Riassunto analitico
The purpose of this thesis is to examine the relationship between the phases of financial cycle and the occurrence banking crises from Italy’s unification to 2013, in order to verify if financial cycle is a good predictor of the probability of crises which have substantial macroeconomic effects. To this aim, I estimate Logit and Probit regressions to examine the link between financial cycle and the probability of banking crises. The financial cycle is extracted by fitting the Hodrick-Prescott filter (1980, 1997) and Baxter-King filter (1995) to total credit in real terms and credit-to-GDP ratio. I find a positive association between credit cycle and banking crises: an increase of the credit cycle by 1 unit in time t-1, increases the probability of bankruptcy at time t. This probability ranges from 3% to 8% depending on the financial variable and filtering procedure used to extract the cycle. At last, I investigate whether the credit cycle components have significant predictive power for business cycle by running a Granger causality analysis. In the pre-WWII period, I find that financial cycle unidirectionality predicts business cycle, instead in the post-WWII period I find causality in both directions between financial and business fluctuations. These suggest that during the first period, episodes of financial distress help to predict economic recessions, while in the second period, the bidirectionality weakens this result.
|
