CFE 2015: Start Registration
View Submission - CFE
A1159
Title: Time-varying transition probabilities for Markov switching copula models Authors:  Anna Czapkiewicz - AGH University of Science and Technology (Poland) [presenting]
Pawel Jamer - Polish Academy of Science (Poland)
Joanna Landmesser - SGGW Warsaw (Poland)
Abstract: Markov switching Copula-Garch model with time varying probabilities for the transitions is proposed to describe the relationships between markets from G6 group. Different copula functions are taken into consideration in this model. It is assumed that the probability of switching from one regime to the other depend on the behavior of some economic indicators such as an unemployment rate, CPI, long-term interest rate, industrial production. The logistic function for the transition probabilities is used to maps the transition variables into the unit interval. The selection of the transition variables is done in two steps. Firstly, the null hypothesis of a standard Markov switching model against a time-varying transition probability Markovian model is tested. Secondly, it is verified that the influence of macroeconomic data is statistically significant. The EM algorithm adopted to estimate unknown parameters of discussed model is presented.