Class RiskParityOptimizer

java.lang.Object

com.quantfinlib.optimization.RiskParityOptimizer

public final class RiskParityOptimizer
extends Object

Risk parity: the portfolio where every asset contributes a set share of total risk (w_i (Σw)_i / w'Σw = b_i; equal shares by default).

Solved through the convex formulation of Spinu (2013, "An algorithm for computing risk parity weights", SSRN 2297383): min_x ½x'Σx − Σ b_i ln x_i over x > 0, whose unique minimiser satisfies x_i (Σx)_i = b_i; then w = x / 1'x. Damped Newton steps with a backtracking line search that keeps x positive converge quadratically, typically in under 20 iterations, including on nearly collinear assets where the multiplicative fixed-point update oscillates. Long-only, deterministic.

Method Summary

Modifier and Type	Method	Description
static PortfolioOptimizer.Allocation	equalRiskContribution(double[] expectedReturns, double[][] covariance)	Equal-risk-contribution weights (expected returns used only for reporting).
static PortfolioOptimizer.Allocation	riskBudget(double[] expectedReturns, double[][] covariance, double[] budgets)	Risk-budgeting weights: asset i carries the fraction budgets[i] of portfolio variance.
static double[]	riskContributions(double[] w, double[][] covariance)	Each asset's fractional contribution to portfolio variance under weights w.

Methods inherited from class Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Method Details

equalRiskContribution

public static PortfolioOptimizer.Allocation equalRiskContribution(double[] expectedReturns,
 double[][] covariance)

Equal-risk-contribution weights (expected returns used only for reporting).

riskBudget

public static PortfolioOptimizer.Allocation riskBudget(double[] expectedReturns,
 double[][] covariance,
 double[] budgets)

Risk-budgeting weights: asset i carries the fraction budgets[i] of portfolio variance.

Parameters:

budgets - positive, summing to 1

riskContributions

public static double[] riskContributions(double[] w,
 double[][] covariance)

Each asset's fractional contribution to portfolio variance under weights w.