RiskContributionAnalysis

class qf_lib.common.timeseries_analysis.risk_contribution_analysis.RiskContributionAnalysis[source]

Bases: object

Calculates risk contribution metrics.

Methods:

get_distance_to_equal_risk_contrib(...)

By minimising this function it is possible to calculate Equal Risk Contribution Portfolio.

get_risk_contribution(factors_rets, ...)

Calculates risk contribution of different factors to the portfolio.

get_risk_contribution_optimised(assets_rets, ...)

Calculates risk contribution of each asset of the portfolio.

is_equal_risk_contribution(...)

Tells whether each asset has an equal risk contribution to the portfolio.
classmethod get_distance_to_equal_risk_contrib(assets_returns_covariance: QFDataFrame, weights_of_assets: QFSeries) → float[source]

By minimising this function it is possible to calculate Equal Risk Contribution Portfolio. It has better numerical properties than simple approach ( riskContribution - mean(riskContribution) ) Details: http://www.thierry-roncalli.com/download/erc-slides.pdf

Parameters:

  • assets_returns_covariance – covariance matrix for assets returns data frame

  • weights_of_assets – weight of each asset in the portfolio. It’s indexed with names of assets.

Return type:

L2 (euclidean) distance from equal risk distribution

classmethod get_risk_contribution(factors_rets: SimpleReturnsDataFrame, weigths_of_assets: QFSeries, portfolio_rets: SimpleReturnsSeries) → QFSeries[source]

Calculates risk contribution of different factors to the portfolio. Risk is defined as volatility. Uses x-sigma-rho formula (MSCI Bara paper).

Parameters:

  • factors_rets – dataframe consisted of returns for different assets

  • weigths_of_assets – series of weights of each asset. It’s indexed with names of assets.

  • portfolio_rets – return of the whole portfolio

Returns:

Series of risk contributions (one for each asset) to the portfolio. It’s indexed with names of assets.

Return type:

pandas.Series

classmethod get_risk_contribution_optimised(assets_rets: SimpleReturnsDataFrame, weights_of_assets: QFSeries) → QFSeries[source]

Calculates risk contribution of each asset of the portfolio.

Parameters:

  • assets_rets – returns of assets building the portfolio (each assets in a separate column)

  • weights_of_assets – Series of weights (one for each asset). It’s indexed with names of assets.

Return type:

Series of risk contributions (one for each asset). It’s indexed with names of assets

classmethod is_equal_risk_contribution(returns_covariance: QFDataFrame, weights_of_assets: QFSeries) → bool[source]

Tells whether each asset has an equal risk contribution to the portfolio.

Parameters:

  • returns_covariance – DataFrame which is a covariance matrix. Columns and rows are both indexed with names of assets.

  • weights_of_assets – Series of weights (one weight for each asset). It’s indexed with names of assets.

Return type:

True if each asset has and equal risk contribution. False – otherwise.