TODO.md

In wilsonfreitas/R-fixedincome: Fixed Income Models, Calculations, Data Structures and Instruments

fixedincome

Classes

• SpotRate
• ForwardRate
• SpotRateCurve
• Term
• Compounding
• Daycount
• Interpolation

TODO

Documentation

• [x] Term
• [x] Compounding
• [x] Compounding Methods
• [x] Daycount
• [x] SpotRate
• [x] SpotRateCurve
• [x] ForwardCurve
• [x] Interpolation

Term

• [ ] work with different units: terms(c(1, 6, 1, 2), c("months", "months", "years", "years"))

Daycount

• [ ] create 30/360 daycount
• [ ] Other daycounts https://en.wikipedia.org/wiki/Day_count_convention
• [ ] use DayCount with Term
• [x] Create functions tomonths and todays

SpotRate

• [ ] compound and discount functions accept (character, character) date arguments discount(r, "2017-08-01", "2017-08-21")

SpotRateCurve

• [x] A spotrate should be removed from a spotratecurve: curve[-10] - curve[-n] should return a spotratecurve without the removed spotrate
• [ ] spotratecurve indexing:
  • [x] positional indexing: curve[1] for the first term, curve[2] for the second and so on. If the given index exceeds the curve length this operation returns NA.
  • [x] replacement with positional index curve[1] \<- 0.01
  • [x] associative indexing curve[["1 day"]] for 1 day, curve[["2 days"]] for 2 days, ...
  • [x] If the spotratecurve doesn't have a 2 days index it returns NA
  • [x] removal of associative index
  • [x] replacement with associative index
  • [ ] date indexes
  • [ ] select
  • [ ] remove
  • [ ] replace
  • [x] term indexes
  • [x] select
  • [ ] remove
  • [x] replace
  • [x] sum curve to: numeric, curve
  • [ ] append SpotRateCurve
  • [ ] concatenate SpotRateCurve
• [ ] Naming: SpotRateCurve should have a name? I am not sure about that!
• [ ] create DiscountCurve - interpolations should be applied to it.
• [ ] arithmetic operations between curves respecting terms
• [ ] accept terms with different units: terms(c(1, 6, 1, 2), c("months", "months", "years", "years"))

Multi SpotRateCurve

• [ ] Data structure for multiple SpotRateCurve objects, like a matrix. Useful for historical data and to build historical risk factors.

ForwardRate

• [x] create forward rates
• [x] daily forward rates with interpolated data
• [ ] create a method for daily forward rates with interpolated data

Interpolation

• [x] set interpolation to SpotRateCurve
• [x] unset interpolation to SpotRateCurve
• [x] Interpolation will be used with associative indexing. If interpolation is set and a missing index is passed, then this value is interpolated.
• [ ] Methods
  • [x] FlatForward
  • [x] Linear
  • [x] LogLinear
  • [x] Monotone Cubic Spline (Hyman)
  • [x] Hermite
  • [x] Natural Spline
  • [x] Nelson-Siegel
  • [x] Nelson-Siegel-Svensson
  • [x] FlatForward COPOM
  • [ ] Constrained Spline
  • [ ] Dynamic Nelson-Siegel ref1, ref2
  • [ ] Vasicek (Haug 11.4.2) ref
  • [ ] Rendleman and Bartter (Haug 11.4.1)
  • [ ] Ho and Lee (Haug 11.4.3)
  • [ ] Hull and White (Haug 11.4.4)
  • [ ] Black-Derman-Toy (Haug 11.4.5)
• [x] Improve curve fit stop using object copy
• [ ] Work with terms in different units

Bonds

• [ ] define bonds
• [ ] pricing bonds

Bootstrap

• [ ] bonds bootstrap
• [ ] Curve models: Nelson-Siegel, Diebold-Li, Nelson-Siegel-Svensson

Other classes

• [ ] IndexCurve - for inflation index or IDI index, for example - rbcb integration
• [ ] HistoricalRates - for CDI historical data, for example - rbcb integration

Plotting

• [x] SpotRateCurve: points and lines
• [x] SpotRateCurve: lines with daily interpolated points
• [x] ForwardRate curve with steps
• [x] daily ForwardRate curve with interpolated points
• [ ] plot multiple SpotRateCurve objects
• [x] use daycount to create x-axis ticks

Other issues

• [ ] IR functions: https://support.google.com/docs/table/25273?hl=en
• [ ] Risk measures: YTM, Duration, Convexity
• [ ] short rate models: Vasicek, CIR, ...

