ISDA CDS Recovery Rates Are Pricing Conventions, Not Default Outcomes
Summary
The document asks why ISDA uses different assumed recovery rates for senior unsecured, subordinated, and emerging-market debt in credit default swap conventions. The answers clarify that these standardized rates are conventions used in quoting and pricing CDS contracts, rather than forecasts of what a particular bond will recover after a credit event. Standardization lets market participants price contracts without specifying a separate recovery assumption for every deal.
The discussion distinguishes the conventional rate used in pricing from an analyst’s expected recovery assumption and from the recovery actually realized after default, which determines the eventual payout. It also notes that, for a given CDS market price, assuming a higher recovery rate implies a higher market-implied default probability. The document does not establish the historical or empirical rationale for the particular category-specific percentages raised in the question; its useful contribution is explaining how to interpret those conventions and keep the three recovery concepts separate.
Key ideas
- ISDA recovery rates are standard pricing conventions for CDS contracts.
- A conventional recovery assumption does not predict the recovery from a particular default.
- Expected recovery is a modeling input, while realized recovery reflects what creditors ultimately receive.
- The recovery realized after a credit event affects the CDS payout.
- For a fixed CDS price, a higher assumed recovery implies a higher implied default probability.
Tags
Full text
# Question about the rationale of applying certain recovery rate by ISDA # Question about the rationale of applying certain recovery rate by ISDA According to ISDA standard (also here), the recovery rate for senior unsecured is 40%, that of subordinate is 20%, and emerging markets is 25% (both senior and subordinate). I wonder the rationale of applying higher recovery rate of emerging markets (subordinate) than that of the subordinate in developed countries. Could anyone explain the rationale of applying certain recovery rate for senior unsecured, subordinate and emerging markets assumed by ISDA? Is there any reference explaining such assumption on recovery rate? Thanks ## Answer by experquisite (score 3) https://quant.stackexchange.com/a/9117 If the contract is actually triggered, then it will pay out depending on the actual recovery rate of the particulars of the circumstance. To recap: - nominal CDS recovery rate in ISDA docs - used as a convention in the pricing of CDS contracts (spread vs upfront etc). Has nothing to do with any particular credit event. - expected recovery rate - modeled when pricing bonds, the expectation of bond recovery in the event of a default/credit event. - actual realized recovery rate - how much the bonds are actually trading for when a credit event is declared, which determines the actual realized payout from the CDS contract. ## Answer by Yugmorf (score -1) https://quant.stackexchange.com/a/9124 Presuming the question refers to CDS pricing: Note that a contract pays out on a credit event (no matter the scale of losses), while the value of the contract also depends on the value that can be recovered (see *NB below). So in agreeing the price of a contract you have two variables to settle on (probability of an event, and the recovery rate). To avoid a having a matrix of different contracts (i believe it started as such), and introduce some standardization, most (all?) CDS contracts will now use a standardized recovery rate as a pricing convention only (says nothing about the actual eventual recovery rate in case of default). Given this, the price depends only on one variable, the probability of default. *NB Given a market price for a CDS, the higher the recovery rate you assume, the higher the default probability implied.
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.