Skip to content
All library documents

How CDS Spread Quotes Relate to Standard Running Coupons

Article Quant Q&A · Author: Guy Kern

Summary

The document explains that conventional CDS spreads are quoted in basis points as an annual rate on the protected notional. It illustrates the quote with a five-year contract and a 75-basis-point spread, where protection on USD 10 million corresponds to USD 75,000 per year under the historical par-spread convention.

It distinguishes this quote from the standardized contracts traded after the CDS market’s Big Bang. Those contracts use a standard running coupon, with an upfront payment that adjusts for the difference between the quoted spread and the standard coupon. Converting a quoted spread to a risk-neutral default probability curve uses a recovery assumption and an interest-rate curve; converting that curve into the upfront amount also depends on the running coupon. The answer says direct comparison of quoted spreads does not require knowing whether the standard coupon is 100 or 500 basis points, while converting an upfront quote into a spread does. These are conventions and model-based conversions, not direct measures of realized default likelihood.

Key ideas

  • Conventional CDS spreads are quoted in basis points per year on notional.
  • Standardized CDS contracts combine a running coupon with an upfront payment.
  • The upfront amount reflects the difference between the quoted spread and the contract’s standard coupon.
  • Converting quotes to risk-neutral default probabilities requires recovery and interest-rate assumptions.
  • Comparing quoted spreads differs from converting an upfront amount into a spread.

Tags

Full text
# In which units the conventional CDS spreads in Markit's data are measured?


# In which units the conventional CDS spreads in Markit's data are measured?












I am trying to understand if the conventional spread column in Markit's CDS database simply represents the CDS spread, measured in bps, or should I make some adjustments (in case I would like to make some comparisons between CDS's with different coupns, e.g., 1% or 5%)?

Thank you!

## Answer by Dimitri Vulis (score 0, accepted)

https://quant.stackexchange.com/a/73693

The CDS spread is quoted in basis points. It represents the fraction of the face value that the credit protection buyer would pay every year to the protection seller. For example "75" basis points for 5Y tenor means that to buy protection on USD 10,000,000 face value, you'd need to pay 0.0075 $\times$ USD 10000000 = USD 75,000 a year for 5 years. There would be zero upfront fee because the swap would have zero msrk to market. This us how CDSs used to be traded before the "big bang".

But this is just a quoting convention. Following the "big bang", what actually gets traded is a "standard" contract, with standardized running spreads. So in this example, the protection seller would actually pay some upfront fee to the buyer (because quoted < standard 100), and then the buyer would pay 25 bps every 3 months. The conversion from the quoted CDS spread, combined with a recovery assumption that it's tagged with and an interest rate curve (which affects the result little) to risk neutral default probability curve using JPM's standard CDS model is unambiguous. Further conversion from the risk neutral default probability curve combined with a standard running spread (usually 100 bps, sometimes 500 bps, or some other) to the upfront fee that actually gets paid at inception is also unambiguous.

Some names on the verge of default are already quoted as upfront in percent. E.g. 40% upfront means the buyer pays 30% of the notional upfront and then 5% of the notional every year. The upfront depends on the running spread.

So if you're comparing quoted CDS spreads to each other, then you don't need to know whether they're for 100 or 500 bps running spread; but if you're converting a not-too-high upfront fee into a spread fir comparison purposes, then you do need the running spread.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.