Title: High-dimensional dynamic covariance modeling via risk factors mapping
Authors: Mike So - The Hong Kong University of Science and Technology (Hong Kong) [presenting]
Abstract: The aim is to explore a modified method of dynamic covariance estimation via risk factors mapping. One important feature of the method is to be able to handle dependence estimation of assets of a large portfolio with high computational efficiency. The main idea is to apply a multivariate generalized autoregressive conditional heteroscedasticity (MGARCH) model to a small number of risk factors, which explain the movement of portfolio returns. The idea of risk mapping is demonstrated by an empirical study with a focus on the Hong Kong stock market. Assessment using portfolio risk calculation is also discussed.