APPLYING NEXT GENERATION TOOLS, DATA, AND ECONOMIC COMPLEXITY IDEAS

LIPARI SUMMER SCHOOL: Temporal Asymmetries in Italian Interbank network reproduced by a data-driven Agent-Based Model

08-09-2015
Location: 
Lipari Italy

Prof. Vinko Zlatic: 16:00 - 17:30

Abstract:

We study the changes in the topology of the structure of the Italian interbank market in the period from January 1st 1999 to September 1st 2009. We find temporal irreversibility in the growth of the largest component of the interbank trading network, a feature not common to any of the usual network growth models. Such asymmetry, (also detected on the growth of the clustering and reciprocity coefficient) reveals that the trading mechanism changes along the day. We are able to recover the statistical properties of the system by means of a simple Agent-Based Model (ABM) where the probability of matching between counter parties depends on a time varying vertex fitness (or attractiveness) describing banks liquidity needs. We show that temporal irreversibility is associated with heterogeneity in the banking system and emerges when the distribution of banks liquidity shocks is broad.