CMStatistics 2016: Start Registration
View Submission - CFE
A1583
Title: CISS in a time-varying environment: How frequent are systemic distress? Authors:  Eddie Gerba - Bank of Spain and London School of Economics (LSE) (Spain) [presenting]
Manfred Kremer - European Central Bank (Germany)
Abstract: The CISS indicator is a composite indicator of systemic stress in an economy, which is highly time-varying. We incorporate this index in multiple macro-financial time-varying parameter VAR models of the Euro area to evaluate its performance and measure its predictability of systemic events. Only during times of financial distress, the indicator increases. Moreover during systemic events such as the post-2007 crises both the indicator as well as the uncertainty regarding the future path of key macro-financial variables increases in our model. This increase in the indicator precedes the increase in volatility of other variables in the model by 1-2 quarters, making it a good forecaster of crises events. Furthermore, systemic shocks (or shocks to the CISS indicator) have a stronger impact on the rest of the economy than any of the other shocks examined. Lastly we compare our model to a Markov-switching framework (ceteris paribus) and find that while the number of regime switches are fewer, increases in the mean and volatility of coefficients associated with each crisis event is higher in our current model.