Hi all,I use the bootstrapping method for a while but by using the discrete compounding and also the date convention (30/360, exact/365, etc.) for eonia, futures and swaps.However, according to the Hull (Options, Futures and other derivatives 5th edition, page 96 for example) using the continuous compounding it seems that it still valid to ignore the date conventions (since we are using the continuous compounding method instead of the discrete compounding method).Also, I would like to know which method you use in your workspace. Was the continuous compounding method in Hull's books simplified by ignoring the date conventions?Thanks in advance.

I think they are different things.In continuous compounding, your discounting function would be exp(-rt).The variable t is year fraction which take 3 parameters: start date, end date and day count convention.You might also need day count convention to project the cash flow. Say in USD, your interest is calculated as Notional x Par rate x YearFrac(Start, End, 30/360).Just my 2 cents. Happy to discuss.

QuoteYou might also need day count convention to project the cash flow. Say in USD, your interest is calculated as Notional x Par rate x YearFrac(Start, End, 30/360).You might or you must? Because, as I think due to the fact it is a continuous compounding there is not approximation (and Hull does not talk about the convention date in the example I refered to).Anyway I will just still using my usual way so.

chilun is correct, they are different things. The day count determines the time. The compounding convention is an arbitrary choice given the day count (or time). You can convert a discount factor into an equivalent zero coupon rate with any compounding convention you choose. Academics typically use continuous compounding because that math is a a bit more familiar to them and also because they may not be aware of the actual market conventions, that is detail missing in Hull's otherwise fine introductory book. It is my personal preference to express zero coupon rates using the market convention for the input swaps or bonds (semi-annual or annual compound, depends on the market you are dealing with). There are better books and papers than Hull for fixed income details. Uri Ron's Bank of Canada working paper from a way's back is closer to a practitioner's implementation.There is a lot of art in curve building, especially these days. I am not aware if any texts have yet incorporated the curve building paradigm shift required since the advent of OTC derivative collateralization.

QuoteOriginally posted by: DavidJNAcademics typically use continuous compounding because that math is a a bit more familiar to them and also because they may not be aware of the actual market conventions, that is detail missing in Hull's otherwise fine introductory book. It is my personal preference to express zero coupon rates using the market convention for the input swaps or bonds (semi-annual or annual compound, depends on the market you are dealing with). There are better books and papers than Hull for fixed income details. Uri Ron's Bank of Canada working paper from a way's back is closer to a practitioner's implementation.There is a lot of art in curve building, especially these days. I am not aware if any texts have yet incorporated the curve building paradigm shift required since the advent of OTC derivative collateralization.Then we are saying the same thing. What is the way of taking into account the OTC derivative collateralization?

"What is the way of taking into account the OTC derivative collateralization?"You have a lot of research and reading to do. A great deal.

Can you recommend me some papers?

You often need to get the accrual periods, so you need the date convention, at least for calculating the fixed leg PV of your swap.

There are separate issues here:Theories and methodologies of curve building - a quick google will bring up many papers - Hagan West is a good starting term to search on.Instrument Day Basis - this is vital. You need to know the day basis, calendar and date roll conventions for all instruments you are using as observations in your curve build. A first test of any curve is to recover the observed instruments used to generate the curve - whether it is built as a set of discrete discount factors or a parametric representation of the evolution of the curve. To do this you must know the dates and cashflow amounts of the observed instruments.Collateralisation is a separate but related topic, the discount (funding) curve used will depend on the collateralisation of the trade. over this funding curve you should fit a forward curve which recovers the observed market prices of your inputs given the funding basis you are using. In a cross currency trade this becomes much more complicated as the currency of the funding curve may be different to the currency of the observed instruments used to build the forward curve.Königspudel