Credit Trading Desk: How IG Spreads, CS01 and the CDS Basis Work

Ask where a bank trades investment grade corporate bonds, IG for short, and you will rarely hear "the investment grade desk". The market calls it the credit desk, or just credit, while the seats beside it, high yield, distressed debt and CLOs, each go by a name of their own.

What the desk trades is one number. A corporate bond's price moves with the Treasury curve underneath it, and the credit desk's business is the gap on top: on 10 September 2026 the broad US investment grade index paid 80 basis points over Treasuries.

What follows is the desk as it runs: who it trades with, how it earns, how screens and a public tape changed it, and three worked examples with the arithmetic shown: a corporate bond credit spread, a CS01 on a $10 million ticket and the bond versus CDS basis.

What the credit desk actually trades

The core product is the US investment grade corporate bond: senior unsecured debt, mostly bullets that repay in one piece at maturity, with some callable issues among them. A client can trade one bond at a time or, more and more often, a portfolio trade, a basket of bonds priced and executed as a single ticket.

Around the bonds sit the tools the desk hedges with or expresses a view through:

• Single-name credit default swaps, which buy or sell protection on one issuer's debt.
• Index credit default swaps, CDX for North American names, which cover a basket of issuers in one contract.
• Credit futures, relaunched in 2024 by two of the big exchange groups, with far more systematic credit trading firms around to use them than at the last attempt after Dodd-Frank.

The credit floor splits from here into its own seats. Bonds rated below investment grade belong to the high yield desk. Debt of companies in or near default goes to distressed debt trading, and leveraged loans pooled into tranches go to CLO trading. The analysis underneath all of them, leverage, coverage and recovery, is tested in credit analysis interview questions.

Who the clients are, and what they want

Two kinds of money use the desk, and they want different things. Real money, meaning asset managers, insurers and pension funds, mostly buys and holds or tracks an index. Hedge funds trade relative value: one bond against another, or a bond against its own credit default swap, the basis trade worked through further down.

• Asset managers, trading to track or beat an investment grade index.
• Insurers and pension funds, largely buy-and-hold owners of what they buy.
• Hedge funds, trading relative value, including the bond versus CDS basis.

Demand is rarely what limits the desk. The limit is how much inventory a dealer can carry on its balance sheet until the other side of a trade turns up, and every bond held overnight uses some of it.

How the credit desk makes money

The desk earns the way every market maker on the floor does. It quotes a bid and an offer, buys from one client, sells to another and carries the inventory in between. The Volcker Rule bans banks from short-term proprietary trading but exempts market making, and its text spells out the job: a desk must "routinely stand ready to purchase and sell," with inventory kept within "the reasonably expected near-term demands of customers."

Flow and risk share the desk. Flow is the stream of client orders, priced fast and turned over quickly. Risk is inventory the desk keeps, because a client left it behind or because the desk wants the position, and it is the part the market-making exemption asks a desk to justify as serving clients rather than betting for the bank.

Where the money comes from is not where the volume is. Across full-year 2024, 48 percent of US investment grade corporate bond volume traded electronically, yet electronic execution earned dealers only about 26 percent of their investment grade trading revenue. High yield showed the same gap: 33 percent of volume, 10 percent of revenue.

Investment grade, share of volume

48%

Investment grade, share of dealer revenue

26%

High yield, share of volume

33%

High yield, share of dealer revenue

10%

Full-year 2024 averages; revenue means dealer revenue from US corporate bond trading.

The research firm that published both numbers drew the conclusion itself: with nearly half of investment grade volume on screens, "the value of the voice business to dealers is pretty clear." A screen prices the trades anyone could price. The revenue sits in the trades where a dealer has to decide how much risk to take, and those still run through a trader.

The market by the numbers

US corporate bonds outstanding reached $11.7 trillion in the first quarter of 2026, up 3.0 percent on the year, the second-largest US fixed income asset class after Treasuries. Issuance has run hot: $1,899.8 billion through August 2026, 29.8 percent more than in the same months of 2025.

Trading has grown too. Average daily volume across US corporate bonds hit a record $46.2 billion in 2024, up 21 percent and the third straight year of growth. The growth has been an investment grade story: daily volume there rose from $22 billion in 2019 to $36 billion in 2024, while high yield stayed roughly flat at about $10 billion a day.

