Posted by
Toyin Akin-4 on
Oct 22, 2002; 2:49am
URL: http://quantlib.414.s1.nabble.com/RE-Delta-on-Capfloor-tp2259p2261.html
Hi Again,
> A swaption is an option to enter into a swap at a given date.
> for example if you have the option to enter in a 5Y swap in 1Y then if you
> exercice this option once you enter in the swap there is no option
anymore.
> In 1Y you enter in a 5Y swap and whatever are the futures rates there is
no
> option anymore.
> This is not exaclty the payoff of a cap: the cap is sum( max(forward -
> strike,0)) but the payoff of a swaption is max(Swap NPV,0).
Agreed.
> Now you already know that the put-call parity between cap and floor is Cap
> = Floor + Swap, this is why the Cap delta is expressed in term of Swap
BPV.
Again agreed, however when we actually price each part of it, we price a
portfolio of caplets, floorlets and
fra's. Where the underlying of each caplet/floorlet is a FRA. It just so
happens that the sum of all these parts maps
onto the definition you have provided.
> As you said the swap 1Y is a swap with 4 FRA's but taking a swap at the
> break even (NPV=0) the swap is also:
> A Notional exchange + fix flows.
> From that you can show that the sensitivity of a swap is very close to
> (swap maturity)*discount factor.
>
> Take a 5 Years swap the sensitivity is roughtly 5*exp(-5*0.03)*notional /
> BPV = around 43 000 for a 100M 5Y swap.
Okay, my feeling was that it was something like
Sum(DayCountFraction*DiscountFactor).
Forgive me, I'm probably missing the point somewhere here, but looking at
your example again.
> Taking your example:
> Caplet 1 (6 months) has delta of 0.6 => delta* swap 6M bpv = 0.6* 0.45 *
> notional/1 BP (around that)
> Caplet 2 (9 month) has a delta of 0.5 => delta* swap 9M bpv = 0.5* 0.67 *
> notional/1BP
> Caplet 3 has a delta of 0.4 => delta* swap 1Y bpv = 0.4*0.9 * notional/
1BP
> Summing these gives 0.965*notional / 1BP that is to say almost 100% of the
> 1Y swap sensitivity isn't it?
If the caplets were deep ITM (delta of 1) you would now have 2.02*notional /
1BP.
Am I correct to read this as I would need close to 2 units of my swap to
hedge the cap where the swap is the underlying
(or the swap expressed in the call-put parity above).
If so, this does not make sense if the swap is the underlying. If the swap
is the underlying, I would need 1 unit.
Again I'm probably missing the point somewhere and I apologise for rambling
on!!
Best Regards,
Toyin Akin.
----- Original Message -----
From: <
[hidden email]>
To: "Toyin Akin" <
[hidden email]>
Cc: "'Andre Louw'" <
[hidden email]>; "Perissin Francesco"
<
[hidden email]>; "QuantlibUsers (E-mail)"
<
[hidden email]>;
<
[hidden email]>
Sent: Tuesday, October 22, 2002 9:20 AM
Subject: Re: [Quantlib-users] RE:Delta on Capfloor
>
> Hi Toyin,
> A swaption is an option to enter into a swap at a given date.
> for example if you have the option to enter in a 5Y swap in 1Y then if you
> exercice this option once you enter in the swap there is no option
anymore.
> In 1Y you enter in a 5Y swap and whatever are the futures rates there is
no
> option anymore.
> This is not exaclty the payoff of a cap: the cap is sum( max(forward -
> strike,0)) but the payoff of a swaption is max(Swap NPV,0).
>
> Now you already know that the put-call parity between cap and floor is Cap
> = Floor + Swap, this is why the Cap delta is expressed in term of Swap
BPV.
>
> I absolutely agree that the cap is the sum of caplet and each caplet has
to
> be valued separately but if you take the B&S valuation for each caplet,
the
> B&S RAW delta is between 0 and 1 but is expressed in term of Swap BPV. (at
> least this is what banks are doing in France and Germany!)
>
> For example a delta of 0.5 for one caplet is not 0.5 EUR but 0.5 EUR per
> Swap BPV. It means that from a delta of 0.5 you find the delta in currency
> multiplying the B&S delta by the swap BVP then you can sum the delta of
> each caplet and compare this value to the delta of a swap or a future or
> even a swaption if you wish. I think this is what Andre was trying to do.
>
> How I found the swap BPV: simple.
