Correlation between BM&FBOVESPA stock returns: an analysis via dynamic copula

Abstract

The behavior of the dependence among returns on financial assets is important for the understanding of finance issues. In this context, the dynamic copulas theory is an important tool in the multivariate analysis of financial series, thanks to its flexibility to construct multivariate distribution functions that reproduce several types of dependencies and measure the movement of this relationship over time. In this way, this work aims to investigate and measure the structure and the movement of the correlation between pairs of stock returns referring to the companies Ambev, Ita ́u Unibanco, Petrobras listed on the São Paulo Stock Exchange (BM&FBOVESPA) at the period from January 3, 2011, to June 21, 2017. For this, he used the conditional copula Normal with time-variant parameter specified by Patton (2006) that quantifies and captures the temporal trajectory of the linear correlation coefficient, which configures a relevant measure of dependency. The results of the magnitude of the dependence and the trajectory over time measured between these pairs of stocks reflect the peculiarities of the sectors of performance of each company, as well as the influence of market uncertainty, which demonstrates the importance of the diversification of assets in investment analysis.