$11.7tn

US corporate bonds outstanding

first quarter of 2026

$1.9tn

issued through August 2026

up 29.8% on the year

$46.2bn

traded per day across 2024

a record, up 21%

$36bn

investment grade traded per day, 2024

up from $22bn in 2019

2008, the Volcker Rule and a thinner balance sheet

The broad US investment grade index's spread peaked at 656 basis points in December 2008. On 10 September 2026 it stood at 80, roughly an eighth of that peak, and most of how the desk runs was rebuilt in the years between.

The Volcker Rule came first. Passed as part of Dodd-Frank in 2010, its final rules took effect on April 1, 2014, and the Federal Reserve extended the conformance period to July 21, 2015. From then on, every bond a bank desk held had to be explainable as market making, underwriting or hedging, not as the bank's own bet.

Dealers held less. Corporate bond inventories plunged during the crisis and stagnated afterwards, even as issuance and debt outstanding set records: $1.4 trillion of new bonds and $7.8 trillion outstanding in 2014 alone.

Screens, portfolio trades and all-to-all trading

Electronic trading has climbed in every year of the full-year series: 34 percent of US investment grade volume in 2021, 41 percent in 2022, 45 percent in 2023 and 48 percent in 2024. Single months ran higher than the average, 49 percent in May 2024 and 50 percent that September, so a monthly high and the full-year figure for the same year sit a point or two apart.

2021

34%

2022

41%

2023

45%

2024

48%

Full-year averages built on FINRA TRACE volume. Single months ran higher: 49 percent in May 2024 and 50 percent in September.

High yield moved more slowly: 25 percent in 2021, 33 percent by 2023, and 33 percent again across 2024, with 32 percent in May of that year.

Two ways of trading grew up beside the screen request and the phone call:

• A portfolio trade prices a whole basket of bonds as one ticket, so a client can move a list in a single trade rather than line by line.
• All-to-all trading opens a request to every participant on a venue, other investors included, so an asset manager selling a bond can meet another asset manager rather than only a dealer.

Two venues dominate the screens. MarketAxess handled 37 percent of US corporate bond electronic trading in 2024 and Tradeweb 34 percent, with Bloomberg at 11 percent and a handful of smaller platforms sharing the rest.

TRACE and the 15-minute tape

Every US corporate bond trade ends up on a public tape. FINRA introduced TRACE, its Trade Reporting and Compliance Engine, in July 2002, and dealers must report each corporate bond trade to it within 15 minutes of execution.

The detail a trader watches is size. The tape shows an investment grade trade's exact size only up to $5 million, and anything larger prints as 5MM+. High yield caps lower, at $1 million, shown as 1MM+:

• An investment grade trade of $3 million prints with its price and its exact size.
• An investment grade trade of $10 million prints with its price and a size of 5MM+.
• A high yield trade of $3 million prints with its price and a size of 1MM+.

So a $10 million block, like the ticket worked through below, shows the market the level it traded at but not how much the dealer now owns.

New-issue days and the syndicate desk

A new deal is the flow desk's future inventory. Practitioners describe the syndicate desk as the seat that runs the deal itself: it gathers investors' orders, agrees the final spread with the issuer and allocates the bonds, work the Volcker Rule exempts as underwriting. Once the deal is done, the bonds trade like any other, and the flow desk makes the market in them. No bank publishes its own step-by-step version of that process.

New issuance moves the whole desk. The first week of January 2025 was the busiest first week of any year for US corporate bond issuance, with $83 billion of new bonds, and secondary trading that week ran at $56 billion a day, nearly 25 percent above the first week of 2024.

The corporate bond credit spread, worked two ways

The textbook definition is simple: a bond's yield minus the yield of a Treasury of matching maturity. In practice two different numbers both get called the credit spread, and they need not agree.

Take the single-A slice of the US investment grade index on 10 September 2026. Its effective yield was 5.54 percent, and the 10-year Treasury par yield that day was 4.95 percent. Subtract one from the other and the nominal spread, or G-spread, is 0.59 percentage points: 59 basis points.

