Solving the high-dimensional Markowitz Optimization Problem: a tale of sparse solutions

We consider the high-dimensional Markowitz optimization problem. A new approach combining sparse regression and estimation of optimal returns based on random matrix theory is proposed to solve the problem. We prove that under some sparsity assumptions on the underlying optimal portfolio, our novel approach asymptotically yields the theoretical optimal return, and in the meanwhile satisfies the risk constraint. To the best of our knowledge, this is the first method that can achieve these two goals simultaneously in the high-dimensional setting. We further conduct simulation and empirical studies to compare our method with some benchmark methods, including the equally weighted portfolio, the bootstrap-corrected method by Bai et al. (2009) and the covariance-shrinkage method by Ledoit and Wolf (2004). The results demonstrate substantial advantage of our method, which attains high level of returns while keeping the risk well controlled by the given constraint.

Based on joint work with Mengmeng Ao and Yingying Li.