Mathis Mourey

Composition of the Jury

Abstract

This thesis first proposes a conceptual framework describing in detail how systemic events (SEs) arise in the financial system, thus enabling the construction of a robust definition of SEs. Second, it gathers empirical evidence on the validity of the framework to legitimize our definition and allow for the construction of a coherent measure of Systemic Risk (SR). The first study presents the Systemic Events Hypothesis (SEH). The SEH is a description, built upon findings in the existing literature, of the steps involved in the unfolding of a SE. We begin with the structural shift as proposed by Kindelberger (1978) and Minsky (1991), which creates a latent risk. Financial firms unknowingly accumulate a common exposure to this latent risk, making the financial system vulnerable to a "not so unusual" event. This triggering event causes the latent risk to materialize and plunges the financial system into a financial crisis. The crisis worsens through contagion and amplification mechanisms before ultimately negatively impacting the real economy. Using well-established measures from the literature, we provide empirical evidence of these stages, particularly in European and American financial systems during the subprime mortgage crisis. The second study demonstrates the multidimensional nature of the social redistribution (SR) and confirms the sequence of stages described in the Higher Economics (HES). By performing a principal component analysis (PCA) on three SR measures, we propose a new measure based on the inertia of the first principal component. We show that the correlation between the SR measures is, in fact, a signal for the SR. Furthermore, we demonstrate that a time lag in the SR measures, as proposed by the HES, improves the final measure, thus further validating the presence of a sequence of stages in an economic system. Finally, using these same SR measures on European and American samples, we show that different types of financial firms exhibit different types of SR. We show that insurance companies tend to be more exposed to losses, while asset management firms are more interconnected than other financial institutions. Furthermore, large financial firms are more likely to suffer significant losses and be highly interconnected at the beginning of a crisis, whereas small firms are more affected by high illiquidity towards the end of the crisis.

Read the thesis