Upfront Fees vs. Periodic Fees in credit derivatives

[Nonius]

So, in some sense banks, when dealing interbank, should be sort of indifferent between the two.  The periodic fee should have a credit risky pv of what the upfront fee is.  Funny how banks probably do not account for the default probability of the bank paying the periodic fee.  or that is how I see it.



Now, zoom to Nonius Re.  Nonius Re would rather have cashish upfront.  Most banks probably would not want to give cash up front to Nonius Re, since, from their perspective, there is less exposure to Nonius Re if they pay over time.  If Nonius Re blows up, then, the MtM of the insurance contract has less value since it is the net of the value of the protection and the value of the future fees owed to Nonius Re.



So, how to get around this?  One thing I schemed about was setting up an "Up Front Fee" facility.  What this facility would do would it would "fund" the upfront fee so that the bank could pay the periodic fees to the facility. The facility, in turn, would issue notes backed by the banks fees and would do repo.  The notes would be highly rated because there would be essentially no risk (except for minimal counterparty risk on repo under margin calls) on the principal.  The only risk would be the risk of default of the bank paying the fee, in which case the notes would unwind and pay back par OR, the default of the reference obligors, in which case the bank stops paying the fees and, again, the notes unwind and pay back par.



what do you guys think?

[FDAXHunter]

I don't think you'll be able to do a lump sum fee, as for this particular deal the default probabilities of the reference obligors and Nonius Re are highly correlated.



So, for that your PVed fee stream would be your best bet, I bet.



However, I don't understand this part: " in which case the bank stops paying the fees and, again, the notes unwind and pay back par"

How can you pay back par if you PV'ed the total stream of fees and the bank stops paying the fees halfway through? You'd end up short a bit, no? What am I missing?

[Nonius]

stupid thought, forget it the lame idea.

[Nonius]

fuk it.  It won't work.



In fact, it would be like getting loan backed by your salary stream.....

[kr]

I don't see why this has to be a big deal - as far as I'm concerned, it's an accounting issue, really the counterparty risk on the fees is a small number.  You bill them upfront, they expense the insurance premium over time, looks no different.  They do not have problems scrounging up the actual money.  How they handle N.Re's ability to provide insurance services is up to their risk management department, and if they don't consult their own RM in doing the deals to begin with, they are stupid.  I mean, I can collateralize my own little insurance pool right here with $5 and a pile of pocket lint, if they believe I can insure $10MM Ford at marketrate, that's their issue.

[Johnny]

It comes back to the same place as all the rest of these discussions. It's a matter of people accepting the credit risk of NoniusRe, which in turn depends on (1) the risks taken by NoniusRe and (2) the strength of the balance sheet of NoniusRe. As ever the answers lie in insurance.



... ... ...

[Martingale]

I have seen something similar, usually we charge some fee ( premium ) based on unpaid balance, which is IO strip, then some times we have some clients paying upfront fees instead ( and I heard it is not a small portion either).

[Nonius]

KR, depends on the size.  10bps on 5Bill of derivatives that have an average life of 10 years is 25MM upfront, assuming a credit risk PVO1 of 5.  Compare 25MM current exposure to Nonius Re with 0 current exposure to Nonius Re.  It definitely makes a difference.  The 5Bill type trades on ports of lots of names at a time would be the most interesting.  If the excess spread on that, ie, my scalp is 2bps, that is 5MM in my pocket on PV basis.  This, to me, is why the business could be interesting.  Do 4 of those trades a year, that is, locate 4 banks that are willing to do it, then, that ain't bad money...on a pv basis.  Of course, I think 2 bp scalp is conservative.

[kr]

You're still talking about 25MM against what should be AAA risk to them... that's not really an inconvenient number is it?  Worst case scenario, they buy 25MM credit protection on Ace, might cost them 10y x 50bps = 1.25MM.  Keep it in perspective - these guys are buying perhaps 500MM of insurance for 10 years, that's nontrivial.

[Nonius]

Tell you the truth, I ain't sure we would give Ace 25MM bucks at this point.  AAA and yet, if Ace and/or any of the monolines blow up....I'm heading for the hills.



but, I see your point, if Nonius Re has a AAA wrap, then, what the fuk?  AAA risk......