Calibrating Initial Shocks in Bank Stress Test Scenarios: An Outlier Detection Based Approach.

We propose a rigorous and flexible methodological framework to select and calibrate initial shocks to be used in bank stress test scenarios based on statistical techniques for detecting outliers in time series of risk factors. Our approach allows us to characterize not only the magnitude, but also the persistence of the initial shock. The stress testing exercises regularly conducted by supervisors distinguish between two types of shocks, transitory and permanent. One of the main advantages of our framework, particularly relevant as regards the calibration of transitory shocks, is that it allows considering various reverting patterns for the stressed variables and informs the choice of the appropriate stress horizon. We illustrate the proposed methodology by implementing outlier detection algorithms to several time series of (macro)economic and financial variables typically used in bank stress testing.