Estimation of covariance matrix with equal shrinkage weights in portfolio selection: An empirical study on the Vietnamese stock market

Abstract

Portfolio optimization based on the process of estimating the covariance matrix is one of the
new approaches. In this article, the author has conducted an analysis and evaluation of various covariance
matrix estimators to measure their effectiveness in investment activities. The results of the empirical study
on over 400 listed stocks on the Vietnamese stock market from January 2011 to January 2024 have shown
that the covariance matrix estimator with balanced shrinkage weight yields significantly superior results
across all three portfolio performance metrics: Sharpe ratio, Sortino ratio, and Calmar ratio, compared to the
other seven estimators divided into three groups including the LW Shrinkage method group, the factor model
group, and the traditional method group. The study’s findings also aim to further encourage investors to
expand research on covariance matrix estimation in selecting optimal investment portfolios.