February 18, 2005: Junyoep Park, CIMS
Dynamic Risk Factor Model for Correlation Matrices:
From Random Matrix Theory to Econometrics
Modelling the dynamics of correlations of a large market has been a very
challenging problem in finance. This is due to the
enormous noise of individual assets and the dimensionality of the
system. With the help of limit theorems in Random Matrix Theory (RMT),
the Generalized Autoregressive Conditional
Heteroscedasticity (GARCH) model, and Principal Component
Analysis (PCA), the Dynamic Risk Factor (DRF) model for the
dynamics of the cross correlation of a large system is proposed.
