QuantLib › quantlib-users

UK Holiday for USDLibor/PiecewiseYieldCurve

Classic

List

Threaded

12 messages Options

George Cowie

UK Holiday for USDLibor/PiecewiseYieldCurve

Hello,

I'm using PiecewiseYieldCurve to build a YieldTermStructure which is ultimately used by BlackScholesMertonProcess in my option pricing. I believe I'm running into an issue on days that are holidays in the UK (but not the US). Here's the error I'm seeing when trying to run the code below for May 5, 2014 (Early May Bank Holiday):

terminate called after throwing an instance of 'QuantLib::Error'

what(): 1st iteration: failed at 1st alive instrument, maturity May 14th, 2014, reference date May 7th, 2014: Missing no-fix1W Actual/360 fixing for May 2nd, 2014

Aborted

Everything seems to work fine on other days... Am I doing something wrong with setting up the PiecewiseYieldCurve?

Thanks,

George

#include <ql/quantlib.hpp>

using namespace std ;

using namespace QuantLib ;

int main(int argc, char** argv)

{

vector<Rate> liborRates ;

liborRates.push_back(0.12075 / 100.0) ; //1W

liborRates.push_back(0.1515 / 100.0) ; //1M

liborRates.push_back(0.1935 / 100.0) ; //2M

liborRates.push_back(0.22535 / 100.0) ; //3M

liborRates.push_back(0.323 / 100.0) ; //6M

liborRates.push_back(0.549 / 100.0) ; //12M

vector<Period> periods ;

periods.push_back(Period(1,Weeks)) ;

periods.push_back(Period(1,Months)) ;

periods.push_back(Period(2,Months)) ;

periods.push_back(Period(3,Months)) ;

periods.push_back(Period(6,Months)) ;

periods.push_back(Period(12,Months)) ;

vector<boost::shared_ptr<RateHelper> > instruments ;

for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)

{

boost::shared_ptr<Quote> liborQuote(new SimpleQuote(liborRates[i])) ;

Handle<Quote> liborHandle(liborQuote) ;

boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods[i])) ;

boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper(liborHandle,liborIndex)) ;

instruments.push_back(rateHelper) ;

}

Calendar calendar = UnitedStates();

Date today (5 , May ,2014);

Settings::instance().evaluationDate() = today ;

Natural settlementDays = 2;

Date settlement = calendar.advance ( today , settlementDays , Days );

DayCounter dc = Actual360 ();

Date maturity(20, Jun, 2014) ;

double strikePrice = 100.0 ;

Option::Type optionType = Option::Call ;

double dividendYield = 0.0 ;

double impliedVolatility = 0.15 ;

double underlyingPrice = 100.0 ;

Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new SimpleQuote(underlyingPrice)));

Handle<YieldTermStructure> rateTermStructure(boost::shared_ptr<YieldTermStructure>(new PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments, dc))) ;

Handle<YieldTermStructure>

dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new FlatForward(today,

dividendYield,

dc)));

Handle<BlackVolTermStructure>

volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new BlackConstantVol(today,

calendar,

impliedVolatility,

dc)));

boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new BlackScholesMertonProcess(underlyingHandle,

dividendTermStructure,

rateTermStructure,

volTermStructure));

boost::shared_ptr<Exercise> americanExercise(new AmericanExercise(today, maturity)) ;

boost::shared_ptr<StrikedTypePayoff> payoff(new PlainVanillaPayoff(optionType, strikePrice));

VanillaOption americanOption(payoff, americanExercise);

americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(

new FDAmericanEngine<CrankNicolson>(bsmProcess,

101,100)));

cout << americanOption.NPV() << endl ;

return 0 ;

}

------------------------------------------------------------------------------
Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds
_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

Andres Hernandez

Re: UK Holiday for USDLibor/PiecewiseYieldCurve

Hi George,

the problem occurs because you are trying to bootstrap on a UnitedKingdom
holiday, which is implicitly included in the definition of USDLibor. The
Libor class makes a joint calendar between the given financial center
calendar and the UK calendar. Since Libor is a London poll, there are only
fixings for it on UK business days. If one wants to bootstrap the curve on
a holiday, the indice's historical fixings (actually just the closest
fixing to the given date) need to be provided.

Mit freundlichen Grüßen / Kind regards

Dr. Andres Hernandez

Senior Financial Engineer
Business Analytics
Risk Analytics

Phone: +49-69-6645-1351 IBM Deutschland (Embedded
image moved
to file:
pic65048.gif)

Mobile: +49-163-609-3280 Wilhelm-Fay-Str. 30-34

E-Mail: [hidden email] 65936 Frankfurt am Main

Germany

IBM Deutschland
GmbH /
Vorsitzender des
Aufsichtsrats:
Martin Jetter
Geschäftsführung:
Martina Koederitz
(Vorsitzende),
Reinhard Reschke,
Dieter Scholz,
Gregor Pillen,
Christian Noll,
Ivo Koerner
Sitz der
Gesellschaft:
Ehningen /
Registergericht:
Amtsgericht
Stuttgart, HRB
14562 /
WEEE-Reg.-Nr. DE
99369940

From: George Cowie <[hidden email]>
To: QuantLib users <[hidden email]>
Date: 22/07/2014 14:10
Subject: [Quantlib-users] UK Holiday for USDLibor/PiecewiseYieldCurve

Hello,

I'm using PiecewiseYieldCurve to build a YieldTermStructure which is
ultimately used by BlackScholesMertonProcess in my option pricing. I
believe I'm running into an issue on days that are holidays in the UK (but
not the US). Here's the error I'm seeing when trying to run the code below
for May 5, 2014 (Early May Bank Holiday):

terminate called after throwing an instance of 'QuantLib::Error'
what(): 1st iteration: failed at 1st alive instrument, maturity May
14th, 2014, reference date May 7th, 2014: Missing no-fix1W Actual/360
fixing for May 2nd, 2014
Aborted

Everything seems to work fine on other days... Am I doing something
wrong with setting up the PiecewiseYieldCurve?

Thanks,
George

#include <ql/quantlib.hpp>

using namespace std ;
using namespace QuantLib ;

int main(int argc, char** argv)
{
vector<Rate> liborRates ;
liborRates.push_back(0.12075 / 100.0) ; //1W
liborRates.push_back(0.1515 / 100.0) ; //1M
liborRates.push_back(0.1935 / 100.0) ; //2M
liborRates.push_back(0.22535 / 100.0) ; //3M
liborRates.push_back(0.323 / 100.0) ; //6M
liborRates.push_back(0.549 / 100.0) ; //12M

vector<Period> periods ;
periods.push_back(Period(1,Weeks)) ;
periods.push_back(Period(1,Months)) ;
periods.push_back(Period(2,Months)) ;
periods.push_back(Period(3,Months)) ;
periods.push_back(Period(6,Months)) ;
periods.push_back(Period(12,Months)) ;

vector<boost::shared_ptr<RateHelper> > instruments ;

for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)
{
boost::shared_ptr<Quote> liborQuote(new SimpleQuote(liborRates[i])) ;
Handle<Quote> liborHandle(liborQuote) ;
boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods[i])) ;
boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper
(liborHandle,liborIndex)) ;
instruments.push_back(rateHelper) ;
}

Calendar calendar = UnitedStates();
Date today (5 , May ,2014);
Settings::instance().evaluationDate() = today ;
Natural settlementDays = 2;
Date settlement = calendar.advance ( today , settlementDays , Days );
DayCounter dc = Actual360 ();

Date maturity(20, Jun, 2014) ;
double strikePrice = 100.0 ;
Option::Type optionType = Option::Call ;
double dividendYield = 0.0 ;
double impliedVolatility = 0.15 ;
double underlyingPrice = 100.0 ;

Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new SimpleQuote
(underlyingPrice)));

Handle<YieldTermStructure> rateTermStructure
(boost::shared_ptr<YieldTermStructure>(new
PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments, dc))) ;

Handle<YieldTermStructure>
dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new
FlatForward(today,
dividendYield,
dc)));
Handle<BlackVolTermStructure>
volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
BlackConstantVol(today,
calendar,
impliedVolatility,
dc)));

boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
BlackScholesMertonProcess(underlyingHandle,
dividendTermStructure,
rateTermStructure,
volTermStructure));

boost::shared_ptr<Exercise> americanExercise(new AmericanExercise(today,
maturity)) ;
boost::shared_ptr<StrikedTypePayoff> payoff(new PlainVanillaPayoff
(optionType, strikePrice));

VanillaOption americanOption(payoff, americanExercise);

americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(
new FDAmericanEngine<CrankNicolson>(bsmProcess,
101,100)));

cout << americanOption.NPV() << endl ;

return 0 ;
}
------------------------------------------------------------------------------

Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

------------------------------------------------------------------------------
Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds
_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

pic65048.gif (2K) Download Attachment

George Cowie

Re: UK Holiday for USDLibor/PiecewiseYieldCurve

Thank you Andres. What you're describing makes sense, but I can't seem to avoid an error when trying to value the option on May 5, 2014. I've copied the updated code snippet with changed section highlighted/bolded below. I seem to always get an exception of the form:

terminate called after throwing an instance of 'QuantLib::Error'

what(): 1st iteration: failed at 1st alive instrument, maturity May 14th, 2014, reference date May 6th, 2014: Missing no-fix1W Actual/360 fixing for May 2nd, 2014

Is there something wrong with my approach here?

Thanks,
George

#include <ql/quantlib.hpp>

using namespace std ;

using namespace QuantLib ;

int main(int argc, char** argv)

{

vector<Rate> liborRates ;

liborRates.push_back(0.12075 / 100.0) ; //1W

liborRates.push_back(0.1515 / 100.0) ; //1M

liborRates.push_back(0.1935 / 100.0) ; //2M

liborRates.push_back(0.22535 / 100.0) ; //3M

liborRates.push_back(0.323 / 100.0) ; //6M

liborRates.push_back(0.549 / 100.0) ; //12M

vector<Period> periods ;

periods.push_back(Period(1,Weeks)) ;

periods.push_back(Period(1,Months)) ;

periods.push_back(Period(2,Months)) ;

periods.push_back(Period(3,Months)) ;

periods.push_back(Period(6,Months)) ;

periods.push_back(Period(12,Months)) ;

vector<boost::shared_ptr<RateHelper> > instruments ;

for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)

{

boost::shared_ptr<Quote> liborQuote(new SimpleQuote(liborRates[i])) ;

Handle<Quote> liborHandle(liborQuote) ;

boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods[i])) ;

boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper(liborHandle,liborIndex)) ;

instruments.push_back(rateHelper) ;

}

Calendar calendar = UnitedStates();

Date today (5 , May ,2014);

Date liborDate (2, May, 2014) ;

Settings::instance().evaluationDate() = today ;

Natural settlementDays = 2;

Date settlement = calendar.advance ( liborDate , settlementDays , Days );

DayCounter dc = Actual360 ();

Date maturity(20, Jun, 2014) ;

double strikePrice = 100.0 ;

Option::Type optionType = Option::Call ;

double dividendYield = 0.0 ;

double impliedVolatility = 0.15 ;

double underlyingPrice = 100.0 ;

Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new SimpleQuote(underlyingPrice)));

Handle<YieldTermStructure> rateTermStructure(boost::shared_ptr<YieldTermStructure>(new PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments, dc))) ;

Handle<YieldTermStructure>

dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new FlatForward(today,

dividendYield,

dc)));

Handle<BlackVolTermStructure>

volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new BlackConstantVol(today,

calendar,

impliedVolatility,

dc)));

boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new BlackScholesMertonProcess(underlyingHandle,

dividendTermStructure,

rateTermStructure,

volTermStructure));

boost::shared_ptr<Exercise> americanExercise(new AmericanExercise(today,

maturity)) ;

boost::shared_ptr<StrikedTypePayoff> payoff(new PlainVanillaPayoff(optionType, strikePrice));

VanillaOption americanOption(payoff, americanExercise);

americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(

new FDAmericanEngine<CrankNicolson>(bsmProcess,

101,100)));

cout << americanOption.NPV() << endl ;

return 0 ;

}

On Tue, Jul 22, 2014 at 7:39 AM, Andres Hernandez <[hidden email]> wrote:

Hi George,

the problem occurs because you are trying to bootstrap on a UnitedKingdom
holiday, which is implicitly included in the definition of USDLibor. The
Libor class makes a joint calendar between the given financial center
calendar and the UK calendar. Since Libor is a London poll, there are only
fixings for it on UK business days. If one wants to bootstrap the curve on
a holiday, the indice's historical fixings (actually just the closest
fixing to the given date) need to be provided.

Mit freundlichen Grüßen / Kind regards

Dr. Andres Hernandez

Senior Financial Engineer
Business Analytics
Risk Analytics

Phone: <a href="tel:%2B49-69-6645-1351" value="+496966451351">+49-69-6645-1351 IBM Deutschland (Embedded
image moved
to file:
pic20635.gif)

Mobile: <a href="tel:%2B49-163-609-3280" value="+491636093280">+49-163-609-3280 Wilhelm-Fay-Str. 30-34

E-Mail: [hidden email] 65936 Frankfurt am Main

Germany

IBM Deutschland
GmbH /
Vorsitzender des
Aufsichtsrats:
Martin Jetter
Geschäftsführung:
Martina Koederitz
(Vorsitzende),
Reinhard Reschke,
Dieter Scholz,
Gregor Pillen,
Christian Noll,
Ivo Koerner
Sitz der
Gesellschaft:
Ehningen /
Registergericht:
Amtsgericht
Stuttgart, HRB
14562 /
WEEE-Reg.-Nr. DE
99369940

From: George Cowie <[hidden email]>
To: QuantLib users <[hidden email]>
Date: 22/07/2014 14:10
Subject: [Quantlib-users] UK Holiday for USDLibor/PiecewiseYieldCurve

Hello,

I'm using PiecewiseYieldCurve to build a YieldTermStructure which is
ultimately used by BlackScholesMertonProcess in my option pricing. I
believe I'm running into an issue on days that are holidays in the UK (but
not the US). Here's the error I'm seeing when trying to run the code below
for May 5, 2014 (Early May Bank Holiday):

terminate called after throwing an instance of 'QuantLib::Error'
what(): 1st iteration: failed at 1st alive instrument, maturity May
14th, 2014, reference date May 7th, 2014: Missing no-fix1W Actual/360
fixing for May 2nd, 2014
Aborted

Everything seems to work fine on other days... Am I doing something
wrong with setting up the PiecewiseYieldCurve?

Thanks,
George

#include <ql/quantlib.hpp>

using namespace std ;
using namespace QuantLib ;

int main(int argc, char** argv)
{
vector<Rate> liborRates ;
liborRates.push_back(0.12075 / 100.0) ; //1W
liborRates.push_back(0.1515 / 100.0) ; //1M
liborRates.push_back(0.1935 / 100.0) ; //2M
liborRates.push_back(0.22535 / 100.0) ; //3M
liborRates.push_back(0.323 / 100.0) ; //6M
liborRates.push_back(0.549 / 100.0) ; //12M

vector<Period> periods ;
periods.push_back(Period(1,Weeks)) ;
periods.push_back(Period(1,Months)) ;
periods.push_back(Period(2,Months)) ;
periods.push_back(Period(3,Months)) ;
periods.push_back(Period(6,Months)) ;
periods.push_back(Period(12,Months)) ;

vector<boost::shared_ptr<RateHelper> > instruments ;

for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)
{
boost::shared_ptr<Quote> liborQuote(new SimpleQuote(liborRates[i])) ;
Handle<Quote> liborHandle(liborQuote) ;
boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods[i])) ;
boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper
(liborHandle,liborIndex)) ;
instruments.push_back(rateHelper) ;
}

Calendar calendar = UnitedStates();
Date today (5 , May ,2014);
Settings::instance().evaluationDate() = today ;
Natural settlementDays = 2;
Date settlement = calendar.advance ( today , settlementDays , Days );
DayCounter dc = Actual360 ();

Date maturity(20, Jun, 2014) ;
double strikePrice = 100.0 ;
Option::Type optionType = Option::Call ;
double dividendYield = 0.0 ;
double impliedVolatility = 0.15 ;
double underlyingPrice = 100.0 ;

Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new SimpleQuote
(underlyingPrice)));

Handle<YieldTermStructure> rateTermStructure
(boost::shared_ptr<YieldTermStructure>(new
PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments, dc))) ;

Handle<YieldTermStructure>
dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new
FlatForward(today,
dividendYield,
dc)));
Handle<BlackVolTermStructure>
volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
BlackConstantVol(today,
calendar,
impliedVolatility,
dc)));

boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
BlackScholesMertonProcess(underlyingHandle,
dividendTermStructure,
rateTermStructure,
volTermStructure));

boost::shared_ptr<Exercise> americanExercise(new AmericanExercise(today,
maturity)) ;
boost::shared_ptr<StrikedTypePayoff> payoff(new PlainVanillaPayoff
(optionType, strikePrice));

VanillaOption americanOption(payoff, americanExercise);

americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(
new FDAmericanEngine<CrankNicolson>(bsmProcess,
101,100)));

cout << americanOption.NPV() << endl ;

return 0 ;
}

------------------------------------------------------------------------------

Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
<a href="http://p.sf.net/sfu/bds_______________________________________________ QuantLib-users" target="_blank">http://p.sf.net/sfu/bds_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

Andres Hernandez

Re: UK Holiday for USDLibor/PiecewiseYieldCurve

Hi George,

you are still setting the evaluation date to a holiday:

Date today (5 , May ,2014);
  Date liborDate (2, May, 2014) ;
Settings::instance().evaluationDate() = today ;

Let's follow the logic:
During bootstrapping, the method impliedQuote() of the rate helper is
called to compare against the market quote. The DepositRateHelper simply
returns the fixing corresponding to the fixingDate_ as a return value.
The fixingDate_ is set during initialization as the return value of the
function InterestRate::fixingDate(earliestDate_). The function
InterestRate::fixingDate sets back the date given as input parameter by the
number of fixing days according to the index' calendar.
The parameter earliestDate_ being passed to fixingDate is set during
initialization to the evaluation date (which you set to the 5th of May)
moved forward by the number of fixing days according to the index'
calendar.
Note that if the evaluation date is a holiday you do not end up with the
same date, but rather with the last business day prior to evaluation date.
The first act moves the date from the 5th of May to the 7th and checks
whether it is a business day, which it is. The second move sets it back two
days to the 5th, and checks whether it is a business day, which it isn't,
so it sets it back until it finds a business day, i.e. the 2nd of May.
Since the value is in the past, it requests a historical fixing, which you
did not provide and so the QL_REQUIRE throws.

Mit freundlichen Grüßen / Kind regards

Dr. Andres Hernandez

Senior Financial Engineer
Business Analytics
Risk Analytics

Phone: +49-69-6645-1351 IBM Deutschland (Embedded
image moved
to file:
pic64586.gif)

Mobile: +49-163-609-3280 Wilhelm-Fay-Str. 30-34

E-Mail: [hidden email] 65936 Frankfurt am Main

Germany

IBM Deutschland
GmbH /
Vorsitzender des
Aufsichtsrats:
Martin Jetter
Geschäftsführung:
Martina Koederitz
(Vorsitzende),
Reinhard Reschke,
Dieter Scholz,
Gregor Pillen,
Christian Noll,
Ivo Koerner
Sitz der
Gesellschaft:
Ehningen /
Registergericht:
Amtsgericht
Stuttgart, HRB
14562 /
WEEE-Reg.-Nr. DE
99369940

From: George Cowie <[hidden email]>
To: Andres Hernandez/Germany/IBM@IBMDE
Cc: QuantLib users <[hidden email]>
Date: 23/07/2014 11:51
Subject: Re: [Quantlib-users] UK Holiday for
USDLibor/PiecewiseYieldCurve

Thank you Andres. What you're describing makes sense, but I can't seem to
avoid an error when trying to value the option on May 5, 2014. I've copied
the updated code snippet with changed section highlighted/bolded below. I
seem to always get an exception of the form:

terminate called after throwing an instance of 'QuantLib::Error'
what(): 1st iteration: failed at 1st alive instrument, maturity May
14th, 2014, reference date May 6th, 2014: Missing no-fix1W Actual/360
fixing for May 2nd, 2014

Is there something wrong with my approach here?

Thanks,
George

#include <ql/quantlib.hpp>

using namespace std ;
using namespace QuantLib ;

int main(int argc, char** argv)
{
vector<Rate> liborRates ;
liborRates.push_back(0.12075 / 100.0) ; //1W
liborRates.push_back(0.1515 / 100.0) ; //1M
liborRates.push_back(0.1935 / 100.0) ; //2M
liborRates.push_back(0.22535 / 100.0) ; //3M
liborRates.push_back(0.323 / 100.0) ; //6M
liborRates.push_back(0.549 / 100.0) ; //12M

vector<Period> periods ;
periods.push_back(Period(1,Weeks)) ;
periods.push_back(Period(1,Months)) ;
periods.push_back(Period(2,Months)) ;
periods.push_back(Period(3,Months)) ;
periods.push_back(Period(6,Months)) ;
periods.push_back(Period(12,Months)) ;

vector<boost::shared_ptr<RateHelper> > instruments ;

for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)
{
boost::shared_ptr<Quote> liborQuote(new SimpleQuote(liborRates[i])) ;
Handle<Quote> liborHandle(liborQuote) ;
boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods[i])) ;
boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper
(liborHandle,liborIndex)) ;
instruments.push_back(rateHelper) ;
}

Calendar calendar = UnitedStates();
Date today (5 , May ,2014);
  Date liborDate (2, May, 2014) ;
Settings::instance().evaluationDate() = today ;
Natural settlementDays = 2;
Date settlement = calendar.advance ( liborDate , settlementDays , Days );
DayCounter dc = Actual360 ();

Date maturity(20, Jun, 2014) ;
double strikePrice = 100.0 ;
Option::Type optionType = Option::Call ;
double dividendYield = 0.0 ;
double impliedVolatility = 0.15 ;
double underlyingPrice = 100.0 ;

Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new SimpleQuote
(underlyingPrice)));

Handle<YieldTermStructure> rateTermStructure
(boost::shared_ptr<YieldTermStructure>(new
PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments, dc))) ;

Handle<YieldTermStructure>
dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new
FlatForward(today,
   dividendYield,
   dc)));
Handle<BlackVolTermStructure>
volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
BlackConstantVol(today,
   calendar,
   impliedVolatility,
   dc)));

boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
BlackScholesMertonProcess(underlyingHandle,
dividendTermStructure,
rateTermStructure,
volTermStructure));

boost::shared_ptr<Exercise> americanExercise(new AmericanExercise(today,
   maturity)) ;
boost::shared_ptr<StrikedTypePayoff> payoff(new PlainVanillaPayoff
(optionType, strikePrice));

VanillaOption americanOption(payoff, americanExercise);

americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(
   new FDAmericanEngine<CrankNicolson>(bsmProcess,
  101,100)));

cout << americanOption.NPV() << endl ;

return 0 ;
}

On Tue, Jul 22, 2014 at 7:39 AM, Andres Hernandez <
[hidden email]> wrote:

Hi George,

the problem occurs because you are trying to bootstrap on a
UnitedKingdom
holiday, which is implicitly included in the definition of USDLibor.
The
Libor class makes a joint calendar between the given financial center
calendar and the UK calendar. Since Libor is a London poll, there are
only
fixings for it on UK business days. If one wants to bootstrap the
curve on
a holiday, the indice's historical fixings (actually just the closest
fixing to the given date) need to be provided.

