Hello all, apologies for the dumb question, but for some reason I don't quite understand the official interpretation of base correlation.Let's say I know the base correlation for 0-3, 0-7, and 0-10, and I also know the BE spread for 0-3 and 3-7. How do I calculate the BE spread of 7-10? And if it requires the value of a Risky01 or default leg from an earlier tranche, it would be helpful to know at which spread and at which correlation that should be done at.Rgds

My €0.02: Since you have the correlation for the 0-10% tranche, you can work out the breakeven spread on that tranche (via Monte Carlo or whatnot). Now, since the 0-10 tranche is just the sum of the 0-3, 3-7, and 7-10 tranches, the breakeven spread on 0-10 is just the appropriate weighted average of the spreads on those three tranches. Since you know spread(0-3), spread(3-7) and spread(0-10), you can solve for spread(7-10). htha.p.

You have the base correlation for the 0-3 , 0-7 and 0-10NPV(0-3,s) + NPV(3-7,s) = NPV(0-7,s)First Method : you find your Base correlation when you take s = Spread[3-7] so NPV(0-3, s) = NPV(0-7,s)Default Leg [7-10] = DefaultLeg[0-10, BC(0-10)] - DefaultLeg[0-7,BC(0-7)]PremiumLeg[7-10] = (10 * PremiumLeg[0-10, BC(0-10)] - 7 * PremiumLeg[0-7, BC(0-7)])/(10 - 7)s[7-10] = Default Leg [7-10] / PremiumLeg[7-10]Second Method :s = 0 (JPM Method)(EL = DefaultLeg)EL[0-7] + EL[7-10] = EL[0-10]You have the Base correlation 0-7 so you know EL[0-7]You have the Base correlation 0-10 so you know EL[0-10]So EL[7-10] = EL[0-10] - EL[0-7]You find the correlation r which mathc the EL[7-10].You price your CDO 7-10 with this correlation and you have your spread 7-10

Hi Jonathan81, thanks very much for your reply. The 1st method sounds easiest so that is the one I will go with. Your insight is much appreciated.Actually I have a small related question, if you have just 1 more minute. Obviously with semi-analytic methods for single tranche pricing we have the prob distribution for defaults only at specified times (i.e. at the fixed leg pay dates) as opposed to continuously throughout the life of the tranche. What is typically assumed about the timing of defaults when calculating the risky01 of a tranche?I had been leaning toward Risky01 = SUM(1 to N) 0.01% * DayCountFraction(i) * Discount Factor(i) * (ExpectedNotionalOutstanding(i-1)+ExpectedNotionalOutstanding(i))/2 / InitialNotional, which I guess assumes mid-point defaults. Is this the typical protcol?Thanks again and regards,Hunting

Sorry but i don't know exactly but as far as i am concerned i use:Risky01 = SUM(1 to N) 0.01% * DayCountFraction(i) * Discount Factor(i) * ExpectedNotionalOutstanding(i) / InitialNotionalRegards,Jonathan

Jonathan, thx for your help! Last one I promise ...On the default leg, do you use DL = SUM(1 to N) ExpectedDefaultLosses(i) * DiscountFactor(i) orDL = SUM(1 to N) ExpectedDefaultLosses(i) * [DiscountFactor(i-1)+ DiscountFactor(i)]/2?Rgds

Hi Hunting,I assumed mid-point (between payment days) for defaults and the results calibrates to the market farily well.The problem with "mid-point" default assumption is the treatment of the "accrued leg" calculation, which is not as obvious as the case for a regular CDS contract.I have also known some other dealers who just assume "end point" default and thus no accrued leg calculation is needed.Not sure if I answered your questio. If not, just ignore my post.

On the default Leg, there is a little mistakeI use http://www.damianobrigo.it/gpl.pdf (p9-12)

Hey Guys, I appreciate the responses.Jonathan, thanks for the reply on the DefaultLeg. I should have written ExpectedDefaultLossesInPeriod(i) which is what I intended, which should equate to the EL(i) - EL(i-1). I appreciate that heads up.Quoted, you are absolutely correct about what I was asking. Can you provide a few comments regarding how to calculate the 3 legs under the mid-period default assumption? Rgds