A0812
Title: International stock co-movements and time-varying risks
Authors: Michael Owyang - Federal Reserve Bank of St Louis (United States) [presenting]
Laura Coroneo - University of York (United Kingdom)
Laura Jackson Young - Bentley University (United States)
Abstract: International stock return comovements of country-industry portfolios are examined. The model allowed comovements to be driven by a global and a cluster factor, with the cluster membership endogenously determined. It is found that country-industry portfolios tend to cluster mainly within geographical areas. The cluster compositions substantially changed over time, with the emergence of clusters among European countries from the early 2000s. The cluster component was the main driver of country-industry portfolio returns for most of the sample, except from the mid-2000s to the mid-2010s, when the global component had a more prominent role. To this model, asymmetric risk is accounted for by adding time-varying skewness to the factors. Because the risk of interest is assumed to be relatively persistent, the skewness is modeled as Markov-switching. It is found that using a standard normal distribution always results in larger volatilities, no matter how the clusters are specified; using the skew-normal distribution results in lower volatilities, with a marginal improvement in the definitions of the endogenous clusters. Changes in volatility and factor loadings for all the specifications are also accounted for.
