wilmott.com

hi all,These days is it safe to assume that every bank on the street using OIS discounting as standard and when they face corporates (e.g. tesco or insurance companies) they simply add a CVA charge to it.I understand that some banks have counterparty specific discount curves where there is not collateral posting (CSA)?As a result do IR option and swap traders these days trade swaps but also trade OIS to hedge discounting risk, in order to reduce the build up of I suppose FRA-OIS basis risk?Apologies these are broad sweeping questions, and in addition to understanding the big picture, I want to get my head round the nitty gritty of actually building out a curve.I know curve construction remains a quant developed highly proprietary piece of work at each bank, but is there any book or resource that comes close to explaining the reality of constructing a curve.Everywhere I look its just high-level papers (e.g. lehman research etc. or other stuff people recommend) without the real granularity on stuff.On an even more practical level, how are swap curves stored, do you use a program, or can it be done on a spreadsheet without macros. How do you construct a curve and then actually use it as input to price a non-standard forward starting swap for example?Appreciate some help and patience please..thanks a lot

Hello Mr GammaSlide, you can see the CSA as a contract which indicates what are the rules of collateral posting when the agreed expected PV between two entities is non null.This agreement includes rules about currency and the overnight rate for the collateral postings.Using this information you have to build the corresponding discount curve in order to price/value/hedge correctly your instruments.

There are various degrees of maturity as to whether OIS discounting is taking place or not. But as Daveangel as often pointed out, traders make markets so that the issues comes to risk management and valutions rather than actually making prices. The issue is that some people are making prices blind...There are lots of papers around how to looks at OIS and FVA at the moment, just my understanding is that there is no real consensus across the market. The standard paper on constructing yield curves is one by the Candian Central bank (I seem to remember). In terms of what you use to construct a curve, that is real up to how you see the market, how liquid you think various products are etc.

thanks guys but rmax that was kinda my point in that the canadian central bank paper is like a very very high level theory look at this. Arent there any books resources which even come close to a swap traders curve building theory.and you raise the point, counterparty specific discount curves for counterparties with no CSA logically is beyond the scope of a trader, it would presumably get passed onto credit/CVA desk.the trader should only have two curves, one with libor discounting and one with ois for all counterparties with CSAs. I assumed 90% of market volume of swaps today were with counterparties using OIS discounting. (LCH switched to this two years ago??)Back to the question anyone know of any resource to learn more about the nuts and bolts of this?

anyone please?

Everything has been said, you build different curves for different LIBORs and one curve for OIS. Then you need a mechanism to either compute or simulate 1 curve for each type of CSA ...There is nothing more to add ... EDIT: I did not mean to imply any order in stripping/boostrapping in the previous sentences. It is likely for most of developed markets that you will need to solve a system of constraints to obtain each curve. If you happen to have assumptions of the market using a particular CSA in a foreign currency you will may to solve a market from another currency to obtain the FXed discount curve to price the market you are interested in ... Good luck !

gammaslide, you should look at the Bank of Canada paper by Uri Ron. It is a fairly detailed how-to manual for building the LIBOR swap curve. The only thing I can think of that is better (apart from my own private scribblings!) is the 1990 Euromoney book by Miron and Swannell.