Register

• [ ] naming objects and register (aka container)
  • Objects like HistoricalRates, SpotRateCurve, IndexCurve and Calendar should be referenced by their names. These objects are used by many others and their copies should imply unnecessary memory use.
  • Integrate with rbcb package to transform series into IndexCurve or HistoricalRates.
  • Methods which work with these objects should work with characters that represent the objects.

Legacy

Methods to be implemented in the term class:

• initialize
• show
• format.S3
• xtfrm.S3
• c
• rep
• [, [[ (x,i, j, ...)
• [\<-, [[\<- (x,i, j, ..., value)
• \$, \$\<- (x, name, value)
• summary.S3
• Arith
  • (., ANY), (ANY, .)
• Compare
  • (., character), (character, .)
  • (., numeric), (numeric, .)
• as.term
  • character
  • numeric

Bonds can be created using a data.frame and spotrate can be set as one of columns.

# Declaring bonds
dc <- as.daycount('actual/360') # daycount rule
comp <- as.compounding('continuous') # compounding regime
days <- c(97, 242, 321) # days to maturity
sr <- as.spotrate(rep(0.06, length(days)), comp, dc) # discount rate

bonds <- data.frame(DaysToMaturity=days, Rate=sr, Notional=100000)
bonds
#   DaysToMaturity                              Rate Notional
# 1             97 0.06 discrete business/252 ANBIMA    1e+05
# 2            242 0.06 discrete business/252 ANBIMA    1e+05
# 3            321 0.06 discrete business/252 ANBIMA    1e+05

# Pricing bond -- discounting their notional value
within(bonds, {
  PV <- Notional*discount(Rate, DaysToMaturity)
})
#   DaysToMaturity                       Rate Notional       PV
# 1             97 0.06 continuous actual/360    1e+05 98396.33
# 2            242 0.06 continuous actual/360    1e+05 96046.92
# 3            321 0.06 continuous actual/360    1e+05 94790.59

Pricing bonds using a calendar

library(bizdays)
cal <- Calendar(holidays=holidaysANBIMA, name='ANBIMA', weekdays=c('saturday', 'sunday'))
dc <- as.daycount('business/252')
comp <- as.compounding('discrete')
dates <- as.Date(c('2014-09-07', '2015-03-07', '2015-09-07'))
sr <- as.spotrate(rep(0.06, length(dates)), comp, dc, cal)
bonds <- data.frame(RefDate=as.Date('2014-03-21'), Maturity=dates, Rate=sr, Notional=100000)
bonds
#      RefDate   Maturity                              Rate Notional
# 1 2014-03-21 2014-09-07 0.06 discrete business/252 ANBIMA    1e+05
# 2 2014-03-21 2015-03-07 0.06 discrete business/252 ANBIMA    1e+05
# 3 2014-03-21 2015-09-07 0.06 discrete business/252 ANBIMA    1e+05
within(bonds, {
  PV <- Notional*discount(Rate, from=RefDate, to=Maturity)
})
#      RefDate   Maturity                              Rate Notional       PV
# 1 2014-03-21 2014-09-07 0.06 discrete business/252 ANBIMA    1e+05 97353.43
# 2 2014-03-21 2015-03-07 0.06 discrete business/252 ANBIMA    1e+05 94558.01
# 3 2014-03-21 2015-09-07 0.06 discrete business/252 ANBIMA    1e+05 91842.86

wilsonfreitas/R-fixedincome documentation built on June 30, 2023, 7:46 a.m.