A0935
Title: A multiple chains hidden Markov model for a sector index and its comovement with the market
Authors: Markus Haas - University of Kiel (Germany) [presenting]
Abstract: Bivariate multiple chains hidden Markov model is considered and applied to characterize the regime-switching dynamics of the US airline industry and its comovement with the S\&P 500 during the Covid-19 pandemic. The model allows for independent regime switches of the sector-specific and the market factors, with the prevailing states of both chains determining the parameters of the conditional sector return distribution. Shocks are drawn from a (possibly regime-specific) Student's $t$ distribution to capture the fat-tailed nature of the returns and to avoid the otherwise distorting effect of outliers on the identification of the regime processes. An EM-type algorithm is proposed to estimate the model parameters.
