Cointegration Analysis: Impact on Stochastic Properties of Ecuindex

Abstract

This text investigates the temporal relationship between the Ecuindex (BVQ) and representative stock indices of Latin America and the United States, during the period 2001-2020. A correlation analysis was performed in order to determine the association between series. To test the stationarity hypothesis, the Augmented Dickey Fuller model was applied and corroborated with the Phillips-Perron test. Finally, the presence of cointegration was verified, using the Engler-Granger, Phillips-Outliers and Johansen tests. The results confirmed the empirical assertion that the correlation increases during bear market periods. The unit root evaluation showed that the BVQ series is stationary in intercept levels. Trace and eigenvalue analysis revealed the presence of two significant cointegration vectors over time, between the BVQ, BOVESPA and NYSE stock indices. Based on these results, it is concluded that the North American stock market, represented by the NYSE, would have its long-term impact on the Ecuindex, inferring that the effects of volatility and contagion affect the diversification of stock risk in country markets. emerging dollar acceptors in the region.