Remember how a week ago people went around bothering themselves about Bank of America's derivatives? Specifically how if they get downgraded, as seems plausible, they will have to come up with a zillion more dollars for derivative collateral? And how earlier this week they did the same for Morgan Stanley?
Anyway we talked about it a bit and I put up a table that I figured I'd update when it was complete and now it is so here it is. Also a JPMorgan downgrade, which looked hilariously unlikely 25 hours ago, looks more likely so I guess this is relevant even where it wasn't before. So here is how much cash various banks will need to stump up - to post as collateral on OTC derivatives or to clearinhouses, or to pay on termination of trades - if they are downgraded two notches:
It is ... am I right in thinking it is harder to get worked up about these numbers today than it was 25 hours ago? I feel like they are interesting - I'd highlight the second-to-last line as a rough-and-ready indicator of something between "how strong counterparties think these banks are" and "how tough these banks are as negotiators looking out for their own liquidity"; lower is better and I suspect the differences do tell you something. But the total number here, $23-ish billion split among five banks if they're all downgraded two notches, doesn't average out to that much more of a liquidity drain than the one that JPMorgan is so manfully taking as we speak.
Just for fun though let's calculate the "cost" of posting all that cash, in terms of foregone earnings rather than just foregone liquidity. Here's one way to do it: treat it as a DVA loss. You go from having uncollateralized exposure at whatever your credit spread is, to having collateralized exposure at a +0 credit spread, so you've lost your credit spread times the duration of your derivatives. Here are some imagined numbers:
These numbers don't seem that intimidating when you're busy wrestling a whale carcass to your boat, though you might gulp hard at that MS number. (They also include numbers for a three-notch downgrade in their Q. Enjoy that!) They are also a lot more predictable, insofar as everyone knows there is some likelihood of these downgrades coming, and everyone knows roughly what the costs will be. Unless, of course, someone discovers that their derivative valuation model was wrong and they actually have bigger liabilities - and thus bigger triggered collateral posting obligations - than they thought. But what are the odds of that?