QuantLib cpibond bond example in Python

Question

I am trying to get the official C++ cpibond example working in Python. The original example is here: https://github.com/lballabio/quantlib/blob/master/QuantLib/test-suite/inflationcpibond.cpp and for scala here: https://github.com/lballabio/quantlib/blob/master/QuantLib-SWIG/Scala/examples/CPIBond.scala

When I run what I have attempted I get this error:

RuntimeError: 1st iteration: failed at 1st alive instrument, maturity September 1st, 2010, reference date September 1st, 2009: 2nd leg: Missing UK RPI fixing for September 1st, 2009

Here is my attempt:

import QuantLib as ql
calendar = ql.UnitedKingdom()
dayCounter = ql.ActualActual();
convention = ql.ModifiedFollowing

today = ql.Date(20, 11, 2009)
evaluationDate = calendar.adjust(today)
ql.Settings.instance().setEvaluationDate(evaluationDate)        
yTS = ql.YieldTermStructureHandle(ql.FlatForward(evaluationDate, 0.05, dayCounter))

from_date = ql.Date(20, ql.July, 2007);
to_date   = ql.Date(20, ql.November, 2009);
tenor = ql.Period(1, ql.Months)
rpiSchedule = ql.Schedule(from_date, to_date, tenor, calendar, 
                               convention, convention,
                               ql.DateGeneration.Backward, False)
cpiTS = ql.RelinkableZeroInflationTermStructureHandle()
inflationIndex = ql.UKRPI(False, cpiTS)
fixData = [206.1, 207.3, 208.0, 208.9, 209.7, 210.9,
                    209.8, 211.4, 212.1, 214.0, 215.1, 216.8,
                    216.5, 217.2, 218.4, 217.7, 216,
                    212.9, 210.1, 211.4, 211.3, 211.5,
                    212.8, 213.4, 213.4, 213.4, 214.4,213.4, 214.4]
dte_fixings=[dtes for dtes in rpiSchedule]
print len(dte_fixings)
print len(fixData)
#must be the same length
inflationIndex.addFixings(dte_fixings, fixData) 
observationLag = ql.Period(2, ql.Months)
zciisData =[( ql.Date(25, ql.November, 2010), 3.0495 ),
              ( ql.Date(25, ql.November, 2011), 2.93 ),
              ( ql.Date(26, ql.November, 2012), 2.9795 ),
              ( ql.Date(25, ql.November, 2013), 3.029 ),
              ( ql.Date(25, ql.November, 2014), 3.1425 ),
              ( ql.Date(25, ql.November, 2015), 3.211 ),
              ( ql.Date(25, ql.November, 2016), 3.2675 ),
              ( ql.Date(25, ql.November, 2017), 3.3625 ),
              ( ql.Date(25, ql.November, 2018), 3.405 ),
              ( ql.Date(25, ql.November, 2019), 3.48 ),
              ( ql.Date(25, ql.November, 2021), 3.576 ),
              ( ql.Date(25, ql.November, 2024), 3.649 ),
              ( ql.Date(26, ql.November, 2029), 3.751 ),
              ( ql.Date(27, ql.November, 2034), 3.77225),
              ( ql.Date(25, ql.November, 2039), 3.77 ),
              ( ql.Date(25, ql.November, 2049), 3.734 ),
              ( ql.Date(25, ql.November, 2059), 3.714 )]

lRates=[rtes/100.0 for rtes in zip(*zciisData)[1]]
baseZeroRate = lRates[0]

zeroSwapHelpers = [ql.ZeroCouponInflationSwapHelper(a[1]/100,observationLag,
    a[0], calendar, convention, dayCounter, inflationIndex) for a in zciisData]


cpiTS.linkTo(ql.PiecewiseZeroInflation(          
  evaluationDate, calendar, dayCounter, observationLag, 
  inflationIndex.frequency(), inflationIndex.interpolated(), 
  baseZeroRate,
  yTS, zeroSwapHelpers, 1.0e-12, ql.Linear()))


notional = 1000000


fixedRates = [0.1]    

