Re: [Quantlib-dev] Inflation curves

[Chris K. <chr...@ya...>]

Hi Simon,

some answers to your questions - let me know if they don't make sense.

date lags
- the assumption is not that inflation indexes are published on the first of a month, but that they are valid for the whole of the period that they refer to.  If they are monthly then the fixing is valid for every day of the month.  
- if you need interpolated fixings (in the past) then you must calculate this separately.

base date
- this is the date at the start of the period that the inflation term structure starts from
- since inflation indexes are lagged this date is in the past
- thus the baseDate() of an inflation term structure is always before the referenceDate() of the corresponding nominal term structure.  The referenceDate() of a nominal term structure is the date when the discount factor is 1 (one).

baseDate() versus trueBaseDate
- baseDate is the start of the last known fixing period
- trueBaseDate is the end of the last known fixing period
- the forecastFixing takes the zero inflation term structure as starting from the end of the period in which the base date falls.  N.B. there must be a fixing in the period when the base date falls, or else there is no way to forecast the index forward with the inflation term structure.  

initial values (values at base date)
- yes, these never change
- in the market, the earliest that ZCIIS and YYIIS quotes are available is one year
- hence to have a curve that goes from the base date to the first market point the user must take a view on the value at the base date (i.e. there is no market data so the user must supply a number).  If you have an inflation instrument with a payoff at less than one year you must take a view.  One view would be the last fixing (for YoY Inf) or some suitable extrapolation (for Zero Inf).  However, I decided that the creation of such a view belonged outside the class, not in it.
- clearly for non-integer years seasonality is important ... this is in progress - see other posts on -dev.


Best regards,
Chris

> From: Simon Ibbotson - Straumur <Sim...@st...>
> To: qua...@li...
> Subject: [Quantlib-dev] Inflation curves
> Date: Thu, 15 May 2008 13:28:24 -0000
> 
> Hi,
> 
>  
> 
> I’m trying to develop an inflation swap (with intermediate coupons)
> where the fixed payments are scaled by a factor (inflation index /
> base value).
> 
> At the moment, I simply cannot understand the way in which inflation
> curves & inflation indexes work in QuantLib. 
> 
>  
> 
> I understand that there has to be a distinction between inflation
> rates based upon year-on-year (either synthetic or derived from an
> index). However, I don’t understand the date lag mechanism. There are
> several assumptions in place: for instance, the inflation numbers are
> assumed to always be published on the 1st of the month. Also, in
> ZeroInflationIndex::forecastFixing, why is there a difference between
> the baseDate, the trueBaseDate, and the curve reference date?
> 
>  
> 
> Can anyone explain the reasoning behind these dates – also, why does
> the initial zero rate (the value at the base date) not change during
> the bootstrapping?
> 
>  
> 
> Cheers,
> 
>  
> 
> Simon
> 
>  
> 
>  
> 
> Simon Ibbotson
> 
> Quantitative Analytics
> 
> Capital Markets
> 
> Straumur
> 
>  
> 
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft 
> Defy all challenges. Microsoft(R) Visual Studio 2008. 
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________ QuantLib-dev mailing list Qua...@li... https://lists.sourceforge.net/lists/listinfo/quantlib-dev

-- 

Better to remain silent and be thought a fool than to speak out and 
remove all doubt. 
-- Abraham Lincoln