Mit freundlichen Grüßen / Kind regards

Dr. Andres Hernandez

Senior Financial Engineer
Business Analytics
Risk Analytics

Phone: +49-69-6645-1351 IBM Deutschland
(Embedded
image moved
to file:
pic20635.gif)

Mobile: +49-163-609-3280 Wilhelm-Fay-Str.
30-34

E-Mail: [hidden email] 65936 Frankfurt am
Main

Germany

IBM Deutschland
GmbH /
Vorsitzender des
Aufsichtsrats:
Martin Jetter
Geschäftsführung:
Martina Koederitz
(Vorsitzende),
Reinhard Reschke,
Dieter Scholz,
Gregor Pillen,
Christian Noll,
Ivo Koerner
Sitz der
Gesellschaft:
Ehningen /
Registergericht:
Amtsgericht
Stuttgart, HRB
14562 /
WEEE-Reg.-Nr. DE
99369940

From: George Cowie <[hidden email]>
To: QuantLib users <[hidden email]>
Date: 22/07/2014 14:10
Subject: [Quantlib-users] UK Holiday for
USDLibor/PiecewiseYieldCurve

Hello,

I'm using PiecewiseYieldCurve to build a YieldTermStructure which
is
ultimately used by BlackScholesMertonProcess in my option pricing. I
believe I'm running into an issue on days that are holidays in the UK
(but
not the US). Here's the error I'm seeing when trying to run the code
below
for May 5, 2014 (Early May Bank Holiday):

terminate called after throwing an instance of 'QuantLib::Error'
what(): 1st iteration: failed at 1st alive instrument, maturity
May
14th, 2014, reference date May 7th, 2014: Missing no-fix1W Actual/360
fixing for May 2nd, 2014
Aborted

Everything seems to work fine on other days... Am I doing
something
wrong with setting up the PiecewiseYieldCurve?

Thanks,
George

#include <ql/quantlib.hpp>

using namespace std ;
using namespace QuantLib ;

int main(int argc, char** argv)
{
vector<Rate> liborRates ;
liborRates.push_back(0.12075 / 100.0) ; //1W
liborRates.push_back(0.1515 / 100.0) ; //1M
liborRates.push_back(0.1935 / 100.0) ; //2M
liborRates.push_back(0.22535 / 100.0) ; //3M
liborRates.push_back(0.323 / 100.0) ; //6M
liborRates.push_back(0.549 / 100.0) ; //12M

vector<Period> periods ;
periods.push_back(Period(1,Weeks)) ;
periods.push_back(Period(1,Months)) ;
periods.push_back(Period(2,Months)) ;
periods.push_back(Period(3,Months)) ;
periods.push_back(Period(6,Months)) ;
periods.push_back(Period(12,Months)) ;

vector<boost::shared_ptr<RateHelper> > instruments ;

for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)
{
boost::shared_ptr<Quote> liborQuote(new SimpleQuote(liborRates
[i])) ;
Handle<Quote> liborHandle(liborQuote) ;
boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods
[i])) ;
boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper
(liborHandle,liborIndex)) ;
instruments.push_back(rateHelper) ;
}

Calendar calendar = UnitedStates();
Date today (5 , May ,2014);
Settings::instance().evaluationDate() = today ;
Natural settlementDays = 2;
Date settlement = calendar.advance ( today , settlementDays ,
Days );
DayCounter dc = Actual360 ();

Date maturity(20, Jun, 2014) ;
double strikePrice = 100.0 ;
Option::Type optionType = Option::Call ;
double dividendYield = 0.0 ;
double impliedVolatility = 0.15 ;
double underlyingPrice = 100.0 ;

Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new
SimpleQuote
(underlyingPrice)));

Handle<YieldTermStructure> rateTermStructure
(boost::shared_ptr<YieldTermStructure>(new
PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments,
dc))) ;

Handle<YieldTermStructure>
dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new
FlatForward(today,
dividendYield,
dc)));
Handle<BlackVolTermStructure>
volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
BlackConstantVol(today,
calendar,
impliedVolatility,
dc)));

boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
BlackScholesMertonProcess(underlyingHandle,
dividendTermStructure,
rateTermStructure,
volTermStructure));

boost::shared_ptr<Exercise> americanExercise(new AmericanExercise
(today,
maturity)) ;
boost::shared_ptr<StrikedTypePayoff> payoff(new PlainVanillaPayoff
(optionType, strikePrice));

VanillaOption americanOption(payoff, americanExercise);

americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(
new FDAmericanEngine<CrankNicolson>(bsmProcess,
101,100)));

cout << americanOption.NPV() << endl ;

return 0 ;
}
------------------------------------------------------------------------------

Want fast and easy access to all the code in your enterprise? Index
and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds_______________________________________________

QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

------------------------------------------------------------------------------
Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds
_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

pic64586.gif (2K) Download Attachment

Peter Caspers-4

Re: UK Holiday for USDLibor/PiecewiseYieldCurve

Hi,

I think this question is an older one but I don't remember that it was
finally decided how to handle this situation (?) - which is, should we
allow to bootstrap a curve on a day which is a holiday for the cash
instruments in the curve or not. If negative, we could maybe throw a
different message ( the current one indeed sounds a bit paradoxical ).
If yes ( which would be my personal preference, because you simply
want to run your daily processes spitting out numbers even if not a
hundred percent sensible all of them ), one has to define how a
deposit which is not traded on that day should be calculated. One
obvious way would be to just treat the evaluation date as a business
day when calculating the helper's fixing date. I did not try, but it
may be enough to join the business days of the helper's calendar with
a new one just having one business day (the evaluation date) before
passing it to the no-fix index to make this kind of solution work.
What do you think ?

Peter

On 23 July 2014 15:05, Andres Hernandez <[hidden email]> wrote:

>
> Hi George,
>
> you are still setting the evaluation date to a holiday:
>
> Date today (5 , May ,2014);
> Date liborDate (2, May, 2014) ;
> Settings::instance().evaluationDate() = today ;
>
>
> Let's follow the logic:
> During bootstrapping, the method impliedQuote() of the rate helper is
> called to compare against the market quote. The DepositRateHelper simply
> returns the fixing corresponding to the fixingDate_ as a return value.
> The fixingDate_ is set during initialization as the return value of the
> function InterestRate::fixingDate(earliestDate_). The function
> InterestRate::fixingDate sets back the date given as input parameter by the
> number of fixing days according to the index' calendar.
> The parameter earliestDate_ being passed to fixingDate is set during
> initialization to the evaluation date (which you set to the 5th of May)
> moved forward by the number of fixing days according to the index'
> calendar.
> Note that if the evaluation date is a holiday you do not end up with the
> same date, but rather with the last business day prior to evaluation date.
> The first act moves the date from the 5th of May to the 7th and checks
> whether it is a business day, which it is. The second move sets it back two
> days to the 5th, and checks whether it is a business day, which it isn't,
> so it sets it back until it finds a business day, i.e. the 2nd of May.
> Since the value is in the past, it requests a historical fixing, which you
> did not provide and so the QL_REQUIRE throws.
>
> Mit freundlichen Grüßen / Kind regards
>
> Dr. Andres Hernandez
>
> Senior Financial Engineer
> Business Analytics
> Risk Analytics
>
>
>
>
>
> Phone: +49-69-6645-1351 IBM Deutschland (Embedded
> image moved
> to file:
> pic64586.gif)
>
> Mobile: +49-163-609-3280 Wilhelm-Fay-Str. 30-34
>
> E-Mail: [hidden email] 65936 Frankfurt am Main
>
> Germany
>
>
>
>
>
> IBM Deutschland
> GmbH /
> Vorsitzender des
> Aufsichtsrats:
> Martin Jetter
> Geschäftsführung:
> Martina Koederitz
> (Vorsitzende),
> Reinhard Reschke,
> Dieter Scholz,
> Gregor Pillen,
> Christian Noll,
> Ivo Koerner
> Sitz der
> Gesellschaft:
> Ehningen /
> Registergericht:
> Amtsgericht
> Stuttgart, HRB
> 14562 /
> WEEE-Reg.-Nr. DE
> 99369940
>
>
>
>
>
>
> From: George Cowie <[hidden email]>
> To: Andres Hernandez/Germany/IBM@IBMDE
> Cc: QuantLib users <[hidden email]>
> Date: 23/07/2014 11:51
> Subject: Re: [Quantlib-users] UK Holiday for
> USDLibor/PiecewiseYieldCurve
>
>
>
>
> Thank you Andres. What you're describing makes sense, but I can't seem to
> avoid an error when trying to value the option on May 5, 2014. I've copied
> the updated code snippet with changed section highlighted/bolded below. I
> seem to always get an exception of the form:
>
> terminate called after throwing an instance of 'QuantLib::Error'
> what(): 1st iteration: failed at 1st alive instrument, maturity May
> 14th, 2014, reference date May 6th, 2014: Missing no-fix1W Actual/360
> fixing for May 2nd, 2014
>
> Is there something wrong with my approach here?
>
> Thanks,
> George
>
>
> #include <ql/quantlib.hpp>
>
> using namespace std ;
> using namespace QuantLib ;
>
> int main(int argc, char** argv)
> {
> vector<Rate> liborRates ;
> liborRates.push_back(0.12075 / 100.0) ; //1W
> liborRates.push_back(0.1515 / 100.0) ; //1M
> liborRates.push_back(0.1935 / 100.0) ; //2M
> liborRates.push_back(0.22535 / 100.0) ; //3M
> liborRates.push_back(0.323 / 100.0) ; //6M
> liborRates.push_back(0.549 / 100.0) ; //12M
>
> vector<Period> periods ;
> periods.push_back(Period(1,Weeks)) ;
> periods.push_back(Period(1,Months)) ;
> periods.push_back(Period(2,Months)) ;
> periods.push_back(Period(3,Months)) ;
> periods.push_back(Period(6,Months)) ;
> periods.push_back(Period(12,Months)) ;
>
> vector<boost::shared_ptr<RateHelper> > instruments ;
>
> for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)
> {
> boost::shared_ptr<Quote> liborQuote(new SimpleQuote(liborRates[i])) ;
> Handle<Quote> liborHandle(liborQuote) ;
> boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods[i])) ;
> boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper
> (liborHandle,liborIndex)) ;
> instruments.push_back(rateHelper) ;
> }
>
> Calendar calendar = UnitedStates();
> Date today (5 , May ,2014);
> Date liborDate (2, May, 2014) ;
> Settings::instance().evaluationDate() = today ;
> Natural settlementDays = 2;
> Date settlement = calendar.advance ( liborDate , settlementDays , Days );
> DayCounter dc = Actual360 ();
>
> Date maturity(20, Jun, 2014) ;
> double strikePrice = 100.0 ;
> Option::Type optionType = Option::Call ;
> double dividendYield = 0.0 ;
> double impliedVolatility = 0.15 ;
> double underlyingPrice = 100.0 ;
>
> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new SimpleQuote
> (underlyingPrice)));
>
> Handle<YieldTermStructure> rateTermStructure
> (boost::shared_ptr<YieldTermStructure>(new
> PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments, dc))) ;
>
> Handle<YieldTermStructure>
> dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new
> FlatForward(today,
> dividendYield,
> dc)));
> Handle<BlackVolTermStructure>
> volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
> BlackConstantVol(today,
> calendar,
> impliedVolatility,
> dc)));
>
> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
> BlackScholesMertonProcess(underlyingHandle,
> dividendTermStructure,
> rateTermStructure,
> volTermStructure));
>
> boost::shared_ptr<Exercise> americanExercise(new AmericanExercise(today,
> maturity)) ;
> boost::shared_ptr<StrikedTypePayoff> payoff(new PlainVanillaPayoff
> (optionType, strikePrice));
>
> VanillaOption americanOption(payoff, americanExercise);
>
> americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(
> new FDAmericanEngine<CrankNicolson>(bsmProcess,
> 101,100)));
>
> cout << americanOption.NPV() << endl ;
>
> return 0 ;
> }
>
>
>
> On Tue, Jul 22, 2014 at 7:39 AM, Andres Hernandez <
> [hidden email]> wrote:
>
> Hi George,
>
> the problem occurs because you are trying to bootstrap on a
> UnitedKingdom
> holiday, which is implicitly included in the definition of USDLibor.
> The
> Libor class makes a joint calendar between the given financial center
> calendar and the UK calendar. Since Libor is a London poll, there are
> only
> fixings for it on UK business days. If one wants to bootstrap the
> curve on
> a holiday, the indice's historical fixings (actually just the closest
> fixing to the given date) need to be provided.
>
>
> Mit freundlichen Grüßen / Kind regards
>
> Dr. Andres Hernandez
>
> Senior Financial Engineer
> Business Analytics
> Risk Analytics
>
>
>
>
>
> Phone: +49-69-6645-1351 IBM Deutschland
> (Embedded
> image moved
> to file:
> pic20635.gif)
>
> Mobile: +49-163-609-3280 Wilhelm-Fay-Str.
> 30-34
>
> E-Mail: [hidden email] 65936 Frankfurt am
> Main
>
> Germany
>
>
>
>
>
> IBM Deutschland
> GmbH /
> Vorsitzender des
> Aufsichtsrats:
> Martin Jetter
> Geschäftsführung:
> Martina Koederitz
> (Vorsitzende),
> Reinhard Reschke,
> Dieter Scholz,
> Gregor Pillen,
> Christian Noll,
> Ivo Koerner
> Sitz der
> Gesellschaft:
> Ehningen /
> Registergericht:
> Amtsgericht
> Stuttgart, HRB
> 14562 /
> WEEE-Reg.-Nr. DE
> 99369940
>
>
>
>
>
>
> From: George Cowie <[hidden email]>
> To: QuantLib users <[hidden email]>
> Date: 22/07/2014 14:10
> Subject: [Quantlib-users] UK Holiday for
> USDLibor/PiecewiseYieldCurve
>
>
>
> Hello,
>
> I'm using PiecewiseYieldCurve to build a YieldTermStructure which
> is
> ultimately used by BlackScholesMertonProcess in my option pricing. I
> believe I'm running into an issue on days that are holidays in the UK
> (but
> not the US). Here's the error I'm seeing when trying to run the code
> below
> for May 5, 2014 (Early May Bank Holiday):
>
> terminate called after throwing an instance of 'QuantLib::Error'
> what(): 1st iteration: failed at 1st alive instrument, maturity
> May
> 14th, 2014, reference date May 7th, 2014: Missing no-fix1W Actual/360
> fixing for May 2nd, 2014
> Aborted
>
> Everything seems to work fine on other days... Am I doing
> something
> wrong with setting up the PiecewiseYieldCurve?
>
> Thanks,
> George
>
> #include <ql/quantlib.hpp>
>
> using namespace std ;
> using namespace QuantLib ;
>
> int main(int argc, char** argv)
> {
> vector<Rate> liborRates ;
> liborRates.push_back(0.12075 / 100.0) ; //1W
> liborRates.push_back(0.1515 / 100.0) ; //1M
> liborRates.push_back(0.1935 / 100.0) ; //2M
> liborRates.push_back(0.22535 / 100.0) ; //3M
> liborRates.push_back(0.323 / 100.0) ; //6M
> liborRates.push_back(0.549 / 100.0) ; //12M
>
> vector<Period> periods ;
> periods.push_back(Period(1,Weeks)) ;
> periods.push_back(Period(1,Months)) ;
> periods.push_back(Period(2,Months)) ;
> periods.push_back(Period(3,Months)) ;
> periods.push_back(Period(6,Months)) ;
> periods.push_back(Period(12,Months)) ;
>
> vector<boost::shared_ptr<RateHelper> > instruments ;
>
> for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)
> {
> boost::shared_ptr<Quote> liborQuote(new SimpleQuote(liborRates
> [i])) ;
> Handle<Quote> liborHandle(liborQuote) ;
> boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods
> [i])) ;
> boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper
> (liborHandle,liborIndex)) ;
> instruments.push_back(rateHelper) ;
> }
>
> Calendar calendar = UnitedStates();
> Date today (5 , May ,2014);
> Settings::instance().evaluationDate() = today ;
> Natural settlementDays = 2;
> Date settlement = calendar.advance ( today , settlementDays ,
> Days );
> DayCounter dc = Actual360 ();
>
> Date maturity(20, Jun, 2014) ;
> double strikePrice = 100.0 ;
> Option::Type optionType = Option::Call ;
> double dividendYield = 0.0 ;
> double impliedVolatility = 0.15 ;
> double underlyingPrice = 100.0 ;
>
> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new
> SimpleQuote
> (underlyingPrice)));
>
> Handle<YieldTermStructure> rateTermStructure
> (boost::shared_ptr<YieldTermStructure>(new
> PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments,
> dc))) ;
>
> Handle<YieldTermStructure>
> dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new
> FlatForward(today,
> dividendYield,
> dc)));
> Handle<BlackVolTermStructure>
> volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
> BlackConstantVol(today,
> calendar,
> impliedVolatility,
> dc)));
>
> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
> BlackScholesMertonProcess(underlyingHandle,
> dividendTermStructure,
> rateTermStructure,
> volTermStructure));
>
> boost::shared_ptr<Exercise> americanExercise(new AmericanExercise
> (today,
> maturity)) ;
> boost::shared_ptr<StrikedTypePayoff> payoff(new PlainVanillaPayoff
> (optionType, strikePrice));
>
> VanillaOption americanOption(payoff, americanExercise);
>
> americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(
> new FDAmericanEngine<CrankNicolson>(bsmProcess,
> 101,100)));
>
> cout << americanOption.NPV() << endl ;
>
> return 0 ;
> }
> ------------------------------------------------------------------------------
>
>
> Want fast and easy access to all the code in your enterprise? Index
> and
> search up to 200,000 lines of code with a free copy of Black Duck
> Code Sight - the same software that powers the world's largest code
> search on Ohloh, the Black Duck Open Hub! Try it now.
> http://p.sf.net/sfu/bds_______________________________________________
>
> QuantLib-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/quantlib-users
>
> ------------------------------------------------------------------------------
> Want fast and easy access to all the code in your enterprise? Index and
> search up to 200,000 lines of code with a free copy of Black Duck
> Code Sight - the same software that powers the world's largest code
> search on Ohloh, the Black Duck Open Hub! Try it now.
> http://p.sf.net/sfu/bds
> _______________________________________________
> QuantLib-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/quantlib-users
>

Andres Hernandez

Re: UK Holiday for USDLibor/PiecewiseYieldCurve

Hi Peter,

I agree that it would be nice to at least have it as an option. We
currently do a small workaround, whereby we set in our system a parameter
for the curve specifying the holiday calendar for that curve. The only use
of that calendar is to determine if it is a holiday for the curve or not
and if yes to make a fallback to yesterday's values. In the case of the EUR
curves, this is arguably a good behaviour as the index and the geographical
region have the same calendar, but for the USD case, we are basically
forcing a fallback for a holiday on a different country. The issue is
compounded by the fact that in this case the deposit is traded on that day,
it is just the index that gets no fixing.

I took a look and I am almost sure that removing the evaluation date from
the fixing calendar in InterestRate constructor would be enough to make
deposits go through. This could simply be accomplished with an extra public
member function that one would call if one wanted it to go through on
holidays. A more involved and important problem would be to get the
SwapRateHelper to go though on holidays (if desired), as I am positive that
that was the reason for our instituting the fallback. I will see if I have
the time to change it, and would then probably subscribe to the developers
list to get some feedback in how to go about changing it.

Mit freundlichen Grüßen / Kind regards

Dr. Andres Hernandez

Senior Financial Engineer
Business Analytics
Risk Analytics

Phone: +49-69-6645-1351 IBM Deutschland (Embedded
image moved
to file:
pic08098.gif)

Mobile: +49-163-609-3280 Wilhelm-Fay-Str. 30-34

E-Mail: [hidden email] 65936 Frankfurt am Main

Germany

IBM Deutschland
GmbH /
Vorsitzender des
Aufsichtsrats:
Martin Jetter
Geschäftsführung:
Martina Koederitz
(Vorsitzende),
Reinhard Reschke,
Dieter Scholz,
Gregor Pillen,
Christian Noll,
Ivo Koerner
Sitz der
Gesellschaft:
Ehningen /
Registergericht:
Amtsgericht
Stuttgart, HRB
14562 /
WEEE-Reg.-Nr. DE
99369940

From: Peter Caspers <[hidden email]>
To: Andres Hernandez/Germany/IBM@IBMDE
Cc: George Cowie <[hidden email]>, QuantLib users
<[hidden email]>
Date: 23/07/2014 21:56
Subject: Re: [Quantlib-users] UK Holiday for
USDLibor/PiecewiseYieldCurve

Hi,

I think this question is an older one but I don't remember that it was
finally decided how to handle this situation (?) - which is, should we
allow to bootstrap a curve on a day which is a holiday for the cash
instruments in the curve or not. If negative, we could maybe throw a
different message ( the current one indeed sounds a bit paradoxical ).
If yes ( which would be my personal preference, because you simply
want to run your daily processes spitting out numbers even if not a
hundred percent sensible all of them ), one has to define how a
deposit which is not traded on that day should be calculated. One
obvious way would be to just treat the evaluation date as a business
day when calculating the helper's fixing date. I did not try, but it
may be enough to join the business days of the helper's calendar with
a new one just having one business day (the evaluation date) before
passing it to the no-fix index to make this kind of solution work.
What do you think ?

