Thank you, Fodao. In my understanding, BS defines base correlations as correlation inputs of series of equity tranches that can be used to replicate pay-offs in mezzanine or senior tranches. So the base correlation of [0,7] tranche is the correlation input which makes the equation below true.PV[3,7](rho, S) = PV[0,7](base7, S) – PV[0,3](base3, S)*S is the 3-7 fair market premium On the other hand, in my understanding, JP defines base correlations as correlation inputs of series of equity tranches that can be used to replicate expected losses of mezzanine or senior tranches at maturity. So the base correlation of [0,7] tranche is the correlation input which makes the equation below true.EL[3,7](rho) = EL[0,7](base 7) – EL[0,3](base 3) The BS approach seems more precise and intuitive in the way that base correlations are defined to replicate pay-offs rather than expected losses at maturity. So it seems natural for a financial institution to use the BS approach for that reason. I guess that JP’s approach can only be applicable if we use homogeneous model (assuming all spreads are flat and equal through the whole period). Because under the homogeneous condition, base correlations of JP’s approach can replicate pay-offs. Is it why you mentioned that the BS approach is better if we are not using homogeneous model?Regards,

Nakano,I've mentioned that the BS approach is better when we do not use the homog. model, because the JP approachstill requires that we calculate the implied corr. for each tranche, while BS doesn't. I suggest you take also a look at the threadBase vs Compound Correlation in the Technical forum.

How do we price the bespoke tranches? Can we use Gaussian Copulas for this?Thanks.

Yes, why not? The only issue is what corr. you are going to use.

Bespoke tranches will have portfolios of different quality. Do we have to modify the specification of gaussian correlation parameter? Or does multifactor representation fit better for the bespoke? Is there any standards in quoting the prices for the bespoke (can we quote the prices of base correlation curve)?Thanks.

Hi guys,My question may be very basic but I would appreciate very much if someone could shed me some light.The question is:How to calculate the expected loss of the tranche EL(k1, k2) from the tranche's market spread?Is the expected loss really independent of correlation? If so, what is the exact formula that relates the expected loss to the spread? If not, how would one calculate EL(K1, K2) then which is central in JPM's base correlation framework.I am quite stuck at this point and your help is very much appreciated.Thanks in advance,Ted

The expected loss of a tranche is NOT independent of correlation. In the JPM approach you still have to usethe concept of compund corr. to bootstrap the Base corr. In the BS approach you don't. Given the Base corr. , calculating the exp. loss of a tranche is trivial.

Hi there,Latest publication paper suggests that in order to calculate the base correlation for the 0-6%: Add PV of 0-3% and 3-6%. Then reproduce PV of 0-6% with a correlation, this is the base correlation for 0-6%. And also to calculate for 0-9%: Add PV of 0-3%, 3-6% and 6-9%. Then reproduce PV of 0-9% with a correlation, this is the base correlation. And similarly for the other tranches e.g. 0-12% etc..However, when I do this no matter what correlation I input I do not get the required PV (sum of PV).Can anyone tell me please what I do wrong? Or point me to some papers that shows the base correlation calcs step by step?Many thanks,

Did you get the correct implied correlation for the 0-3 tranche? If not, focus on getting this right first.If so, it is odd that you aren't getting the correct NPV for the 0-6 tranche.

Any chance you are forgetting to add the notionals of the two tranches in your PV calculation for the fixed leg?

Hi all,Does anyone know how can I get the implied correlation for North America standard tranches from bloomberg? The Morgan Stanly page on bloomberg only give the base correaltion for different tranches - but I need the implied correalation in order to test my pricing model.Regards,

CDO2 if it is what you are after BPCI 12 shows compound correlation levels.Y

Hi CDO2,You basically got two choices:1. Implement the base correlation scheme, which is not that difficult.2. If you really want to have tranche implied corr. you can try to contact the individual dealer and if your company has business with them, they will be happy to send you daily emails containing CDO tranche prices, etc. Ironically, I happen to know that Bear Stearn's price email for their CDO market has quotes with respect to tranche implied corr. Ted

Hi guys,I have the following question regarding the calculation of CDO tranche delta and I would really appreciate if some expert could shed me some lights:To calculate the Delta, one would need to manually shift the CDS spreads of the underlying reference entities in order to compute the MTM change of the tranche NPV. However, I am not too sure how exactly the shifting is done. Do I simply need to add some small but fixed amount (say 1 or 2 or 5 bps) to each spread or perform a proportional shift (i.e. increase each spread by 1%) or some other shifting scheme? I personally found the resulting delta quite sensitive to what shifting scheme is used and I am having a hard time calibrating my delta to the market (for example, the US equity tranche typically has a Delta around 15x). Has anyone encountered similiar issues before or could some dealer side people provide some information.Thank you very much.Ted

On the first point, depends, but I think standard approach is absolute, fixed bp across each. Given the different schemes, sounds like convexity is playing a role.