Relative Value in Credit: CDS Basis and Credit Term Structure
This is a retelling of Chapter 7 (Part 1) from “Trading the Fixed Income, Inflation and Credit Markets: A Relative Value Guide” by Neil C. Schofield and Troy Bowler (Wiley, 2011, ISBN: 978-0-470-74229-7).
Chapter 7 takes the relative value framework from earlier chapters and applies it to credit markets. The first half focuses on the relationship between bonds and CDS, credit indices, forward CDS spreads, credit volatility, and how to trade the credit term structure.
The Bond-CDS Relationship
The most important spread measure for cash bonds is the Z-spread. It’s the flat spread you add to the entire zero-coupon curve so that discounted cash flows match the bond’s market price. Think of it as the market’s pricing of an issuer’s credit risk.
The CDS spread does something similar. It’s taken as the “true” measure of an issuer’s default risk, priced relative to the LIBOR curve.
Since both measure credit risk, you’d expect them to be equal. They’re not. The gap between them is called the CDS basis:
CDS basis = CDS spread - Z-spread (or asset swap spread)
You can also use the asset swap spread (ASW) instead of the Z-spread.
Why the Basis Exists
Intuitively, the basis should be positive. The cost of insuring against default should exceed what you earn from taking that credit risk. A negative basis would seem like free money: buy the bond, buy CDS protection, and still earn a spread.
But negative bases do happen. Here’s why the two markets can disagree.
Why the CDS spread might be higher than the bond spread:
- Protection buyers get a delivery option. They can choose the cheapest bond to deliver from a basket of eligible assets.
- Restructuring can trigger a CDS credit event even when it doesn’t constitute a bond default.
- It’s harder to short a corporate bond than to buy CDS protection, so CDS can carry a premium.
- CDS liquidity is concentrated in specific tenors (3, 5, 7, 10 years). Off-the-run maturities are wider.
Why the CDS spread might be lower:
- If banks fund bond purchases above LIBOR, the bond’s net return drops. Selling CDS protection looks better by comparison, pushing CDS spreads down.
- If the reference entity and the CDS seller are correlated in default risk, CDS protection is worth less. The spread falls.
- Heavy issuance of structured credit products referencing CDS contracts creates a surplus of protection sellers.
- Basis traders who hold bonds in asset swap form get stuck with a residual interest rate swap if a credit event hits. They demand a higher ASW to compensate.
In trading terms: a long basis trade means you buy the bond and buy CDS protection. A short basis trade means you short the bond and sell CDS protection. Don’t confuse a short basis trade with a negative CDS basis. They’re different things.
Credit Indices
CDS indices like iTraxx Main and CDX let you trade baskets of single-name CDS contracts in a single trade. The iTraxx Main, for example, has 125 equally weighted investment grade names.
A few things to know:
- The index composition refreshes every 6 months on “roll dates” around March 20th and September 20th.
- Each series has a consecutive number. The latest is “on the run.” Older ones are “off the run.”
- Indices trade at fixed maturities (3, 5, 7, 10 years). A new 5-year contract actually matures in 5.25 years.
- They trade with fixed coupons and upfront payments, just like single-name CDS.
- If a name defaults, it gets removed and the notional shrinks. Each name in iTraxx Main is 0.8%, so a default on a 10 million euro contract reduces the notional by 80,000 euros.
Most participants roll their index positions when a new series launches. Market makers offer the roll as a single trade, quoted as a bid/offer spread on the net basis points exchanged.
The Index CDS Basis (Skew)
The fair value of an index should equal the aggregate value of its constituent single-name CDS contracts. It doesn’t, because markets are never perfectly efficient.
The difference is called the skew or basis:
Basis = Theoretical index spread - Market index spread
You can’t just add up the par spreads to calculate the theoretical value. You need to account for upfront cash flows on all the constituent contracts.
When the skew gets big enough, an arbitrage trade becomes possible: buy the cheap side, sell the expensive side. But executing this on the iTraxx Main means dealing in 125 individual names. That’s why traders prefer indices with fewer components, like HiVol or Crossover. The skew is typically widest during volatility spikes, but that’s also when bid/offer spreads are at their worst.