Peter

On 23 July 2014 15:05, Andres Hernandez <[hidden email]>
wrote:

the
> number of fixing days according to the index' calendar.
> The parameter earliestDate_ being passed to fixingDate is set during
> initialization to the evaluation date (which you set to the 5th of May)
> moved forward by the number of fixing days according to the index'
> calendar.
> Note that if the evaluation date is a holiday you do not end up with the
> same date, but rather with the last business day prior to evaluation

date.
> The first act moves the date from the 5th of May to the 7th and checks
> whether it is a business day, which it is. The second move sets it back
two
> days to the 5th, and checks whether it is a business day, which it isn't,
> so it sets it back until it finds a business day, i.e. the 2nd of May.
> Since the value is in the past, it requests a historical fixing, which
you

> did not provide and so the QL_REQUIRE throws.
>
> Mit freundlichen Grüßen / Kind regards
>
> Dr. Andres Hernandez
>
> Senior Financial Engineer
> Business Analytics
> Risk Analytics
>
>
>
>
>
> Phone: +49-69-6645-1351 IBM Deutschland

(Embedded
>

image moved
>
to file:
>
pic64586.gif)

>
> Mobile: +49-163-609-3280 Wilhelm-Fay-Str. 30-34
>
> E-Mail: [hidden email] 65936 Frankfurt am Main
>
> Germany
>
>
>
>
>
> IBM Deutschland
> GmbH /
> Vorsitzender des
> Aufsichtsrats:
> Martin Jetter
> Geschäftsführung:
> Martina Koederitz
> (Vorsitzende),
> Reinhard Reschke,
> Dieter Scholz,
> Gregor Pillen,
> Christian Noll,
> Ivo Koerner
> Sitz der
> Gesellschaft:
> Ehningen /
> Registergericht:
> Amtsgericht
> Stuttgart, HRB
> 14562 /
> WEEE-Reg.-Nr. DE
> 99369940
>
>
>
>
>
>
> From: George Cowie <[hidden email]>
> To: Andres Hernandez/Germany/IBM@IBMDE
> Cc: QuantLib users <[hidden email]>
> Date: 23/07/2014 11:51
> Subject: Re: [Quantlib-users] UK Holiday for
> USDLibor/PiecewiseYieldCurve
>
>
>
>
> Thank you Andres. What you're describing makes sense, but I can't seem

to
> avoid an error when trying to value the option on May 5, 2014. I've

copied
> the updated code snippet with changed section highlighted/bolded below.
I

> seem to always get an exception of the form:
>
> terminate called after throwing an instance of 'QuantLib::Error'
> what(): 1st iteration: failed at 1st alive instrument, maturity May
> 14th, 2014, reference date May 6th, 2014: Missing no-fix1W Actual/360
> fixing for May 2nd, 2014
>
> Is there something wrong with my approach here?
>
> Thanks,
> George
>
>
> #include <ql/quantlib.hpp>
>
> using namespace std ;
> using namespace QuantLib ;
>
> int main(int argc, char** argv)
> {
> vector<Rate> liborRates ;
> liborRates.push_back(0.12075 / 100.0) ; //1W
> liborRates.push_back(0.1515 / 100.0) ; //1M
> liborRates.push_back(0.1935 / 100.0) ; //2M
> liborRates.push_back(0.22535 / 100.0) ; //3M
> liborRates.push_back(0.323 / 100.0) ; //6M
> liborRates.push_back(0.549 / 100.0) ; //12M
>
> vector<Period> periods ;
> periods.push_back(Period(1,Weeks)) ;
> periods.push_back(Period(1,Months)) ;
> periods.push_back(Period(2,Months)) ;
> periods.push_back(Period(3,Months)) ;
> periods.push_back(Period(6,Months)) ;
> periods.push_back(Period(12,Months)) ;
>
> vector<boost::shared_ptr<RateHelper> > instruments ;
>
> for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)
> {
> boost::shared_ptr<Quote> liborQuote(new SimpleQuote(liborRates

[i])) ;

> Handle<Quote> liborHandle(liborQuote) ;
> boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods[i])) ;
> boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper
> (liborHandle,liborIndex)) ;
> instruments.push_back(rateHelper) ;
> }
>
> Calendar calendar = UnitedStates();
> Date today (5 , May ,2014);
> Date liborDate (2, May, 2014) ;
> Settings::instance().evaluationDate() = today ;
> Natural settlementDays = 2;
> Date settlement = calendar.advance ( liborDate , settlementDays ,

Days );

> DayCounter dc = Actual360 ();
>
> Date maturity(20, Jun, 2014) ;
> double strikePrice = 100.0 ;
> Option::Type optionType = Option::Call ;
> double dividendYield = 0.0 ;
> double impliedVolatility = 0.15 ;
> double underlyingPrice = 100.0 ;
>
> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new SimpleQuote
> (underlyingPrice)));
>
> Handle<YieldTermStructure> rateTermStructure
> (boost::shared_ptr<YieldTermStructure>(new
> PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments, dc))) ;
>
> Handle<YieldTermStructure>
> dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new
> FlatForward(today,
> dividendYield,
> dc)));
> Handle<BlackVolTermStructure>
> volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
> BlackConstantVol(today,
> calendar,
> impliedVolatility,
> dc)));
>
> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
> BlackScholesMertonProcess(underlyingHandle,
> dividendTermStructure,
> rateTermStructure,
> volTermStructure));
>
> boost::shared_ptr<Exercise> americanExercise(new AmericanExercise

(today,

> maturity)) ;
> boost::shared_ptr<StrikedTypePayoff> payoff(new PlainVanillaPayoff
> (optionType, strikePrice));
>
> VanillaOption americanOption(payoff, americanExercise);
>
> americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(
> new FDAmericanEngine<CrankNicolson>(bsmProcess,
> 101,100)));
>
> cout << americanOption.NPV() << endl ;
>
> return 0 ;
> }
>
>
>
> On Tue, Jul 22, 2014 at 7:39 AM, Andres Hernandez <
> [hidden email]> wrote:
>
> Hi George,
>
> the problem occurs because you are trying to bootstrap on a
> UnitedKingdom
> holiday, which is implicitly included in the definition of

USDLibor.
> The
> Libor class makes a joint calendar between the given financial

center
> calendar and the UK calendar. Since Libor is a London poll, there
are
> only
> fixings for it on UK business days. If one wants to bootstrap the
> curve on
> a holiday, the indice's historical fixings (actually just the
closest

> fixing to the given date) need to be provided.
>
>
> Mit freundlichen Grüßen / Kind regards
>
> Dr. Andres Hernandez
>
> Senior Financial Engineer
> Business Analytics
> Risk Analytics
>
>
>
>
>
> Phone: +49-69-6645-1351 IBM Deutschland
> (Embedded
> image moved
> to file:
> pic20635.gif)
>
> Mobile: +49-163-609-3280 Wilhelm-Fay-Str.
> 30-34
>
> E-Mail: [hidden email] 65936 Frankfurt

> Main
>
> Germany
>
>
>
>
>
> IBM Deutschland
> GmbH /
> Vorsitzender des
> Aufsichtsrats:
> Martin Jetter
> Geschäftsführung:
> Martina Koederitz
> (Vorsitzende),
> Reinhard Reschke,
> Dieter Scholz,
> Gregor Pillen,
> Christian Noll,
> Ivo Koerner
> Sitz der
> Gesellschaft:
> Ehningen /
> Registergericht:
> Amtsgericht
> Stuttgart, HRB
> 14562 /
> WEEE-Reg.-Nr. DE
> 99369940
>
>
>
>
>
>
> From: George Cowie <[hidden email]>
> To: QuantLib users <[hidden email]>
> Date: 22/07/2014 14:10
> Subject: [Quantlib-users] UK Holiday for
> USDLibor/PiecewiseYieldCurve
>
>
>
> Hello,
>
> I'm using PiecewiseYieldCurve to build a YieldTermStructure which
> is
> ultimately used by BlackScholesMertonProcess in my option pricing.

I
> believe I'm running into an issue on days that are holidays in the

UK
> (but
> not the US). Here's the error I'm seeing when trying to run the
code
> below
> for May 5, 2014 (Early May Bank Holiday):
>
> terminate called after throwing an instance of 'QuantLib::Error'
> what(): 1st iteration: failed at 1st alive instrument, maturity
> May
> 14th, 2014, reference date May 7th, 2014: Missing no-fix1W
Actual/360

> fixing for May 2nd, 2014
> Aborted
>
> Everything seems to work fine on other days... Am I doing
> something
> wrong with setting up the PiecewiseYieldCurve?
>
> Thanks,
> George
>
> #include <ql/quantlib.hpp>
>
> using namespace std ;
> using namespace QuantLib ;
>
> int main(int argc, char** argv)
> {
> vector<Rate> liborRates ;
> liborRates.push_back(0.12075 / 100.0) ; //1W
> liborRates.push_back(0.1515 / 100.0) ; //1M
> liborRates.push_back(0.1935 / 100.0) ; //2M
> liborRates.push_back(0.22535 / 100.0) ; //3M
> liborRates.push_back(0.323 / 100.0) ; //6M
> liborRates.push_back(0.549 / 100.0) ; //12M
>
> vector<Period> periods ;
> periods.push_back(Period(1,Weeks)) ;
> periods.push_back(Period(1,Months)) ;
> periods.push_back(Period(2,Months)) ;
> periods.push_back(Period(3,Months)) ;
> periods.push_back(Period(6,Months)) ;
> periods.push_back(Period(12,Months)) ;
>
> vector<boost::shared_ptr<RateHelper> > instruments ;
>
> for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)
> {
> boost::shared_ptr<Quote> liborQuote(new SimpleQuote

(liborRates
> [i])) ;
> Handle<Quote> liborHandle(liborQuote) ;
> boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods
> [i])) ;
> boost::shared_ptr<RateHelper> rateHelper(new

DepositRateHelper

> (liborHandle,liborIndex)) ;
> instruments.push_back(rateHelper) ;
> }
>
> Calendar calendar = UnitedStates();
> Date today (5 , May ,2014);
> Settings::instance().evaluationDate() = today ;
> Natural settlementDays = 2;
> Date settlement = calendar.advance ( today , settlementDays ,
> Days );
> DayCounter dc = Actual360 ();
>
> Date maturity(20, Jun, 2014) ;
> double strikePrice = 100.0 ;
> Option::Type optionType = Option::Call ;
> double dividendYield = 0.0 ;
> double impliedVolatility = 0.15 ;
> double underlyingPrice = 100.0 ;
>
> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new
> SimpleQuote
> (underlyingPrice)));
>
> Handle<YieldTermStructure> rateTermStructure
> (boost::shared_ptr<YieldTermStructure>(new
> PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments,
> dc))) ;
>
> Handle<YieldTermStructure>
> dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new
> FlatForward(today,
> dividendYield,
> dc)));
> Handle<BlackVolTermStructure>
> volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
> BlackConstantVol(today,
> calendar,
> impliedVolatility,
> dc)));
>
> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
> BlackScholesMertonProcess(underlyingHandle,
> dividendTermStructure,
> rateTermStructure,
> volTermStructure));
>
> boost::shared_ptr<Exercise> americanExercise(new AmericanExercise
> (today,
> maturity)) ;
> boost::shared_ptr<StrikedTypePayoff> payoff(new

PlainVanillaPayoff

> (optionType, strikePrice));
>
> VanillaOption americanOption(payoff, americanExercise);
>
> americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(
> new FDAmericanEngine<CrankNicolson>(bsmProcess,
> 101,100)));
>
> cout << americanOption.NPV() << endl ;
>
> return 0 ;
> }
>

------------------------------------------------------------------------------

>
>
> Want fast and easy access to all the code in your enterprise? Index
> and
> search up to 200,000 lines of code with a free copy of Black Duck
> Code Sight - the same software that powers the world's largest code
> search on Ohloh, the Black Duck Open Hub! Try it now.
>
http://p.sf.net/sfu/bds_______________________________________________
>
> QuantLib-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/quantlib-users
>
>
------------------------------------------------------------------------------

> Want fast and easy access to all the code in your enterprise? Index and
> search up to 200,000 lines of code with a free copy of Black Duck
> Code Sight - the same software that powers the world's largest code
> search on Ohloh, the Black Duck Open Hub! Try it now.
> http://p.sf.net/sfu/bds
> _______________________________________________
> QuantLib-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/quantlib-users
>

pic08098.gif (2K) Download Attachment

Peter Caspers-4

Re: UK Holiday for USDLibor/PiecewiseYieldCurve

Hi Andres,

yes, the index in the deposit rate helper seems to be merely a
convenient way for coding - with the one exception that deposits do
not (directly) have anything to do with actual index fixings, and this
is also where the code breaks. And yes, it gets even worse in your
example when you use the USD Libor index for USD cash deposits on a UK
but non-NY holiday (you could use the index-free constructor though
and pass the NY calendar directly instead, right, solving half of the
problem ?). So just treating the deposit helper's index and its fixing
as a computational tool to get the implied Quote seems a reasonable
solution, in your USD / Libor example yielding the correct conventions
even, if I am not mistaken ? I wouldn't introduce any flag or method
to activate this behaviour, it should be the only one present (because
checking for holidays can easily be done in the user's code if
necessary at all).

For swaps on the other hand I think the current behaviour is fine -
what fixing would you take for the first period in a Euribor swap on a
TARGET holiday ? Probably the same as you take after the fixing around
11am on a trading day, i.e. the last known fixing. That is exactly the
logic applied in the library. I wouldn't change anything here. It
might be that you get an exception because you just don't provide the
historic fixing, but in this case it is a legitimate demand of the
library and working around this should only be done in the user code
imo.

best
Peter

On 24 July 2014 09:23, Andres Hernandez <[hidden email]> wrote:

>
> Hi Peter,
>
> I agree that it would be nice to at least have it as an option. We
> currently do a small workaround, whereby we set in our system a parameter
> for the curve specifying the holiday calendar for that curve. The only use
> of that calendar is to determine if it is a holiday for the curve or not
> and if yes to make a fallback to yesterday's values. In the case of the EUR
> curves, this is arguably a good behaviour as the index and the geographical
> region have the same calendar, but for the USD case, we are basically
> forcing a fallback for a holiday on a different country. The issue is
> compounded by the fact that in this case the deposit is traded on that day,
> it is just the index that gets no fixing.
>
> I took a look and I am almost sure that removing the evaluation date from
> the fixing calendar in InterestRate constructor would be enough to make
> deposits go through. This could simply be accomplished with an extra public
> member function that one would call if one wanted it to go through on
> holidays. A more involved and important problem would be to get the
> SwapRateHelper to go though on holidays (if desired), as I am positive that
> that was the reason for our instituting the fallback. I will see if I have
> the time to change it, and would then probably subscribe to the developers
> list to get some feedback in how to go about changing it.
>
>
> Mit freundlichen Grüßen / Kind regards
>
> Dr. Andres Hernandez
>
> Senior Financial Engineer
> Business Analytics
> Risk Analytics
>
>
>
>
>
> Phone: +49-69-6645-1351 IBM Deutschland (Embedded
> image moved
> to file:
> pic02235.gif)
>
> Mobile: +49-163-609-3280 Wilhelm-Fay-Str. 30-34
>
> E-Mail: [hidden email] 65936 Frankfurt am Main
>
> Germany
>
>
>
>
>
> IBM Deutschland
> GmbH /
> Vorsitzender des
> Aufsichtsrats:
> Martin Jetter
> Geschäftsführung:
> Martina Koederitz
> (Vorsitzende),
> Reinhard Reschke,
> Dieter Scholz,
> Gregor Pillen,
> Christian Noll,
> Ivo Koerner
> Sitz der
> Gesellschaft:
> Ehningen /
> Registergericht:
> Amtsgericht
> Stuttgart, HRB
> 14562 /
> WEEE-Reg.-Nr. DE
> 99369940
>
>
>
>
>
>
> From: Peter Caspers <[hidden email]>
> To: Andres Hernandez/Germany/IBM@IBMDE
> Cc: George Cowie <[hidden email]>, QuantLib users
> <[hidden email]>
> Date: 23/07/2014 21:56
> Subject: Re: [Quantlib-users] UK Holiday for
> USDLibor/PiecewiseYieldCurve
>
>
>
> Hi,
>
> I think this question is an older one but I don't remember that it was
> finally decided how to handle this situation (?) - which is, should we
> allow to bootstrap a curve on a day which is a holiday for the cash
> instruments in the curve or not. If negative, we could maybe throw a
> different message ( the current one indeed sounds a bit paradoxical ).
> If yes ( which would be my personal preference, because you simply
> want to run your daily processes spitting out numbers even if not a
> hundred percent sensible all of them ), one has to define how a
> deposit which is not traded on that day should be calculated. One
> obvious way would be to just treat the evaluation date as a business
> day when calculating the helper's fixing date. I did not try, but it
> may be enough to join the business days of the helper's calendar with
> a new one just having one business day (the evaluation date) before
> passing it to the no-fix index to make this kind of solution work.
> What do you think ?
>
> Peter
>
>
> On 23 July 2014 15:05, Andres Hernandez <[hidden email]>
> wrote:
>>
>> Hi George,
>>
>> you are still setting the evaluation date to a holiday:
>>
>> Date today (5 , May ,2014);
>> Date liborDate (2, May, 2014) ;
>> Settings::instance().evaluationDate() = today ;
>>
>>
>> Let's follow the logic:
>> During bootstrapping, the method impliedQuote() of the rate helper is
>> called to compare against the market quote. The DepositRateHelper simply
>> returns the fixing corresponding to the fixingDate_ as a return value.
>> The fixingDate_ is set during initialization as the return value of the
>> function InterestRate::fixingDate(earliestDate_). The function
>> InterestRate::fixingDate sets back the date given as input parameter by
> the
>> number of fixing days according to the index' calendar.
>> The parameter earliestDate_ being passed to fixingDate is set during
>> initialization to the evaluation date (which you set to the 5th of May)
>> moved forward by the number of fixing days according to the index'
>> calendar.
>> Note that if the evaluation date is a holiday you do not end up with the
>> same date, but rather with the last business day prior to evaluation
> date.
>> The first act moves the date from the 5th of May to the 7th and checks
>> whether it is a business day, which it is. The second move sets it back
> two
>> days to the 5th, and checks whether it is a business day, which it isn't,
>> so it sets it back until it finds a business day, i.e. the 2nd of May.
>> Since the value is in the past, it requests a historical fixing, which
> you
>> did not provide and so the QL_REQUIRE throws.
>>
>> Mit freundlichen Grüßen / Kind regards
>>
>> Dr. Andres Hernandez
>>
>> Senior Financial Engineer
>> Business Analytics
>> Risk Analytics
>>
>>
>>
>>
>>
>> Phone: +49-69-6645-1351 IBM Deutschland
> (Embedded
>>
> image moved
>>
> to file:
>>
> pic64586.gif)
>>
>> Mobile: +49-163-609-3280 Wilhelm-Fay-Str. 30-34
>>
>> E-Mail: [hidden email] 65936 Frankfurt am Main
>>
>> Germany
>>
>>
>>
>>
>>
>> IBM Deutschland
>> GmbH /
>> Vorsitzender des
>> Aufsichtsrats:
>> Martin Jetter
>> Geschäftsführung:
>> Martina Koederitz
>> (Vorsitzende),
>> Reinhard Reschke,
>> Dieter Scholz,
>> Gregor Pillen,
>> Christian Noll,
>> Ivo Koerner
>> Sitz der
>> Gesellschaft:
>> Ehningen /
>> Registergericht:
>> Amtsgericht
>> Stuttgart, HRB
>> 14562 /
>> WEEE-Reg.-Nr. DE
>> 99369940
>>
>>
>>
>>
>>
>>
>> From: George Cowie <[hidden email]>
>> To: Andres Hernandez/Germany/IBM@IBMDE
>> Cc: QuantLib users <[hidden email]>
>> Date: 23/07/2014 11:51
>> Subject: Re: [Quantlib-users] UK Holiday for
>> USDLibor/PiecewiseYieldCurve
>>
>>
>>
>>
>> Thank you Andres. What you're describing makes sense, but I can't seem
> to
>> avoid an error when trying to value the option on May 5, 2014. I've
> copied
>> the updated code snippet with changed section highlighted/bolded below.
> I
>> seem to always get an exception of the form:
>>
>> terminate called after throwing an instance of 'QuantLib::Error'
>> what(): 1st iteration: failed at 1st alive instrument, maturity May
>> 14th, 2014, reference date May 6th, 2014: Missing no-fix1W Actual/360
>> fixing for May 2nd, 2014
>>
>> Is there something wrong with my approach here?
>>
>> Thanks,
>> George
>>
>>
>> #include <ql/quantlib.hpp>
>>
>> using namespace std ;
>> using namespace QuantLib ;
>>
>> int main(int argc, char** argv)
>> {
>> vector<Rate> liborRates ;
>> liborRates.push_back(0.12075 / 100.0) ; //1W
>> liborRates.push_back(0.1515 / 100.0) ; //1M
>> liborRates.push_back(0.1935 / 100.0) ; //2M
>> liborRates.push_back(0.22535 / 100.0) ; //3M
>> liborRates.push_back(0.323 / 100.0) ; //6M
>> liborRates.push_back(0.549 / 100.0) ; //12M
>>
>> vector<Period> periods ;
>> periods.push_back(Period(1,Weeks)) ;
>> periods.push_back(Period(1,Months)) ;
>> periods.push_back(Period(2,Months)) ;
>> periods.push_back(Period(3,Months)) ;
>> periods.push_back(Period(6,Months)) ;
>> periods.push_back(Period(12,Months)) ;
>>
>> vector<boost::shared_ptr<RateHelper> > instruments ;
>>
>> for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)
>> {
>> boost::shared_ptr<Quote> liborQuote(new SimpleQuote(liborRates
> [i])) ;
>> Handle<Quote> liborHandle(liborQuote) ;
>> boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods[i])) ;
>> boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper
>> (liborHandle,liborIndex)) ;
>> instruments.push_back(rateHelper) ;
>> }
>>
>> Calendar calendar = UnitedStates();
>> Date today (5 , May ,2014);
>> Date liborDate (2, May, 2014) ;
>> Settings::instance().evaluationDate() = today ;
>> Natural settlementDays = 2;
>> Date settlement = calendar.advance ( liborDate , settlementDays ,
> Days );
>> DayCounter dc = Actual360 ();
>>
>> Date maturity(20, Jun, 2014) ;
>> double strikePrice = 100.0 ;
>> Option::Type optionType = Option::Call ;
>> double dividendYield = 0.0 ;
>> double impliedVolatility = 0.15 ;
>> double underlyingPrice = 100.0 ;
>>
>> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new SimpleQuote
>> (underlyingPrice)));
>>
>> Handle<YieldTermStructure> rateTermStructure
>> (boost::shared_ptr<YieldTermStructure>(new
>> PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments, dc))) ;
>>
>> Handle<YieldTermStructure>
>> dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new
>> FlatForward(today,
>> dividendYield,
>> dc)));
>> Handle<BlackVolTermStructure>
>> volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
>> BlackConstantVol(today,
>> calendar,
>> impliedVolatility,
>> dc)));
>>
>> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
>> BlackScholesMertonProcess(underlyingHandle,
>> dividendTermStructure,
>> rateTermStructure,
>> volTermStructure));
>>
>> boost::shared_ptr<Exercise> americanExercise(new AmericanExercise
> (today,
>> maturity)) ;
>> boost::shared_ptr<StrikedTypePayoff> payoff(new PlainVanillaPayoff
>> (optionType, strikePrice));
>>
>> VanillaOption americanOption(payoff, americanExercise);
>>
>> americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(
>> new FDAmericanEngine<CrankNicolson>(bsmProcess,
>> 101,100)));
>>
>> cout << americanOption.NPV() << endl ;
>>
>> return 0 ;
>> }
>>
>>
>>
>> On Tue, Jul 22, 2014 at 7:39 AM, Andres Hernandez <
>> [hidden email]> wrote:
>>
>> Hi George,
>>
>> the problem occurs because you are trying to bootstrap on a
>> UnitedKingdom
>> holiday, which is implicitly included in the definition of
> USDLibor.
>> The
>> Libor class makes a joint calendar between the given financial
> center
>> calendar and the UK calendar. Since Libor is a London poll, there
> are
>> only
>> fixings for it on UK business days. If one wants to bootstrap the
>> curve on
>> a holiday, the indice's historical fixings (actually just the
> closest
>> fixing to the given date) need to be provided.
>>
>>
>> Mit freundlichen Grüßen / Kind regards
>>
>> Dr. Andres Hernandez
>>
>> Senior Financial Engineer
>> Business Analytics
>> Risk Analytics
>>
>>
>>
>>
>>
>> Phone: +49-69-6645-1351 IBM Deutschland
>> (Embedded
>> image moved
>> to file:
>> pic20635.gif)
>>
>> Mobile: +49-163-609-3280 Wilhelm-Fay-Str.
>> 30-34
>>
>> E-Mail: [hidden email] 65936 Frankfurt
> am
>> Main
>>
>> Germany
>>
>>
>>
>>
>>
>> IBM Deutschland
>> GmbH /
>> Vorsitzender des
>> Aufsichtsrats:
>> Martin Jetter
>> Geschäftsführung:
>> Martina Koederitz
>> (Vorsitzende),
>> Reinhard Reschke,
>> Dieter Scholz,
>> Gregor Pillen,
>> Christian Noll,
>> Ivo Koerner
>> Sitz der
>> Gesellschaft:
>> Ehningen /
>> Registergericht:
>> Amtsgericht
>> Stuttgart, HRB
>> 14562 /
>> WEEE-Reg.-Nr. DE
>> 99369940
>>
>>
>>
>>
>>
>>
>> From: George Cowie <[hidden email]>
>> To: QuantLib users <[hidden email]>
>> Date: 22/07/2014 14:10
>> Subject: [Quantlib-users] UK Holiday for
>> USDLibor/PiecewiseYieldCurve
>>
>>
>>
>> Hello,
>>
>> I'm using PiecewiseYieldCurve to build a YieldTermStructure which
>> is
>> ultimately used by BlackScholesMertonProcess in my option pricing.
> I
>> believe I'm running into an issue on days that are holidays in the
> UK
>> (but
>> not the US). Here's the error I'm seeing when trying to run the
> code
>> below
>> for May 5, 2014 (Early May Bank Holiday):
>>
>> terminate called after throwing an instance of 'QuantLib::Error'
>> what(): 1st iteration: failed at 1st alive instrument, maturity
>> May
>> 14th, 2014, reference date May 7th, 2014: Missing no-fix1W
> Actual/360
>> fixing for May 2nd, 2014
>> Aborted
>>
>> Everything seems to work fine on other days... Am I doing
>> something
>> wrong with setting up the PiecewiseYieldCurve?
>>
>> Thanks,
>> George
>>
>> #include <ql/quantlib.hpp>
>>
>> using namespace std ;
>> using namespace QuantLib ;
>>
>> int main(int argc, char** argv)
>> {
>> vector<Rate> liborRates ;
>> liborRates.push_back(0.12075 / 100.0) ; //1W
>> liborRates.push_back(0.1515 / 100.0) ; //1M
>> liborRates.push_back(0.1935 / 100.0) ; //2M
>> liborRates.push_back(0.22535 / 100.0) ; //3M
>> liborRates.push_back(0.323 / 100.0) ; //6M
>> liborRates.push_back(0.549 / 100.0) ; //12M
>>
>> vector<Period> periods ;
>> periods.push_back(Period(1,Weeks)) ;
>> periods.push_back(Period(1,Months)) ;
>> periods.push_back(Period(2,Months)) ;
>> periods.push_back(Period(3,Months)) ;
>> periods.push_back(Period(6,Months)) ;
>> periods.push_back(Period(12,Months)) ;
>>
>> vector<boost::shared_ptr<RateHelper> > instruments ;
>>
>> for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)
>> {
>> boost::shared_ptr<Quote> liborQuote(new SimpleQuote
> (liborRates
>> [i])) ;
>> Handle<Quote> liborHandle(liborQuote) ;
>> boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods
>> [i])) ;
>> boost::shared_ptr<RateHelper> rateHelper(new
> DepositRateHelper
>> (liborHandle,liborIndex)) ;
>> instruments.push_back(rateHelper) ;
>> }
>>
>> Calendar calendar = UnitedStates();
>> Date today (5 , May ,2014);
>> Settings::instance().evaluationDate() = today ;
>> Natural settlementDays = 2;
>> Date settlement = calendar.advance ( today , settlementDays ,
>> Days );
>> DayCounter dc = Actual360 ();
>>
>> Date maturity(20, Jun, 2014) ;
>> double strikePrice = 100.0 ;
>> Option::Type optionType = Option::Call ;
>> double dividendYield = 0.0 ;
>> double impliedVolatility = 0.15 ;
>> double underlyingPrice = 100.0 ;
>>
>> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new
>> SimpleQuote
>> (underlyingPrice)));
>>
>> Handle<YieldTermStructure> rateTermStructure
>> (boost::shared_ptr<YieldTermStructure>(new
>> PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments,
>> dc))) ;
>>
>> Handle<YieldTermStructure>
>> dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new
>> FlatForward(today,
>> dividendYield,
>> dc)));
>> Handle<BlackVolTermStructure>
>> volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
>> BlackConstantVol(today,
>> calendar,
>> impliedVolatility,
>> dc)));
>>
>> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
>> BlackScholesMertonProcess(underlyingHandle,
>> dividendTermStructure,
>> rateTermStructure,
>> volTermStructure));
>>
>> boost::shared_ptr<Exercise> americanExercise(new AmericanExercise
>> (today,
>> maturity)) ;
>> boost::shared_ptr<StrikedTypePayoff> payoff(new
> PlainVanillaPayoff
>> (optionType, strikePrice));
>>
>> VanillaOption americanOption(payoff, americanExercise);
>>
>> americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(
>> new FDAmericanEngine<CrankNicolson>(bsmProcess,
>> 101,100)));
>>
>> cout << americanOption.NPV() << endl ;
>>
>> return 0 ;
>> }
>>
> ------------------------------------------------------------------------------
>
>>
>>
>> Want fast and easy access to all the code in your enterprise? Index
>> and
>> search up to 200,000 lines of code with a free copy of Black Duck
>> Code Sight - the same software that powers the world's largest code
>> search on Ohloh, the Black Duck Open Hub! Try it now.
>>
> http://p.sf.net/sfu/bds_______________________________________________
>>
>> QuantLib-users mailing list
>> [hidden email]
>> https://lists.sourceforge.net/lists/listinfo/quantlib-users
>>
>>
> ------------------------------------------------------------------------------
>
>> Want fast and easy access to all the code in your enterprise? Index and
>> search up to 200,000 lines of code with a free copy of Black Duck
>> Code Sight - the same software that powers the world's largest code
>> search on Ohloh, the Black Duck Open Hub! Try it now.
>> http://p.sf.net/sfu/bds
>> _______________________________________________
>> QuantLib-users mailing list
>> [hidden email]
>> https://lists.sourceforge.net/lists/listinfo/quantlib-users
>>
>

