Modelando a volatilidade dos retornos de Petrobrás usando dados de alta frequência

The aim of this paper is to assess the empirical characteristics of a high-frequency return series of one of the main assets traded at the São Paulo Stock Exchange. We are interested in modeling the conditional volatility of these return series, particularly testing for long-memory. Our findings reveal that besides long memory, there is strong intraday periodicity, but we found no evidence of leverage effect. We use models that are able to account for the long memory in the conditional variance of the seasonally adjusted returns, yielding superior results when compared to traditional short-memory volatility models, with important implications to option pricing and risk management