Hi all,A question on the end of month convention for dates in the interest rates swap space. (Sorry for that, I know it's boring ! )Assume that we're in a market where 28,29 and 30/04/2012 and 01/05/2012 are non business day, and that we want to add two months to the date 29/02/2012, with the following convention : end of month (EMC for short) is "last" and date moving convention is "forward".Then 29/02/2012 + 2M = 29/04/2012 is a non business day so that so that EMC "last" should give the last business day of april 2012 (whatever the position of this last day wrt 29/04/2012 may be), which is then 27/04/2012. And the date moving convention is useless in this case.The question is : should EMC "last" give the last day of month, or the last business day of month ? (And then the result in the previous example would be 02/05/2012) Or is there a convention on what "last" for EMC should mean according to the market we're in ?Thx a lotR

As far as I'm aware the end of month convention is for generating the dates initially rather than adjusting them for holidays. So with EoM 'Last' you'd get 30/04/2012 not 29/04/2012. Then adjust that using your dayroll convention.

QuoteOriginally posted by: rockinsquatThe question is : should EMC "last" give the last day of month, or the last business day of month ? (And then the result in the previous example would be 02/05/2012) Or is there a convention on what "last" for EMC should mean according to the market we're in ?To my understanding, the end-of-month rule is "Where the start date of a period is on the final business day of a particular calendar month, the end date is on the final business day of the end month". In your case the end date would be 27-Apr-12. The business day convention (like "Following" or "Modified Following") is in that case irrelevant.I have collected a lot of those rules in a document: Interest Rate Instruments and Market Conventions Guide . It is available under a Creative Commons CC BY 3.0 (which means roughly that you can use it freely as long as you reference the original document).

Thx for your interest and response ! 29/02/2012 + 2M = 29/04/2012 when you have no conventions at all, so that you say that in EoM 'Last', the last is for last calendar (and not business) day of month, which is asfaik the opposite of usual market conventions... Here the calendar in the chinese one, and for IRS's, so that there's maybe something, but can hardly believe it....QuoteSo with EoM 'Last' you'd get 30/04/2012 not 29/04/2012

Thx for your interest and response also ! That's what I thought, yes. Do you know if there are exceptions to this most common interpretation of EoM Last, and where the last means "last calendar day" and not "last business day" ? Here I'm working with chinese NDIRS's.QuoteOriginally posted by: mathmarcQuoteOriginally posted by: rockinsquatThe question is : should EMC "last" give the last day of month, or the last business day of month ? (And then the result in the previous example would be 02/05/2012) Or is there a convention on what "last" for EMC should mean according to the market we're in ?To my understanding, the end-of-month rule is "Where the start date of a period is on the final business day of a particular calendar month, the end date is on the final business day of the end month". In your case the end date would be 27-Apr-12. The business day convention (like "Following" or "Modified Following") is in that case irrelevant.I have collected a lot of those rules in a document: Interest Rate Instruments and Market Conventions Guide . It is available under a Creative Commons CC BY 3.0 (which means roughly that you can use it freely as long as you reference the original document).

QuoteOriginally posted by: mathmarcQuoteOriginally posted by: rockinsquatThe question is : should EMC "last" give the last day of month, or the last business day of month ? (And then the result in the previous example would be 02/05/2012) Or is there a convention on what "last" for EMC should mean according to the market we're in ?To my understanding, the end-of-month rule is "Where the start date of a period is on the final business day of a particular calendar month, the end date is on the final business day of the end month". In your case the end date would be 27-Apr-12. The business day convention (like "Following" or "Modified Following") is in that case irrelevant.I have collected a lot of those rules in a document: Interest Rate Instruments and Market Conventions Guide . It is available under a Creative Commons CC BY 3.0 (which means roughly that you can use it freely as long as you reference the original document).Unfortunately this definition is subtly wrong, although with dayroll conventions of Modified Following or Preceding it would come to the same thing. EoM Last means that when the unadjusted start date is the last calendar day of the month then the unadjusted end date is also the last (calendar) of the month. It's got nothing to do with holiday calendars or dayroll conventions.So in the example given we have a start date of 29/2/2012 and EoM true (or "Last"). That means we ignore the 29 and just note it's the last day of the month and generate 2 months on as 30/4/2012.Now we find that 30/4/2012 is a holiday and so adjust it using the dayroll convention - in this case following - to 2/5/2012.When you (rockinsquat) claim it's "the opposite of usual market conventions" it's worrying as it suggests you haven't understood my explanation properly. It's not the opposite of anything. It's a method of generating (unadjusted) dates by just using the last day in the month.