Peter Caspers-4

Re: UK Holiday for USDLibor/PiecewiseYieldCurve

reminding myself of what workaround we actually implemented for swap
helper :o ... and what Murex offers here (always a good reference),
the full set of options would probably be

default - require historic fixings, take today's fixing if available
ignore today's tixing - require historic fixings, always ignore today's fixing
ignore all fixings - ignore all fixings (i.e. estimate forward rates
always on the curve - this is technically ok even if the fixing date
is in the past, since only the index estimation period's
start and end date enter the estimation, and these two dates are
always greater or equal today)

these options to be applied only in the context of yield curve
bootstrapping, ideally to be set on the level of the curve itself.

On 24 July 2014 10:26, Peter Caspers <[hidden email]> wrote:

> Hi Andres,
>
> yes, the index in the deposit rate helper seems to be merely a
> convenient way for coding - with the one exception that deposits do
> not (directly) have anything to do with actual index fixings, and this
> is also where the code breaks. And yes, it gets even worse in your
> example when you use the USD Libor index for USD cash deposits on a UK
> but non-NY holiday (you could use the index-free constructor though
> and pass the NY calendar directly instead, right, solving half of the
> problem ?). So just treating the deposit helper's index and its fixing
> as a computational tool to get the implied Quote seems a reasonable
> solution, in your USD / Libor example yielding the correct conventions
> even, if I am not mistaken ? I wouldn't introduce any flag or method
> to activate this behaviour, it should be the only one present (because
> checking for holidays can easily be done in the user's code if
> necessary at all).
>
> For swaps on the other hand I think the current behaviour is fine -
> what fixing would you take for the first period in a Euribor swap on a
> TARGET holiday ? Probably the same as you take after the fixing around
> 11am on a trading day, i.e. the last known fixing. That is exactly the
> logic applied in the library. I wouldn't change anything here. It
> might be that you get an exception because you just don't provide the
> historic fixing, but in this case it is a legitimate demand of the
> library and working around this should only be done in the user code
> imo.
>
> best
> Peter
>
>
>
>
>
> On 24 July 2014 09:23, Andres Hernandez <[hidden email]> wrote:
>>
>> Hi Peter,
>>
>> I agree that it would be nice to at least have it as an option. We
>> currently do a small workaround, whereby we set in our system a parameter
>> for the curve specifying the holiday calendar for that curve. The only use
>> of that calendar is to determine if it is a holiday for the curve or not
>> and if yes to make a fallback to yesterday's values. In the case of the EUR
>> curves, this is arguably a good behaviour as the index and the geographical
>> region have the same calendar, but for the USD case, we are basically
>> forcing a fallback for a holiday on a different country. The issue is
>> compounded by the fact that in this case the deposit is traded on that day,
>> it is just the index that gets no fixing.
>>
>> I took a look and I am almost sure that removing the evaluation date from
>> the fixing calendar in InterestRate constructor would be enough to make
>> deposits go through. This could simply be accomplished with an extra public
>> member function that one would call if one wanted it to go through on
>> holidays. A more involved and important problem would be to get the
>> SwapRateHelper to go though on holidays (if desired), as I am positive that
>> that was the reason for our instituting the fallback. I will see if I have
>> the time to change it, and would then probably subscribe to the developers
>> list to get some feedback in how to go about changing it.
>>
>>
>> Mit freundlichen Grüßen / Kind regards
>>
>> Dr. Andres Hernandez
>>
>> Senior Financial Engineer
>> Business Analytics
>> Risk Analytics
>>
>>
>>
>>
>>
>> Phone: +49-69-6645-1351 IBM Deutschland (Embedded
>> image moved
>> to file:
>> pic02235.gif)
>>
>> Mobile: +49-163-609-3280 Wilhelm-Fay-Str. 30-34
>>
>> E-Mail: [hidden email] 65936 Frankfurt am Main
>>
>> Germany
>>
>>
>>
>>
>>
>> IBM Deutschland
>> GmbH /
>> Vorsitzender des
>> Aufsichtsrats:
>> Martin Jetter
>> Geschäftsführung:
>> Martina Koederitz
>> (Vorsitzende),
>> Reinhard Reschke,
>> Dieter Scholz,
>> Gregor Pillen,
>> Christian Noll,
>> Ivo Koerner
>> Sitz der
>> Gesellschaft:
>> Ehningen /
>> Registergericht:
>> Amtsgericht
>> Stuttgart, HRB
>> 14562 /
>> WEEE-Reg.-Nr. DE
>> 99369940
>>
>>
>>
>>
>>
>>
>> From: Peter Caspers <[hidden email]>
>> To: Andres Hernandez/Germany/IBM@IBMDE
>> Cc: George Cowie <[hidden email]>, QuantLib users
>> <[hidden email]>
>> Date: 23/07/2014 21:56
>> Subject: Re: [Quantlib-users] UK Holiday for
>> USDLibor/PiecewiseYieldCurve
>>
>>
>>
>> Hi,
>>
>> I think this question is an older one but I don't remember that it was
>> finally decided how to handle this situation (?) - which is, should we
>> allow to bootstrap a curve on a day which is a holiday for the cash
>> instruments in the curve or not. If negative, we could maybe throw a
>> different message ( the current one indeed sounds a bit paradoxical ).
>> If yes ( which would be my personal preference, because you simply
>> want to run your daily processes spitting out numbers even if not a
>> hundred percent sensible all of them ), one has to define how a
>> deposit which is not traded on that day should be calculated. One
>> obvious way would be to just treat the evaluation date as a business
>> day when calculating the helper's fixing date. I did not try, but it
>> may be enough to join the business days of the helper's calendar with
>> a new one just having one business day (the evaluation date) before
>> passing it to the no-fix index to make this kind of solution work.
>> What do you think ?
>>
>> Peter
>>
>>
>> On 23 July 2014 15:05, Andres Hernandez <[hidden email]>
>> wrote:
>>>
>>> Hi George,
>>>
>>> you are still setting the evaluation date to a holiday:
>>>
>>> Date today (5 , May ,2014);
>>> Date liborDate (2, May, 2014) ;
>>> Settings::instance().evaluationDate() = today ;
>>>
>>>
>>> Let's follow the logic:
>>> During bootstrapping, the method impliedQuote() of the rate helper is
>>> called to compare against the market quote. The DepositRateHelper simply
>>> returns the fixing corresponding to the fixingDate_ as a return value.
>>> The fixingDate_ is set during initialization as the return value of the
>>> function InterestRate::fixingDate(earliestDate_). The function
>>> InterestRate::fixingDate sets back the date given as input parameter by
>> the
>>> number of fixing days according to the index' calendar.
>>> The parameter earliestDate_ being passed to fixingDate is set during
>>> initialization to the evaluation date (which you set to the 5th of May)
>>> moved forward by the number of fixing days according to the index'
>>> calendar.
>>> Note that if the evaluation date is a holiday you do not end up with the
>>> same date, but rather with the last business day prior to evaluation
>> date.
>>> The first act moves the date from the 5th of May to the 7th and checks
>>> whether it is a business day, which it is. The second move sets it back
>> two
>>> days to the 5th, and checks whether it is a business day, which it isn't,
>>> so it sets it back until it finds a business day, i.e. the 2nd of May.
>>> Since the value is in the past, it requests a historical fixing, which
>> you
>>> did not provide and so the QL_REQUIRE throws.
>>>
>>> Mit freundlichen Grüßen / Kind regards
>>>
>>> Dr. Andres Hernandez
>>>
>>> Senior Financial Engineer
>>> Business Analytics
>>> Risk Analytics
>>>
>>>
>>>
>>>
>>>
>>> Phone: +49-69-6645-1351 IBM Deutschland
>> (Embedded
>>>
>> image moved
>>>
>> to file:
>>>
>> pic64586.gif)
>>>
>>> Mobile: +49-163-609-3280 Wilhelm-Fay-Str. 30-34
>>>
>>> E-Mail: [hidden email] 65936 Frankfurt am Main
>>>
>>> Germany
>>>
>>>
>>>
>>>
>>>
>>> IBM Deutschland
>>> GmbH /
>>> Vorsitzender des
>>> Aufsichtsrats:
>>> Martin Jetter
>>> Geschäftsführung:
>>> Martina Koederitz
>>> (Vorsitzende),
>>> Reinhard Reschke,
>>> Dieter Scholz,
>>> Gregor Pillen,
>>> Christian Noll,
>>> Ivo Koerner
>>> Sitz der
>>> Gesellschaft:
>>> Ehningen /
>>> Registergericht:
>>> Amtsgericht
>>> Stuttgart, HRB
>>> 14562 /
>>> WEEE-Reg.-Nr. DE
>>> 99369940
>>>
>>>
>>>
>>>
>>>
>>>
>>> From: George Cowie <[hidden email]>
>>> To: Andres Hernandez/Germany/IBM@IBMDE
>>> Cc: QuantLib users <[hidden email]>
>>> Date: 23/07/2014 11:51
>>> Subject: Re: [Quantlib-users] UK Holiday for
>>> USDLibor/PiecewiseYieldCurve
>>>
>>>
>>>
>>>
>>> Thank you Andres. What you're describing makes sense, but I can't seem
>> to
>>> avoid an error when trying to value the option on May 5, 2014. I've
>> copied
>>> the updated code snippet with changed section highlighted/bolded below.
>> I
>>> seem to always get an exception of the form:
>>>
>>> terminate called after throwing an instance of 'QuantLib::Error'
>>> what(): 1st iteration: failed at 1st alive instrument, maturity May
>>> 14th, 2014, reference date May 6th, 2014: Missing no-fix1W Actual/360
>>> fixing for May 2nd, 2014
>>>
>>> Is there something wrong with my approach here?
>>>
>>> Thanks,
>>> George
>>>
>>>
>>> #include <ql/quantlib.hpp>
>>>
>>> using namespace std ;
>>> using namespace QuantLib ;
>>>
>>> int main(int argc, char** argv)
>>> {
>>> vector<Rate> liborRates ;
>>> liborRates.push_back(0.12075 / 100.0) ; //1W
>>> liborRates.push_back(0.1515 / 100.0) ; //1M
>>> liborRates.push_back(0.1935 / 100.0) ; //2M
>>> liborRates.push_back(0.22535 / 100.0) ; //3M
>>> liborRates.push_back(0.323 / 100.0) ; //6M
>>> liborRates.push_back(0.549 / 100.0) ; //12M
>>>
>>> vector<Period> periods ;
>>> periods.push_back(Period(1,Weeks)) ;
>>> periods.push_back(Period(1,Months)) ;
>>> periods.push_back(Period(2,Months)) ;
>>> periods.push_back(Period(3,Months)) ;
>>> periods.push_back(Period(6,Months)) ;
>>> periods.push_back(Period(12,Months)) ;
>>>
>>> vector<boost::shared_ptr<RateHelper> > instruments ;
>>>
>>> for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)
>>> {
>>> boost::shared_ptr<Quote> liborQuote(new SimpleQuote(liborRates
>> [i])) ;
>>> Handle<Quote> liborHandle(liborQuote) ;
>>> boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods[i])) ;
>>> boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper
>>> (liborHandle,liborIndex)) ;
>>> instruments.push_back(rateHelper) ;
>>> }
>>>
>>> Calendar calendar = UnitedStates();
>>> Date today (5 , May ,2014);
>>> Date liborDate (2, May, 2014) ;
>>> Settings::instance().evaluationDate() = today ;
>>> Natural settlementDays = 2;
>>> Date settlement = calendar.advance ( liborDate , settlementDays ,
>> Days );
>>> DayCounter dc = Actual360 ();
>>>
>>> Date maturity(20, Jun, 2014) ;
>>> double strikePrice = 100.0 ;
>>> Option::Type optionType = Option::Call ;
>>> double dividendYield = 0.0 ;
>>> double impliedVolatility = 0.15 ;
>>> double underlyingPrice = 100.0 ;
>>>
>>> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new SimpleQuote
>>> (underlyingPrice)));
>>>
>>> Handle<YieldTermStructure> rateTermStructure
>>> (boost::shared_ptr<YieldTermStructure>(new
>>> PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments, dc))) ;
>>>
>>> Handle<YieldTermStructure>
>>> dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new
>>> FlatForward(today,
>>> dividendYield,
>>> dc)));
>>> Handle<BlackVolTermStructure>
>>> volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
>>> BlackConstantVol(today,
>>> calendar,
>>> impliedVolatility,
>>> dc)));
>>>
>>> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
>>> BlackScholesMertonProcess(underlyingHandle,
>>> dividendTermStructure,
>>> rateTermStructure,
>>> volTermStructure));
>>>
>>> boost::shared_ptr<Exercise> americanExercise(new AmericanExercise
>> (today,
>>> maturity)) ;
>>> boost::shared_ptr<StrikedTypePayoff> payoff(new PlainVanillaPayoff
>>> (optionType, strikePrice));
>>>
>>> VanillaOption americanOption(payoff, americanExercise);
>>>
>>> americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(
>>> new FDAmericanEngine<CrankNicolson>(bsmProcess,
>>> 101,100)));
>>>
>>> cout << americanOption.NPV() << endl ;
>>>
>>> return 0 ;
>>> }
>>>
>>>
>>>
>>> On Tue, Jul 22, 2014 at 7:39 AM, Andres Hernandez <
>>> [hidden email]> wrote:
>>>
>>> Hi George,
>>>
>>> the problem occurs because you are trying to bootstrap on a
>>> UnitedKingdom
>>> holiday, which is implicitly included in the definition of
>> USDLibor.
>>> The
>>> Libor class makes a joint calendar between the given financial
>> center
>>> calendar and the UK calendar. Since Libor is a London poll, there
>> are
>>> only
>>> fixings for it on UK business days. If one wants to bootstrap the
>>> curve on
>>> a holiday, the indice's historical fixings (actually just the
>> closest
>>> fixing to the given date) need to be provided.
>>>
>>>
>>> Mit freundlichen Grüßen / Kind regards
>>>
>>> Dr. Andres Hernandez
>>>
>>> Senior Financial Engineer
>>> Business Analytics
>>> Risk Analytics
>>>
>>>
>>>
>>>
>>>
>>> Phone: +49-69-6645-1351 IBM Deutschland
>>> (Embedded
>>> image moved
>>> to file:
>>> pic20635.gif)
>>>
>>> Mobile: +49-163-609-3280 Wilhelm-Fay-Str.
>>> 30-34
>>>
>>> E-Mail: [hidden email] 65936 Frankfurt
>> am
>>> Main
>>>
>>> Germany
>>>
>>>
>>>
>>>
>>>
>>> IBM Deutschland
>>> GmbH /
>>> Vorsitzender des
>>> Aufsichtsrats:
>>> Martin Jetter
>>> Geschäftsführung:
>>> Martina Koederitz
>>> (Vorsitzende),
>>> Reinhard Reschke,
>>> Dieter Scholz,
>>> Gregor Pillen,
>>> Christian Noll,
>>> Ivo Koerner
>>> Sitz der
>>> Gesellschaft:
>>> Ehningen /
>>> Registergericht:
>>> Amtsgericht
>>> Stuttgart, HRB
>>> 14562 /
>>> WEEE-Reg.-Nr. DE
>>> 99369940
>>>
>>>
>>>
>>>
>>>
>>>
>>> From: George Cowie <[hidden email]>
>>> To: QuantLib users <[hidden email]>
>>> Date: 22/07/2014 14:10
>>> Subject: [Quantlib-users] UK Holiday for
>>> USDLibor/PiecewiseYieldCurve
>>>
>>>
>>>
>>> Hello,
>>>
>>> I'm using PiecewiseYieldCurve to build a YieldTermStructure which
>>> is
>>> ultimately used by BlackScholesMertonProcess in my option pricing.
>> I
>>> believe I'm running into an issue on days that are holidays in the
>> UK
>>> (but
>>> not the US). Here's the error I'm seeing when trying to run the
>> code
>>> below
>>> for May 5, 2014 (Early May Bank Holiday):
>>>
>>> terminate called after throwing an instance of 'QuantLib::Error'
>>> what(): 1st iteration: failed at 1st alive instrument, maturity
>>> May
>>> 14th, 2014, reference date May 7th, 2014: Missing no-fix1W
>> Actual/360
>>> fixing for May 2nd, 2014
>>> Aborted
>>>
>>> Everything seems to work fine on other days... Am I doing
>>> something
>>> wrong with setting up the PiecewiseYieldCurve?
>>>
>>> Thanks,
>>> George
>>>
>>> #include <ql/quantlib.hpp>
>>>
>>> using namespace std ;
>>> using namespace QuantLib ;
>>>
>>> int main(int argc, char** argv)
>>> {
>>> vector<Rate> liborRates ;
>>> liborRates.push_back(0.12075 / 100.0) ; //1W
>>> liborRates.push_back(0.1515 / 100.0) ; //1M
>>> liborRates.push_back(0.1935 / 100.0) ; //2M
>>> liborRates.push_back(0.22535 / 100.0) ; //3M
>>> liborRates.push_back(0.323 / 100.0) ; //6M
>>> liborRates.push_back(0.549 / 100.0) ; //12M
>>>
>>> vector<Period> periods ;
>>> periods.push_back(Period(1,Weeks)) ;
>>> periods.push_back(Period(1,Months)) ;
>>> periods.push_back(Period(2,Months)) ;
>>> periods.push_back(Period(3,Months)) ;
>>> periods.push_back(Period(6,Months)) ;
>>> periods.push_back(Period(12,Months)) ;
>>>
>>> vector<boost::shared_ptr<RateHelper> > instruments ;
>>>
>>> for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)
>>> {
>>> boost::shared_ptr<Quote> liborQuote(new SimpleQuote
>> (liborRates
>>> [i])) ;
>>> Handle<Quote> liborHandle(liborQuote) ;
>>> boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods
>>> [i])) ;
>>> boost::shared_ptr<RateHelper> rateHelper(new
>> DepositRateHelper
>>> (liborHandle,liborIndex)) ;
>>> instruments.push_back(rateHelper) ;
>>> }
>>>
>>> Calendar calendar = UnitedStates();
>>> Date today (5 , May ,2014);
>>> Settings::instance().evaluationDate() = today ;
>>> Natural settlementDays = 2;
>>> Date settlement = calendar.advance ( today , settlementDays ,
>>> Days );
>>> DayCounter dc = Actual360 ();
>>>
>>> Date maturity(20, Jun, 2014) ;
>>> double strikePrice = 100.0 ;
>>> Option::Type optionType = Option::Call ;
>>> double dividendYield = 0.0 ;
>>> double impliedVolatility = 0.15 ;
>>> double underlyingPrice = 100.0 ;
>>>
>>> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new
>>> SimpleQuote
>>> (underlyingPrice)));
>>>
>>> Handle<YieldTermStructure> rateTermStructure
>>> (boost::shared_ptr<YieldTermStructure>(new
>>> PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments,
>>> dc))) ;
>>>
>>> Handle<YieldTermStructure>
>>> dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new
>>> FlatForward(today,
>>> dividendYield,
>>> dc)));
>>> Handle<BlackVolTermStructure>
>>> volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
>>> BlackConstantVol(today,
>>> calendar,
>>> impliedVolatility,
>>> dc)));
>>>
>>> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
>>> BlackScholesMertonProcess(underlyingHandle,
>>> dividendTermStructure,
>>> rateTermStructure,
>>> volTermStructure));
>>>
>>> boost::shared_ptr<Exercise> americanExercise(new AmericanExercise
>>> (today,
>>> maturity)) ;
>>> boost::shared_ptr<StrikedTypePayoff> payoff(new
>> PlainVanillaPayoff
>>> (optionType, strikePrice));
>>>
>>> VanillaOption americanOption(payoff, americanExercise);
>>>
>>> americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(
>>> new FDAmericanEngine<CrankNicolson>(bsmProcess,
>>> 101,100)));
>>>
>>> cout << americanOption.NPV() << endl ;
>>>
>>> return 0 ;
>>> }
>>>
>> ------------------------------------------------------------------------------
>>
>>>
>>>
>>> Want fast and easy access to all the code in your enterprise? Index
>>> and
>>> search up to 200,000 lines of code with a free copy of Black Duck
>>> Code Sight - the same software that powers the world's largest code
>>> search on Ohloh, the Black Duck Open Hub! Try it now.
>>>
>> http://p.sf.net/sfu/bds_______________________________________________
>>>
>>> QuantLib-users mailing list
>>> [hidden email]
>>> https://lists.sourceforge.net/lists/listinfo/quantlib-users
>>>
>>>
>> ------------------------------------------------------------------------------
>>
>>> Want fast and easy access to all the code in your enterprise? Index and
>>> search up to 200,000 lines of code with a free copy of Black Duck
>>> Code Sight - the same software that powers the world's largest code
>>> search on Ohloh, the Black Duck Open Hub! Try it now.
>>> http://p.sf.net/sfu/bds
>>> _______________________________________________
>>> QuantLib-users mailing list
>>> [hidden email]
>>> https://lists.sourceforge.net/lists/listinfo/quantlib-users
>>>
>>

George Cowie

Re: UK Holiday for USDLibor/PiecewiseYieldCurve

Interesting discussion... sorry to change the subject, but I'm still trying to figure out how to adapt my code to this situation. How do I force my YieldTermStructure to bootstrap with a different date? I have to value the option on May 5th, but as you pointed out, changing the evaluation date to a holiday has the described consequences.

