Dear All:An CDS (protection) buyer pays a periodic fee for thel compensation when a predefined credit event occurs. Most books say the compensation equals (1 - recovery rate) times the face value of the bond (plus accrued interest). My question is, does the CDS premium varies with the different bonds issued by the same company? The current quote of 5-yr CDS on IBM is around 24 bps, but what is the underlying bond for this contract ? Assuming zero recovery rate and the default compensation euqals the bond's face value (say $100), then the same protection fee (24bps) should offer quite different protection values for an IBM zero coupon bond and an IBM 5% coupon bond, right? Thanks for your helpful answer.Paka

I thought it was an easy question when I posted it yesterday, but so far no reply yet...........For actively traded single-name CDS, such as GE, Ford, IBM, GM, etc., What is the underlying for the contract? If I have a 5-year bond issued by Ford, can I just buy the 5-yr Ford CDS to hedge out the credit risk? Will the CDS spread on my Ford bond be different from that on Ford CDS actively traded in the market?If anyone knows the answer, please respond. Thanks

You usually have some referencebond as underlying for your CDS, but I can't tell you which one it is for IBM 5Y.Anyway. If a credit event occurs, the protection holder (i.e. buyer of protection/CDS) has the right to exchange any corporate bond that meets the definitions for the par value to the protection seller. The definitions usually are about the seniority, which makes it a cheapest to deliver contract in case of default and multiple possible bonds.(sorry, don't check the forum during the weekend) [edit] those zerobonds are tricky. I'm not sure about those either. Check the ISDA def's[/edit]

QuoteOriginally posted by: PakaDear All:Most books say the compensation equals (1 - recovery rate) times the face value of the bond (plus accrued interest)PakaThere is no explicit reference to a recovery rate is CDSs settlement mechanism. Either the CDS is physically settled and a bond is delivered vs payment of the notional, or the CDS is cash settled and market value of a bond is paid vs payment of notional.Y

unfortunately there is no (or very rarely a) liquid market for postdefault bonds. The usual procedure for cash-settled CDSs is that you start a dealer poll after a period of time (say one month to calm things down) to estimate a "market value" or recovery of the bond. As soon as both parties agree on the conditions (huge field of work for whole armies of lawyers the CDS is settled .Since this is even more costly and strenous than the physical settlement CDSs are usually settled in the latter way.

Something is rather confusing about CDS (at least for me). If the default probability of a certain name is increasing (hihger for longer term), then its CDS spread should be higher, the longer the CDS maturity. For example, 1-yr CDS is 20bps, and 2-yr CDS is 30bps for the same name. Now after one year, what will be the mark-to-market value of the original 2-yr CDS? Assuming other things held constant, the CDS with one year remaining maturiing must have a positive market value for the protection seller, right? (since the new 1-yr CDS on this name has again a spread of 20 bps, but the protection seller still receives 30bps on the original 2-yr CDS. In short, for protection seller in this type of name (increaseing default probabiity), it can expects a increasing market value for its CDS position? Can this be right? Please advise on this. ThanksPaka

I think you compare the 40bp (=20+20) for the two 1Ys with the "30bp" for the one 2Ys, right?The point is, that in the CDS-market you don't quote upfront payments (in this case you were right) but annual payments. That means, if we consider 100% Discont Factor, that you receive 40 bps for the two 1Y and 60bps for the one 2Y. This is reasonable, because at the two 1Y you can recalibrate the payments after one year, wich you can't do for the 2Y.Hope that was, what was confusing you...

The survival probability is clearly non-increasing in time. This, however, does not imply that spreads have to be non-decreasing, but rather they must only satisfy a set of no-arb conditions determined by the other marks on the curve. In the situation you described, yes, the protection seller would have a positive MTM. This is due to the fact that the implied probability of default in the second year is decreasing since you are "rolling down" the curve (your assumption was that "other things held constant").Hope this helps.