The banking sector has been on the spotlight in both academic and policy circles since the outburst of the subprime bubble. The crisis has its roots in the US, but there were spillover effects around the world. We study the behavior of the banking sector of 40 countries during the period 2007–2010, using a new clustering methodology. Our methodology combines regime switching models in the modeling of longitudinal variations with cluster analysis that identifies groups of countries with similar profiles. Our results show that although there were periods of intense contagion, the impact was uneven among sample countries. The crisis had episodic effects on some countries, while others had severe devaluations after the Lehman Brothers bankruptcy. Finally, a small group of banking systems has plunged into a long severe crisis.

--

Banking sector; Clustering methods; Time series data; Regime switching models; Hidden Markov model

• Economia e Gestão - Ciências Sociais