Computation of Portfolio VaRs with GARCH-Type Volatility
Posted: 2 Apr 2013
Date Written: March 1, 2013
Abstract
In this paper, we explore the use of Independent Component Analysis (ICA) from the field of signal processing to model and estimate the dynamics of multivariate volatilities of financial asset returns in the GARCH framework. The resulting ICA-GARCH approach is shown to provide a computationally tractable method for constructing Value at Risk (VaR) of portfolios consisting of a large number of assets that are typically characterized by nonlinearity and nonnormality. In addition, it is also shown to be effective in capturing the time-varying features of volatilities and is more stable than other comparable models.
Keywords: Multivariate Models, Independent Component Analysis, Principal Component Analysis, GARCH, Value at Risk
JEL Classification: C22, C53, G17
Suggested Citation: Suggested Citation