May 23rd, 2005, 7:26 am
This is what you would call a "highly non-trivial problem"! (Or the 64 million dollar question).A portfolio that is very much like iTraxx would be priced like iTraxx, and similarly for CDX. But if you were to for example combine the iTraxx and CDX portfolios it's not obvious to me that you would neccessarily price tranches using interpolated parameters.Bespoke synthetic tranche pricing is big money for banks right now, so you can be sure that many intelligent people have spent a lot of time thinking about this issue. I believe that everyone has their own opinion as to how to do this, and practitioners are unlikely to give their secrets away lightly.But the thing is, these banks need to lay off some of the risk with hedge funds, and if hedge fund guys are clueless as to how to do this (that was "if"!) then the banks would be stuck with the bespoke equity tranches. Banks would then have to take basis risk, keeping the bespoke equity tranche and trying to lay off some combination of iTraxx and CDX equity tranches, or combinations of the standardised FTD basket.One thing that has reportedly been tried is teamwork between a big hedge fund and a bank. The big hedge fund comes up with an equity tranche it could live with, and the bank tries to market the resulting junior mezz (BBB rated) tranche. Such a trade can work as long as the hedge fund is not too highly levered and is not too likely to have redemptions. Ideally a real money (zero leverage long term investor) should take this risk, not a hedge fund.If such a hedge fund were to get redemptions, it would have to try and lay off a bespoke equity tranche that nobody has any interest in, or maybe it would have to pay a bank to take the risk. In that case the bank may have to dump huge amounts of standardised equity tranche on the market, destabilising the tranche market.This is of course all just a thought experiment. Who knows if it could happen in real life?