wilmott.com

I'm looking for contract specifications for market quoted CDO tranche contracts. The ideal package would be a snapshot on a specific point in time of the following:1) The underlying index members (and preferably single name spreads)2) The loss protection attachment points of the tranche(s) and other contract specifics.3) Market quotes and other risk figures of the tranche contracts available from distributorsCan someone point me in the right direction? I would be forever grateful for any input.Thanks in advance

for index members, visit markit.com

How do we price the index tranches? Is the underlying spreads and the index have one to one relationship?Thanks in advance.

No, it is not enough to know all the single name spreads. You need a correlation assumption. Actually, correlation doesn't begin to describe the complexity of the relationship among credits, so you really need an implied correlation assumption, which is another way of saying you need the price. No one knows how to price index tranches from first principles.

Would you mind expanding on this relationship between implied correlation and price, and why the former is often said to suffice in place of the latter?Thx.

If you have all the single name spreads, you know the spread of the basket, but not of tranches on the basket. To compute the price of tranches, you need to know the entire joint distribution of prices of the underlyings.Instead of that, you can make the highly simplified assumption that all pairwise correlations are the same and there are no higher order effects. The market convention is to use the Gaussian copula time to default correlation. If you do that, you can plug in a value for correlation and get a price for any tranche. Or you can plug in the price of any tranche and solve for the correlation that implies it (subject to the unpleasantness that you will sometimes get multiple or no solutions, which is one big reason people use base correlations instead).Therefore, as long as the relation between implied correlations and tranche prices is one-to-one, prices and correlations are equivalent ways to express the same thing, like Farenheit and Celsius temperature. You can convert from one to the other easily. You will get a different implied correlation for each tranche, and different correlations for the same tranche in different baskets.

Thanks, Aaron. Once I run the Gaussian Copula, how do I convert my output of implied correlation into price? Conversely, how am I to determine implied correlation based on a set price I'm willing to pay for a given tranche?Thanks in advance

This is not simple. I recommend the help screens to the CDSM function in Bloomberg for the technical details.

Will do. Thanks Aaron.WCM