The network econometrics of financial concentration

The financial system is becoming more and more interconnected at an international level. This interconnection can lead to widespread financial shocks and crises around the globe. However, there is not a clear and unique understanding of the impact of financial concentration over interconnectedness and systemic risk. In the last decades, there has been a considerable development in the applications of network theory to finance. These advances allow us to study, at a network level, the effect of financial concentration on the degree of financial interconnection, leading to robust evidence about this relationship. This is specially relevant in the current context of fusion waves happening in some European countries, fostered by the supervisory and regulatory agencies. The objective of this paper is to unveil the relationship between financial concentration and financial interconnectedness by employing network models. This paper applies the Exponential Random Graph Model to a multiplex financial network connecting some of the main countries of the international financial system in different layers. As each layer represents a different set of monetary and financial institutions, this approach leads to a rich understanding of the relationship between the variables because it is possible to see how it operates at many levels. We find that financial concentration decreases the number of relationships between the agents of the international financial network. We also find that the volume of assets that a country has leads to a similar result, whilst the number of monetary and financial institutions increases it. Finally, we find that the ERGM is a valid methodology to inquire the behavior of a relationship of this kind. The model does not allow to study the dynamic behavior of the network. Thus, other kind of methodologies are necessary in order to achieve results about how the relationship evolves over time.