I never wrote anything about adjustment. I thought that the market definition was, in case of eom, starting from a date D_1 to first do a "+2 then mod 12 then +1" on the month number of D_1 to obtain a date D_2, and then apply the eom to D_2 by replacing it by the last day of month, definition of last that may depend on market convention.QuoteOriginally posted by: MattFWhen you (rockinsquat) claim it's "the opposite of usual market conventions" it's worrying as it suggests you haven't understood my explanation properly. It's not the opposite of anything. It's a method of generating (unadjusted) dates by just using the last day in the month.Now yourQuoteOriginally posted by: MattFEoM Last means that when the unadjusted start date is the last calendar day of the month then the unadjusted end date is also the last (calendar) of the month. It's got nothing to do with holiday calendars or dayroll conventions.is clear, as opposed to your earlierQuoteOriginally posted by: MattFAs far as I'm aware the end of month convention is for generating the dates initially rather than adjusting them for holidays.Thank you.

QuoteOriginally posted by: MattFQuoteOriginally posted by: mathmarcTo my understanding, the end-of-month rule is "Where the start date of a period is on the final business day of a particular calendar month, the end date is on the final business day of the end month". In your case the end date would be 27-Apr-12. The business day convention (like "Following" or "Modified Following") is in that case irrelevant.Unfortunately this definition is subtly wrong, although with dayroll conventions of Modified Following or Preceding it would come to the same thing. EoM Last means that when the unadjusted start date is the last calendar day of the month then the unadjusted end date is also the last (calendar) of the month. It's got nothing to do with holiday calendars or dayroll conventions.The definition I quoted is the BBA Libor one (under their Fixing, Value and Maturity page): "Where a deposit is made on the final business day of a particular calendar month, the maturity of the deposit shall be on the final business day of the month in which it matures (not the corresponding date in the month of maturity). Or in other words, in line with market convention, bbalibor rates are dealt on an end-end basis."If you have another source with a different definition, I would be happy to refer to it in a subsequent version of our guide.

QuoteOriginally posted by: mathmarcThe definition I quoted is the BBA Libor one (under their Fixing, Value and Maturity page): "Where a deposit is made on the final business day of a particular calendar month, the maturity of the deposit shall be on the final business day of the month in which it matures (not the corresponding date in the month of maturity). Or in other words, in line with market convention, bbalibor rates are dealt on an end-end basis."I believe my definition is the FpML/ISDA way which allows complete separation of dayroll and business day conventions for EndOfMonth however I can't trawl through the documents now as both organisations seem to have made it extremely difficult to download anything of use from their websites.The definition you cite is fine for the unadjusted dates but they haven't yet defined what we have to do when the end-of-month is not a working day. Here it comes (from the same page): "If we then added the fact that February 28th was a weekend, all the maturity dates above would roll back to the latest available business day in February." So you see that adjusting for non-business days is not really part of the EoM definition. The BBA are using EoM plus a "Preceding" business day convention. You are essentially embedding that Preceding convention in your definition of EoM and that logically forces you to declare the business day convention "irrelevant".

If one speaks about business days ("...maturity of the deposit shall be on the final business day of the month in which it matures"), how could it speak about unadjusted days ? ("unadjusted dates but they haven't yet defined what we have to do when the end-of-month is not a working day.") Working day and business day to me, is the same

MattF is correct.End of Month (roll) convention tells you how to generate the unadjusted calculation period dates on swap leg. The confusion arises when first regular period start date or last regular period end date is on the end of a month. A roll convention decides, either it is a specific day of the month, say 30, in which case the calculation period dates are on the 30th of each month, except february where you are forced to use the last day of the month, or it is EOM roll convention, in which case, every calculation period start and end date is the last calendar in the month. After generating the calculation period unadjusted dates, they are then adjusted according to a holiday calendar and business day convention, often modified following.Marc's point about LIBOR End-End convention is technically correct, but I would not confuse End of Month roll convention on swap schedules which determine calculation periods with the determination of the LIBOR forward start and end dates. The two are separate, first one constructs the swap legs calculation period dates, adjusted them for holidays, then one starts to look at LIBOR conventions, to determined the fixing date and the forward start and end dates.So maybe the original question needs to be slightly more specific, are you trying to determine the calculation periods on the accrual schedule or are you trying to figure the forward start and end dates associated with a LIBOR fixing on a swap leg accrual schedule.