duggar/dcf

Python library dcf

A Python library for generating discounted cashflows (dcf). Typical banking business methods are provided like interpolation, compounding, discounting and fx.

Example Usage

>>> from datetime import date
>>> from dcf import ZeroRateCurve

>>> start = date(2014,1,1)
>>> mid = date(2015,1,1)
>>> end = date(2016,1,1)

>>> ZeroRateCurve([start, end], [.03, .05]).get_zero_rate(start, mid)
0.04

>>> ZeroRateCurve([start, end], [.03, .05]).get_discount_factor(start, mid)
0.9608157444936446

The framework works fine with native datetime but we recommend businessdate package for more convenient functionality to roll out date schedules.

>>> from businessdate import BusinessDate, BusinessSchedule

>>> today = BusinessDate(20201031)

>>> schedule = BusinessSchedule(today, today + "8q", step="1q")
>>> schedule
[BusinessDate(20201031), BusinessDate(20210131), BusinessDate(20210430), BusinessDate(20210731), BusinessDate(20211031), BusinessDate(20220131), BusinessDate(20220430), BusinessDate(20220731), BusinessDate(20221031)]

To build payment plans for, e.g. annuity loans, pick a plan function and generate an redemption amount list for paying back the loan notional amount.

>>> from dcf import annuity, outstanding

>>> number_of_payments = 8
>>> interest_rate = 0.02
>>> notional = 1000.

>>> plan = annuity(number_of_payments, amount=notional, fixed_rate=interest_rate)
>>> plan
[116.50979913376267, 118.83999511643792, 121.21679501876667, 123.64113091914203, 126.11395353752485, 128.63623260827535, 131.20895726044085, 133.83313640564967]


>>> sum(plan)
1000.0

>>> out = outstanding(plan, amount=notional)
>>> out
[1000.0, 883.4902008662373, 764.6502057497994, 643.4334107310327, 519.7922798118907, 393.6783262743659, 265.0420936660905, 133.83313640564967]

>>> compound = [o * interest_rate + p for o, p in zip(out, plan)]
>>> compound
[136.50979913376267, 136.50979913376267, 136.50979913376267, 136.50979913376267, 136.50979913376267, 136.50979913376267, 136.50979913376267, 136.50979913376267]

Putting all together and feeding the plan into a FixedCashFlowList and the list of outstanding into a RateCashflowList gives the legs of a loan.

>>> from businessdate import BusinessDate, BusinessSchedule
>>> from dcf import amortize, outstanding
>>> from dcf import FixedCashFlowList, RateCashFlowList

Again, build a date schedule.

>>> today = BusinessDate(20201031)

>>> schedule = BusinessSchedule(today, today + "8q", step="1q")

>>> start_date, payment_dates = schedule[0], schedule[1:]

Fixing the properties of the product and rolling out the payment plan and list of notional outstanding.

>>> number_of_payments = 8
>>> interest_rate = 0.01
>>> notional = 1000.

>>> plan = amortize(number_of_payments, amount=notional)
>>> out = outstanding(plan, amount=notional)

Finally, create for each leg a CashFlowList.

>>> principal = FixedCashFlowList([start_date], [-notional], origin=start_date)
>>> print(principal)
FixedCashFlowList([BusinessDate(20201031) ... BusinessDate(20201031)], [-1000.0 ... -1000.0], origin=BusinessDate(20201031), day_count=day_count)

>>> redemption = FixedCashFlowList(payment_dates, plan, origin=start_date)
>>> print(redemption)
FixedCashFlowList([BusinessDate(20210131) ... BusinessDate(20221031)], [125.0 ... 125.0], origin=BusinessDate(20201031), day_count=day_count)

>>> interest = RateCashFlowList(payment_dates, out, origin=start_date, fixed_rate=interest_rate)
>>> print(interest)
RateCashFlowList([BusinessDate(20210131) ... BusinessDate(20221031)], [1000.0 ... 125.0], origin=BusinessDate(20201031), day_count=day_count)

Add those legs to CashFlowLegList provides a smart container for valuation (get_present_value()).

>>> from dcf import CashFlowLegList, ZeroRateCurve, get_present_value

>>> loan = CashFlowLegList([principal, redemption, interest])
>>> curve = ZeroRateCurve([today, today + '2y'], [-.005, .01])
>>> pv = get_present_value(cashflow_list=loan, discount_curve=curve, valuation_date=today)
>>> pv
4.935421637918779

Install

The latest stable version can always be installed or updated via pip:

$ pip install dcf

Development Version

The latest development version can be installed directly from GitHub:

$ pip install --upgrade git+https://github.com/sonntagsgesicht/dcf.git

Contributions

Issues and Pull Requests are always welcome.

License

Code and documentation are available according to the Apache Software License (see LICENSE).

duggar / dcf