Data-driven multi-layer network approaches to qualify the spreading of systemic risk

M-LAYER

We use the availability of daily empirical financial (asset-liability) network data to explore the relation of network topology and systemic risk (SR). The central idea of this project is to analyze and understand the spreading of SR across multiple layers. In particular, we want to explore how the understanding of financial multi-layer networks helps to quantify and ultimately manage systemic risk. We propose to do this in three directions. First, building on earlier work, we want to design an improved SR indicator (SRI) that is able to relate the detailed information of the networks to the expected systemic losses that would occur in case of a system-wide crisis. We will systematically study the daily time series (2007 to 2014) of the SRI and relate it to historical economic events and a series of instruments. Aim of this part is to estimate the predictive power of SR build-up in networks as reflected in the SRI. Second, SR is not limited to direct exposure networks

(such as credit networks), but applies similarly to other financial networks. In particular, networks that link financial agents through overlapping portfolios, derivatives, foreign exchange, or securities are potentially relevant sources of SR. We want to use a unique dataset to study the relative contributions of these different networks to the total SR in the system. Special focus will be on the mutual correlations between the network layers and SRI. The third direction addresses the management of SR through the idea of a SR tax, that was proposed in the context of inter-bank borrowing-lending networks. This tax taxes transactions that increase the SR of the system. By trying to evade the tax, agents look for counterparties and transactions that do not increase SR. Here we propose to extend the notion of the SR tax from the credit market to all layers of financial interactions, including derivatives, securities, overlapping portfolios and FX. This extension is highly non-trivial, since through netting of positions between financial layers non-linear effects are introduced. The technical framework is a large-scale macro-financial simulator that we are planning to extend massively in the direction of financial networks along this project. We have ongoing network analysis projects with the Banco de Mexico and can use these for realistic scenarios in the simulator.

Duration: 

01.08.2018 – 
31.12.2021
Stefan Thurner @ Franziska Liehl, President of the Complexity Science Hub

Stefan Thurner

Funded by

Project Partners

0 Pages 0 Press 0 News 0 Events 0 Projects 0 Publications 0 Person 0 Visualisation 0 Art

Signup

CSH Newsletter

Choose your preference
   
Data Protection*