Early warning signals of financial stress: a “wavelet-based” composite indicators approach

In this paper we exploit the usefulness of wavelet multi resolution analysis in providing early warning signals of financial stress conditions. The proposed "wavelet-based" approach gives rise to a composite indicator obtained by aggregating several "scale-based" sub-indexes whose individual components are selected on the basis of their cross-correlations properties at different frequency bands. The performance of the "wavelet-based" composite indicator is assessed by evaluating its predictive power relative to the individual financial variables taken in isolation through an out-of-sample forecasting exercise for the US financial stress index. The findings indicate that the wavelet-based composite indicator largely outperforms any individual financial variable taken in isolation in early detecting nancial stress at every horizon and that the gain tends to increase as the time horizon increases.