Expected Duration as a Leading Index for Systemic Risk

This paper empirically analyzes the determinants of systemic risk using dynamic panel data regressions, because they allow controlling for unobserved heterogeneity and omitted variables, decreasing the bias in the coefficient estimates. Additionally, it analyzes the recurrence of high systemic risk events using a duration models approach, in which the time spent in a state of financial stability is probabilistically characterized, as well as the transition probability to an unstable state, in which systemic risk is high. Moreover, it suggests that the expected duration of financial stability can be used as a leading index for systemic risk