The index's own published option-adjusted spread for the same day was 68 basis points. It is built differently: each bond's cash flows are priced against the whole Treasury curve, not one tenor, and the results are weighted by value.

The 9 basis point gap is the benchmark, not a mistake. A broad index blends many maturities, and its blended maturity is not exactly ten years. Measured against the 7-year Treasury at 4.84 percent, the same 5.54 percent gives 70 basis points, within 2 of the published figure: an index whose blended maturity sits closer to seven years than ten. Picking the right benchmark matters as much as the subtraction.

Spreads also widen as credit quality falls. On the same day the broad investment grade index paid 80 basis points and its BBB slice 98.

Single-A

68bp

All investment grade

80bp

BBB

98bp

The broad index and two of its rating slices, same day.

Spread duration and CS01 on a $10 million ticket

Spread duration measures how far a bond's price moves when its credit spread changes. For a plain fixed-rate bond it is the same number as modified duration, because the whole bond is exposed to the spread. CS01 turns it into dollars: the change in a position's value for a one basis point move in spread, the credit desk's version of the DV01 a rates desk uses.

The duration arithmetic underneath is worked from scratch in bond math interview questions. The bond here is illustrative, from a single-A rated utility: a 10-year bullet with a 5.50 percent coupon paid semiannually, priced to yield 5.63 percent. That yield is the 10-year Treasury's 4.95 percent plus the single-A index's 68 basis points, both on 10 September 2026. The ticket is $10 million face value.

1. Price. Twenty half-yearly coupons of $27.50 per $1,000, discounted at 2.815 percent a period, are worth $416.22, and the $1,000 principal discounted the same way is worth $573.95. The bond is worth $990.16 per $1,000, just under par because the coupon sits below the yield, so the ticket is worth about $9.90 million.
2. Duration by formula. Macaulay duration is 15.62 half-year periods, or 7.81 years. Divided by 1.02815, modified duration, and so spread duration, is 7.60.
3. Duration by reprice. One basis point wider, at 5.64 percent, the price falls to $989.41; one basis point tighter, at 5.62 percent, it rises to $990.91. The price change across that two basis point range, divided by the price and by 0.0002, gives 7.60 again, working from unrounded prices.
4. CS01. Spread duration times market value times 0.0001: 7.60 times $9.90 million times 0.0001 is about $7,520 per basis point.

Scale it up to a real move. If this bond's spread widened from the single-A level to the BBB index level of the same day, 98 basis points, the 30 basis point move would cost roughly 30 times $7,520, or $225,600. Repricing the bond in full at 5.93 percent gives a loss of about $222,500, a little smaller, which is convexity again. The bond calculator reproduces every step from the same inputs.

The bond versus CDS basis, worked through

A bond and a credit default swap on the same company carry the same credit risk, so their spreads should line up. The gap between them is the basis. The New York Fed defines it as the difference between the running spread on the CDS and "the theoretical (par-equivalent) CDS spread implied by the yield on the cash bond."

When the basis is negative, the bond pays more for the risk than protection costs. In the paper's words, the return "is earned by purchasing the cash bond and purchasing protection in the CDS market." Its worked example has a dealer put that trade on in size: a $10 million bond purchase financed in repo with a 5 percent haircut, the haircut borrowed short-term and unsecured at 0.5 percent, and the repo at 0.48 percent.

1. Haircut. 5 percent of $10 million is $500,000, borrowed unsecured.
2. Repo. The other $9.5 million is financed in repo against the bond.
3. Carrying cost. A year of the $500,000 at 0.5 percent is $2,500, and a year of the $9.5 million at 0.48 percent is $45,600: $48,100 a year before the CDS leg.
4. Protection. Buying the CDS lifts short-term borrowing to $700,000 and adds $166,385 of derivative assets and payables, taking the balance sheet to $10,366,385.

A hedged position still eats balance sheet: $10.4 million of it here, for a trade whose whole point is a spread difference. The paper calls its numbers illustrative, because there is no standardized version of the arbitrage, and the financing leg is the repo market's business, covered on the repo desk.

