Forecasting Systemic Risk in the European Banking Industry: A Machine Learning Approach

The aim of this article is to forecast the systemic risk contribution and exposure measured by the delta conditional value at risk (ΔCoVaR) and the marginal expected shortfall (MES), respectively. We first estimate the ΔCoVaR and MES for banks in 16 European countries for the 2002–2016 period. We then predict systemic risk measures using machine learning techniques, such as artificial neural network (ANN) and support vector machine (SVM), and we use AR-GARCH specification. Finally, we compare the methods’ forecasting values and the actual values. Our results show that two hidden layers of artificial neural networks perform efficiently in forecasting systemic risk.