Contágio financeiro entre o petróleo e os mercados de ações durante a pandemia da COVID-19: Evidências via cópulas bivariadas estáticas, dinâmicas e Markovianas

Abstract

The present study used static, dynamic and Markovian copula functions to examine the existence of a financial contagion effect between the energy market, represented by oil prices, and stock indices, of a set of emerging economies, represented by Brazil , India, Mexico and China, and a group of developed economies, represented by the United States, United Kingdom, Germany and Japan, during the recent economic crisis of 2020, caused by the COVID-19 pandemic. The ARMA-GJR-GARCH model was used with errors following an asymmetric t-Student distribution function for the marginal distributions and Gaussian, t-Student, Gumbel and Clayton copulas for the joint distribution. Using the quantiles of the standardized residuals of spot prices for Brent oil and the IBOVESPA, NIFTY-500, S&P/BMV-IPC, SSE, S&P-500, FTSE-100, DAX and NIKKEI-225 indices, it was found, through static/dynamic copulas, significant increase in quantile dependence between oil and most of the markets analyzed, after the advent of COVID-19, which constitutes evidence of contagion in the sense of shift contagion. The increase in correlation and the change in the structure of dependence, with a predominance of lower and symmetric tail dependence, characterizes the majority of markets analyzed during the crisis period. This indicates that significant increases in tail dependency represent a real dimension of contagious phenomena. We deepen our analysis, associating the copulas with Markov regime change models, a method capable of accommodating structural breaks in variance and, therefore, does not depend on ad hoc determinations relating to the crisis period. Hereby, we once again highlighted contagious events for the majority of cases analyzed, however, the Asian markets, consisting of China, Japan and India, did not show evidence of contagion, as no significant changes were identified in their respective quantile dependence on oil. Our empirical analyzes have potentially important implications for risk management and the development of appropriate public policies to manage and mitigate the side effects of contagion.