A0448
Title: Network GARCH model
Authors: Dong Li - Tsinghua University (China) [presenting]
Abstract: The multivariate GARCH (MGARCH) models are popularly used for analyzing financial time series data. However, statistical inference for MGARCH models is quite challenging due to high dimensionality. To circumvent this difficulty, we propose here a network GARCH model, which can substantially reduce computational complexity. The proposed model makes use of information derived from an appropriately defined network structure. By doing so, the number of unknown parameters in the new model is much reduced. Strict and weak stationarity of the network GARCH model is rigorously established.In order to estimate the model, a quasi-maximum likelihood estimator (QMLE) is developed, and its asymptotics are investigated. Simulation studies are carried out to assess the performance of the QMLEin finite samples and empirical examples are analyzed to illustrate the usefulness of network GARCH models.
