Abstract:Portfolio optimization aims at constructing a realistic portfolio with significant out-of-sample performance, which is typically measured by the out-of-sample Sharpe ratio. However, due to in-sample optimism, it is inappropriate to use the in-sample estimated covariance to evaluate the out-of-sample Sharpe, especially in the high dimensional settings. In this paper, we propose a novel method to estimate the out-of-sample Sharpe ratio using only in-sample data, based on random matrix theory. Furthermore, portfolio managers can use the estimated out-of-sample Sharpe as a criterion to decide the best tuning for constructing their portfolios. Specifically, we consider the classical framework of Markowits mean-variance portfolio optimization with known mean vector and the high dimensional regime of $p/n \to c \in (0,\infty)$, where $p$ is the portfolio dimension and $n$ is the number of samples or time points. We propose to correct the sample covariance by a regularization matrix and provide a consistent estimator of its Sharpe ratio. The new estimator works well under either of three conditions: (1) bounded covariance spectrum, (2) arbitrary number of diverging spikes when $c < 1$, and (3) fixed number of diverging spikes when $c \ge 1$. We can also extend the results to construct global minimum variance portfolio and correct out-of-sample efficient frontier. We demonstrate the effectiveness of our approach through comprehensive simulations and real data experiments. Our results highlight the potential of this methodology as a powerful tool for portfolio optimization in high dimensional settings.

Portfolio Allocation under Asymmetric Dependence in Asset Returns using Local Gaussian Correlations

The Locally Gaussian Partial Correlation

Portfolio Optimization Rules beyond the Mean-Variance Approach

Portfolio Selection Via Graph-Aware Gaussian Processes with Generalized Gaussian Likelihood

New general dependence measures: construction, estimation and application to high-frequency stock returns

Introduction into "Local Correlation Modelling"

Robust Portfolio Optimization with Respect to Spectral Risk Measures Under Correlation Uncertainty

Local Gaussian Autocorrelation and Tests for Serial Independence

Is optimum always optimal? A revisit of the mean‐variance method under nonlinear measures of dependence and non‐normal liquidity constraints

Measuring the Dependence between Non-Gaussian Financial Assets Using Copulae: Risk Management, Option Pricing and Default Risk

Optimising portfolio diversification and dimensionality

General framework for a portfolio theory with non-Gaussian risks and non-linear correlations

Portfolio optimization under multivariate affine generalized hyperbolic distributions

Smart network based portfolios

Cross-sectional Learning of Extremal Dependence among Financial Assets

On allocations to portfolios of assets with statistically dependent potential risk returns

On estimating covariances between many assets with histories of highly variable length

Kendall Correlation Coefficients for Portfolio Optimization

Portfolio Selection Using Multivariate Semiparametric Estimators and a Copula PCA-Based Approach

Risk parity portfolio optimization under heavy‐tailed returns and dynamic correlations

Estimation of Out-of-Sample Sharpe Ratio for High Dimensional Portfolio Optimization