A Fast Algorithm for Computing High-dimensional Risk Parity Portfolios

TL;DR

This paper introduces a rapid cyclical coordinate descent algorithm for high-dimensional risk parity portfolio optimization, demonstrating superior speed and convergence even with large covariance matrices.

Contribution

The paper presents a novel CCD algorithm specifically designed for high-dimensional risk parity problems, improving computational efficiency over existing methods.

Findings

Algorithm converges reliably for large covariance matrices

Outperforms existing algorithms in speed and efficiency

Effective for portfolios with over 500 assets

Abstract

In this paper we propose a cyclical coordinate descent (CCD) algorithm for solving high dimensional risk parity problems. We show that this algorithm converges and is very fast even with large covariance matrices (n > 500). Comparison with existing algorithms also shows that it is one of the most efficient algorithms.