Credit Default Swaps (CDS): Spreads, Probability, and Payouts

A CDS is a contract between a protection buyer and a protection seller that references a third party, called the reference entity. The buyer pays for credit protection; the seller takes on the specified credit risk in return for that premium. The contract itself is separate from a bond or loan issued by the reference entity. A buyer may use a CDS to hedge a debt holding, but the CDS does not transfer ownership of that debt or remove the holder’s obligation to fund it.

The ISDA glossary defines the protection buyer and seller by which side pays for protection and which side assumes reference-credit risk. The exact result still depends on the confirmation and incorporated definitions: reference entity, seniority, currency, maturity, credit events, deliverable obligations, and settlement method all matter. A corporate name alone is not enough to identify the risk being traded.

A CDS spread is usually quoted in basis points per year against notional. One basis point is 0.01 percentage point, so 100 bp equals 1% per year. In a simplified running-premium example, a buyer of $10 million of protection at 100 bp pays about $100,000 per year while the contract remains in force. If payments are quarterly, that is roughly $25,000 per quarter before the contract’s day-count and accrual details.

That annualized price is not the probability of default. It is the price of the protection leg relative to the premium leg under the contract and market’s pricing conventions. In standardized CDS markets, the scheduled coupon can differ from the quoted market spread; an upfront amount may reconcile the two. Check whether a screen shows a running spread, a fixed coupon, an upfront payment, or a combination rather than treating every displayed number as cash due each quarter.

A rating downgrade or a falling bond price does not, by itself, trigger a CDS payment. A triggering event must meet the definitions in the contract. The relevant ISDA Determinations Committee may resolve questions such as whether a covered Credit Event occurred, whether an auction will be held, and which obligations may be delivered under documentation that incorporates that process. ISDA’s review of the committee process explains why the determination follows contractual definitions and documented facts rather than a rating label.

If settlement is by auction, the final price is used to calculate a cash amount under the transaction terms. With a hypothetical $10 million notional and a 25% auction final price, a simple protection amount is $10 million × (1 − 0.25) = $7.5 million. Physical settlement instead involves delivery of an eligible obligation against payment at the agreed amount. Not every contract uses the same terms, and an auction is not automatic for every event. The ISDA auction-settlement materials describe how auction terms and determinations were incorporated into standard documentation.

The basic intuition is that the CDS premium compensates the protection seller for expected credit loss, after accounting for the contract’s payment timing and discounting. A common simplified relationship is spread ≈ default intensity × loss given default. If the assumed recovery rate is R, loss given default is 1 − R, so a flat-hazard approximation is λ ≈ s ÷ (1 − R). The BIS pricing discussion derives the pricing relationship by equating the present value of expected premium payments and expected protection payments.

Suppose a hypothetical five-year CDS spread is 100 bp per year (s = 0.01), and the model assumes 40% recovery (R = 0.40). The simplified annual risk-neutral default intensity is 0.01 ÷ 0.60 = 1.67%. Under a constant intensity, the five-year cumulative probability is 1 − exp(−λ × 5), or about 8.0%. This is a model illustration, not a probability read directly from the quote. A term structure of spreads, discount factors, payment dates, accrued premium, recovery conventions, and upfront cash flows require a fuller valuation.

Holding the spread at 100 bp (s = 0.01), a 40% recovery assumption gives an implied intensity of λ ≈ 0.01 ÷ 0.60 = 1.67%; a 20% recovery assumption gives 0.01 ÷ 0.80 = 1.25%. Lower recovery means a larger loss per default, so a lower default intensity is needed to explain the same spread. If each intensity stays constant for five years, the corresponding cumulative estimates are about 8.0% and 6.1%. This shows sensitivity to a model input; it does not show that an issuer’s real-world risk has fallen. A CDS curve also contains information across maturities, while one five-year spread cannot show when losses are expected or whether conditional risk is equal each year.

The spread-implied quantity is usually described as risk-neutral or market-implied. It is the probability measure used to price contingent cash flows, not the historical or physical frequency of default investors should expect to observe. The Basel Framework explicitly says market-implied default probability generally differs from real-world likelihood; BIS research also uses CDS prices to extract risk-neutral distributions. Those estimates can incorporate the price investors demand for bearing credit risk, alongside expected loss.

The spread can move because expected loss changes, but also because of risk appetite, liquidity, funding and collateral conditions, contract terms, or buying and selling pressure. A higher quote can therefore mean protection became more expensive without proving that an issuer’s real-world default probability rose by the same amount. A lower quote is not proof that the issuer is safe. The BIS discussion of CDS terms and pricing also explains that differences in contract provisions can be priced into quoted spreads.

A CDS spread and a corporate bond’s yield spread are related credit-market measures, not interchangeable readings. Bond spreads can reflect the bond’s cash-market liquidity, financing, taxes, coupon and benchmark conventions. A CDS quote reflects its own reference entity, documentation, maturity, seniority, collateral and settlement terms. The difference is often called the CDS-bond basis, but the label does not make a mismatch-free arbitrage. Align currency, maturity, seniority, deliverability, timestamps, and price conventions before interpreting a gap.

Hedging a bond with a CDS can leave basis risk if the bond is not a deliverable obligation or does not match the contract’s reference entity, seniority, currency, or maturity. The protection buyer also takes counterparty and collateral risk; the seller may face a large contingent payment compared with the premiums collected. A hedge changes exposures; it does not erase funding, liquidity, or settlement risk.

Before interpreting a quoted spread, record the reference entity and obligation, protection side, notional, maturity, currency, seniority, and quote timestamp. Identify whether the quote is a par spread or a standard coupon plus upfront amount. Then read the governing credit-event and restructuring terms, eligible deliverable obligations, settlement method, and collateral arrangements. If converting the quote to a probability, state the recovery assumption, hazard-curve shape, discounting and payment conventions, and whether the estimate is risk-neutral.

For related fixed-income context, compare the G-spread, I-spread, Z-spread, and OAS guide and the corporate bond yield and Treasury spread guide. This article explains market mechanics; it does not recommend a CDS trade or estimate any issuer’s current default risk.