Volatility of intraday financial data: Multiscale Ibovespa behavior under to the COVID-19 pandemic

Abstract

In financial markets, volatility modeling has been a strategy widely used because it reflects uncertainties about changes in asset prices. Incorporating peculiarities of financial series, this study estimated the volatility for the intraday index of the Brazilian stock market (Ibovespa) using ARIMA-APARCH models in different time frequencies with the aid of the wavelet MODWT decomposition technique. This work proposes an analysis of the impacts of the frequency components on the behavior of the volatility of intraday returns using the series of details wavelet in different time horizons, in an atypical period in the global financial markets, generated by the COVID-19 pandemic. The empirical results suggest low unconditional volatility and strong signs of persistence in all analyzed frequencies. The asymmetry in volatility is evidenced in the higher frequencies, the leverage effect being present only in the series of details with variations of 15-120 min., which is corroborated with the results obtained with the reconstructed series. The evidenced behaviors have an impact on the elaboration of short-term investment strategies and risk management, since the positive and negative shocks, such as those given by the world pandemic of COVID-19, have different impacts on the volatility of returns in shorter periods. The information obtained can contribute to the analysis of future atypical events in the Brazilian stock market, supporting the decision-making of economic agents.