On Thu, Jul 24, 2014 at 6:00 AM, Peter Caspers <[hidden email]> wrote:

reminding myself of what workaround we actually implemented for swap
helper :o ... and what Murex offers here (always a good reference),
the full set of options would probably be

default - require historic fixings, take today's fixing if available
ignore today's tixing - require historic fixings, always ignore today's fixing
ignore all fixings - ignore all fixings (i.e. estimate forward rates
always on the curve - this is technically ok even if the fixing date
is in the past, since only the index estimation period's
start and end date enter the estimation, and these two dates are
always greater or equal today)

these options to be applied only in the context of yield curve
bootstrapping, ideally to be set on the level of the curve itself.

On 24 July 2014 10:26, Peter Caspers <[hidden email]> wrote:
> Hi Andres,
>
> yes, the index in the deposit rate helper seems to be merely a
> convenient way for coding - with the one exception that deposits do
> not (directly) have anything to do with actual index fixings, and this
> is also where the code breaks. And yes, it gets even worse in your
> example when you use the USD Libor index for USD cash deposits on a UK
> but non-NY holiday (you could use the index-free constructor though
> and pass the NY calendar directly instead, right, solving half of the
> problem ?). So just treating the deposit helper's index and its fixing
> as a computational tool to get the implied Quote seems a reasonable
> solution, in your USD / Libor example yielding the correct conventions
> even, if I am not mistaken ? I wouldn't introduce any flag or method
> to activate this behaviour, it should be the only one present (because
> checking for holidays can easily be done in the user's code if
> necessary at all).
>
> For swaps on the other hand I think the current behaviour is fine -
> what fixing would you take for the first period in a Euribor swap on a
> TARGET holiday ? Probably the same as you take after the fixing around
> 11am on a trading day, i.e. the last known fixing. That is exactly the
> logic applied in the library. I wouldn't change anything here. It
> might be that you get an exception because you just don't provide the
> historic fixing, but in this case it is a legitimate demand of the
> library and working around this should only be done in the user code
> imo.
>
> best
> Peter
>
>
>
>
>
> On 24 July 2014 09:23, Andres Hernandez <[hidden email]> wrote:
>>
>> Hi Peter,
>>
>> I agree that it would be nice to at least have it as an option. We
>> currently do a small workaround, whereby we set in our system a parameter
>> for the curve specifying the holiday calendar for that curve. The only use
>> of that calendar is to determine if it is a holiday for the curve or not
>> and if yes to make a fallback to yesterday's values. In the case of the EUR
>> curves, this is arguably a good behaviour as the index and the geographical
>> region have the same calendar, but for the USD case, we are basically
>> forcing a fallback for a holiday on a different country. The issue is
>> compounded by the fact that in this case the deposit is traded on that day,
>> it is just the index that gets no fixing.
>>
>> I took a look and I am almost sure that removing the evaluation date from
>> the fixing calendar in InterestRate constructor would be enough to make
>> deposits go through. This could simply be accomplished with an extra public
>> member function that one would call if one wanted it to go through on
>> holidays. A more involved and important problem would be to get the
>> SwapRateHelper to go though on holidays (if desired), as I am positive that
>> that was the reason for our instituting the fallback. I will see if I have
>> the time to change it, and would then probably subscribe to the developers
>> list to get some feedback in how to go about changing it.
>>
>>
>> Mit freundlichen Grüßen / Kind regards
>>
>> Dr. Andres Hernandez
>>
>> Senior Financial Engineer
>> Business Analytics
>> Risk Analytics
>>
>>
>>
>>
>>
>> Phone: <a href="tel:%2B49-69-6645-1351" value="+496966451351">+49-69-6645-1351 IBM Deutschland (Embedded
>> image moved
>> to file:
>> pic02235.gif)
>>
>> Mobile: <a href="tel:%2B49-163-609-3280" value="+491636093280">+49-163-609-3280 Wilhelm-Fay-Str. 30-34
>>
>> E-Mail: [hidden email] 65936 Frankfurt am Main
>>
>> Germany
>>
>>
>>
>>
>>
>> IBM Deutschland
>> GmbH /
>> Vorsitzender des
>> Aufsichtsrats:
>> Martin Jetter
>> Geschäftsführung:
>> Martina Koederitz
>> (Vorsitzende),
>> Reinhard Reschke,
>> Dieter Scholz,
>> Gregor Pillen,
>> Christian Noll,
>> Ivo Koerner
>> Sitz der
>> Gesellschaft:
>> Ehningen /
>> Registergericht:
>> Amtsgericht
>> Stuttgart, HRB
>> 14562 /
>> WEEE-Reg.-Nr. DE
>> 99369940
>>
>>
>>
>>
>>
>>
>> From: Peter Caspers <[hidden email]>
>> To: Andres Hernandez/Germany/IBM@IBMDE
>> Cc: George Cowie <[hidden email]>, QuantLib users
>> <[hidden email]>
>> Date: 23/07/2014 21:56
>> Subject: Re: [Quantlib-users] UK Holiday for
>> USDLibor/PiecewiseYieldCurve
>>
>>
>>
>> Hi,
>>
>> I think this question is an older one but I don't remember that it was
>> finally decided how to handle this situation (?) - which is, should we
>> allow to bootstrap a curve on a day which is a holiday for the cash
>> instruments in the curve or not. If negative, we could maybe throw a
>> different message ( the current one indeed sounds a bit paradoxical ).
>> If yes ( which would be my personal preference, because you simply
>> want to run your daily processes spitting out numbers even if not a
>> hundred percent sensible all of them ), one has to define how a
>> deposit which is not traded on that day should be calculated. One
>> obvious way would be to just treat the evaluation date as a business
>> day when calculating the helper's fixing date. I did not try, but it
>> may be enough to join the business days of the helper's calendar with
>> a new one just having one business day (the evaluation date) before
>> passing it to the no-fix index to make this kind of solution work.
>> What do you think ?
>>
>> Peter
>>
>>
>> On 23 July 2014 15:05, Andres Hernandez <[hidden email]>
>> wrote:
>>>
>>> Hi George,
>>>
>>> you are still setting the evaluation date to a holiday:
>>>
>>> Date today (5 , May ,2014);
>>> Date liborDate (2, May, 2014) ;
>>> Settings::instance().evaluationDate() = today ;
>>>
>>>
>>> Let's follow the logic:
>>> During bootstrapping, the method impliedQuote() of the rate helper is
>>> called to compare against the market quote. The DepositRateHelper simply
>>> returns the fixing corresponding to the fixingDate_ as a return value.
>>> The fixingDate_ is set during initialization as the return value of the
>>> function InterestRate::fixingDate(earliestDate_). The function
>>> InterestRate::fixingDate sets back the date given as input parameter by
>> the
>>> number of fixing days according to the index' calendar.
>>> The parameter earliestDate_ being passed to fixingDate is set during
>>> initialization to the evaluation date (which you set to the 5th of May)
>>> moved forward by the number of fixing days according to the index'
>>> calendar.
>>> Note that if the evaluation date is a holiday you do not end up with the
>>> same date, but rather with the last business day prior to evaluation
>> date.
>>> The first act moves the date from the 5th of May to the 7th and checks
>>> whether it is a business day, which it is. The second move sets it back
>> two
>>> days to the 5th, and checks whether it is a business day, which it isn't,
>>> so it sets it back until it finds a business day, i.e. the 2nd of May.
>>> Since the value is in the past, it requests a historical fixing, which
>> you
>>> did not provide and so the QL_REQUIRE throws.
>>>
>>> Mit freundlichen Grüßen / Kind regards
>>>
>>> Dr. Andres Hernandez
>>>
>>> Senior Financial Engineer
>>> Business Analytics
>>> Risk Analytics
>>>
>>>
>>>
>>>
>>>
>>> Phone: +49-69-6645-1351 IBM Deutschland
>> (Embedded
>>>
>> image moved
>>>
>> to file:
>>>
>> pic64586.gif)
>>>
>>> Mobile: +49-163-609-3280 Wilhelm-Fay-Str. 30-34
>>>
>>> E-Mail: [hidden email] 65936 Frankfurt am Main
>>>
>>> Germany
>>>
>>>
>>>
>>>
>>>
>>> IBM Deutschland
>>> GmbH /
>>> Vorsitzender des
>>> Aufsichtsrats:
>>> Martin Jetter
>>> Geschäftsführung:
>>> Martina Koederitz
>>> (Vorsitzende),
>>> Reinhard Reschke,
>>> Dieter Scholz,
>>> Gregor Pillen,
>>> Christian Noll,
>>> Ivo Koerner
>>> Sitz der
>>> Gesellschaft:
>>> Ehningen /
>>> Registergericht:
>>> Amtsgericht
>>> Stuttgart, HRB
>>> 14562 /
>>> WEEE-Reg.-Nr. DE
>>> 99369940
>>>
>>>
>>>
>>>
>>>
>>>
>>> From: George Cowie <[hidden email]>
>>> To: Andres Hernandez/Germany/IBM@IBMDE
>>> Cc: QuantLib users <[hidden email]>
>>> Date: 23/07/2014 11:51
>>> Subject: Re: [Quantlib-users] UK Holiday for
>>> USDLibor/PiecewiseYieldCurve
>>>
>>>
>>>
>>>
>>> Thank you Andres. What you're describing makes sense, but I can't seem
>> to
>>> avoid an error when trying to value the option on May 5, 2014. I've
>> copied
>>> the updated code snippet with changed section highlighted/bolded below.
>> I
>>> seem to always get an exception of the form:
>>>
>>> terminate called after throwing an instance of 'QuantLib::Error'
>>> what(): 1st iteration: failed at 1st alive instrument, maturity May
>>> 14th, 2014, reference date May 6th, 2014: Missing no-fix1W Actual/360
>>> fixing for May 2nd, 2014
>>>
>>> Is there something wrong with my approach here?
>>>
>>> Thanks,
>>> George
>>>
>>>
>>> #include <ql/quantlib.hpp>
>>>
>>> using namespace std ;
>>> using namespace QuantLib ;
>>>
>>> int main(int argc, char** argv)
>>> {
>>> vector<Rate> liborRates ;
>>> liborRates.push_back(0.12075 / 100.0) ; //1W
>>> liborRates.push_back(0.1515 / 100.0) ; //1M
>>> liborRates.push_back(0.1935 / 100.0) ; //2M
>>> liborRates.push_back(0.22535 / 100.0) ; //3M
>>> liborRates.push_back(0.323 / 100.0) ; //6M
>>> liborRates.push_back(0.549 / 100.0) ; //12M
>>>
>>> vector<Period> periods ;
>>> periods.push_back(Period(1,Weeks)) ;
>>> periods.push_back(Period(1,Months)) ;
>>> periods.push_back(Period(2,Months)) ;
>>> periods.push_back(Period(3,Months)) ;
>>> periods.push_back(Period(6,Months)) ;
>>> periods.push_back(Period(12,Months)) ;
>>>
>>> vector<boost::shared_ptr<RateHelper> > instruments ;
>>>
>>> for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)
>>> {
>>> boost::shared_ptr<Quote> liborQuote(new SimpleQuote(liborRates
>> [i])) ;
>>> Handle<Quote> liborHandle(liborQuote) ;
>>> boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods[i])) ;
>>> boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper
>>> (liborHandle,liborIndex)) ;
>>> instruments.push_back(rateHelper) ;
>>> }
>>>
>>> Calendar calendar = UnitedStates();
>>> Date today (5 , May ,2014);
>>> Date liborDate (2, May, 2014) ;
>>> Settings::instance().evaluationDate() = today ;
>>> Natural settlementDays = 2;
>>> Date settlement = calendar.advance ( liborDate , settlementDays ,
>> Days );
>>> DayCounter dc = Actual360 ();
>>>
>>> Date maturity(20, Jun, 2014) ;
>>> double strikePrice = 100.0 ;
>>> Option::Type optionType = Option::Call ;
>>> double dividendYield = 0.0 ;
>>> double impliedVolatility = 0.15 ;
>>> double underlyingPrice = 100.0 ;
>>>
>>> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new SimpleQuote
>>> (underlyingPrice)));
>>>
>>> Handle<YieldTermStructure> rateTermStructure
>>> (boost::shared_ptr<YieldTermStructure>(new
>>> PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments, dc))) ;
>>>
>>> Handle<YieldTermStructure>
>>> dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new
>>> FlatForward(today,
>>> dividendYield,
>>> dc)));
>>> Handle<BlackVolTermStructure>
>>> volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
>>> BlackConstantVol(today,
>>> calendar,
>>> impliedVolatility,
>>> dc)));
>>>
>>> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
>>> BlackScholesMertonProcess(underlyingHandle,
>>> dividendTermStructure,
>>> rateTermStructure,
>>> volTermStructure));
>>>
>>> boost::shared_ptr<Exercise> americanExercise(new AmericanExercise
>> (today,
>>> maturity)) ;
>>> boost::shared_ptr<StrikedTypePayoff> payoff(new PlainVanillaPayoff
>>> (optionType, strikePrice));
>>>
>>> VanillaOption americanOption(payoff, americanExercise);
>>>
>>> americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(
>>> new FDAmericanEngine<CrankNicolson>(bsmProcess,
>>> 101,100)));
>>>
>>> cout << americanOption.NPV() << endl ;
>>>
>>> return 0 ;
>>> }
>>>
>>>
>>>
>>> On Tue, Jul 22, 2014 at 7:39 AM, Andres Hernandez <
>>> [hidden email]> wrote:
>>>
>>> Hi George,
>>>
>>> the problem occurs because you are trying to bootstrap on a
>>> UnitedKingdom
>>> holiday, which is implicitly included in the definition of
>> USDLibor.
>>> The
>>> Libor class makes a joint calendar between the given financial
>> center
>>> calendar and the UK calendar. Since Libor is a London poll, there
>> are
>>> only
>>> fixings for it on UK business days. If one wants to bootstrap the
>>> curve on
>>> a holiday, the indice's historical fixings (actually just the
>> closest
>>> fixing to the given date) need to be provided.
>>>
>>>
>>> Mit freundlichen Grüßen / Kind regards
>>>
>>> Dr. Andres Hernandez
>>>
>>> Senior Financial Engineer
>>> Business Analytics
>>> Risk Analytics
>>>
>>>
>>>
>>>
>>>
>>> Phone: <a href="tel:%2B49-69-6645-1351" value="+496966451351">+49-69-6645-1351 IBM Deutschland
>>> (Embedded
>>> image moved
>>> to file:
>>> pic20635.gif)
>>>
>>> Mobile: <a href="tel:%2B49-163-609-3280" value="+491636093280">+49-163-609-3280 Wilhelm-Fay-Str.
>>> 30-34
>>>
>>> E-Mail: [hidden email] 65936 Frankfurt
>> am
>>> Main
>>>
>>> Germany
>>>
>>>
>>>
>>>
>>>
>>> IBM Deutschland
>>> GmbH /
>>> Vorsitzender des
>>> Aufsichtsrats:
>>> Martin Jetter
>>> Geschäftsführung:
>>> Martina Koederitz
>>> (Vorsitzende),
>>> Reinhard Reschke,
>>> Dieter Scholz,
>>> Gregor Pillen,
>>> Christian Noll,
>>> Ivo Koerner
>>> Sitz der
>>> Gesellschaft:
>>> Ehningen /
>>> Registergericht:
>>> Amtsgericht
>>> Stuttgart, HRB
>>> 14562 /
>>> WEEE-Reg.-Nr. DE
>>> 99369940
>>>
>>>
>>>
>>>
>>>
>>>
>>> From: George Cowie <[hidden email]>
>>> To: QuantLib users <[hidden email]>
>>> Date: 22/07/2014 14:10
>>> Subject: [Quantlib-users] UK Holiday for
>>> USDLibor/PiecewiseYieldCurve
>>>
>>>
>>>
>>> Hello,
>>>
>>> I'm using PiecewiseYieldCurve to build a YieldTermStructure which
>>> is
>>> ultimately used by BlackScholesMertonProcess in my option pricing.
>> I
>>> believe I'm running into an issue on days that are holidays in the
>> UK
>>> (but
>>> not the US). Here's the error I'm seeing when trying to run the
>> code
>>> below
>>> for May 5, 2014 (Early May Bank Holiday):
>>>
>>> terminate called after throwing an instance of 'QuantLib::Error'
>>> what(): 1st iteration: failed at 1st alive instrument, maturity
>>> May
>>> 14th, 2014, reference date May 7th, 2014: Missing no-fix1W
>> Actual/360
>>> fixing for May 2nd, 2014
>>> Aborted
>>>
>>> Everything seems to work fine on other days... Am I doing
>>> something
>>> wrong with setting up the PiecewiseYieldCurve?
>>>
>>> Thanks,
>>> George
>>>
>>> #include <ql/quantlib.hpp>
>>>
>>> using namespace std ;
>>> using namespace QuantLib ;
>>>
>>> int main(int argc, char** argv)
>>> {
>>> vector<Rate> liborRates ;
>>> liborRates.push_back(0.12075 / 100.0) ; //1W
>>> liborRates.push_back(0.1515 / 100.0) ; //1M
>>> liborRates.push_back(0.1935 / 100.0) ; //2M
>>> liborRates.push_back(0.22535 / 100.0) ; //3M
>>> liborRates.push_back(0.323 / 100.0) ; //6M
>>> liborRates.push_back(0.549 / 100.0) ; //12M
>>>
>>> vector<Period> periods ;
>>> periods.push_back(Period(1,Weeks)) ;
>>> periods.push_back(Period(1,Months)) ;
>>> periods.push_back(Period(2,Months)) ;
>>> periods.push_back(Period(3,Months)) ;
>>> periods.push_back(Period(6,Months)) ;
>>> periods.push_back(Period(12,Months)) ;
>>>
>>> vector<boost::shared_ptr<RateHelper> > instruments ;
>>>
>>> for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i++)
>>> {
>>> boost::shared_ptr<Quote> liborQuote(new SimpleQuote
>> (liborRates
>>> [i])) ;
>>> Handle<Quote> liborHandle(liborQuote) ;
>>> boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods
>>> [i])) ;
>>> boost::shared_ptr<RateHelper> rateHelper(new
>> DepositRateHelper
>>> (liborHandle,liborIndex)) ;
>>> instruments.push_back(rateHelper) ;
>>> }
>>>
>>> Calendar calendar = UnitedStates();
>>> Date today (5 , May ,2014);
>>> Settings::instance().evaluationDate() = today ;
>>> Natural settlementDays = 2;
>>> Date settlement = calendar.advance ( today , settlementDays ,
>>> Days );
>>> DayCounter dc = Actual360 ();
>>>
>>> Date maturity(20, Jun, 2014) ;
>>> double strikePrice = 100.0 ;
>>> Option::Type optionType = Option::Call ;
>>> double dividendYield = 0.0 ;
>>> double impliedVolatility = 0.15 ;
>>> double underlyingPrice = 100.0 ;
>>>
>>> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new
>>> SimpleQuote
>>> (underlyingPrice)));
>>>
>>> Handle<YieldTermStructure> rateTermStructure
>>> (boost::shared_ptr<YieldTermStructure>(new
>>> PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments,
>>> dc))) ;
>>>
>>> Handle<YieldTermStructure>
>>> dividendTermStructure(boost::shared_ptr<YieldTermStructure>(new
>>> FlatForward(today,
>>> dividendYield,
>>> dc)));
>>> Handle<BlackVolTermStructure>
>>> volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
>>> BlackConstantVol(today,
>>> calendar,
>>> impliedVolatility,
>>> dc)));
>>>
>>> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
>>> BlackScholesMertonProcess(underlyingHandle,
>>> dividendTermStructure,
>>> rateTermStructure,
>>> volTermStructure));
>>>
>>> boost::shared_ptr<Exercise> americanExercise(new AmericanExercise
>>> (today,
>>> maturity)) ;
>>> boost::shared_ptr<StrikedTypePayoff> payoff(new
>> PlainVanillaPayoff
>>> (optionType, strikePrice));
>>>
>>> VanillaOption americanOption(payoff, americanExercise);
>>>
>>> americanOption.setPricingEngine(boost::shared_ptr<PricingEngine>(
>>> new FDAmericanEngine<CrankNicolson>(bsmProcess,
>>> 101,100)));
>>>
>>> cout << americanOption.NPV() << endl ;
>>>
>>> return 0 ;
>>> }
>>>
>> ------------------------------------------------------------------------------
>>
>>>
>>>
>>> Want fast and easy access to all the code in your enterprise? Index
>>> and
>>> search up to 200,000 lines of code with a free copy of Black Duck
>>> Code Sight - the same software that powers the world's largest code
>>> search on Ohloh, the Black Duck Open Hub! Try it now.
>>>
>> http://p.sf.net/sfu/bds_______________________________________________
>>>
>>> QuantLib-users mailing list
>>> [hidden email]
>>> https://lists.sourceforge.net/lists/listinfo/quantlib-users
>>>
>>>
>> ------------------------------------------------------------------------------
>>
>>> Want fast and easy access to all the code in your enterprise? Index and
>>> search up to 200,000 lines of code with a free copy of Black Duck
>>> Code Sight - the same software that powers the world's largest code
>>> search on Ohloh, the Black Duck Open Hub! Try it now.
>>> http://p.sf.net/sfu/bds
>>> _______________________________________________
>>> QuantLib-users mailing list
>>> [hidden email]
>>> https://lists.sourceforge.net/lists/listinfo/quantlib-users
>>>
>>

Andres Hernandez

Re: UK Holiday for USDLibor/PiecewiseYieldCurve

Hi George,

as Peter mentioned, you can use another of the constructors. If you want to
use the USDLibor conventions you could use something like the following
inside the loop:

boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods[i])) ;

Calendar calendar = liborIndex->fixingCalendar();
calendar.removeHoliday(today);
boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper
( liborHandle,
liborIndex->tenor(),
liborIndex->fixingDays(),
calendar,
liborIndex->
businessDayConvention(),
liborIndex->endOfMonth(),
liborIndex->dayCounter
()) );
instruments.push_back(rateHelper);

@Peter, you are right,no modification would be needed and the behaviour for
the swaps is the correct one.

cheers,
Andres

From: George Cowie <[hidden email]>
To: Peter Caspers <[hidden email]>
Cc: Andres Hernandez/Germany/IBM@IBMDE, QuantLib users
<[hidden email]>
Date: 24/07/2014 13:52
Subject: Re: [Quantlib-users] UK Holiday for
USDLibor/PiecewiseYieldCurve

Interesting discussion... sorry to change the subject, but I'm still trying
to figure out how to adapt my code to this situation. How do I force my
YieldTermStructure to bootstrap with a different date? I have to value the
option on May 5th, but as you pointed out, changing the evaluation date to
a holiday has the described consequences.

On Thu, Jul 24, 2014 at 6:00 AM, Peter Caspers <[hidden email]>
wrote:
reminding myself of what workaround we actually implemented for swap
helper :o ... and what Murex offers here (always a good reference),
the full set of options would probably be

default - require historic fixings, take today's fixing if available
ignore today's tixing - require historic fixings, always ignore
today's fixing
ignore all fixings - ignore all fixings (i.e. estimate forward rates
always on the curve - this is technically ok even if the fixing date
is in the past, since only the index estimation period's
start and end date enter the estimation, and these two dates are
always greater or equal today)

these options to be applied only in the context of yield curve
bootstrapping, ideally to be set on the level of the curve itself.

