No Risk
PortfolioOptimisers.NoRisk — Type
struct NoRisk{__T_settings} <: RiskMeasureMeasures no risk: its value is zero for every portfolio.
The JuMP formulation adds a zero risk expression and no row, so a linear programme stays a linear programme. MeanRisk needs a risk measure, and without NoRisk an objective that ignores the risk still builds the default Variance term. That term adds second-order cone rows, so the model needs a conic solver. Problems 9, 14 and 21 of [8] are linear programmes of this kind. They give the robust $1/N$ and quintile portfolios, the dollar-neutral quintile portfolio and the inverse-volatility quintile portfolio. The global maximum return portfolio of Table I of [8] is another one.
NoRisk is valid only under an objective that does not read the risk.
MeanRiskaccepts it underMaximumReturnand underMaximumUtility. UnderMaximumUtilitythe risk term is zero, so the two objectives give the same weights.MeanRiskrefuses it underMinimumRiskand underMaximumRatio, throughassert_no_risk_objective_compatibility.RiskBudgeting,FactorRiskContribution,NearOptimalCentering,HierarchicalRiskParityandHierarchicalEqualRiskContributionrefuse it, throughassert_risk_measure_required.
A refusal is an ArgumentError from the constructor of the optimiser, and zero_risk_expression_cause names the cause in its message. A bound settings.ub has no effect, because RiskMeasureSettings refuses a negative bound and the risk is always zero.
Mathematical definition
\[\begin{align} \mathrm{NoRisk}(\boldsymbol{x}) &= 0\,. \end{align}\]
Where:
- $\mathrm{NoRisk}(\boldsymbol{x})$: Risk of the portfolio, which is zero.
- $\boldsymbol{x}$: Portfolio returns vector $T \times 1$.
Fields
settings: Risk measure settings.
Constructors
NoRisk(; settings::RiskMeasureSettings = RiskMeasureSettings()) -> NoRiskKeywords correspond to the struct's fields.
Functor
(r::NoRisk)(x::VecNum)Returns zero, with the element type of the portfolio returns vector x.
Arguments
x::VecNum: Portfolio returns vector.
Examples
julia> r = NoRisk()NoRisk settings ┼ RiskMeasureSettings │ scale ┼ Float64: 1.0 │ ub ┼ nothing │ rke ┴ Bool: truejulia> r([0.1, -0.2, 0.05])0.0julia> r(Float32[0.1, -0.2])0.0f0Related
RiskMeasureRiskMeasureSettingsMaximumReturnCharacteristicUncertaintySet: The uncertainty set of the robust portfolios of [8].norisk_flagzero_risk_expression_flag
References
- [8] R. Zhou and D. P. Palomar. Understanding the Quintile Portfolio. IEEE Transactions on Signal Processing 68, 4030–4040 (2020). Problems 9, 14 and 21, and Table I.
References
- [8]
- R. Zhou and D. P. Palomar. Understanding the Quintile Portfolio. IEEE Transactions on Signal Processing 68, 4030–4040 (2020).