WSEAS Transactions on Business and Economics
Print ISSN: 1109-9526, E-ISSN: 2224-2899
Volume 11, 2014
Volatility Forecast Combination of Brazilian Selic Interest Rate and Exchange Rate by Means of Principal Component Analysis
Author(s): Lizandra Salau Da Rocha, Adriano Mendonça Souza, Roselaine Ruviaro Zaninni, Meire Mezzomo
Abstract: Using volatility models for macroeconomic variables can provide more efficient results than models which estimate the average of the process. In this context, the purpose of this research was to evaluate the efficiency of individual models and combination models in forecasting the Brazilian SELIC interest rate and exchange rate between January, 1974 and June, 2012 and from January, 1980 to May, 2012, respectively. The analysis of the series confirmed the presence of volatility in those periods, where Brazil’s economic scenario, marked by both internal and external crises, was decisive for such variance. For this purpose, joint modeling was used for the average (ARIMA) and variance (ARCH, GARCH, EGARCH, TARCH) of the process. The results showed that, in general, the performance measures considered (MAPE, MSE, and U-THEIL) are better for forecast combinations. In addition, forecast combinations by PCA using different kinds of weighting were not conclusive for the kinds of weighting used. This shows that when forecast combination by the PCA method is performed, the best alternative is to use more than one type of PCA in order to obtain the best results.
Keywords: Time Series, Volatility Models, Selic Interest Rate, Exchange Rate, Forecast Combination, Pca Method
Pages: 514-524WSEAS Transactions on Business and Economics, ISSN / E-ISSN: 1109-9526 / 2224-2899, Volume 11, 2014, Art. #48