Search results

This paper presents a new database for financial crises in European countries, which serves as an important step towards establishing a common ground for macroprudential oversight and policymaking in the EU. The database focuses on providing precise chronological definitions of crisis periods to support the calibration of models in macroprudential analysis. An important contribution of this work is the identification of financial crises by combining a quantitative approach based on a financial stress index with expert judgement from national and European authorities. Key innovations of this database are (i) the inclusion of qualitative information about events and policy responses, (ii) the introduction of a broad set of non-exclusive categories to classify events, and (iii) a distinction between event and post-event adjustment periods. The paper explains the two-step approach for identifying crises and other key choices in the construction of the dataset. Moreover, stylised facts about the systemic crises in the dataset are presented together with estimations of output losses and fiscal costs associated with these crises. A preliminary assessment of the performance of standard early warning indicators based on the new crises dataset confirms findings in the literature that multivariate models can improve compared to univariate signalling models.

This paper presents a new database for financial crises in European countries, which serves as an important step towards establishing a common ground for macroprudential oversight and policymaking in the EU. The database focuses on providing precise chronological definitions of crisis periods to support the calibration of models in macroprudential analysis. An important contribution of this work is the identification of financial crises by combining a quantitative approach based on a financial stress index with expert judgement from national and European authorities. Key innovations of this database are (i) the inclusion of qualitative information about events and policy responses, (ii) the introduction of a broad set of non-exclusive categories to classify events, and (iii) a distinction between event and post-event adjustment periods. The paper explains the two-step approach for identifying crises and other key choices in the construction of the dataset. Moreover, stylised facts about the systemic crises in the dataset are presented together with estimations of output losses and fiscal costs associated with these crises. A preliminary assessment of the performance of standard early warning indicators based on the new crises dataset confirms findings in the literature that multivariate models can improve compared to univariate signalling models.