This paper proposes a methodology to identify deposits' outflows in real-time using daily TARGET2 data from 100 Italian banks. The ability of the methodology to timely identify anomalous outflows is assessed using Monte Carlo methods and comparing it with alternative methodologies.
The methodology identified a limited number of anomalous outflows for the period between August 2012 and August 2019. About 3 percent of deposits was withdrawn on average, 95 percent of which was transferred to other banks. The remaining 5 percent was converted into cash. Bilateral transfers to other banks were mostly directed to large domestic banks, without contagion effects.
Forthcoming in: Journal of Financial Econometrics.
