Neural Nonlinear Shrinkage of Covariance Matrices for Minimum Variance Portfolio Optimization

TL;DR

This paper presents a neural network-based nonlinear shrinkage method for covariance matrix estimation, improving minimum variance portfolio optimization by reducing out-of-sample risk through a hybrid statistical and machine learning approach.

Contribution

It introduces a neural eigenvalue shrinkage function trained with portfolio risk as the loss, enhancing covariance estimation for portfolio optimization.

Findings

Achieves lower out-of-sample realized risk compared to benchmarks.

Scalable across different sample sizes and asset universes.

Demonstrates effectiveness on S&P 500 stock data.

Abstract

This paper introduces a neural network-based nonlinear shrinkage estimator of covariance matrices for the purpose of minimum variance portfolio optimization. It is a hybrid approach that integrates statistical estimation with machine learning. Starting from the Ledoit-Wolf (LW) shrinkage estimator, we decompose the LW covariance matrix into its eigenvalues and eigenvectors, and apply a lightweight transformer-based neural network to learn a nonlinear eigenvalue shrinkage function. Trained with portfolio risk as the loss function, the resulting precision matrix (the inverse covariance matrix) estimator directly targets portfolio risk minimization. By conditioning on the sample-to-dimension ratio, the approach remains scalable across different sample sizes and asset universes. Empirical results on stock daily returns from Standard & Poor's 500 Index (S&P500) demonstrate that the proposed method consistently achieves lower out-of-sample realized risk than benchmark approaches. This highlights the promise of integrating structural statistical models with data-driven learning.