A1293
Title: Modelling endogenous regime-switching: A copula-based latent factor approach
Authors: Ruijun Bu - University of Liverpool (United Kingdom)
Jie Cheng - Keele University (United Kingdom)
Yuyi Li - University of Liverpool (United Kingdom) [presenting]
Abstract: A copula-based latent factor-driven endogenous regime-switching model is proposed with a general dependence structure between the innovation of the regime-dependent process and that of the latent factor assumed to be driving the switching of the regimes. In the spirit of copula modelling, we first specify the marginal distribution of regime-dependent innovation and that of the latent factor independently, and then model their dependence either by a copula function or by directly specifying their conditional behaviour. Consequently, our model can accommodate potentially non-Gaussian regime-dependent dynamics and nonlinear endogenous dependence, which are typically observed in empirical data. Similar to existing models in the literature, the parameters of the proposed model can be estimated by maximum likelihood using a modified Markov switching filter and the unobservable latent factor can be extracted using a smoothing technique. The flexibility and usefulness of our model are illustrated by numerical examples.
