Source code for qf_lib.portfolio_construction.covariance_estimation.robust_covariance

#     Copyright 2016-present CERN – European Organization for Nuclear Research
#
#     Licensed under the Apache License, Version 2.0 (the "License");
#     you may not use this file except in compliance with the License.
#     You may obtain a copy of the License at
#
#         http://www.apache.org/licenses/LICENSE-2.0
#
#     Unless required by applicable law or agreed to in writing, software
#     distributed under the License is distributed on an "AS IS" BASIS,
#     WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#     See the License for the specific language governing permissions and
#     limitations under the License.

import numpy as np
from arch.univariate.volatility import VolatilityProcess

from qf_lib.common.utils.volatility.volatility_forecast import VolatilityForecast
from qf_lib.containers.dataframe.qf_dataframe import QFDataFrame


[docs]class RobustCovariance: """ Creates a class used for vol forecasting and covariance matrix estimation NOTE: this method has a tendency to make decrease the volatility Parameters ---------- simple_returns: QFDataFrame QFDataFrame of simple returns of the assets """ def __init__(self, simple_returns: QFDataFrame): self.returns = simple_returns
[docs] def calculate_covariance(self, vol_process: VolatilityProcess, horizon: int, method: str = 'analytic') -> QFDataFrame: """ Calculates the covariance matrix using foretasted volatility for each asset and rank correlation between the assets. Covariance matrix contains NOT annualised values. They are expressed in the frequency of the input returns Parameters ----------- vol_process: VolatilityProcess volatility proces used for forecats calculation. For example EGARCH(p=p, o=o, q=q) horizon: int horizon for the volatility forecast. It is expressed in the frequency of the returns provided method (optional): str method of the volatility forecast calculation. Possible: 'analytic', 'simulation' or 'bootstrap' 'analytic' is the fastest but is not available for EGARCH and possibly some other models. For details check arch documentation Returns ---------- QFDataFrame cov_matrix of the assets. """ vol_forecast_array = self._calculate_expected_volatilities(vol_process, horizon, method) # use spearman rank correlation instead of simple pearson correlation corr_matrix = self.returns.corr(method='spearman') cov_matrix = corr_matrix.copy() for i, _ in enumerate(vol_forecast_array): for j, _ in enumerate(vol_forecast_array): cov_matrix.iloc[i, j] = corr_matrix.iloc[i, j] * vol_forecast_array[i] * vol_forecast_array[j] return cov_matrix
def _calculate_expected_volatilities(self, vol_process, horizon, method): vol_forecast_array = np.zeros(self.returns.shape[1]) # size = nr o assets for i, asset in enumerate(self.returns): asset_returns = self.returns[asset] vol_forecast = VolatilityForecast(asset_returns, vol_process, method, horizon, annualise=False) vol_forecast_array[i] = vol_forecast.calculate_single_forecast() return vol_forecast_array