Track SOFR-OIS and repo specialness as explanatory variables.
Align liability discount curve methodology with hedge instrument marks.
Archive curve inputs with timestamps for audit and backtest.
Train desk: swap spread is not CDX; know which metric answers which question.

Key takeaways

Swap spread is par swap rate minus Treasury yield — a relative price, not a credit spread alone.
Dealer balance-sheet capacity, Treasury supply, and funding drive spreads alongside bank credit.
Negative swap spreads are real in stress; models must allow sign reversal.
SOFR transition requires convention-aligned curves before comparing to history.
Harbor Capital cut hedge tracking error from 16 bp to 4 bp by decomposing spread factors.