Title: An Asymmetric Block Dynamic ... model Author: Gregorio A. Vargas III

advertisement
Title: An Asymmetric Block Dynamic Conditional Correlation Multivariate GARCH
model
Author: Gregorio A. Vargas III
Degree: Master of Science in Statistics
Date: April 2006
Abstract:
A new dynamic conditional correlation model is proposed in this paper. The Block DCC
model for determining dynamic correlations between groups of financial assets is
extended to account for the asymmetric effect between groups. Simulation results show
that the maximum likelihood estimator of the model is biased but consistent. Empirical
results show that the model is able to capture the behavior of some blocks of currencies in
Asia in the turbulent years of the late 1990s. The volatilities of the peso-baht and ringgitSG dollar blocks are positively conditionally correlated. This dynamic conditional
correlation rose sharply during the 1997-98 Asian Financial Crisis and is highly persistent
but not asymmetric. On the other hand, the conditional correlation of the volatility of HK
dollar-yuan block saw periods of high positive and negative dynamic conditional
correlations with the peso-baht and ringgit-SG dollar blocks. Furthermore, the dynamic
conditional correlation of the volatility of this block is highly asymmetric with the pesobaht and highly persistent with the ringgit-SG dollar especially during the crisis years.
Keywords:
Asymmetric block effect, block dynamic conditional correlation, multivariate GARCH
Download