# -*- coding: utf-8 -*-
"""
Created on Sun Feb 21 16:59:39 2021
@author: Ib
Function to calculate expected credit loss using ifrs9 rules, experimental
"""
import pandas as pd
import numpy as np
[docs]
def lt_ifrs9(maturity,discount_rate,_lgd,_pd,debug=False):
'''
Parameters
----------
maturity : integer or float
maturity over which the exposure is amortised - by equal instalments.
discount_rate : float
discount rate
_lgd : array of float
list of loss given default
_pd : array of float
propability of defaults
debug : bool, optional
calculate a intermidiately dataframes . The default is False.
Returns
-------
float the long term credit loss
'''
startstock = 100
_mat = float(int(maturity))
time = list(range(int(_mat)+1))
cashflow = pd.Series(float(startstock)/_mat,index=time,name='cashflow')
accumulated_cashflow = cashflow.cumsum().shift(fill_value=0.0)
accumulated_cashflow.name='accumulated_cashflow'
stock = startstock-accumulated_cashflow
stock.name = 'Outstanding'
stock_lag = stock.shift().fillna(0.0)
stock_lag.name = 'outstanding_lagged'
stock_average = stock.rolling(2).mean().fillna(0)
stock_average.name = 'Outstanding average'
assert cashflow.sum() != startstock,'Amortisation does not match outstanding '
discount_series = pd.Series([1./(discount_rate+1)**t for t in time],index=time,name='discount')
lgd_series = pd.Series(_lgd[:len(cashflow)] if type(_lgd) == np.ndarray else _lgd,index = time,name='lgd')
pd_series = pd.Series(_pd[:len(cashflow)] if type(_pd) == np.ndarray else _pd, index=time,name='pd')
pd_series[0] = 0.0
# breakpoint()
pd_survival = 1-pd_series
pd_survival.name = 'pd_survival'
pd_incremental = pd_series*pd_survival.shift().cumprod()
pd_incremental.name = 'pd_incremental'
lgd_decay = stock_average*lgd_series/100.
lgd_decay.name = 'lgd_decay '
lt_loss = pd_incremental* lgd_decay*stock_lag
lt_loss.name='Loss'
lt_loss_discounted = lt_loss*discount_series
lt_loss_discounted.name = 'Discounted Loss'
lt_loss_discounted_sum = lt_loss_discounted.sum()
# breakpoint()
if debug:
debugout = pd.concat([stock,stock_average,pd_series,pd_survival,pd_incremental,
lgd_series,lgd_decay,lt_loss,discount_series,lt_loss_discounted],axis=1)
print(f'Values:\n{debugout}')
return lt_loss_discounted_sum/startstock
if __name__ == '__main__':
pd_fut = np.append([0],np.full(20,0.006831))
lgd_fut = np.full(20,0.2)
xx = lt_ifrs9(maturity=5,discount_rate=0.015,_lgd=lgd_fut,_pd=pd_fut,debug=1)
yy = lt_ifrs9(maturity=5,discount_rate=0.015,_lgd=0.2, _pd=0.006831,debug=1)