>
> As you said the swap 1Y is a swap with 4 FRA's but taking a swap at the
> break even (NPV=0) the swap is also:
> A Notional exchange + fix flows.
> From that you can show that the sensitivity of a swap is very close to
> (swap maturity)*discount factor.
>
> Take a 5 Years swap the sensitivity is roughtly 5*exp(-5*0.03)*notional /
> BPV = around 43 000 for a 100M 5Y swap.
>
> bye
> Xavier
>
>
>
>
>
> "Toyin Akin"
> <
[hidden email]> To:
<
[hidden email]>
> Sent by: cc: "'Andre
Louw'" <
[hidden email]>, "Perissin Francesco"
>
[hidden email]
<
[hidden email]>, "QuantlibUsers \(E-mail\)"
> eforge.net
<
[hidden email]>,
>
<
[hidden email]>
> Subject:
Re: [Quantlib-users] RE:Delta on Capfloor
> 21/10/2002 16:01
>
>
>
>
>
>
> Hi all,
>
> > Andre, you're right to compare the swap sensitivity to the cap's one.
> > This is the same structure: a Swap is serie of FRA and a caplet is an
> > option on FRA. Exactly like comparing the option on Equity and the
> Equity.
> > this is what traders are doing: hedging CAP/FLOOR with Futures and Swap
> or
> > the opposite.
>
> I agree that in the equity case comparing the option on Equity and the
> Equity makes sense, the underlying of the
> equity option *IS* the equity.
>
> Maybe I'm missing the point somewhere but...
>
> The underlying for a caplet is a single FRA, not a series of FRAs.
> An option on a series of FRA's is a swaption.
> Several options where the underlyings are FRA's is a cap.
>
> The underlying of a cap is not a swap.
>
> Now I haven't looked at the history of this exchange, but... if the
> question
> you are all asking is that you want to hedge the cap's delta with a swap
> and
> not why does the cap's delta's sum, sum to more than 1, then yes, you
could
> hedge with any instrument, the most liquid being FRA's, futures, swaps or,
> if one really wanted to, any option where you want to introduce
gamma/vega.
> You just need to take into account the hedge instruments sensitivity.
>
> The reason why we pick Deposit,s FRA's, futures, swaps to hedge is because
> these are the most liquid instruments and
> our yield curve is composed of these instruments. Thus all we need to do
is
> to find the right ratio to hedge.
>
> But in answer to the original question... as to why the total delta of a
> cap
> sum could be more than 1... you really have a portfolio of options.
>
> > Taking your example:
> > Caplet 1 (6 months) has delta of 0.6 => delta* swap 6M bpv = 0.6* 0.45
*
> > notional/1 BP (around that)
> > Caplet 2 (9 month) has a delta of 0.5 => delta* swap 9M bpv = 0.5* 0.67
*
> > notional/1BP
>
> > Caplet 3 has a delta of 0.4 => delta* swap 1Y bpv = 0.4*0.9 * notional/
> 1BP
> > Summing these gives 0.965*notional / 1BP that is to say almost 100% of
> the
> > 1Y swap sensitivity isn't it?
>
> I don't really have all the calculations for the above and not really sure
> how the swap bpv is calculated, but lets say we have a very in the money
> option (strikes at 0.5, not assuming JPY), assuming that the swap bpv
> remains the same above, and the deltas are all 1.0 we'll now have
> 3*notional
> / 1BP as the final value. Again (protecting myself here!!) I'm not sure
how
> you have computed the 0.45 in the 6M swaps's bpv calc above, but I assume
> it's independant of the option's strike.
>
> Also is the swap 1Y bpv an ATM swap from spot up to one year? If so you
can
> see immediatly that we are dealing
> with different structures. Caplet 3 is an option on a 3M FRA... one
fixing.
> The swap 1Y is a swap with 4 FRA's.
>
> Regards,
> Toyin Akin.
>
> ----- Original Message -----
> From: <
[hidden email]>
> To: "Toyin Akin" <
[hidden email]>
> Cc: "'Andre Louw'" <
[hidden email]>; "Perissin Francesco"
> <
[hidden email]>; "QuantlibUsers (E-mail)"
> <
[hidden email]>;
> <
[hidden email]>
> Sent: Monday, October 21, 2002 1:26 PM
> Subject: Re: [Quantlib-users] RE:Delta on Capfloor
>
>
> >
> > Hi All,
> > Andre, you're right to compare the swap sensitivity to the cap's one.