fixedDayCounter = ql.Actual365Fixed()
fixedPaymentConvention = ql.ModifiedFollowing
fixedPaymentCalendar = ql.UnitedKingdom()
contractObservationLag = ql.Period(3, ql.Months)
observationInterpolation = ql.CPI.Flat
settlementDays = 3
growthOnly = True

baseCPI = 206.1
startDate = ql.Date(2, 10, 2007)
endDate = ql.Date(2, 10, 2052)

fixedSchedule = ql.Schedule(startDate, endDate, 
                  ql.Period(6, ql.Months), fixedPaymentCalendar,
                  ql.Unadjusted,
                  ql.Unadjusted,
                  ql.DateGeneration.Backward, False)

bond = ql.CPIBond(settlementDays, notional, growthOnly,
                       baseCPI, contractObservationLag,
                       inflationIndex, observationInterpolation,
                       fixedSchedule, fixedRates, fixedDayCounter, 
                       fixedPaymentConvention)

bondEngine=ql.DiscountingBondEngine(yTS)
bond.setPricingEngine(bondEngine)
print bond.NPV() 
print bond.cleanPrice()

Most of my problem is that I am finding it difficult to get to grips with how the objects fit together.

Answer 1

got the above example working:

import QuantLib as ql
import datetime as dt
calendar = ql.UnitedKingdom()
dayCounter = ql.ActualActual();
convention = ql.ModifiedFollowing
lag = 3

today = ql.Date(5,3,2008)
evaluationDate = calendar.adjust(today)
issue_date = calendar.advance(evaluationDate,-1, ql.Years)
maturity_date = ql.Date(2,9,2052)
fixing_date = calendar.advance(evaluationDate,-lag, ql.Months)

ql.Settings.instance().setEvaluationDate(evaluationDate)        
yTS = ql.YieldTermStructureHandle(ql.FlatForward(evaluationDate, 0.05, dayCounter))

tenor = ql.Period(1, ql.Months)

from_date = ql.Date(20, ql.July, 2007);
to_date   = ql.Date(20, ql.November, 2009);
rpiSchedule = ql.Schedule(from_date, to_date, tenor, calendar, 
                               convention, convention,
                               ql.DateGeneration.Backward, False)

# this is the going to be holder the inflation curve.
cpiTS = ql.RelinkableZeroInflationTermStructureHandle()
inflationIndex = ql.UKRPI(False, cpiTS)
fixData = [206.1, 207.3, 208.0, 208.9, 209.7, 210.9,
            209.8, 211.4, 212.1, 214.0, 215.1, 216.8,
            216.5, 217.2, 218.4, 217.7, 216,
            212.9, 210.1, 211.4, 211.3, 211.5,
            212.8, 213.4, 213.4, 213.4, 214.4]
dte_fixings=[dtes for dtes in rpiSchedule]
print len(fixData)
print len(dte_fixings[:len(fixData)])
#must be the same length
#inflationIndex.addFixings(dte_fixings[:len(fixData)], fixData) 
#Current CPI level
#last observed rate
fixing_rate = 214.4


inflationIndex.addFixing(fixing_date, fixing_rate)

observationLag = ql.Period(lag, ql.Months)
zciisData =[( ql.Date(25, ql.November, 2010), 3.0495 ),
              ( ql.Date(25, ql.November, 2011), 2.93 ),
              ( ql.Date(26, ql.November, 2012), 2.9795 ),
              ( ql.Date(25, ql.November, 2013), 3.029 ),
              ( ql.Date(25, ql.November, 2014), 3.1425 ),
              ( ql.Date(25, ql.November, 2015), 3.211 ),
              ( ql.Date(25, ql.November, 2016), 3.2675 ),
              ( ql.Date(25, ql.November, 2017), 3.3625 ),
              ( ql.Date(25, ql.November, 2018), 3.405 ),
              ( ql.Date(25, ql.November, 2019), 3.48 ),
              ( ql.Date(25, ql.November, 2021), 3.576 ),
              ( ql.Date(25, ql.November, 2024), 3.649 ),
              ( ql.Date(26, ql.November, 2029), 3.751 ),
              ( ql.Date(27, ql.November, 2034), 3.77225),
              ( ql.Date(25, ql.November, 2039), 3.77 ),
              ( ql.Date(25, ql.November, 2049), 3.734 ),
              ( ql.Date(25, ql.November, 2059), 3.714 )]