The basis moves, too. In the same paper's data the investment grade basis reached its most negative levels of the 2005 to 2016 sample in 2008 and 2009, then worked back toward zero. Over the month to 28 January 2016 it fell 13.2 basis points, and over the six months to that date 24.3, both among the bottom 10 to 15 percent of its historical changes.

Who the dealers are

The New York Fed's list of primary dealers ran to 26 firms in 2026. Ten of them are the global dealer banks named throughout these desk guides, each on the list under its own legal entity:

What the banks publish stops at the division line. Goldman Sachs reported FICC intermediation net revenues of $10,271 million for 2025, up from $9,564 million for 2024. J.P. Morgan reported Fixed Income Markets revenue of $5.4 billion for the fourth quarter of 2025 alone, up 7 percent on the year. Both lines bundle rates, credit, currencies, commodities and securitized products together.

None of the ten breaks out investment grade credit on its own or publishes the desk's headcount or bid-offer economics, and none publishes a structured account of how its credit desk interviews. The only credit-only revenue figure is the industry estimate above, not a number any bank reports.

A junior's day on the credit desk

Practitioners and training providers describe a day set by the handoff between Europe and New York, and by whatever the news does to single names. The pattern they describe runs like this:

• Before the US open, the desk reviews overnight runs from European dealers and any rating actions that landed overnight.
• A junior updates the desk's bid and offer runs and the price levels clients see.
• The morning meeting covers axes, the bonds the desk wants to buy or sell, along with client feedback and the headlines and earnings likely to move names.
• Liquidity improves while the US session overlaps with Europe's, and it can shift fast on bad news about a single company.
• Through the day, a first-year analyst helps cover institutional accounts and monitor risk, liquidity and exposure, while senior traders at VP and MD level take most of the risk decisions.
• At the close, the desk finalises marks and quotes for the day's positions.

Where a rates junior's day turns on a data release, a credit junior's turns on company news: earnings, rating actions and headlines that move one name at a time. How that compares with the other seats on the floor is its own subject, in a day in the life of a fixed income trader.

What the credit trading interview tests

The credit questions themselves, leverage, coverage, covenants and recovery, belong to credit analysis interview questions, and the desk-agnostic core sits in fixed income interview questions. A credit trading interview adds the desk's own mechanics on top, and a strong candidate can work each from a blank page:

• Quote a corporate bond credit spread two ways, and explain why a spread to one Treasury and an option-adjusted spread disagree.
• Size spread duration and CS01 on a real-sized position, by formula and by reprice.
• Give the sign of the bond versus CDS basis and the trade a negative basis calls for.
• Separate flow from risk, and explain how the desk earns inside the Volcker Rule's market-making exemption.
• Hold a view on where investment grade spreads sit against their own history.

Market-view questions of this kind get their own treatment in sales and trading technical questions.

Where credit traders go next, and how to prepare

Career guides describe a consistent set of exits, and the most-cited one needs the longest runway:

• Credit-focused hedge funds, both dedicated credit funds and multi-manager pods, usually after three to five years of building a track record with the trader's own P&L.
• Private credit, meaning direct lending and mezzanine funds.
• Another trading desk, as a lateral move or a step up.
• Asset management, in a buy-side portfolio role.
• Research or strategy, for a smaller number.

The exits compare across desks on fixed income exit opportunities, and what each seat pays on fixed income trader salary. Preparation is the worked material above, rebuilt rather than reread:

• Pull a fresh day's single-A effective yield, option-adjusted spread and Treasury par curve, then redo the two-route spread and explain the gap.
• Rework the CS01 on a different coupon or maturity in the bond calculator, by formula and by reprice.
• Say the basis sign aloud until it is automatic: negative means the bond is cheap to protection, and the trade is long the bond and long protection.
• Know what a 5MM+ print on the tape does and does not tell you.

A broader interview readiness check covers the rest of the desk-agnostic material.

The bottom line

The credit desk makes markets in the second-largest US bond market, trading a spread rather than a price for asset managers, insurers, pension funds and hedge funds, and holding inventory for clients rather than for the bank. Screens carried nearly half its volume across 2024, and a public tape prints every trade within 15 minutes. What a trader lives by stays short: a spread two ways, a CS01 two ways and the sign of the basis.