> > This is the same structure: a Swap is serie of FRA and a caplet is an
> > option on FRA. Exactly like comparing the option on Equity and the
> Equity.
> > this is what traders are doing: hedging CAP/FLOOR with Futures and Swap
> or
> > the opposite.
> >
> > Now what you forgot is to multiply each caplet sensitivity by the swap
> BPV
> > (the sensi for 1BP change in the IR curve) because the delta of a caplet
> is
> > not in % but calculated for a move of the underlying.
> >
> > Taking your example:
> > Caplet 1 (6 months) has delta of 0.6 => delta* swap 6M bpv = 0.6* 0.45
*
> > notional/1 BP (around that)
> > Caplet 2 (9 month) has a delta of 0.5 => delta* swap 9M bpv = 0.5* 0.67
*
> > notional/1BP
>
> > Caplet 3 has a delta of 0.4 => delta* swap 1Y bpv = 0.4*0.9 * notional/
> 1BP
> > Summing these gives 0.965*notional / 1BP that is to say almost 100% of
> the
> > 1Y swap sensitivity isn't it?
> >
> > Bye
> > Xavier
> >
> >
> >
> >
> > "Toyin Akin"
> > <
[hidden email]> To:
> "Perissin Francesco" <
[hidden email]>, "'Andre
> > Sent by: Louw'"
> <
[hidden email]>, "QuantlibUsers \(E-mail\)"
> >
[hidden email]
> <
[hidden email]>
> > eforge.net cc:
> > Subject:
> Re: [Quantlib-users] RE:Delta on Capfloor
> >
> > 21/10/2002 13:13
> >
> >
> >
> >
> >
> >
> > Hi again,
> >
> > In addition, a cap structure being a portfolio of caplets does not
> include
> > information regarding the correlation being
> > the individual FRA's. You'll need to look at swaptions for this.
> >
> > I may be wrong here but I wouldn't try to compare a swap's sensitivity
to
> > that of the cap's.
> > The 2 numbers really account for different measures of risk.
> >
> > compare swap sensitivity with that of swaptions
> > compare a FRA's sensitivity to that of caplet's
> > a series of FRA's sensitivity (which really is a swap) should thus be
> > compared to swaptions.
> >
> > Regards,
> > Toy.
> >
> > ----- Original Message -----
> > From: "Toyin Akin" <
[hidden email]>
> > To: "Perissin Francesco" <
[hidden email]>; "'Andre
> Louw'"
> > <
[hidden email]>; "QuantlibUsers (E-mail)"
> > <
[hidden email]>
> > Sent: Monday, October 21, 2002 11:54 AM
> > Subject: Re: [Quantlib-users] RE:Delta on Capfloor
> >
> >
> > > Hi all,
> > >
> > > Remember, a cap is a series of caplets. thus you are buying a
portfolio
> > of
> > > options.
> > > Like any portfolio of options, the sum of the deltas, which will be
> your
> > > total sensitivity if rates move can
> > > be well over 1.0. However the delta of any *ONE* caplet will be <= +-
> > 1.0.
> > >
> > > If each caplet is well in the money, then each caplet will have a
delta
> > of
> > > 1. Thus the sum will be above one.
> > >
> > > Also, a caplet is an option on an individual FRA. Thus you cannot
> really
> > > bring swaps into this.
> > > You may be confusing swaptions with caps here.
> > >
> > > A swaption is an option on a swap.
> > >
> > > Regards,
> > > Toyin Akin.
> > >
> > >
> > > ----- Original Message -----
> > > From: "Perissin Francesco" <
[hidden email]>
> > > To: "'Andre Louw'" <
[hidden email]>; "QuantlibUsers (E-mail)"
> > > <
[hidden email]>
> > > Sent: Monday, October 21, 2002 11:13 AM
> > > Subject: RE: [Quantlib-users] RE:Delta on Capfloor
> > >
> > >
> > > >
> > > > Hi,
> > > >
> > > > >The answer given originally to my question on calculating the delta
> on
> > a
> > > > >capfloor structure was that it is simply the sum of the delta's on
> all
> > > the
> > > > >caplets (assuming flat nominals, there is no weighting needed).
> > > >
> > > >
> > > > I gave you this answer assuming another definition of delta for an
> > > interest
> > > > rate instrument, i.e. the change in present value given a parallel
> > shift
> > > of
> > > > 1bp in the used curve. Also, this is quite close to the change
> obtaine
> d
> > by
> > > > shifting of 1bp the prices of the benchmarks used to bootstrap the
> > yield
> > > > curve.
