A Bayesian panel stochastic volatility measure of financial stability

We propose to model financial stability, opting for an alternative bank profit function whose volatility is measured within a framework of panel stochastic volatility. Within this model financial stability and volatility are latent variables. To observe financial stability and volatility we employ Bayesian inference procedures organized around Sequential Monte Carlo (SMC) technique and particle filtering. We do so in a single stage that controls also for non‐linearities, while we also allow for some key bank and country‐specific variables to impact upon financial stability and volatility. Thus, we provide a new measure of financial stability by country, over time and also at a global level. In an empirical application, we derive financial stability indexes for a plethora of countries, as well as the global financial stability index that acts as an early warning index. Our results suggest that the financial cycle is subject to non‐linearities. We argue that the global financial system should closely monitor large, systemic, banks as key to support financial stability.