Robust Risk Parity Under Covariance Estimation Uncertainty

Summary

The article presents a robust version of equal risk contribution (risk parity) portfolio construction. It treats the estimated covariance matrix as uncertain and builds a bounded uncertainty set around the estimate. The optimization then accounts for uncertainty in both total portfolio risk and each asset’s marginal risk contribution, using a second order cone formulation. A factor model is offered as one way to estimate covariance uncertainty from factor loading errors.

The authors compare the robust portfolios with nominal risk parity and a worst case variance approach in rolling out of sample tests on US equities. Across the reported experiments, robust portfolios generally had higher excess returns and Sharpe ratios, including during market downturns, while retaining some diversification. The tradeoffs were higher turnover and departures from perfectly equal risk contributions; the most conservative setting did not perform best at every portfolio size. The evidence depends on historical data, a particular factor model, and a universe with survivorship bias, so it does not establish that the method will perform similarly in other markets or periods.

Key ideas

• Risk parity allocates capital to equalize each asset’s contribution to portfolio risk without estimating expected returns.
• The proposed method models covariance estimates as a set of possible values and incorporates that uncertainty into risk contribution constraints.
• A factor model’s loading standard errors can help set the size of covariance perturbations.
• Historical tests report stronger risk adjusted results for robust portfolios, alongside higher turnover and less exact risk balance.
• The robustness parameter controls conservatism, and its best setting can vary with portfolio size.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.