| Summary: | The main objective of this dissertation is to investigate the asymmetric effects of shocks on volatility during the Global Financial Crisis of 2007 – 2009. Using daily logarithmic returns, we estimate univariate EGARCH and GJR models assuming three different conditional distributions: the Gaussian normal, Student’s t and Generalized Error Distribution. The stock indices under analysis, which include largest companies in the world, are S&P 500, NASDAQ, FTSE 100, DAX, CAC 40, NIKKEI 225 and HSI. The data ranges from September 15, 2006 to September 15, 2010, being split in two subsamples by the collapse of Lehman Brothers on September 15, 2008. Our results suggest that asymmetric effects are present in all stock markets analysed. In most cases, the impact becomes weaker after the Lehman Brothers bankruptcy, indicating that the negative shocks did not raise volatility as much as they did before the bankruptcy. EGARCH model with fatter tailed distributions appears to be the best in-sample predictive model. Moreover, we test the statistical significance of the change between asymmetry coefficient estimates of the EGARCH model, and conclude that the majority are not statistically significant, suggesting that the asymmetry coefficients do not depend on the sample period.
|
|---|