CDOS: Correlation-EL equity tranche relationships

eschewier · Post by **eschewier** » Tue Jun 26, 2007 12:27 pm

Why does the EL decrease as the correlation increases in the equity tranche of a CDO? Why does the correlation affect the EL? I've missed something here. Was looking at this paper:

Fitch Corporate: A Primer on Valuing Synthetic CDOs of Corporates

CDO/credit novice

MadMax · Post by **MadMax** » Tue Jun 26, 2007 1:00 pm

Because, increasing correlation should normally increase probability of having no-defaults at all, or many defaults together. Basically it takes out proba from the center of the distribution and distributes it the extremes. Given that an equity tranche does not care what happens after a certain number of defaults as it gets all wiped out, increasing correlation is good because it increases the probability of survival of all underlying names...

Cheng · Post by **Cheng** » Tue Jun 26, 2007 2:10 pm

I think the best thing is to build a simple model in Excel using the Large Pool Approximation formula and play with the correlation. Have a look at the loss distribution of the portfolio, this should help. MadMax is absolutly right in what he says but I know this result is counterintuitive at a first glance.

Junior · Post by **Junior** » Tue Jun 26, 2007 2:13 pm

Impossible to give a better answer...

I would just add those two carton I think are fun to explain the situation: look at the position of the mousetraps, the dispersion will symbolize the correlation... when you are in the lower tranch, then you'll be among the first to pay. you are in the same situation than Tom, you don't want to get hit by any mousetrap, which is more likely in the right picture, i.e when the correlation is high

it's the opposite when you are in the higher tranch, high correlation => more likely for a lot of firm to default together => more likely to pay for the "safer" tranch...

[url=/User%20Files/3011/correl%20cartoon.doc]Attached File: correl cartoon.doc[/url]

(source: JPmorgan research)

Cheng · Post by **Cheng** » Tue Jun 26, 2007 2:40 pm

Junior,

could it be that we already met in meat space ? Just wondering...

Junior · Post by **Junior** » Tue Jun 26, 2007 2:55 pm

Well, to be honnest I strongly doubt...I haven't been in the practical world yet, so I' haven't been able to meet a lot of of people except the newbies in my library Confused

but I'll be out there in the shark pool around september, so perhaps in a couple of months!

eschewier · Post by **eschewier** » Tue Jun 26, 2007 3:50 pm

Cheng,

I believe that you might have recommended a Christian Bluhm book to me.

Cheng · Post by **Cheng** » Tue Jun 26, 2007 3:53 pm

Bluhm's books are indeed good for beginners to get an overview. And the large pool approximation is simple enough as a toy model to get an idea how things work. Nothing more, nothing less.

eschewier · Post by **eschewier** » Tue Jun 26, 2007 4:02 pm

But it's increasing probability of default of underlying names as well? Increasing the the equity tranche's volatility? but reducing the portfolio EL by improving the chances of credit quality improving? Or CDO EL, rather?

Wait.That makes no sense.

Why is the EL and not the UL affected by the increased correlation?

eschewier · Post by **eschewier** » Tue Jun 26, 2007 4:10 pm

I haven't yet looked at Bluhm - waiting for the library to receive the order. Do you know any good books/papers on credit portfolio management? I feel as if I'm drowning in the pool of all the knowledge out there.