Selecting appropriate countercyclical capital buffer for banking sector

Abstract

The paper proposes formal econometric framework, with background in economic and finance theories, that supports macroprudential policy decision-makers in the process of choosing appropriate value of the countercyclical buffer and proper timing for countercyclical buffer introduction (build-up) and resolution. The dedicated dataset, which consists of time series describing banking sector (e.g. share of wholesale financing), market risk (e.g. CDS of banks and sovereigns), real estate prices, and macroeconomic measures, was used to select the group of time series signalling, within the Multivariate Markov-Switching Model with Distributed Lags (MMS-DL) with an appropriate lead, the beginning and the end of the financial crisis phase, and to build the group of composite leading indicators of the private debt cycle. Different data transformation methods and different statistical data definitions were used to analyse the process of early warning signal extraction useful for countercyclical buffer operationalization. The constructed indicators were confronted with two kinds of competitors: the naive univariate indicators and composite leading measures prepared with the help of Logistic Regression (LR) approach allowing to choose the most efficient analytical structure to support decision-makers.(original abstract)