On 24 July 2014 10:26, Peter Caspers <[hidden email]> wrote:
> Hi Andres,
>
> yes, the index in the deposit rate helper seems to be merely a
> convenient way for coding - with the one exception that deposits do
> not (directly) have anything to do with actual index fixings, and
this
> is also where the code breaks. And yes, it gets even worse in your
> example when you use the USD Libor index for USD cash deposits on a
UK
> but non-NY holiday (you could use the index-free constructor though
> and pass the NY calendar directly instead, right, solving half of
the
> problem ?). So just treating the deposit helper's index and its
fixing
> as a computational tool to get the implied Quote seems a reasonable
> solution, in your USD / Libor example yielding the correct
conventions
> even, if I am not mistaken ? I wouldn't introduce any flag or
method
> to activate this behaviour, it should be the only one present
(because
> checking for holidays can easily be done in the user's code if
> necessary at all).
>
> For swaps on the other hand I think the current behaviour is fine -
> what fixing would you take for the first period in a Euribor swap
on a
> TARGET holiday ? Probably the same as you take after the fixing
around
> 11am on a trading day, i.e. the last known fixing. That is exactly
the
> logic applied in the library. I wouldn't change anything here. It
> might be that you get an exception because you just don't provide
the
> historic fixing, but in this case it is a legitimate demand of the
> library and working around this should only be done in the user
code
> imo.
>
> best
> Peter
>
>
>
>
>
> On 24 July 2014 09:23, Andres Hernandez <
[hidden email]> wrote:
>>
>> Hi Peter,
>>
>> I agree that it would be nice to at least have it as an option. We
>> currently do a small workaround, whereby we set in our system a
parameter
>> for the curve specifying the holiday calendar for that curve. The
only use
>> of that calendar is to determine if it is a holiday for the curve
or not
>> and if yes to make a fallback to yesterday's values. In the case
of the EUR
>> curves, this is arguably a good behaviour as the index and the
geographical
>> region have the same calendar, but for the USD case, we are
basically
>> forcing a fallback for a holiday on a different country. The issue
is
>> compounded by the fact that in this case the deposit is traded on
that day,
>> it is just the index that gets no fixing.
>>
>> I took a look and I am almost sure that removing the evaluation
date from
>> the fixing calendar in InterestRate constructor would be enough to
make
>> deposits go through. This could simply be accomplished with an
extra public
>> member function that one would call if one wanted it to go through
on
>> holidays. A more involved and important problem would be to get
the
>> SwapRateHelper to go though on holidays (if desired), as I am
positive that
>> that was the reason for our instituting the fallback. I will see
if I have
>> the time to change it, and would then probably subscribe to the
developers
>> list to get some feedback in how to go about changing it.
>>
>>
>> Mit freundlichen Grüßen / Kind regards
>>
>> Dr. Andres Hernandez
>>
>> Senior Financial Engineer
>> Business Analytics
>> Risk Analytics
>>
>>
>>
>>
>>
>> Phone: +49-69-6645-1351 IBM Deutschland
(Embedded
>>
image moved
>>
to file:
>>
pic02235.gif)
>>
>> Mobile: +49-163-609-3280 Wilhelm-Fay-Str.
30-34
>>
>> E-Mail: [hidden email] 65936 Frankfurt
am Main
>>
>> Germany
>>
>>
>>
>>
>>
>> IBM Deutschland
>> GmbH /
>> Vorsitzender des
>> Aufsichtsrats:
>> Martin Jetter
>> Geschäftsführung:
>> Martina Koederitz
>> (Vorsitzende),
>> Reinhard Reschke,
>> Dieter Scholz,
>> Gregor Pillen,
>> Christian Noll,
>> Ivo Koerner
>> Sitz der
>> Gesellschaft:
>> Ehningen /
>> Registergericht:
>> Amtsgericht
>> Stuttgart, HRB
>> 14562 /
>> WEEE-Reg.-Nr. DE
>> 99369940
>>
>>
>>
>>
>>
>>
>> From: Peter Caspers <[hidden email]>
>> To: Andres Hernandez/Germany/IBM@IBMDE
>> Cc: George Cowie <[hidden email]>, QuantLib users
>> <[hidden email]>
>> Date: 23/07/2014 21:56
>> Subject: Re: [Quantlib-users] UK Holiday for
>> USDLibor/PiecewiseYieldCurve
>>
>>
>>
>> Hi,
>>
>> I think this question is an older one but I don't remember that it
was
>> finally decided how to handle this situation (?) - which is,
should we
>> allow to bootstrap a curve on a day which is a holiday for the
cash
>> instruments in the curve or not. If negative, we could maybe throw
a
>> different message ( the current one indeed sounds a bit
paradoxical ).
>> If yes ( which would be my personal preference, because you simply
>> want to run your daily processes spitting out numbers even if not
a
>> hundred percent sensible all of them ), one has to define how a
>> deposit which is not traded on that day should be calculated. One
>> obvious way would be to just treat the evaluation date as a
business
>> day when calculating the helper's fixing date. I did not try, but
it
>> may be enough to join the business days of the helper's calendar
with
>> a new one just having one business day (the evaluation date)
before
>> passing it to the no-fix index to make this kind of solution work.
>> What do you think ?
>>
>> Peter
>>
>>
>> On 23 July 2014 15:05, Andres Hernandez <
[hidden email]>
>> wrote:
>>>
>>> Hi George,
>>>
>>> you are still setting the evaluation date to a holiday:
>>>
>>> Date today (5 , May ,2014);
>>> Date liborDate (2, May, 2014) ;
>>> Settings::instance().evaluationDate() = today ;
>>>
>>>
>>> Let's follow the logic:
>>> During bootstrapping, the method impliedQuote() of the rate
helper is
>>> called to compare against the market quote. The DepositRateHelper
simply
>>> returns the fixing corresponding to the fixingDate_ as a return
value.
>>> The fixingDate_ is set during initialization as the return value
of the
>>> function InterestRate::fixingDate(earliestDate_). The function
>>> InterestRate::fixingDate sets back the date given as input
parameter by
>> the
>>> number of fixing days according to the index' calendar.
>>> The parameter earliestDate_ being passed to fixingDate is set
during
>>> initialization to the evaluation date (which you set to the 5th
of May)
>>> moved forward by the number of fixing days according to the
index'
>>> calendar.
>>> Note that if the evaluation date is a holiday you do not end up
with the
>>> same date, but rather with the last business day prior to
evaluation
>> date.
>>> The first act moves the date from the 5th of May to the 7th and
checks
>>> whether it is a business day, which it is. The second move sets
it back
>> two
>>> days to the 5th, and checks whether it is a business day, which
it isn't,
>>> so it sets it back until it finds a business day, i.e. the 2nd of
May.
>>> Since the value is in the past, it requests a historical fixing,
which
>> you
>>> did not provide and so the QL_REQUIRE throws.
>>>
>>> Mit freundlichen Grüßen / Kind regards
>>>
>>> Dr. Andres Hernandez
>>>
>>> Senior Financial Engineer
>>> Business Analytics
>>> Risk Analytics
>>>
>>>
>>>
>>>
>>>
>>> Phone: +49-69-6645-1351 IBM Deutschland
>> (Embedded
>>>
>> image moved
>>>
>> to file:
>>>
>> pic64586.gif)
>>>
>>> Mobile: +49-163-609-3280
Wilhelm-Fay-Str. 30-34
>>>
>>> E-Mail: [hidden email] 65936 Frankfurt
am Main
>>>
>>> Germany
>>>
>>>
>>>
>>>
>>>
>>> IBM Deutschland
>>> GmbH /
>>> Vorsitzender des
>>> Aufsichtsrats:
>>> Martin Jetter
>>> Geschäftsführung:
>>> Martina Koederitz
>>> (Vorsitzende),
>>> Reinhard Reschke,
>>> Dieter Scholz,
>>> Gregor Pillen,
>>> Christian Noll,
>>> Ivo Koerner
>>> Sitz der
>>> Gesellschaft:
>>> Ehningen /
>>> Registergericht:
>>> Amtsgericht
>>> Stuttgart, HRB
>>> 14562 /
>>> WEEE-Reg.-Nr. DE
>>> 99369940
>>>
>>>
>>>
>>>
>>>
>>>
>>> From: George Cowie <[hidden email]>
>>> To: Andres Hernandez/Germany/IBM@IBMDE
>>> Cc: QuantLib users <[hidden email]>
>>> Date: 23/07/2014 11:51
>>> Subject: Re: [Quantlib-users] UK Holiday for
>>> USDLibor/PiecewiseYieldCurve
>>>
>>>
>>>
>>>
>>> Thank you Andres. What you're describing makes sense, but I
can't seem
>> to
>>> avoid an error when trying to value the option on May 5, 2014.
I've
>> copied
>>> the updated code snippet with changed section highlighted/bolded
below.
>> I
>>> seem to always get an exception of the form:
>>>
>>> terminate called after throwing an instance of 'QuantLib::Error'
>>> what(): 1st iteration: failed at 1st alive instrument,
maturity May
>>> 14th, 2014, reference date May 6th, 2014: Missing no-fix1W
Actual/360
>>> fixing for May 2nd, 2014
>>>
>>> Is there something wrong with my approach here?
>>>
>>> Thanks,
>>> George
>>>
>>>
>>> #include <ql/quantlib.hpp>
>>>
>>> using namespace std ;
>>> using namespace QuantLib ;
>>>
>>> int main(int argc, char** argv)
>>> {
>>> vector<Rate> liborRates ;
>>> liborRates.push_back(0.12075 / 100.0) ; //1W
>>> liborRates.push_back(0.1515 / 100.0) ; //1M
>>> liborRates.push_back(0.1935 / 100.0) ; //2M
>>> liborRates.push_back(0.22535 / 100.0) ; //3M
>>> liborRates.push_back(0.323 / 100.0) ; //6M
>>> liborRates.push_back(0.549 / 100.0) ; //12M
>>>
>>> vector<Period> periods ;
>>> periods.push_back(Period(1,Weeks)) ;
>>> periods.push_back(Period(1,Months)) ;
>>> periods.push_back(Period(2,Months)) ;
>>> periods.push_back(Period(3,Months)) ;
>>> periods.push_back(Period(6,Months)) ;
>>> periods.push_back(Period(12,Months)) ;
>>>
>>> vector<boost::shared_ptr<RateHelper> > instruments ;
>>>
>>> for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i+
+)
>>> {
>>> boost::shared_ptr<Quote> liborQuote(new SimpleQuote
(liborRates
>> [i])) ;
>>> Handle<Quote> liborHandle(liborQuote) ;
>>> boost::shared_ptr<IborIndex> liborIndex(new USDLibor
(periods[i])) ;
>>> boost::shared_ptr<RateHelper> rateHelper(new
DepositRateHelper
>>> (liborHandle,liborIndex)) ;
>>> instruments.push_back(rateHelper) ;
>>> }
>>>
>>> Calendar calendar = UnitedStates();
>>> Date today (5 , May ,2014);
>>> Date liborDate (2, May, 2014) ;
>>> Settings::instance().evaluationDate() = today ;
>>> Natural settlementDays = 2;
>>> Date settlement = calendar.advance ( liborDate ,
settlementDays ,
>> Days );
>>> DayCounter dc = Actual360 ();
>>>
>>> Date maturity(20, Jun, 2014) ;
>>> double strikePrice = 100.0 ;
>>> Option::Type optionType = Option::Call ;
>>> double dividendYield = 0.0 ;
>>> double impliedVolatility = 0.15 ;
>>> double underlyingPrice = 100.0 ;
>>>
>>> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new
SimpleQuote
>>> (underlyingPrice)));
>>>
>>> Handle<YieldTermStructure> rateTermStructure
>>> (boost::shared_ptr<YieldTermStructure>(new
>>> PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments,
dc))) ;
>>>
>>> Handle<YieldTermStructure>
>>> dividendTermStructure(boost::shared_ptr<YieldTermStructure>
(new
>>> FlatForward(today,
>>> dividendYield,
>>> dc)));
>>> Handle<BlackVolTermStructure>
>>> volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
>>> BlackConstantVol(today,
>>> calendar,
>>> impliedVolatility,
>>> dc)));
>>>
>>> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
>>> BlackScholesMertonProcess(underlyingHandle,
>>> dividendTermStructure,
>>> rateTermStructure,
>>> volTermStructure));
>>>
>>> boost::shared_ptr<Exercise> americanExercise(new
AmericanExercise
>> (today,
>>> maturity)) ;
>>> boost::shared_ptr<StrikedTypePayoff> payoff(new
PlainVanillaPayoff
>>> (optionType, strikePrice));
>>>
>>> VanillaOption americanOption(payoff, americanExercise);
>>>
>>> americanOption.setPricingEngine
(boost::shared_ptr<PricingEngine>(
>>> new FDAmericanEngine<CrankNicolson>(bsmProcess,
>>> 101,100)));
>>>
>>> cout << americanOption.NPV() << endl ;
>>>
>>> return 0 ;
>>> }
>>>
>>>
>>>
>>> On Tue, Jul 22, 2014 at 7:39 AM, Andres Hernandez <
>>> [hidden email]> wrote:
>>>
>>> Hi George,
>>>
>>> the problem occurs because you are trying to bootstrap on a
>>> UnitedKingdom
>>> holiday, which is implicitly included in the definition of
>> USDLibor.
>>> The
>>> Libor class makes a joint calendar between the given
financial
>> center
>>> calendar and the UK calendar. Since Libor is a London poll,
there
>> are
>>> only
>>> fixings for it on UK business days. If one wants to
bootstrap the
>>> curve on
>>> a holiday, the indice's historical fixings (actually just
the
>> closest
>>> fixing to the given date) need to be provided.
>>>
>>>
>>> Mit freundlichen Grüßen / Kind regards
>>>
>>> Dr. Andres Hernandez
>>>
>>> Senior Financial Engineer
>>> Business Analytics
>>> Risk Analytics
>>>
>>>
>>>
>>>
>>>
>>> Phone: +49-69-6645-1351 IBM
Deutschland
>>> (Embedded
>>> image moved
>>> to file:
>>> pic20635.gif)
>>>
>>> Mobile: +49-163-609-3280
Wilhelm-Fay-Str.
>>> 30-34
>>>
>>> E-Mail: [hidden email] 65936
Frankfurt
>> am
>>> Main
>>>
>>> Germany
>>>
>>>
>>>
>>>
>>>
>>> IBM Deutschland
>>> GmbH /
>>> Vorsitzender des
>>> Aufsichtsrats:
>>> Martin Jetter
>>> Geschäftsführung:
>>> Martina Koederitz
>>> (Vorsitzende),
>>> Reinhard Reschke,
>>> Dieter Scholz,
>>> Gregor Pillen,
>>> Christian Noll,
>>> Ivo Koerner
>>> Sitz der
>>> Gesellschaft:
>>> Ehningen /
>>> Registergericht:
>>> Amtsgericht
>>> Stuttgart, HRB
>>> 14562 /
>>> WEEE-Reg.-Nr. DE
>>> 99369940
>>>
>>>
>>>
>>>
>>>
>>>
>>> From: George Cowie <[hidden email]>
>>> To: QuantLib users <
[hidden email]>
>>> Date: 22/07/2014 14:10
>>> Subject: [Quantlib-users] UK Holiday for
>>> USDLibor/PiecewiseYieldCurve
>>>
>>>
>>>
>>> Hello,
>>>
>>> I'm using PiecewiseYieldCurve to build a
YieldTermStructure which
>>> is
>>> ultimately used by BlackScholesMertonProcess in my option
pricing.
>> I
>>> believe I'm running into an issue on days that are holidays
in the
>> UK
>>> (but
>>> not the US). Here's the error I'm seeing when trying to
run the
>> code
>>> below
>>> for May 5, 2014 (Early May Bank Holiday):
>>>
>>> terminate called after throwing an instance of
'QuantLib::Error'
>>> what(): 1st iteration: failed at 1st alive instrument,
maturity
>>> May
>>> 14th, 2014, reference date May 7th, 2014: Missing no-fix1W
>> Actual/360
>>> fixing for May 2nd, 2014
>>> Aborted
>>>
>>> Everything seems to work fine on other days... Am I
doing
>>> something
>>> wrong with setting up the PiecewiseYieldCurve?
>>>
>>> Thanks,
>>> George
>>>
>>> #include <ql/quantlib.hpp>
>>>
>>> using namespace std ;
>>> using namespace QuantLib ;
>>>
>>> int main(int argc, char** argv)
>>> {
>>> vector<Rate> liborRates ;
>>> liborRates.push_back(0.12075 / 100.0) ; //1W
>>> liborRates.push_back(0.1515 / 100.0) ; //1M
>>> liborRates.push_back(0.1935 / 100.0) ; //2M
>>> liborRates.push_back(0.22535 / 100.0) ; //3M
>>> liborRates.push_back(0.323 / 100.0) ; //6M
>>> liborRates.push_back(0.549 / 100.0) ; //12M
>>>
>>> vector<Period> periods ;
>>> periods.push_back(Period(1,Weeks)) ;
>>> periods.push_back(Period(1,Months)) ;
>>> periods.push_back(Period(2,Months)) ;
>>> periods.push_back(Period(3,Months)) ;
>>> periods.push_back(Period(6,Months)) ;
>>> periods.push_back(Period(12,Months)) ;
>>>
>>> vector<boost::shared_ptr<RateHelper> > instruments ;
>>>
>>> for (vector<int>::size_type i = 0 ; i < liborRates.size
() ; i++)
>>> {
>>> boost::shared_ptr<Quote> liborQuote(new SimpleQuote
>> (liborRates
>>> [i])) ;
>>> Handle<Quote> liborHandle(liborQuote) ;
>>> boost::shared_ptr<IborIndex> liborIndex(new USDLibor
(periods
>>> [i])) ;
>>> boost::shared_ptr<RateHelper> rateHelper(new
>> DepositRateHelper
>>> (liborHandle,liborIndex)) ;
>>> instruments.push_back(rateHelper) ;
>>> }
>>>
>>> Calendar calendar = UnitedStates();
>>> Date today (5 , May ,2014);
>>> Settings::instance().evaluationDate() = today ;
>>> Natural settlementDays = 2;
>>> Date settlement = calendar.advance ( today ,
settlementDays ,
>>> Days );
>>> DayCounter dc = Actual360 ();
>>>
>>> Date maturity(20, Jun, 2014) ;
>>> double strikePrice = 100.0 ;
>>> Option::Type optionType = Option::Call ;
>>> double dividendYield = 0.0 ;
>>> double impliedVolatility = 0.15 ;
>>> double underlyingPrice = 100.0 ;
>>>
>>> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>
(new
>>> SimpleQuote
>>> (underlyingPrice)));
>>>
>>> Handle<YieldTermStructure> rateTermStructure
>>> (boost::shared_ptr<YieldTermStructure>(new
>>> PiecewiseYieldCurve<ZeroYield,Linear>(settlement,
instruments,
>>> dc))) ;
>>>
>>> Handle<YieldTermStructure>
>>> dividendTermStructure
(boost::shared_ptr<YieldTermStructure>(new
>>> FlatForward(today,
>>> dividendYield,
>>> dc)));
>>> Handle<BlackVolTermStructure>
>>> volTermStructure
(boost::shared_ptr<BlackVolTermStructure>(new
>>> BlackConstantVol(today,
>>> calendar,
>>> impliedVolatility,
>>> dc)));
>>>
>>> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess
(new
>>> BlackScholesMertonProcess(underlyingHandle,
>>> dividendTermStructure,
>>> rateTermStructure,
>>> volTermStructure));
>>>
>>> boost::shared_ptr<Exercise> americanExercise(new
AmericanExercise
>>> (today,
>>> maturity)) ;
>>> boost::shared_ptr<StrikedTypePayoff> payoff(new
>> PlainVanillaPayoff
>>> (optionType, strikePrice));
>>>
>>> VanillaOption americanOption(payoff, americanExercise);
>>>
>>> americanOption.setPricingEngine
(boost::shared_ptr<PricingEngine>(
>>> new FDAmericanEngine<CrankNicolson>(bsmProcess,
>>> 101,100)));
>>>
>>> cout << americanOption.NPV() << endl ;
>>>
>>> return 0 ;
>>> }
>>>
>>
------------------------------------------------------------------------------

>>
>>>
>>>
>>> Want fast and easy access to all the code in your
enterprise? Index
>>> and
>>> search up to 200,000 lines of code with a free copy of
Black Duck
>>> Code Sight - the same software that powers the world's
largest code
>>> search on Ohloh, the Black Duck Open Hub! Try it now.
>>>
>>
http://p.sf.net/sfu/bds_______________________________________________

>>>
>>> QuantLib-users mailing list
>>> [hidden email]
>>> https://lists.sourceforge.net/lists/listinfo/quantlib-users
>>>
>>>
>>
------------------------------------------------------------------------------

>>
>>> Want fast and easy access to all the code in your enterprise?
Index and
>>> search up to 200,000 lines of code with a free copy of Black Duck
>>> Code Sight - the same software that powers the world's largest
code
>>> search on Ohloh, the Black Duck Open Hub! Try it now.
>>> http://p.sf.net/sfu/bds
>>> _______________________________________________
>>> QuantLib-users mailing list
>>> [hidden email]
>>> https://lists.sourceforge.net/lists/listinfo/quantlib-users
>>>
>>

------------------------------------------------------------------------------
Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds
_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

Peter Caspers-4

Re: UK Holiday for USDLibor/PiecewiseYieldCurve

thanks a lot, Andres and George. I created a pull request to fix the
deposit helpers in this sense:

https://github.com/lballabio/quantlib/pull/117

There is only one more little complication I think, namely if we have
zero fixing days and are on a holiday in which case we should adjust
the start date of the deposit just as it would have been done under
the original calendar.

The code looks not as neat as in Andres' example above, mainly because
I don't know how the hell I can clone a calendar and remove a holiday
without impacting all other instances of the same type ...

