Dynamique des réseaux financiers

Dynamics in financial networks

• Auteur / Autrice : Lucia Bellenzier
• Direction : Jorgen-Vitting Andersen, Silvana Stefani
• Type : Thèse de doctorat
• Discipline(s) : Sciences économiques
• Date : Soutenance le 25/11/2013
• Etablissement(s) : Paris 1 en cotutelle avec Università degli studi di Milano - Bicocca
• Ecole(s) doctorale(s) : École doctorale d'Économie (Paris ; 2004-....)
• Partenaire(s) de recherche : Equipe de recherche : Centre d'économie de la Sorbonne (Paris ; 2006-....)
• Partenaire(s) de recherche : Etablissement d'accueil : Università degli studi di Milano - Bicocca
• Jury : Président / Présidente : Serge Galam
• Jury : Examinateurs / Examinatrices : Jorgen-Vitting Andersen, Silvana Stefani, Giulia Rotundo
• DOI : 10.70675/b3aa88ffzb303z4ed5z8e6az1475dc8262e8

Mots clés

Mots clés contrôlés

Systèmes dynamiques

Graphes dynamiques

Finances

Réseaux sociaux

Finances -- Modèles mathématiques

Mathématiques financières

Théorie des graphes

Mots clés libres

Réseaux financiers

Dynamiques

Italie

Theorie des graphes

Financial networks

Dynamics

Italy

Graph theory

Résumé

Dynamique des réseaux financiers.

The aim of this work is to examine the dynamics in financial networks. We propose two cases of study : the network of world’s stock exchanges and the network of Interlocking Directorates in Italy. In the first case we study the dynamics on the network and in the second one the dynamics of the network, i.e. in the first case the network topology does not change and we study the dynamic information that passes through the structure, instead, in the second case, we study how the network topology evolves over time. In ’Prices-Quakes Shaking the World’s Stock Exchanges’ (2011) Vitting Andersen etal. propose a model of the World’s Stock Exchanges that predicts how an individual stock exchange should be priced in terms of performance of the global exchange market. In the present work this model is adopted to describe the financial network. Understanding how disruption can propagate across financial markets is indeed of theutmost importance, hence our aim is to study the dynamics of the World’s Stock Exchange network, inter alia we are interested in the study of the avalanches of price, disturbances propagating in the world financial network of stock exchanges. In fact the model has a direct correspondence to models of earth tectonic plate movements developed in physics. In tectonic plate movement stresses are slowly build up over centuries only to be released in a quick snap, lasting from seconds to at most some few minutes, that we feel as an earth quake. The main idea is to describe a similar slow build-up of ’stresses’ in the world’s financial network of stock exchanges where stresses can be thought of arising from for examplebusiness cycles in the real economy. Just like earth quakes such a slow build up of ’stress’is then followed by a quick release in terms of domino effects where the major part of the worlds stock exchanges resonate with big up or down price movements. The main innovative part of the model that we will introduce is therefore a separation of time scales just as seen in earthquakes. We first replicate the results of the model using empirical data using as source Bloomberg. We then verify that the price dynamics can indeed be described in terms of avalanche dynamics, and we find that such dynamics show power law behavior in agree-ment with what is found in other Self Organized Critical (SOC) models. We extends such studies and give a quantitative as well as qualitative description of the details in the dynamics of the propagation of ”price-quakes (avalanches). The second case is the Interlocking Directorates in Italy, i.e. the situation that occurs when a person affiliated with one company sits on the board of directors of another organization, analyzed by using network theory. We first analyzed the Italian case in orderto investigate the presence of a persistent core. Applying the same methodology used by Milakovic et al. [35], from 1998 to 2010 we found quite different results : the persistent sub-graphs are not connected and so we could not find a core. Instead in the German stock exchange Milakovic et al. have found a small core of directors densely connected among themselves. In order to capture the persistent structure of the Interlocks Network we propose a different approach that allows us to assess the stability of links between companies in Italy. We describe the dynamic board networks by means of a static graph in which an edgeis related with the persistence over time of an interlock between two companies. The results lead to affirm that in the Italian board network a set of stable links is observable, nevertheless a presence of a large turnover between the directors. There are strong ties among firms in the overall period. Most of them are ties due to the ownership of family firms : Berlusconi, Benetton, Agnelli, Caltagirone, De Benedetti.