Selection of Minimum Variance Portfolio Using Intraday Data: An Empirical Comparison Among Different Realized Measures for BM&FBovespa Data

Abstract

This paper explores different covariance matrix estimators, either the conditional or the unconditional versions, obtained via intradaily data and named realized measures, to the minimum variance portfolio selection problem. Intradaily data are sampled in a synchronized manner as well as in a unsynchronized version. For sake of comparison, we also

use daily data estimators. The major contribution of this work has an empirical nature focused on the Brazilian scenario. We evaluate some out-of-sample performance indexes of the obtained portfolios for a set of 30 stocks traded on the São Paulo stock exchange (BM&FBovespa). The results show that the estimator of the conditional covariance matrix of returns coming from a scalar vt-VECH model based on higher frequency data leads to substantial earnings, reducing the portfolio risk, increasing the average adjustedby risk return and decreasing the turnover