We should think about adding the ignore - fixing options for swaps later on !

kind regards
Peter

On 24 July 2014 16:06, Andres Hernandez <[hidden email]> wrote:

> Hi George,
>
> as Peter mentioned, you can use another of the constructors. If you want to
> use the USDLibor conventions you could use something like the following
> inside the loop:
>
> boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods[i])) ;
>
> Calendar calendar = liborIndex->fixingCalendar();
> calendar.removeHoliday(today);
> boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper
> ( liborHandle,
> liborIndex->tenor(),
> liborIndex->fixingDays(),
> calendar,
> liborIndex->
> businessDayConvention(),
> liborIndex->endOfMonth(),
> liborIndex->dayCounter
> ()) );
> instruments.push_back(rateHelper);
>
>
> @Peter, you are right,no modification would be needed and the behaviour for
> the swaps is the correct one.
>
> cheers,
> Andres
>
>
>
> From: George Cowie <[hidden email]>
> To: Peter Caspers <[hidden email]>
> Cc: Andres Hernandez/Germany/IBM@IBMDE, QuantLib users
> <[hidden email]>
> Date: 24/07/2014 13:52
> Subject: Re: [Quantlib-users] UK Holiday for
> USDLibor/PiecewiseYieldCurve
>
>
>
> Interesting discussion... sorry to change the subject, but I'm still trying
> to figure out how to adapt my code to this situation. How do I force my
> YieldTermStructure to bootstrap with a different date? I have to value the
> option on May 5th, but as you pointed out, changing the evaluation date to
> a holiday has the described consequences.
>
>
> On Thu, Jul 24, 2014 at 6:00 AM, Peter Caspers <[hidden email]>
> wrote:
> reminding myself of what workaround we actually implemented for swap
> helper :o ... and what Murex offers here (always a good reference),
> the full set of options would probably be
>
> default - require historic fixings, take today's fixing if available
> ignore today's tixing - require historic fixings, always ignore
> today's fixing
> ignore all fixings - ignore all fixings (i.e. estimate forward rates
> always on the curve - this is technically ok even if the fixing date
> is in the past, since only the index estimation period's
> start and end date enter the estimation, and these two dates are
> always greater or equal today)
>
> these options to be applied only in the context of yield curve
> bootstrapping, ideally to be set on the level of the curve itself.
>
>
> On 24 July 2014 10:26, Peter Caspers <[hidden email]> wrote:
> > Hi Andres,
> >
> > yes, the index in the deposit rate helper seems to be merely a
> > convenient way for coding - with the one exception that deposits do
> > not (directly) have anything to do with actual index fixings, and
> this
> > is also where the code breaks. And yes, it gets even worse in your
> > example when you use the USD Libor index for USD cash deposits on a
> UK
> > but non-NY holiday (you could use the index-free constructor though
> > and pass the NY calendar directly instead, right, solving half of
> the
> > problem ?). So just treating the deposit helper's index and its
> fixing
> > as a computational tool to get the implied Quote seems a reasonable
> > solution, in your USD / Libor example yielding the correct
> conventions
> > even, if I am not mistaken ? I wouldn't introduce any flag or
> method
> > to activate this behaviour, it should be the only one present
> (because
> > checking for holidays can easily be done in the user's code if
> > necessary at all).
> >
> > For swaps on the other hand I think the current behaviour is fine -
> > what fixing would you take for the first period in a Euribor swap
> on a
> > TARGET holiday ? Probably the same as you take after the fixing
> around
> > 11am on a trading day, i.e. the last known fixing. That is exactly
> the
> > logic applied in the library. I wouldn't change anything here. It
> > might be that you get an exception because you just don't provide
> the
> > historic fixing, but in this case it is a legitimate demand of the
> > library and working around this should only be done in the user
> code
> > imo.
> >
> > best
> > Peter
> >
> >
> >
> >
> >
> > On 24 July 2014 09:23, Andres Hernandez <
> [hidden email]> wrote:
> >>
> >> Hi Peter,
> >>
> >> I agree that it would be nice to at least have it as an option. We
> >> currently do a small workaround, whereby we set in our system a
> parameter
> >> for the curve specifying the holiday calendar for that curve. The
> only use
> >> of that calendar is to determine if it is a holiday for the curve
> or not
> >> and if yes to make a fallback to yesterday's values. In the case
> of the EUR
> >> curves, this is arguably a good behaviour as the index and the
> geographical
> >> region have the same calendar, but for the USD case, we are
> basically
> >> forcing a fallback for a holiday on a different country. The issue
> is
> >> compounded by the fact that in this case the deposit is traded on
> that day,
> >> it is just the index that gets no fixing.
> >>
> >> I took a look and I am almost sure that removing the evaluation
> date from
> >> the fixing calendar in InterestRate constructor would be enough to
> make
> >> deposits go through. This could simply be accomplished with an
> extra public
> >> member function that one would call if one wanted it to go through
> on
> >> holidays. A more involved and important problem would be to get
> the
> >> SwapRateHelper to go though on holidays (if desired), as I am
> positive that
> >> that was the reason for our instituting the fallback. I will see
> if I have
> >> the time to change it, and would then probably subscribe to the
> developers
> >> list to get some feedback in how to go about changing it.
> >>
> >>
> >> Mit freundlichen Grüßen / Kind regards
> >>
> >> Dr. Andres Hernandez
> >>
> >> Senior Financial Engineer
> >> Business Analytics
> >> Risk Analytics
> >>
> >>
> >>
> >>
> >>
> >> Phone: +49-69-6645-1351 IBM Deutschland
> (Embedded
> >>
> image moved
> >>
> to file:
> >>
> pic02235.gif)
> >>
> >> Mobile: +49-163-609-3280 Wilhelm-Fay-Str.
> 30-34
> >>
> >> E-Mail: [hidden email] 65936 Frankfurt
> am Main
> >>
> >> Germany
> >>
> >>
> >>
> >>
> >>
> >> IBM Deutschland
> >> GmbH /
> >> Vorsitzender des
> >> Aufsichtsrats:
> >> Martin Jetter
> >> Geschäftsführung:
> >> Martina Koederitz
> >> (Vorsitzende),
> >> Reinhard Reschke,
> >> Dieter Scholz,
> >> Gregor Pillen,
> >> Christian Noll,
> >> Ivo Koerner
> >> Sitz der
> >> Gesellschaft:
> >> Ehningen /
> >> Registergericht:
> >> Amtsgericht
> >> Stuttgart, HRB
> >> 14562 /
> >> WEEE-Reg.-Nr. DE
> >> 99369940
> >>
> >>
> >>
> >>
> >>
> >>
> >> From: Peter Caspers <[hidden email]>
> >> To: Andres Hernandez/Germany/IBM@IBMDE
> >> Cc: George Cowie <[hidden email]>, QuantLib users
> >> <[hidden email]>
> >> Date: 23/07/2014 21:56
> >> Subject: Re: [Quantlib-users] UK Holiday for
> >> USDLibor/PiecewiseYieldCurve
> >>
> >>
> >>
> >> Hi,
> >>
> >> I think this question is an older one but I don't remember that it
> was
> >> finally decided how to handle this situation (?) - which is,
> should we
> >> allow to bootstrap a curve on a day which is a holiday for the
> cash
> >> instruments in the curve or not. If negative, we could maybe throw
> a
> >> different message ( the current one indeed sounds a bit
> paradoxical ).
> >> If yes ( which would be my personal preference, because you simply
> >> want to run your daily processes spitting out numbers even if not
> a
> >> hundred percent sensible all of them ), one has to define how a
> >> deposit which is not traded on that day should be calculated. One
> >> obvious way would be to just treat the evaluation date as a
> business
> >> day when calculating the helper's fixing date. I did not try, but
> it
> >> may be enough to join the business days of the helper's calendar
> with
> >> a new one just having one business day (the evaluation date)
> before
> >> passing it to the no-fix index to make this kind of solution work.
> >> What do you think ?
> >>
> >> Peter
> >>
> >>
> >> On 23 July 2014 15:05, Andres Hernandez <
> [hidden email]>
> >> wrote:
> >>>
> >>> Hi George,
> >>>
> >>> you are still setting the evaluation date to a holiday:
> >>>
> >>> Date today (5 , May ,2014);
> >>> Date liborDate (2, May, 2014) ;
> >>> Settings::instance().evaluationDate() = today ;
> >>>
> >>>
> >>> Let's follow the logic:
> >>> During bootstrapping, the method impliedQuote() of the rate
> helper is
> >>> called to compare against the market quote. The DepositRateHelper
> simply
> >>> returns the fixing corresponding to the fixingDate_ as a return
> value.
> >>> The fixingDate_ is set during initialization as the return value
> of the
> >>> function InterestRate::fixingDate(earliestDate_). The function
> >>> InterestRate::fixingDate sets back the date given as input
> parameter by
> >> the
> >>> number of fixing days according to the index' calendar.
> >>> The parameter earliestDate_ being passed to fixingDate is set
> during
> >>> initialization to the evaluation date (which you set to the 5th
> of May)
> >>> moved forward by the number of fixing days according to the
> index'
> >>> calendar.
> >>> Note that if the evaluation date is a holiday you do not end up
> with the
> >>> same date, but rather with the last business day prior to
> evaluation
> >> date.
> >>> The first act moves the date from the 5th of May to the 7th and
> checks
> >>> whether it is a business day, which it is. The second move sets
> it back
> >> two
> >>> days to the 5th, and checks whether it is a business day, which
> it isn't,
> >>> so it sets it back until it finds a business day, i.e. the 2nd of
> May.
> >>> Since the value is in the past, it requests a historical fixing,
> which
> >> you
> >>> did not provide and so the QL_REQUIRE throws.
> >>>
> >>> Mit freundlichen Grüßen / Kind regards
> >>>
> >>> Dr. Andres Hernandez
> >>>
> >>> Senior Financial Engineer
> >>> Business Analytics
> >>> Risk Analytics
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> Phone: +49-69-6645-1351 IBM Deutschland
> >> (Embedded
> >>>
> >> image moved
> >>>
> >> to file:
> >>>
> >> pic64586.gif)
> >>>
> >>> Mobile: +49-163-609-3280
> Wilhelm-Fay-Str. 30-34
> >>>
> >>> E-Mail: [hidden email] 65936 Frankfurt
> am Main
> >>>
> >>> Germany
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> IBM Deutschland
> >>> GmbH /
> >>> Vorsitzender des
> >>> Aufsichtsrats:
> >>> Martin Jetter
> >>> Geschäftsführung:
> >>> Martina Koederitz
> >>> (Vorsitzende),
> >>> Reinhard Reschke,
> >>> Dieter Scholz,
> >>> Gregor Pillen,
> >>> Christian Noll,
> >>> Ivo Koerner
> >>> Sitz der
> >>> Gesellschaft:
> >>> Ehningen /
> >>> Registergericht:
> >>> Amtsgericht
> >>> Stuttgart, HRB
> >>> 14562 /
> >>> WEEE-Reg.-Nr. DE
> >>> 99369940
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> From: George Cowie <[hidden email]>
> >>> To: Andres Hernandez/Germany/IBM@IBMDE
> >>> Cc: QuantLib users <[hidden email]>
> >>> Date: 23/07/2014 11:51
> >>> Subject: Re: [Quantlib-users] UK Holiday for
> >>> USDLibor/PiecewiseYieldCurve
> >>>
> >>>
> >>>
> >>>
> >>> Thank you Andres. What you're describing makes sense, but I
> can't seem
> >> to
> >>> avoid an error when trying to value the option on May 5, 2014.
> I've
> >> copied
> >>> the updated code snippet with changed section highlighted/bolded
> below.
> >> I
> >>> seem to always get an exception of the form:
> >>>
> >>> terminate called after throwing an instance of 'QuantLib::Error'
> >>> what(): 1st iteration: failed at 1st alive instrument,
> maturity May
> >>> 14th, 2014, reference date May 6th, 2014: Missing no-fix1W
> Actual/360
> >>> fixing for May 2nd, 2014
> >>>
> >>> Is there something wrong with my approach here?
> >>>
> >>> Thanks,
> >>> George
> >>>
> >>>
> >>> #include <ql/quantlib.hpp>
> >>>
> >>> using namespace std ;
> >>> using namespace QuantLib ;
> >>>
> >>> int main(int argc, char** argv)
> >>> {
> >>> vector<Rate> liborRates ;
> >>> liborRates.push_back(0.12075 / 100.0) ; //1W
> >>> liborRates.push_back(0.1515 / 100.0) ; //1M
> >>> liborRates.push_back(0.1935 / 100.0) ; //2M
> >>> liborRates.push_back(0.22535 / 100.0) ; //3M
> >>> liborRates.push_back(0.323 / 100.0) ; //6M
> >>> liborRates.push_back(0.549 / 100.0) ; //12M
> >>>
> >>> vector<Period> periods ;
> >>> periods.push_back(Period(1,Weeks)) ;
> >>> periods.push_back(Period(1,Months)) ;
> >>> periods.push_back(Period(2,Months)) ;
> >>> periods.push_back(Period(3,Months)) ;
> >>> periods.push_back(Period(6,Months)) ;
> >>> periods.push_back(Period(12,Months)) ;
> >>>
> >>> vector<boost::shared_ptr<RateHelper> > instruments ;
> >>>
> >>> for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i+
> +)
> >>> {
> >>> boost::shared_ptr<Quote> liborQuote(new SimpleQuote
> (liborRates
> >> [i])) ;
> >>> Handle<Quote> liborHandle(liborQuote) ;
> >>> boost::shared_ptr<IborIndex> liborIndex(new USDLibor
> (periods[i])) ;
> >>> boost::shared_ptr<RateHelper> rateHelper(new
> DepositRateHelper
> >>> (liborHandle,liborIndex)) ;
> >>> instruments.push_back(rateHelper) ;
> >>> }
> >>>
> >>> Calendar calendar = UnitedStates();
> >>> Date today (5 , May ,2014);
> >>> Date liborDate (2, May, 2014) ;
> >>> Settings::instance().evaluationDate() = today ;
> >>> Natural settlementDays = 2;
> >>> Date settlement = calendar.advance ( liborDate ,
> settlementDays ,
> >> Days );
> >>> DayCounter dc = Actual360 ();
> >>>
> >>> Date maturity(20, Jun, 2014) ;
> >>> double strikePrice = 100.0 ;
> >>> Option::Type optionType = Option::Call ;
> >>> double dividendYield = 0.0 ;
> >>> double impliedVolatility = 0.15 ;
> >>> double underlyingPrice = 100.0 ;
> >>>
> >>> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new
> SimpleQuote
> >>> (underlyingPrice)));
> >>>
> >>> Handle<YieldTermStructure> rateTermStructure
> >>> (boost::shared_ptr<YieldTermStructure>(new
> >>> PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments,
> dc))) ;
> >>>
> >>> Handle<YieldTermStructure>
> >>> dividendTermStructure(boost::shared_ptr<YieldTermStructure>
> (new
> >>> FlatForward(today,
> >>> dividendYield,
> >>> dc)));
> >>> Handle<BlackVolTermStructure>
> >>> volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
> >>> BlackConstantVol(today,
> >>> calendar,
> >>> impliedVolatility,
> >>> dc)));
> >>>
> >>> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
> >>> BlackScholesMertonProcess(underlyingHandle,
> >>> dividendTermStructure,
> >>> rateTermStructure,
> >>> volTermStructure));
> >>>
> >>> boost::shared_ptr<Exercise> americanExercise(new
> AmericanExercise
> >> (today,
> >>> maturity)) ;
> >>> boost::shared_ptr<StrikedTypePayoff> payoff(new
> PlainVanillaPayoff
> >>> (optionType, strikePrice));
> >>>
> >>> VanillaOption americanOption(payoff, americanExercise);
> >>>
> >>> americanOption.setPricingEngine
> (boost::shared_ptr<PricingEngine>(
> >>> new FDAmericanEngine<CrankNicolson>(bsmProcess,
> >>> 101,100)));
> >>>
> >>> cout << americanOption.NPV() << endl ;
> >>>
> >>> return 0 ;
> >>> }
> >>>
> >>>
> >>>
> >>> On Tue, Jul 22, 2014 at 7:39 AM, Andres Hernandez <
> >>> [hidden email]> wrote:
> >>>
> >>> Hi George,
> >>>
> >>> the problem occurs because you are trying to bootstrap on a
> >>> UnitedKingdom
> >>> holiday, which is implicitly included in the definition of
> >> USDLibor.
> >>> The
> >>> Libor class makes a joint calendar between the given
> financial
> >> center
> >>> calendar and the UK calendar. Since Libor is a London poll,
> there
> >> are
> >>> only
> >>> fixings for it on UK business days. If one wants to
> bootstrap the
> >>> curve on
> >>> a holiday, the indice's historical fixings (actually just
> the
> >> closest
> >>> fixing to the given date) need to be provided.
> >>>
> >>>
> >>> Mit freundlichen Grüßen / Kind regards
> >>>
> >>> Dr. Andres Hernandez
> >>>
> >>> Senior Financial Engineer
> >>> Business Analytics
> >>> Risk Analytics
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> Phone: +49-69-6645-1351 IBM
> Deutschland
> >>> (Embedded
> >>> image moved
> >>> to file:
> >>> pic20635.gif)
> >>>
> >>> Mobile: +49-163-609-3280
> Wilhelm-Fay-Str.
> >>> 30-34
> >>>
> >>> E-Mail: [hidden email] 65936
> Frankfurt
> >> am
> >>> Main
> >>>
> >>> Germany
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> IBM Deutschland
> >>> GmbH /
> >>> Vorsitzender des
> >>> Aufsichtsrats:
> >>> Martin Jetter
> >>> Geschäftsführung:
> >>> Martina Koederitz
> >>> (Vorsitzende),
> >>> Reinhard Reschke,
> >>> Dieter Scholz,
> >>> Gregor Pillen,
> >>> Christian Noll,
> >>> Ivo Koerner
> >>> Sitz der
> >>> Gesellschaft:
> >>> Ehningen /
> >>> Registergericht:
> >>> Amtsgericht
> >>> Stuttgart, HRB
> >>> 14562 /
> >>> WEEE-Reg.-Nr. DE
> >>> 99369940
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> From: George Cowie <[hidden email]>
> >>> To: QuantLib users <
> [hidden email]>
> >>> Date: 22/07/2014 14:10
> >>> Subject: [Quantlib-users] UK Holiday for
> >>> USDLibor/PiecewiseYieldCurve
> >>>
> >>>
> >>>
> >>> Hello,
> >>>
> >>> I'm using PiecewiseYieldCurve to build a
> YieldTermStructure which
> >>> is
> >>> ultimately used by BlackScholesMertonProcess in my option
> pricing.
> >> I
> >>> believe I'm running into an issue on days that are holidays
> in the
> >> UK
> >>> (but
> >>> not the US). Here's the error I'm seeing when trying to
> run the
> >> code
> >>> below
> >>> for May 5, 2014 (Early May Bank Holiday):
> >>>
> >>> terminate called after throwing an instance of
> 'QuantLib::Error'
> >>> what(): 1st iteration: failed at 1st alive instrument,
> maturity
> >>> May
> >>> 14th, 2014, reference date May 7th, 2014: Missing no-fix1W
> >> Actual/360
> >>> fixing for May 2nd, 2014
> >>> Aborted
> >>>
> >>> Everything seems to work fine on other days... Am I
> doing
> >>> something
> >>> wrong with setting up the PiecewiseYieldCurve?
> >>>
> >>> Thanks,
> >>> George
> >>>
> >>> #include <ql/quantlib.hpp>
> >>>
> >>> using namespace std ;
> >>> using namespace QuantLib ;
> >>>
> >>> int main(int argc, char** argv)
> >>> {
> >>> vector<Rate> liborRates ;
> >>> liborRates.push_back(0.12075 / 100.0) ; //1W
> >>> liborRates.push_back(0.1515 / 100.0) ; //1M
> >>> liborRates.push_back(0.1935 / 100.0) ; //2M
> >>> liborRates.push_back(0.22535 / 100.0) ; //3M
> >>> liborRates.push_back(0.323 / 100.0) ; //6M
> >>> liborRates.push_back(0.549 / 100.0) ; //12M
> >>>
> >>> vector<Period> periods ;
> >>> periods.push_back(Period(1,Weeks)) ;
> >>> periods.push_back(Period(1,Months)) ;
> >>> periods.push_back(Period(2,Months)) ;
> >>> periods.push_back(Period(3,Months)) ;
> >>> periods.push_back(Period(6,Months)) ;
> >>> periods.push_back(Period(12,Months)) ;
> >>>
> >>> vector<boost::shared_ptr<RateHelper> > instruments ;
> >>>
> >>> for (vector<int>::size_type i = 0 ; i < liborRates.size
> () ; i++)
> >>> {
> >>> boost::shared_ptr<Quote> liborQuote(new SimpleQuote
> >> (liborRates
> >>> [i])) ;
> >>> Handle<Quote> liborHandle(liborQuote) ;
> >>> boost::shared_ptr<IborIndex> liborIndex(new USDLibor
> (periods
> >>> [i])) ;
> >>> boost::shared_ptr<RateHelper> rateHelper(new
> >> DepositRateHelper
> >>> (liborHandle,liborIndex)) ;
> >>> instruments.push_back(rateHelper) ;
> >>> }
> >>>
> >>> Calendar calendar = UnitedStates();
> >>> Date today (5 , May ,2014);
> >>> Settings::instance().evaluationDate() = today ;
> >>> Natural settlementDays = 2;
> >>> Date settlement = calendar.advance ( today ,
> settlementDays ,
> >>> Days );
> >>> DayCounter dc = Actual360 ();
> >>>
> >>> Date maturity(20, Jun, 2014) ;
> >>> double strikePrice = 100.0 ;
> >>> Option::Type optionType = Option::Call ;
> >>> double dividendYield = 0.0 ;
> >>> double impliedVolatility = 0.15 ;
> >>> double underlyingPrice = 100.0 ;
> >>>
> >>> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>
> (new
> >>> SimpleQuote
> >>> (underlyingPrice)));
> >>>
> >>> Handle<YieldTermStructure> rateTermStructure
> >>> (boost::shared_ptr<YieldTermStructure>(new
> >>> PiecewiseYieldCurve<ZeroYield,Linear>(settlement,
> instruments,
> >>> dc))) ;
> >>>
> >>> Handle<YieldTermStructure>
> >>> dividendTermStructure
> (boost::shared_ptr<YieldTermStructure>(new
> >>> FlatForward(today,
> >>> dividendYield,
> >>> dc)));
> >>> Handle<BlackVolTermStructure>
> >>> volTermStructure
> (boost::shared_ptr<BlackVolTermStructure>(new
> >>> BlackConstantVol(today,
> >>> calendar,
> >>> impliedVolatility,
> >>> dc)));
> >>>
> >>> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess
> (new
> >>> BlackScholesMertonProcess(underlyingHandle,
> >>> dividendTermStructure,
> >>> rateTermStructure,
> >>> volTermStructure));
> >>>
> >>> boost::shared_ptr<Exercise> americanExercise(new
> AmericanExercise
> >>> (today,
> >>> maturity)) ;
> >>> boost::shared_ptr<StrikedTypePayoff> payoff(new
> >> PlainVanillaPayoff
> >>> (optionType, strikePrice));
> >>>
> >>> VanillaOption americanOption(payoff, americanExercise);
> >>>
> >>> americanOption.setPricingEngine
> (boost::shared_ptr<PricingEngine>(
> >>> new FDAmericanEngine<CrankNicolson>(bsmProcess,
> >>> 101,100)));
> >>>
> >>> cout << americanOption.NPV() << endl ;
> >>>
> >>> return 0 ;
> >>> }
> >>>
> >>
> ------------------------------------------------------------------------------
>
> >>
> >>>
> >>>
> >>> Want fast and easy access to all the code in your
> enterprise? Index
> >>> and
> >>> search up to 200,000 lines of code with a free copy of
> Black Duck
> >>> Code Sight - the same software that powers the world's
> largest code
> >>> search on Ohloh, the Black Duck Open Hub! Try it now.
> >>>
> >>
> http://p.sf.net/sfu/bds_______________________________________________
>
> >>>
> >>> QuantLib-users mailing list
> >>> [hidden email]
> >>> https://lists.sourceforge.net/lists/listinfo/quantlib-users
> >>>
> >>>
> >>
> ------------------------------------------------------------------------------
>
> >>
> >>> Want fast and easy access to all the code in your enterprise?
> Index and
> >>> search up to 200,000 lines of code with a free copy of Black Duck
> >>> Code Sight - the same software that powers the world's largest
> code
> >>> search on Ohloh, the Black Duck Open Hub! Try it now.
> >>> http://p.sf.net/sfu/bds
> >>> _______________________________________________
> >>> QuantLib-users mailing list
> >>> [hidden email]
> >>> https://lists.sourceforge.net/lists/listinfo/quantlib-users
> >>>
> >>
>
>

George Cowie

Re: UK Holiday for USDLibor/PiecewiseYieldCurve

Thanks again guys, I was able to get my example code working with Andres suggestion above. Peter, I'll try pulling in your fix as well when it's available.

On Thu, Jul 24, 2014 at 12:54 PM, Peter Caspers <[hidden email]> wrote:

thanks a lot, Andres and George. I created a pull request to fix the
deposit helpers in this sense:

https://github.com/lballabio/quantlib/pull/117

There is only one more little complication I think, namely if we have
zero fixing days and are on a holiday in which case we should adjust
the start date of the deposit just as it would have been done under
the original calendar.

The code looks not as neat as in Andres' example above, mainly because
I don't know how the hell I can clone a calendar and remove a holiday
without impacting all other instances of the same type ...

