hey,is there a market standard on how to determine the collateral in a fully collateralized interest rate swap (for simplicity assume daily margining, 0 threshold and MTA). As I see it the (individual) value of a swap for a bank somehow depends on its funding level (as I argued here, I think I payer swap is more likely to have a positive than a negative PV in the future. Consequently, it should be more valuable (at initialization of the deal) to a bank with a high funding spread than to one with a low funding spread).So then how do 2 banks with different funding agree on the amount of collateral to be posted? Do the discount both legs with ois (simply as a convention) - after all they should pretty much agree on the bootstrapped floating forwads, shouldn't they?thanks, bernd

QuoteOriginally posted by: BerndSchmitzhey,is there a market standard on how to determine the collateral in a fully collateralized interest rate swap (for simplicity assume daily margining, 0 threshold and MTA). As I see it the (individual) value of a swap for a bank somehow depends on its funding level (as I argued here, I think I payer swap is more likely to have a positive than a negative PV in the future. Consequently, it should be more valuable (at initialization of the deal) to a bank with a high funding spread than to one with a low funding spread).So then how do 2 banks with different funding agree on the amount of collateral to be posted? Do the discount both legs with ois (simply as a convention) - after all they should pretty much agree on the bootstrapped floating forwads, shouldn't they?thanks, berndThe collateral is at a firm level with a cparty, hence I believe the differences are likely to come out in the wash except where there are significant differences in opinion. Most collateral processes are off the EOD FO valuations without any input from Finance, so the valuation is subject to change anyway. I have seen instances where errenous MTM cause the collateral people get worried and hence there has been a bit of rush to find out the true valuation on a derivative.

The problem with say a 5year swap is that the next day you cannot observe a liquid price for a (5y-1d) swap. So the two parties have to agree on a fair price. But if the price depends on firm specific properties (in this case the funding spread) this is not straightforward. As I see it they need to agree ex ante on some general pricing algorithm - like discounting all swap payments with ois ...

QuoteThe problem with say a 5year swap is that the next day you cannot observe a liquid price for a (5y-1d) swap. So the two parties have to agree on a fair price. But if the price depends on firm specific properties (in this case the funding spread) this is not straightforward. As I see it they need to agree ex ante on some general pricing algorithm - like discounting all swap payments with ois ...Ok- firstly, the next day you have a (4y 364d) swap... you roll down the maturity, not up.Secondly- two parties agreeing to a price on a 4y 364d swap (w/o funding considerations) is very simple indeed- it's basic bootstrapping (either of a single curve or a dual LIBOR/OIS curve). I don't see how obtaining FVA adjustments on a 5y swap is any more complex than obtaining FVA adjustments on a 4y 364d swap...

It all depends on the portfolio that your firm has and what kind of threshold/MTA's you're running in your CSA. Depending on what side you take up, your starting MtM under a CSA will be more or less be zero (IRS under CSA conventions are usually priced with the mid-market price, leading to 0 or close to 0 MtM), and the future development of your MtM (add-on) is dependent upon how the swap curve (or treasury yield curve) moves over time. The margin call process allows for amounts to be disputed and then the disputes to be remedied by various means, as each counterparty has it's own valuation metrics which more often than not are not 1:1 similar to each others. So I may call EUR 15 mn from you for a portfolio of receiver swaps I have with you but your collateral model tells you that you only owe me EUR 13.5 mn - we can then arrange a middle point and then agree to transfer a certain amount in order to cover a majority of the margin call.I've personally seen some interesting collateral cases, one which has stuck with me was that the wrong curve for an underlying in a structured CDS was delivered by the front office system and our counterparty's exposure and MtM went from within limits to something like a trillion euro exposure after the overnight batch, simply due to the program choosing the wrong curve, and IT not being able to deliver a respectable intraday feed.

QuoteOriginally posted by: PvalAnal85QuoteThe problem with say a 5year swap is that the next day you cannot observe a liquid price for a (5y-1d) swap. So the two parties have to agree on a fair price. But if the price depends on firm specific properties (in this case the funding spread) this is not straightforward. As I see it they need to agree ex ante on some general pricing algorithm - like discounting all swap payments with ois ...Ok- firstly, the next day you have a (4y 364d) swap... you roll down the maturity, not up.Secondly- two parties agreeing to a price on a 4y 364d swap (w/o funding considerations) is very simple indeed- it's basic bootstrapping (either of a single curve or a dual LIBOR/OIS curve). I don't see how obtaining FVA adjustments on a 5y swap is any more complex than obtaining FVA adjustments on a 4y 364d swap...Maybe I did not understand something but he said '5Y minus 1D' this is the same than your 4y 364d (well aside from the business days issues).On your second point, it is not hard but you probably have to agree in a valuation methodology (what will be the discount factors used for the valuation, what will be the boostrapping method, ... etc) or use an external/neutral package which you agreed to ?