Page 1 of 1

Bootstrapping the intensity of default from CDS when the CDS market is illiquid

Posted: November 20th, 2011, 7:30 pm
by HeartOfDarkness
Hi, all,I am having a look at how to estimate the intensity of default from CDS prices and it looks like the bootstrapping method is widely used by practitioners. But, it looks to me that it relies on the assumption that the CDS market is liquid. I would be interested to know how this method performs when the market is not liquid (like in the 2008, for instance). Any thoughts/insights?Many thanksHoD

Bootstrapping the intensity of default from CDS when the CDS market is illiquid

Posted: November 23rd, 2011, 2:24 pm
by HNK
In an illiquid CDS market the inferences drawn from the implied hazard rates are just approximations, so you can use simple mid-values. If certain parts of the curve are missing, just fill them in as a flat-rate or interpolate or use some educated guess.Even otherwise, it's good to maintain a margin of safety so if the credit trade (buy/sell CDS) is to be used as a hedge, you should take the worse of the bid/offer for your purpose.

Bootstrapping the intensity of default from CDS when the CDS market is illiquid

Posted: November 24th, 2011, 2:02 am
by katastrofa
When hedging with bid offer spread, I'd rather take the mid and then count the spread towards the cost of the hedge.

Bootstrapping the intensity of default from CDS when the CDS market is illiquid

Posted: November 24th, 2011, 10:21 pm
by hamster
hmm... an illiquid contract where someone will pay 1-R if something else goes down...a try...- take a big bunch of cds contracts from different protection sellers for a set of comparable reference entities over a time period you believe nothing material changed in your illiquid world- check how much of the spread depends on the creditworthiness of the different protection seller (counterparty risk)- check how much of the remaining depends on the bid-ask spread of the cds (illiquidity premium)- all the remaining adjusted data should correspond to the original purpose of the cds how to calculate this depends how much data you have, and how accurate results should be.usually simple regression is enough to adjust data. for an illiquid market you will not have data anyway.might not even enough for meaningful regression. so you might go for an educated guess?