Clustering asset markets based on volatility connectedness to political news

Abstract

To assess similarities in international asset markets’ responses to political news, we construct a political news index using advanced natural language processing. We then examine how the volatility across international asset markets is connected to the development of our political news index by measuring the daily directional connectedness using a VAR-based framework. Finally, we apply an unsupervised algorithm to cluster markets based on their volatility connectedness to political news. Our analysis reveals eight distinct clusters that reflect the markets’ sensitivities to political dynamics. This data-driven analysis offers insights into the influence of political developments on market volatility.