QuantLib › quantlib-users

ActualActual::ISMA bond duration problem (1.0.1 version of bondscalc)

Classic

List

Threaded

4 messages Options

Dagur Gunnarsson-2

ActualActual::ISMA bond duration problem (1.0.1 version of bondscalc)

hello,

Could there be a bug in the calculation of duration for ActualActual::ISMA FixedRateBonds - as I am getting the same Duration for all settlement dates from 21. of Mars to the 28 of Mars.

Here is the program I am running

#include <ql/quantlib.hpp>

#include <iostream>

using namespace QuantLib;

using namespace std;

int main(int, char* [])

{

Date settlementDate = Date(21, Mar, 2012);

Date issueDate = Date(13, Oct, 2010);

Date interestFromDate = Date(13, Oct, 2010);

Date firstCouponDate = Date(13, Oct, 2011);

Date firstInstallmentDate = Date(13, Oct, 2016);

Date maturityDate = Date(13, Oct, 2016);

Real couponRate = 0.06;

Frequency freq(Annual);

DayCounter dayCounter = ActualActual(ActualActual::ISMA);

Compounding compounding(Compounded);

Schedule *schedule = new Schedule(interestFromDate,

maturityDate,

Period(freq),

Iceland(),

Unadjusted,

DateGeneration::Backward,

false,

firstCouponDate);

FixedRateBond* bond = new FixedRateBond(1,

100.0,

*schedule,

vector<Rate>(1, Rate(couponRate)),

dayCounter,

Following,

Real(100.0),

issueDate);

std::vector <boost::shared_ptr<QuantLib::CashFlow>, std::allocator<boost::shared_ptr<QuantLib::CashFlow> > > leg = bond->cashflows();

for (unsigned i = 0; i < leg.size(); i++) {

cout<<"day["<<i<<"] is "<<leg[i]->date()<<" "<<setprecision(15)

<<dayCounter.yearFraction(settlementDate, leg[i]->date())<<", cashflow["<<i<<"] is "<<leg[i]->amount()<<endl;

}

Real yield = 0.05;

Real d1 = BondFunctions::duration(*bond, InterestRate(yield, dayCounter, compounding, Annual), Duration::Macaulay, settlementDate);

Real d2 = BondFunctions::duration(*bond, InterestRate(yield, dayCounter, compounding, Annual), Duration::Macaulay, settlementDate+7);

cout<<"duration for settlement date1 "<<settlementDate<<" is "<<d1<<endl;

cout<<"duration for settlement date1 "<<settlementDate+7<<" is "<<d2<<endl;

}

this gives me

day[0] is October 13th, 2011 -0.416666666666667, cashflow[0] is 6.00000000000001

day[1] is October 15th, 2012 0.583333333333333, cashflow[1] is 6.00000000000001

day[2] is October 14th, 2013 1.58333333333333, cashflow[2] is 6.00000000000001

day[3] is October 13th, 2014 2.58333333333333, cashflow[3] is 6.00000000000001

day[4] is October 13th, 2015 3.58333333333333, cashflow[4] is 6.00000000000001

day[5] is October 13th, 2016 4.58333333333333, cashflow[5] is 6.00000000000001

day[6] is October 13th, 2016 4.58333333333333, cashflow[6] is 100

duration for settlement date1 March 21st, 2012 is 4.06108457642133

duration for settlement date1 March 28th, 2012 is 4.06108457642133

the duration is the same on both dates for ActualActual::ISMA, but not for i.e. ActualActual::AFB - Could this be a bug?

regards

Dagur G

------------------------------------------------------------------------------
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here
http://p.sf.net/sfu/sfd2d-msazure
_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

Guowen Han

Re: ActualActual::ISMA bond duration problem (1.0.1 version of bondscalc)

The implementation for the yearFraction for ActualActual::ISMA daycount was not correct. I would suggest you to use ActualActual::ISDA instead.

From: Dagur Gunnarsson [mailto:[hidden email]]
Sent: Thursday, March 22, 2012 11:16 AM
To: [hidden email]
Subject: [Quantlib-users] ActualActual::ISMA bond duration problem (1.0.1 version of bondscalc)