Dynamic Relationships between Markov-switching Processes

Abstract

Regime shifts in different economic time series may be interconnected in various ways, some of which are investigated within a Markov-switching modeling framework. We derive restrictions implied by the lead-lag relationship and the conditional independence, estimate the restricted models, and test these dynamic relationships by using the standard likelihood ratio test. Our empirical examples suggest that the stock volatility regime leads the business cycle regime and that the evidence for a cross-border volatility spillover effect is sensitive to model assumptions.