Modeling and Forecasting of Realized Volatility: Evidence from Brazil

Abstract

Using intraday data for the most actively traded stocks of BOVESPA, this work has considered two recently developed models in the literature of the estimation and forecasting of realized volatility; The Heterogeneous Autorregressive Model of Realized Volatility (HAR-RV), developed by Corsi (2009) and the Mixed Data Sampling (MIDAS-RV), developed by Ghysels et al. (2004). Through statistical comparison of forecasts in-sample and out-of-sample, it was found that superior results of the MIDAS-RV modeloccurred only for the in-sample forecasting. However, for out-of-sample forecasts no statistically different results were found between the models. Also, there are evidences that the use of realized volatility inducesnormality in standardized returns.