We should think about adding the ignore - fixing options for swaps later on !

kind regards
Peter

On 24 July 2014 16:06, Andres Hernandez <[hidden email]> wrote:
> Hi George,
>
> as Peter mentioned, you can use another of the constructors. If you want to
> use the USDLibor conventions you could use something like the following
> inside the loop:
>
> boost::shared_ptr<IborIndex> liborIndex(new USDLibor(periods[i])) ;
>
> Calendar calendar = liborIndex->fixingCalendar();
> calendar.removeHoliday(today);
> boost::shared_ptr<RateHelper> rateHelper(new DepositRateHelper
> ( liborHandle,
> liborIndex->tenor(),
> liborIndex->fixingDays(),
> calendar,
> liborIndex->
> businessDayConvention(),
> liborIndex->endOfMonth(),
> liborIndex->dayCounter
> ()) );
> instruments.push_back(rateHelper);
>
>
> @Peter, you are right,no modification would be needed and the behaviour for
> the swaps is the correct one.
>
> cheers,
> Andres
>
>
>
> From: George Cowie <[hidden email]>
> To: Peter Caspers <[hidden email]>
> Cc: Andres Hernandez/Germany/IBM@IBMDE, QuantLib users
> <[hidden email]>
> Date: 24/07/2014 13:52
> Subject: Re: [Quantlib-users] UK Holiday for
> USDLibor/PiecewiseYieldCurve
>
>
>
> Interesting discussion... sorry to change the subject, but I'm still trying
> to figure out how to adapt my code to this situation. How do I force my
> YieldTermStructure to bootstrap with a different date? I have to value the
> option on May 5th, but as you pointed out, changing the evaluation date to
> a holiday has the described consequences.
>
>
> On Thu, Jul 24, 2014 at 6:00 AM, Peter Caspers <[hidden email]>
> wrote:
> reminding myself of what workaround we actually implemented for swap
> helper :o ... and what Murex offers here (always a good reference),
> the full set of options would probably be
>
> default - require historic fixings, take today's fixing if available
> ignore today's tixing - require historic fixings, always ignore
> today's fixing
> ignore all fixings - ignore all fixings (i.e. estimate forward rates
> always on the curve - this is technically ok even if the fixing date
> is in the past, since only the index estimation period's
> start and end date enter the estimation, and these two dates are
> always greater or equal today)
>
> these options to be applied only in the context of yield curve
> bootstrapping, ideally to be set on the level of the curve itself.
>
>
> On 24 July 2014 10:26, Peter Caspers <[hidden email]> wrote:
> > Hi Andres,
> >
> > yes, the index in the deposit rate helper seems to be merely a
> > convenient way for coding - with the one exception that deposits do
> > not (directly) have anything to do with actual index fixings, and
> this
> > is also where the code breaks. And yes, it gets even worse in your
> > example when you use the USD Libor index for USD cash deposits on a
> UK
> > but non-NY holiday (you could use the index-free constructor though
> > and pass the NY calendar directly instead, right, solving half of
> the
> > problem ?). So just treating the deposit helper's index and its
> fixing
> > as a computational tool to get the implied Quote seems a reasonable
> > solution, in your USD / Libor example yielding the correct
> conventions
> > even, if I am not mistaken ? I wouldn't introduce any flag or
> method
> > to activate this behaviour, it should be the only one present
> (because
> > checking for holidays can easily be done in the user's code if
> > necessary at all).
> >
> > For swaps on the other hand I think the current behaviour is fine -
> > what fixing would you take for the first period in a Euribor swap
> on a
> > TARGET holiday ? Probably the same as you take after the fixing
> around
> > 11am on a trading day, i.e. the last known fixing. That is exactly
> the
> > logic applied in the library. I wouldn't change anything here. It
> > might be that you get an exception because you just don't provide
> the
> > historic fixing, but in this case it is a legitimate demand of the
> > library and working around this should only be done in the user
> code
> > imo.
> >
> > best
> > Peter
> >
> >
> >
> >
> >
> > On 24 July 2014 09:23, Andres Hernandez <
> [hidden email]> wrote:
> >>
> >> Hi Peter,
> >>
> >> I agree that it would be nice to at least have it as an option. We
> >> currently do a small workaround, whereby we set in our system a
> parameter
> >> for the curve specifying the holiday calendar for that curve. The
> only use
> >> of that calendar is to determine if it is a holiday for the curve
> or not
> >> and if yes to make a fallback to yesterday's values. In the case
> of the EUR
> >> curves, this is arguably a good behaviour as the index and the
> geographical
> >> region have the same calendar, but for the USD case, we are
> basically
> >> forcing a fallback for a holiday on a different country. The issue
> is
> >> compounded by the fact that in this case the deposit is traded on
> that day,
> >> it is just the index that gets no fixing.
> >>
> >> I took a look and I am almost sure that removing the evaluation
> date from
> >> the fixing calendar in InterestRate constructor would be enough to
> make
> >> deposits go through. This could simply be accomplished with an
> extra public
> >> member function that one would call if one wanted it to go through
> on
> >> holidays. A more involved and important problem would be to get
> the
> >> SwapRateHelper to go though on holidays (if desired), as I am
> positive that
> >> that was the reason for our instituting the fallback. I will see
> if I have
> >> the time to change it, and would then probably subscribe to the
> developers
> >> list to get some feedback in how to go about changing it.
> >>
> >>
> >> Mit freundlichen Grüßen / Kind regards
> >>
> >> Dr. Andres Hernandez
> >>
> >> Senior Financial Engineer
> >> Business Analytics
> >> Risk Analytics
> >>
> >>
> >>
> >>
> >>
> >> Phone: <a href="tel:%2B49-69-6645-1351" value="+496966451351">+49-69-6645-1351 IBM Deutschland
> (Embedded
> >>
> image moved
> >>
> to file:
> >>
> pic02235.gif)
> >>
> >> Mobile: <a href="tel:%2B49-163-609-3280" value="+491636093280">+49-163-609-3280 Wilhelm-Fay-Str.
> 30-34
> >>
> >> E-Mail: [hidden email] 65936 Frankfurt
> am Main
> >>
> >> Germany
> >>
> >>
> >>
> >>
> >>
> >> IBM Deutschland
> >> GmbH /
> >> Vorsitzender des
> >> Aufsichtsrats:
> >> Martin Jetter
> >> Geschäftsführung:
> >> Martina Koederitz
> >> (Vorsitzende),
> >> Reinhard Reschke,
> >> Dieter Scholz,
> >> Gregor Pillen,
> >> Christian Noll,
> >> Ivo Koerner
> >> Sitz der
> >> Gesellschaft:
> >> Ehningen /
> >> Registergericht:
> >> Amtsgericht
> >> Stuttgart, HRB
> >> 14562 /
> >> WEEE-Reg.-Nr. DE
> >> 99369940
> >>
> >>
> >>
> >>
> >>
> >>
> >> From: Peter Caspers <[hidden email]>
> >> To: Andres Hernandez/Germany/IBM@IBMDE
> >> Cc: George Cowie <[hidden email]>, QuantLib users
> >> <[hidden email]>
> >> Date: 23/07/2014 21:56
> >> Subject: Re: [Quantlib-users] UK Holiday for
> >> USDLibor/PiecewiseYieldCurve
> >>
> >>
> >>
> >> Hi,
> >>
> >> I think this question is an older one but I don't remember that it
> was
> >> finally decided how to handle this situation (?) - which is,
> should we
> >> allow to bootstrap a curve on a day which is a holiday for the
> cash
> >> instruments in the curve or not. If negative, we could maybe throw
> a
> >> different message ( the current one indeed sounds a bit
> paradoxical ).
> >> If yes ( which would be my personal preference, because you simply
> >> want to run your daily processes spitting out numbers even if not
> a
> >> hundred percent sensible all of them ), one has to define how a
> >> deposit which is not traded on that day should be calculated. One
> >> obvious way would be to just treat the evaluation date as a
> business
> >> day when calculating the helper's fixing date. I did not try, but
> it
> >> may be enough to join the business days of the helper's calendar
> with
> >> a new one just having one business day (the evaluation date)
> before
> >> passing it to the no-fix index to make this kind of solution work.
> >> What do you think ?
> >>
> >> Peter
> >>
> >>
> >> On 23 July 2014 15:05, Andres Hernandez <
> [hidden email]>
> >> wrote:
> >>>
> >>> Hi George,
> >>>
> >>> you are still setting the evaluation date to a holiday:
> >>>
> >>> Date today (5 , May ,2014);
> >>> Date liborDate (2, May, 2014) ;
> >>> Settings::instance().evaluationDate() = today ;
> >>>
> >>>
> >>> Let's follow the logic:
> >>> During bootstrapping, the method impliedQuote() of the rate
> helper is
> >>> called to compare against the market quote. The DepositRateHelper
> simply
> >>> returns the fixing corresponding to the fixingDate_ as a return
> value.
> >>> The fixingDate_ is set during initialization as the return value
> of the
> >>> function InterestRate::fixingDate(earliestDate_). The function
> >>> InterestRate::fixingDate sets back the date given as input
> parameter by
> >> the
> >>> number of fixing days according to the index' calendar.
> >>> The parameter earliestDate_ being passed to fixingDate is set
> during
> >>> initialization to the evaluation date (which you set to the 5th
> of May)
> >>> moved forward by the number of fixing days according to the
> index'
> >>> calendar.
> >>> Note that if the evaluation date is a holiday you do not end up
> with the
> >>> same date, but rather with the last business day prior to
> evaluation
> >> date.
> >>> The first act moves the date from the 5th of May to the 7th and
> checks
> >>> whether it is a business day, which it is. The second move sets
> it back
> >> two
> >>> days to the 5th, and checks whether it is a business day, which
> it isn't,
> >>> so it sets it back until it finds a business day, i.e. the 2nd of
> May.
> >>> Since the value is in the past, it requests a historical fixing,
> which
> >> you
> >>> did not provide and so the QL_REQUIRE throws.
> >>>
> >>> Mit freundlichen Grüßen / Kind regards
> >>>
> >>> Dr. Andres Hernandez
> >>>
> >>> Senior Financial Engineer
> >>> Business Analytics
> >>> Risk Analytics
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> Phone: +49-69-6645-1351 IBM Deutschland
> >> (Embedded
> >>>
> >> image moved
> >>>
> >> to file:
> >>>
> >> pic64586.gif)
> >>>
> >>> Mobile: +49-163-609-3280
> Wilhelm-Fay-Str. 30-34
> >>>
> >>> E-Mail: [hidden email] 65936 Frankfurt
> am Main
> >>>
> >>> Germany
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> IBM Deutschland
> >>> GmbH /
> >>> Vorsitzender des
> >>> Aufsichtsrats:
> >>> Martin Jetter
> >>> Geschäftsführung:
> >>> Martina Koederitz
> >>> (Vorsitzende),
> >>> Reinhard Reschke,
> >>> Dieter Scholz,
> >>> Gregor Pillen,
> >>> Christian Noll,
> >>> Ivo Koerner
> >>> Sitz der
> >>> Gesellschaft:
> >>> Ehningen /
> >>> Registergericht:
> >>> Amtsgericht
> >>> Stuttgart, HRB
> >>> 14562 /
> >>> WEEE-Reg.-Nr. DE
> >>> 99369940
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> From: George Cowie <[hidden email]>
> >>> To: Andres Hernandez/Germany/IBM@IBMDE
> >>> Cc: QuantLib users <[hidden email]>
> >>> Date: 23/07/2014 11:51
> >>> Subject: Re: [Quantlib-users] UK Holiday for
> >>> USDLibor/PiecewiseYieldCurve
> >>>
> >>>
> >>>
> >>>
> >>> Thank you Andres. What you're describing makes sense, but I
> can't seem
> >> to
> >>> avoid an error when trying to value the option on May 5, 2014.
> I've
> >> copied
> >>> the updated code snippet with changed section highlighted/bolded
> below.
> >> I
> >>> seem to always get an exception of the form:
> >>>
> >>> terminate called after throwing an instance of 'QuantLib::Error'
> >>> what(): 1st iteration: failed at 1st alive instrument,
> maturity May
> >>> 14th, 2014, reference date May 6th, 2014: Missing no-fix1W
> Actual/360
> >>> fixing for May 2nd, 2014
> >>>
> >>> Is there something wrong with my approach here?
> >>>
> >>> Thanks,
> >>> George
> >>>
> >>>
> >>> #include <ql/quantlib.hpp>
> >>>
> >>> using namespace std ;
> >>> using namespace QuantLib ;
> >>>
> >>> int main(int argc, char** argv)
> >>> {
> >>> vector<Rate> liborRates ;
> >>> liborRates.push_back(0.12075 / 100.0) ; //1W
> >>> liborRates.push_back(0.1515 / 100.0) ; //1M
> >>> liborRates.push_back(0.1935 / 100.0) ; //2M
> >>> liborRates.push_back(0.22535 / 100.0) ; //3M
> >>> liborRates.push_back(0.323 / 100.0) ; //6M
> >>> liborRates.push_back(0.549 / 100.0) ; //12M
> >>>
> >>> vector<Period> periods ;
> >>> periods.push_back(Period(1,Weeks)) ;
> >>> periods.push_back(Period(1,Months)) ;
> >>> periods.push_back(Period(2,Months)) ;
> >>> periods.push_back(Period(3,Months)) ;
> >>> periods.push_back(Period(6,Months)) ;
> >>> periods.push_back(Period(12,Months)) ;
> >>>
> >>> vector<boost::shared_ptr<RateHelper> > instruments ;
> >>>
> >>> for (vector<int>::size_type i = 0 ; i < liborRates.size() ; i+
> +)
> >>> {
> >>> boost::shared_ptr<Quote> liborQuote(new SimpleQuote
> (liborRates
> >> [i])) ;
> >>> Handle<Quote> liborHandle(liborQuote) ;
> >>> boost::shared_ptr<IborIndex> liborIndex(new USDLibor
> (periods[i])) ;
> >>> boost::shared_ptr<RateHelper> rateHelper(new
> DepositRateHelper
> >>> (liborHandle,liborIndex)) ;
> >>> instruments.push_back(rateHelper) ;
> >>> }
> >>>
> >>> Calendar calendar = UnitedStates();
> >>> Date today (5 , May ,2014);
> >>> Date liborDate (2, May, 2014) ;
> >>> Settings::instance().evaluationDate() = today ;
> >>> Natural settlementDays = 2;
> >>> Date settlement = calendar.advance ( liborDate ,
> settlementDays ,
> >> Days );
> >>> DayCounter dc = Actual360 ();
> >>>
> >>> Date maturity(20, Jun, 2014) ;
> >>> double strikePrice = 100.0 ;
> >>> Option::Type optionType = Option::Call ;
> >>> double dividendYield = 0.0 ;
> >>> double impliedVolatility = 0.15 ;
> >>> double underlyingPrice = 100.0 ;
> >>>
> >>> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>(new
> SimpleQuote
> >>> (underlyingPrice)));
> >>>
> >>> Handle<YieldTermStructure> rateTermStructure
> >>> (boost::shared_ptr<YieldTermStructure>(new
> >>> PiecewiseYieldCurve<ZeroYield,Linear>(settlement, instruments,
> dc))) ;
> >>>
> >>> Handle<YieldTermStructure>
> >>> dividendTermStructure(boost::shared_ptr<YieldTermStructure>
> (new
> >>> FlatForward(today,
> >>> dividendYield,
> >>> dc)));
> >>> Handle<BlackVolTermStructure>
> >>> volTermStructure(boost::shared_ptr<BlackVolTermStructure>(new
> >>> BlackConstantVol(today,
> >>> calendar,
> >>> impliedVolatility,
> >>> dc)));
> >>>
> >>> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess(new
> >>> BlackScholesMertonProcess(underlyingHandle,
> >>> dividendTermStructure,
> >>> rateTermStructure,
> >>> volTermStructure));
> >>>
> >>> boost::shared_ptr<Exercise> americanExercise(new
> AmericanExercise
> >> (today,
> >>> maturity)) ;
> >>> boost::shared_ptr<StrikedTypePayoff> payoff(new
> PlainVanillaPayoff
> >>> (optionType, strikePrice));
> >>>
> >>> VanillaOption americanOption(payoff, americanExercise);
> >>>
> >>> americanOption.setPricingEngine
> (boost::shared_ptr<PricingEngine>(
> >>> new FDAmericanEngine<CrankNicolson>(bsmProcess,
> >>> 101,100)));
> >>>
> >>> cout << americanOption.NPV() << endl ;
> >>>
> >>> return 0 ;
> >>> }
> >>>
> >>>
> >>>
> >>> On Tue, Jul 22, 2014 at 7:39 AM, Andres Hernandez <
> >>> [hidden email]> wrote:
> >>>
> >>> Hi George,
> >>>
> >>> the problem occurs because you are trying to bootstrap on a
> >>> UnitedKingdom
> >>> holiday, which is implicitly included in the definition of
> >> USDLibor.
> >>> The
> >>> Libor class makes a joint calendar between the given
> financial
> >> center
> >>> calendar and the UK calendar. Since Libor is a London poll,
> there
> >> are
> >>> only
> >>> fixings for it on UK business days. If one wants to
> bootstrap the
> >>> curve on
> >>> a holiday, the indice's historical fixings (actually just
> the
> >> closest
> >>> fixing to the given date) need to be provided.
> >>>
> >>>
> >>> Mit freundlichen Grüßen / Kind regards
> >>>
> >>> Dr. Andres Hernandez
> >>>
> >>> Senior Financial Engineer
> >>> Business Analytics
> >>> Risk Analytics
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> Phone: <a href="tel:%2B49-69-6645-1351" value="+496966451351">+49-69-6645-1351 IBM
> Deutschland
> >>> (Embedded
> >>> image moved
> >>> to file:
> >>> pic20635.gif)
> >>>
> >>> Mobile: <a href="tel:%2B49-163-609-3280" value="+491636093280">+49-163-609-3280
> Wilhelm-Fay-Str.
> >>> 30-34
> >>>
> >>> E-Mail: [hidden email] 65936
> Frankfurt
> >> am
> >>> Main
> >>>
> >>> Germany
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> IBM Deutschland
> >>> GmbH /
> >>> Vorsitzender des
> >>> Aufsichtsrats:
> >>> Martin Jetter
> >>> Geschäftsführung:
> >>> Martina Koederitz
> >>> (Vorsitzende),
> >>> Reinhard Reschke,
> >>> Dieter Scholz,
> >>> Gregor Pillen,
> >>> Christian Noll,
> >>> Ivo Koerner
> >>> Sitz der
> >>> Gesellschaft:
> >>> Ehningen /
> >>> Registergericht:
> >>> Amtsgericht
> >>> Stuttgart, HRB
> >>> 14562 /
> >>> WEEE-Reg.-Nr. DE
> >>> 99369940
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> From: George Cowie <[hidden email]>
> >>> To: QuantLib users <
> [hidden email]>
> >>> Date: 22/07/2014 14:10
> >>> Subject: [Quantlib-users] UK Holiday for
> >>> USDLibor/PiecewiseYieldCurve
> >>>
> >>>
> >>>
> >>> Hello,
> >>>
> >>> I'm using PiecewiseYieldCurve to build a
> YieldTermStructure which
> >>> is
> >>> ultimately used by BlackScholesMertonProcess in my option
> pricing.
> >> I
> >>> believe I'm running into an issue on days that are holidays
> in the
> >> UK
> >>> (but
> >>> not the US). Here's the error I'm seeing when trying to
> run the
> >> code
> >>> below
> >>> for May 5, 2014 (Early May Bank Holiday):
> >>>
> >>> terminate called after throwing an instance of
> 'QuantLib::Error'
> >>> what(): 1st iteration: failed at 1st alive instrument,
> maturity
> >>> May
> >>> 14th, 2014, reference date May 7th, 2014: Missing no-fix1W
> >> Actual/360
> >>> fixing for May 2nd, 2014
> >>> Aborted
> >>>
> >>> Everything seems to work fine on other days... Am I
> doing
> >>> something
> >>> wrong with setting up the PiecewiseYieldCurve?
> >>>
> >>> Thanks,
> >>> George
> >>>
> >>> #include <ql/quantlib.hpp>
> >>>
> >>> using namespace std ;
> >>> using namespace QuantLib ;
> >>>
> >>> int main(int argc, char** argv)
> >>> {
> >>> vector<Rate> liborRates ;
> >>> liborRates.push_back(0.12075 / 100.0) ; //1W
> >>> liborRates.push_back(0.1515 / 100.0) ; //1M
> >>> liborRates.push_back(0.1935 / 100.0) ; //2M
> >>> liborRates.push_back(0.22535 / 100.0) ; //3M
> >>> liborRates.push_back(0.323 / 100.0) ; //6M
> >>> liborRates.push_back(0.549 / 100.0) ; //12M
> >>>
> >>> vector<Period> periods ;
> >>> periods.push_back(Period(1,Weeks)) ;
> >>> periods.push_back(Period(1,Months)) ;
> >>> periods.push_back(Period(2,Months)) ;
> >>> periods.push_back(Period(3,Months)) ;
> >>> periods.push_back(Period(6,Months)) ;
> >>> periods.push_back(Period(12,Months)) ;
> >>>
> >>> vector<boost::shared_ptr<RateHelper> > instruments ;
> >>>
> >>> for (vector<int>::size_type i = 0 ; i < liborRates.size
> () ; i++)
> >>> {
> >>> boost::shared_ptr<Quote> liborQuote(new SimpleQuote
> >> (liborRates
> >>> [i])) ;
> >>> Handle<Quote> liborHandle(liborQuote) ;
> >>> boost::shared_ptr<IborIndex> liborIndex(new USDLibor
> (periods
> >>> [i])) ;
> >>> boost::shared_ptr<RateHelper> rateHelper(new
> >> DepositRateHelper
> >>> (liborHandle,liborIndex)) ;
> >>> instruments.push_back(rateHelper) ;
> >>> }
> >>>
> >>> Calendar calendar = UnitedStates();
> >>> Date today (5 , May ,2014);
> >>> Settings::instance().evaluationDate() = today ;
> >>> Natural settlementDays = 2;
> >>> Date settlement = calendar.advance ( today ,
> settlementDays ,
> >>> Days );
> >>> DayCounter dc = Actual360 ();
> >>>
> >>> Date maturity(20, Jun, 2014) ;
> >>> double strikePrice = 100.0 ;
> >>> Option::Type optionType = Option::Call ;
> >>> double dividendYield = 0.0 ;
> >>> double impliedVolatility = 0.15 ;
> >>> double underlyingPrice = 100.0 ;
> >>>
> >>> Handle<Quote> underlyingHandle(boost::shared_ptr<Quote>
> (new
> >>> SimpleQuote
> >>> (underlyingPrice)));
> >>>
> >>> Handle<YieldTermStructure> rateTermStructure
> >>> (boost::shared_ptr<YieldTermStructure>(new
> >>> PiecewiseYieldCurve<ZeroYield,Linear>(settlement,
> instruments,
> >>> dc))) ;
> >>>
> >>> Handle<YieldTermStructure>
> >>> dividendTermStructure
> (boost::shared_ptr<YieldTermStructure>(new
> >>> FlatForward(today,
> >>> dividendYield,
> >>> dc)));
> >>> Handle<BlackVolTermStructure>
> >>> volTermStructure
> (boost::shared_ptr<BlackVolTermStructure>(new
> >>> BlackConstantVol(today,
> >>> calendar,
> >>> impliedVolatility,
> >>> dc)));
> >>>
> >>> boost::shared_ptr<BlackScholesMertonProcess> bsmProcess
> (new
> >>> BlackScholesMertonProcess(underlyingHandle,
> >>> dividendTermStructure,
> >>> rateTermStructure,
> >>> volTermStructure));
> >>>
> >>> boost::shared_ptr<Exercise> americanExercise(new
> AmericanExercise
> >>> (today,
> >>> maturity)) ;
> >>> boost::shared_ptr<StrikedTypePayoff> payoff(new
> >> PlainVanillaPayoff
> >>> (optionType, strikePrice));
> >>>
> >>> VanillaOption americanOption(payoff, americanExercise);
> >>>
> >>> americanOption.setPricingEngine
> (boost::shared_ptr<PricingEngine>(
> >>> new FDAmericanEngine<CrankNicolson>(bsmProcess,
> >>> 101,100)));
> >>>
> >>> cout << americanOption.NPV() << endl ;
> >>>
> >>> return 0 ;
> >>> }
> >>>
> >>
> ------------------------------------------------------------------------------
>
> >>
> >>>
> >>>
> >>> Want fast and easy access to all the code in your
> enterprise? Index
> >>> and
> >>> search up to 200,000 lines of code with a free copy of
> Black Duck
> >>> Code Sight - the same software that powers the world's
> largest code
> >>> search on Ohloh, the Black Duck Open Hub! Try it now.
> >>>
> >>
> http://p.sf.net/sfu/bds_______________________________________________
>
> >>>
> >>> QuantLib-users mailing list
> >>> [hidden email]
> >>> https://lists.sourceforge.net/lists/listinfo/quantlib-users
> >>>
> >>>
> >>
> ------------------------------------------------------------------------------
>
> >>
> >>> Want fast and easy access to all the code in your enterprise?
> Index and
> >>> search up to 200,000 lines of code with a free copy of Black Duck
> >>> Code Sight - the same software that powers the world's largest
> code
> >>> search on Ohloh, the Black Duck Open Hub! Try it now.
> >>> http://p.sf.net/sfu/bds
> >>> _______________________________________________
> >>> QuantLib-users mailing list
> >>> [hidden email]
> >>> https://lists.sourceforge.net/lists/listinfo/quantlib-users
> >>>
> >>
>
>