#lRates=[rtes/100.0 for rtes in zip(*zciisData)[1]]
#baseZeroRate = lRates[0]

zeroSwapHelpers = [ql.ZeroCouponInflationSwapHelper(rate/100,observationLag,
    date, calendar, convention, dayCounter, inflationIndex) for date,rate in zciisData]


# the derived inflation curve
jj=ql.PiecewiseZeroInflation(          
                              evaluationDate, calendar, dayCounter, observationLag, 
                              inflationIndex.frequency(), inflationIndex.interpolated(), 
                              zciisData[0][1],#baseZeroRate,
                              yTS, zeroSwapHelpers, 1.0e-12, ql.Linear())

cpiTS.linkTo(jj)

notional = 1000000

fixedRates = [0.1]    

fixedDayCounter = ql.Actual365Fixed()
fixedPaymentConvention = ql.ModifiedFollowing
fixedPaymentCalendar = ql.UnitedKingdom()
contractObservationLag = ql.Period(3, ql.Months)
observationInterpolation = ql.CPI.Flat
settlementDays = 3
growthOnly = False

baseCPI = 206.1


fixedSchedule = ql.Schedule(issue_date,
                  maturity_date, 
                  ql.Period(ql.Semiannual),
                  fixedPaymentCalendar,
                  ql.Unadjusted,
                  ql.Unadjusted,
                  ql.DateGeneration.Backward,
                  False)

bond = ql.CPIBond(settlementDays,
                    notional,
                    growthOnly,
                    baseCPI,
                    contractObservationLag,
                    inflationIndex,
                    observationInterpolation,
                    fixedSchedule,
                    fixedRates,
                    fixedDayCounter, 
                    fixedPaymentConvention)

#bond2= ql.QuantLib.C

bondEngine=ql.DiscountingBondEngine(yTS)
bond.setPricingEngine(bondEngine)
print bond.NPV() 
print bond.cleanPrice()
compounding = ql.Compounded
yield_rate = bond.bondYield(fixedDayCounter,compounding,ql.Semiannual)
y_curve = ql.InterestRate(yield_rate,fixedDayCounter,compounding,ql.Semiannual)
##Collate results
print "Clean Price:", bond.cleanPrice()
print "Dirty Price:", bond.dirtyPrice()
print "Notional:", bond.notional()
print "Yield:", yield_rate
print "Accrued Amount:", bond.accruedAmount()
print "Settlement Value:", bond.settlementValue()

#suspect there's more to this for TIPS
print "Duration:", ql.BondFunctions.duration(bond,y_curve)
print "Convexity:", ql.BondFunctions.convexity(bond,y_curve)
print "Bps:", ql.BondFunctions.bps(bond,y_curve)
print "Basis Point Value:", ql.BondFunctions.basisPointValue(bond,y_curve)
print "Yield Value Basis Point:", ql.BondFunctions.yieldValueBasisPoint(bond,y_curve)

print "NPV:", bond.NPV()

# get the cash flows:
#cf_list=[(cf.amount(),cf.date()) for cf in bond.cashflows()]

def to_datetime(d):
    return dt.datetime(d.year(),d.month(), d.dayOfMonth())

for cf in bond.cashflows():
    try:
        amt=cf.amount()
        rte=jj.zeroRate(cf.date())
        zc=yTS.zeroRate(cf.date(),fixedDayCounter,compounding,ql.Semiannual).rate()
    except:
        amt=0
        rte=0
        zc=0
    print to_datetime(cf.date()),amt,rte,zc

The issue it seems was that the inflationIndex object needed one date instead of multiple index points. My assumption was it would pull out the latest valid point.

The way the pricer works is that the real coupons are increased by the inflation rate curve term structure, zciisData. The result therefore becomes a nominal future cash flow. To price then the bondpricer simply discounts these by the nominal term structure. I added some additional code to print the determined cash flows and the "growth factor" and then the discount rate.