Q. So here the “recovery” relates to the amount a creditor actually receives on a bond post-default, right?
A. Right. Payouts aren’t binary: bondholders don’t always get either full payment or zero in event of default. In fact, that’s the basis of distressed credit investing… you might be buying a bond for pennies on the dollar, in the expectation that it will default, but you’ll still be able to recover, say, fifty cents, in the bankruptcy proceeding. The actual rates could be in a huge range, 2 cents to 90 cents.

Q. So the actual “recovery” on a defaulted bond can be all over the place…
A. Right. And that’s the point of a “recovery lock”. A plain vanilla credit default swap pays out the full face amount of a bond in the event of default–it’s an insurance policy, and the person who bought the “protection” gets paid a flat amount.

Q. OK, then, how does a “recovery lock” CDS work?
A. It operates like a contingent futures contract. In the simplest case, the seller of the recovery lock agrees to sell a defaulted bond, post credit event, at a specified percentage of its face amount. Say that’s 20%. The buyer of the recover lock agrees to buy the bond for that percentage of face, takes the bond, and will get whatever the actual recovery on the bond is. There are variations on the settlement mechanism– whether by delivery of the bond or in cash– but those are the economic consequences.

Q. So the buyer of the recovery lock is speculating that he’s be able to recover more than the percentage recovery specified in the contract.
A. Right. In that example, he’ll make money if the recovery is actually 30%, and lose money if its only 5%.

Q. So, backing up, what this is doing is creating a more refined CDS market?
A. That’s a good way to look at it. You can separately trade not just whether or not a given bond will default or not, but exactly how much recovery will be paid out if there is a credit event. Needless to say, that allows a whole new range of hedging and speculation opportunities.