Bayesian Nonparametric Panel Markov-Switching GARCH Models

This paper introduces a new model for panel data with Markov-switching GARCH effects. The incorporates series-specific hidden Markov chain process that drives the parameters. To cope high-dimensionality of parameter space, exploits cross-sectional clustering series by first assuming soft pooling through hierarchical prior distribution two-step procedure, and then introducing effects in space nonparametric distribution. proposed inference are evaluated simulation experiment. results suggest is able to recover true value parameters number groups each regime. An empirical application 78 assets SP\&100 index from $6^{th}$ January 2000 $3^{rd}$ October 2020 also carried out using two-regime switching model. findings shows presence 2 3 clusters among constituents second regime, respectively.