Improving Market-Based Systemic Risk Measures - Andreas Fuster
Speaker : Andreas Fuster (EPFL)
We identify a bias in existing systemic risk measures based on the market value of equity (e.g., SRISK). The bias is a function of expected creditor losses and increases in the variance of the realised value of bank assets. This implies that such systemic risk measures decrease when the volatility of a bank’s assets increases. We propose an approach that addresses the bias and, based on the novel concept of a systemic-risk-neutral probability measure, allows to capture the interaction of changes in volatility at the bank and the sectoral level. Then, we estimate and simulate a combined model for equity and CDS prices for a set of global banks. The bias-correcting term is quantitatively important, and, as theory predicts, increases in times of stress. Based on our estimates, we introduce a systemic risk dashboard that allows to decompose systemic risk based on the assumed importance of different externalities.