Sample and realized minimum variance portfolios: Estimation, statistical inference, and tests

Abstract The global minimum variance portfolio (GMVP) is the starting point of Markowitz mean‐variance efficient frontier. estimation GMVP weights therefore much importance for financial investors. depend only on inverse covariance matrix returns risky assets, this reason estimated are subject to substantial risk, especially in high‐dimensional settings. In paper we review recent literature traditional sample estimators unconditional which typically based daily asset returns, as well modern realized conditional intraday high‐frequency returns. We present various types with corresponding stochastic results, discuss statistical tests and some directions further research. Our empirical application illustrates selected properties weights. This article categorized under: Statistical Graphical Methods Data Analysis > Multivariate High Dimensional Models