Forward CDS Spreads
There’s no corporate bond future, so you can’t trade forwards directly. But you can create a synthetic forward CDS position by combining two different maturities.
Buy 10-year protection and sell 5-year protection on the same name with the same notional. This gives you a forward-starting bearish view on the 5-year spread, starting 5 years from now. The trade profits if the par 5-year spread in 5 years is greater than the implied forward spread today.
The forward spread is calculated using expected losses and PV01 values. The numerator represents expected loss, the denominator represents expected risk.
Key takeaways:
- A forward purchase (long 10-year, short 5-year protection) benefits from spread widening or curve steepening. It’s cheaper than outright protection because the short leg subsidizes it.
- A short forward benefits from spread narrowing or curve flattening. It’s default neutral until the short-dated contract expires.
The book walks through a detailed example with 100bp fixed-coupon CDS. The 10-year requires 9.5% upfront and the 5-year trades at 5.0%, so net cost is 4.5%. For the first 5 years, the coupon payments cancel out. The pull-to-par effect (mark-to-market from time decay) roughly offsets the funding cost, so P&L is nearly flat. After year 5, the trade carries negatively.
A credit event shortly after inception would be costly not because of the notional (which nets to zero) but because you lose the 4.5% upfront payment.
Credit Volatility: Options on CDS
Options on CDS (credit swaptions) let you trade the direction of spreads and their volatility. The naming convention follows interest rate swaptions:
- Payer swaption: Right to buy protection at a strike spread. Profits when spreads widen. Think of it as a call on credit spreads.
- Receiver swaption: Right to sell protection at a strike spread. Profits when spreads tighten. Think of it as a put on spreads.
The payoff at expiry isn’t just the intrinsic value. You multiply by the risky PV01 at expiry. So a 10bp in-the-money payer on $10 million with a PV01 of 4.2 is worth $42,000.
Breakevens are trickier than they look. If the premium is 30bp and the PV01 is 4.2, the spread only needs to move 7.14bp from the strike to break even (30/4.2).
Index options are more popular than single-name options. Liquidity concentrates in 1-6 month maturities on 5-year CDS. They’re usually European style, physically settled, and expire on standard CDS roll dates.
To evaluate whether an option is cheap or rich, traders compare implied vol against realized vol, convert implied vol to daily spread moves using the formula (implied vol x spread / 16), or compare straddle breakeven ranges to historical spread ranges.
Trading the Credit Term Structure
CDS contracts have a term structure just like interest rates. You can trade steepeners, flatteners, and butterflies.
Steepening/flattening trades are built the same way as in rates: buy long-dated protection and sell short-dated protection (or vice versa), DV01 neutral. The twist in credit is the fixed-coupon format and upfront payments.
For a steepener (buy 10-year protection, sell 5-year protection), if the DV01 ratio between 5 and 10-year is 1.56, then a 100 million euro 10-year leg needs a 156 million euro 5-year leg. The notional imbalance means coupons don’t cancel, creating carry. In the book’s example, the net coupon carry is 570,000 euros in the trader’s favor.
There’s also a pull-to-par effect. The short 5-year position decays positively (it could be reversed at a lower spread over time), while the long 10-year decays negatively. Initially, the 5-year’s positive decay wins.
Because the notionals aren’t equal, an event of default has a net impact. If you’ve sold 57 million more protection than you’ve bought, that’s your jump-to-default risk.
Butterfly trades work the same conceptually. Using 3-5-7 year maturities, if the 5-year belly looks too high, you buy the wings (3 and 7-year protection) and sell the belly. The trade is neutral to parallel shifts and pivots. Profit comes from the belly rate falling relative to the wings.
Convexity matters more in credit than in rates. A long protection position has negative convexity: profits from widening decelerate as spreads move. A short protection position has positive convexity.
For curve trades, flatteners have positive convexity (profits from flattening exceed losses from equal steepening), while steepeners have negative convexity (the reverse). Any scenario involving 5-year spread changes of 100bp or more needs to account for this.
That covers the first half of Chapter 7. We’ve seen how relative value works in credit: basis trading between bonds and CDS, index arbitrage, forward CDS positions, credit options, and term structure trades. Next up: single-name credit views, credit-linked notes, and portfolio strategies.