> > > > This seems to be equal to what you name "sensititvity", is it right?
> > > >
> > > > Regarding the definition that you are referring to, (i.e. the cap
> > > > sensitivity divided by the IRS sensitivity) I have to think a while
> > about
> > > > it. The example could help.
> > > >
> > > >
> > > >
> > > > >e.g I have a 3Mx12M, 3 monthly cap, which results in 3 caplets,
> > results
> > > as
> > > > >follows:
> > > > >the 1st strikes at a delta of 0.6, the 2nd strikes at a delta of
> 0.5,
> > the
> > > > >3rd at a delta of 0.4. Summing these gives 1.5E
> > > >
> > > >
> > > > Just for understanding, do you mean that the first fixing is in 3
> > months
> > > ans
> > > > last fixing in 9 months? Which curr? Which strike?
> > > > In any case, it's not correct to sum the deltas (according to your
> > > > definition). I would be tempted to say that you should sum the
caplet
> > > > sensitivities in order to get the cap sensitivity. Now you could
> divide
> > > the
> > > > result by the IRS sensitivity, isn't it?
> > > > Let me know it this can help...
> > > >
> > > >
> > > > >Ok, maybe it's on the Delta with respect to spot: the above caplets
> > gives
> > > > >0.13, 0.11, 0.08, giving a total of 0.32, (doesn't 'feel' right!),
> > > > depending
> > > > >on the termstructure and the length of the cap, the sum of the
> > > spot-delta's
> > > > >quite often goes over 1.0!
> > > >
> > > >
> > > > Andre, I could not understand this... could you try to explain again
> > the
> > > > above numbers?
> > > >
> > > >
> > > > Ciao
> > > > Francesco
> > > > --
> > > > ############################### DISCLAIMER
> > > #################################
> > > >
> > > > This message (including any attachments) is confidential and
> > may
> > > be
> > > > privileged. If you have received it by mistake please notify the
> > sender
> > > by
> > > > return e-mail and delete this message from your system. Any
> > > unauthorised
> > > > use or dissemination of this message in whole or in part is
> > > strictly
> > > > prohibited. Please note that e-mails are susceptible to change.
> > Banca
> > > del
> > > > Gottardo (including its group companies) shall not be liable for
> > the
> > > > improper or incomplete transmission of the information contained
> in
> > > this
> > > > communication nor for any delay in its receipt or damage to your
> > > system.
> > > > Banca del Gottardo (or its group companies) does not guarantee that
> > the
> > > > integrity of this communication has been maintained nor
that
> > > this
> > > > communication is free of viruses, interceptions or interference.
> > > >
> > > >
> > >
> >
>
############################################################################
*************************************************************************
> > Ce message et toutes les pieces jointes (ci-apres le "message") sont
> > confidentiels et etablis a l'intention exclusive de ses destinataires.
> > Toute utilisation ou diffusion non autorisee est interdite.
> > Tout message electronique est susceptible d'alteration.
> > La Fimat et ses filiales declinent toute responsabilite au
> > titre de ce message s'il a ete altere, deforme ou falsifie.
> > ********
> > This message and any attachments (the "message") are confidential and
> > intended solely for the addressees.
> > Any unauthorised use or dissemination is prohibited.
> > E-mails are susceptible to alteration.
> > Neither Fimat nor any of its subsidiaries or affiliates shall
> > be liable for the message if altered, changed or falsified.
> >
*************************************************************************
> >
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
>
http://thinkgeek.com/sf> _______________________________________________
> Quantlib-users mailing list
>
[hidden email]
>
https://lists.sourceforge.net/lists/listinfo/quantlib-users>
>
>
>
>
> *************************************************************************
> Ce message et toutes les pieces jointes (ci-apres le "message") sont
> confidentiels et etablis a l'intention exclusive de ses destinataires.
> Toute utilisation ou diffusion non autorisee est interdite.
> Tout message electronique est susceptible d'alteration.
> La Fimat et ses filiales declinent toute responsabilite au
> titre de ce message s'il a ete altere, deforme ou falsifie.
> ********
> This message and any attachments (the "message") are confidential and
> intended solely for the addressees.
> Any unauthorised use or dissemination is prohibited.
> E-mails are susceptible to alteration.
> Neither Fimat nor any of its subsidiaries or affiliates shall
> be liable for the message if altered, changed or falsified.
> *************************************************************************
>