Research
Estimating the Volatility of Brazilian Equities using Garch-Type Models and High-Frequency Volatility Measures
Financial markets require an accurate estimate of asset volatility for various purposes such as risk management, decision-making and portfolio selection. Moreover, for risk management, volatility estimation is critical in Value-at-Risk (VaR) calculation models. However, there is still no consensus on a model that performs best in estimating volatility. This study proposes comparing volatility measures based on high-frequency data, such as RV and RRV, with heteroskedastic volatility models that use squared daily returns and daily closing prices. Four GARCH type models were implemented to estimate heteroskedastic volatility for the two most actively traded shares on the Brazilian stock exchange, using skewed generalized t (SGT) distribution and allowing flexibility for modeling the empirical distribution of these asymmetric financial data. Performed tests indicated no differential between the GARCH models and the high-frequency volatility measures used to estimate the VaR, indicating that both measures could be utilized for risk management purposes.
Future Volatility Forecasting Models: An Analysis of the Brazilian Stock Market
Future volatility forecasting intrigues many scholars, researchers, and people from the financial markets. The model and methodology used for forecasting are fundamental for asset pricing in general, since future volatility deeply influences the final result. Thus, this study uses databases from the companies Vale and Petrobrás, in the period from July 1994 to August 2013, to test the Univariate, Bivariate, GARCH, and EGARCH models (also analyzing the results for the linear and quadratic methods) in order to assess the best model for forecasting future volatility. The results indicate that the quadratic method can better forecast future volatility than the linear method. The Univariate model showed the best results, proving that it is more efficient to use only short-term volatility for future volatility forecasting. If it were necessary to include long-term volatility, the Bivariate model would be the best, despite the GARCH and EGARCH models showing similar results.
