The CDS-Bond Basis

If a company defaults, a bondholder loses money and a CDS protection buyer gets paid — two instruments, insuring against the same event. In theory their prices should be locked together. In practice a corporate bond and a CDS on the same company almost never trade at exactly matching spreads, and that persistent gap is called the CDS-bond basis.

Why the two prices should match, and why they don't

Both a CDS spread and a bond's credit spread over the risk-free rate are, in theory, compensation for the same default risk on the same firm. If the basis moved far from zero, a trader could in principle buy the cheap side and sell the expensive one and lock in the difference regardless of whether the company later defaults.

$$\text{Basis} = \text{CDS spread} - \text{Bond asset-swap spread}$$

In words: subtract what the bond market is charging for the credit risk from what the CDS market is charging for the same risk. The result should hover near zero but rarely does, because the two markets aren't perfect substitutes:

• Funding. Buying a bond ties up cash that has to be funded at your own borrowing rate; buying CDS protection needs almost no upfront cash. When funding is expensive, bonds look artificially cheap (negative basis).
• Counterparty risk. A CDS contract depends on the protection seller actually paying out, so CDS carries its own layer of risk the bond doesn't.
• Technical demand. A wave of investors wanting to hedge without selling the physical bond (blocked by holding restrictions, say) pushes CDS spreads up on its own, unrelated to the bond.

Worked example

A corporate bond trades at an asset-swap spread of 150 bps over the reference rate. The 5-year CDS on the same company trades at 120 bps.

1. Basis. $120 - 150 = -30$ bps. This is a negative basis.
2. Interpretation. Protection is cheaper than the bond spread implies — the market is effectively saying the bond is offering more compensation for default risk than the CDS market is charging to insure it.
3. The trade. A basis trader would buy the bond (collect 150 bps of spread) and buy CDS protection (pay 120 bps), pocketing roughly 30 bps of carry that should be close to riskless — except it isn't fully riskless, because the trade needs financing the bond position and taking on counterparty exposure to the protection seller.

What this means in practice

Basis trading desks exist specifically to lean against these gaps, but the basis rarely closes cleanly because the frictions behind it — funding cost, counterparty risk, restricted holders — are structural, not accidental mispricings. Watching the basis is still useful even for someone not running the trade: a basis that's widening sharply negative is often an early signal of funding stress or a technical bid for protection building in the market, ahead of it showing up plainly in either instrument on its own.