Måling og Regulering af Systemisk Risiko og Bankers Systemiske Vigtighed

Abstract

The purpose of this thesis is to gain insight on the phenomenon of systemic risk that has been adopted in the wake of the global financial crisis. Systemic risk implies that not only may the real economy suffer losses from the failure of a single institution. A failure of a single institution or impairment of parts of the financial system may even cause a disruption to financial services in an extent that losses suffered by the real economy are much more serious than in the first case, due to the financial system itself. At first, we show that although it has often been assumed that the size of an institution defines it’s systemically importance, there is indeed cases in which smaller institutions are more systemically important. As size is not the sole determinant of systemic importance, we look to the literature for ways to model and measure systemic risk, and to develop a measure of a single institutions systemic importance in order to identify a systemically important financial institution - a SIFI. We do this in a developing manner by firstly presenting an initial measure of cascade effects in a indirectly connected system and then lastly by introducing network theory in order to also model direct exposures that arise through the interbank market. Along with the introduction of networks, we present the shapley value. The shapley value decomposes total systemic risk into each bank’s marginal contribution to systemic risk. We use this measure as an approximation of a banks systemic importance. We analyze the drivers of both systemic risk and bank’s systemic importance. For systemic risk, we find that systemic risk is mainly driven by the soundness of institutions, but that dependency structures play a crucial role. These include asset correlation, architecture of the interbank market and capital requirements, but also the number, concentration and heterogeneity of banks. For a bank’s systemic importance, we find that size is indeed a main driver, but that asset correlation and composition are also important factors, along with activity and position in the interbank market, although the positional effects are ambiguous. We compare these findings with the recently implemented Danish SIFI framework, as the purpose of the framework is to identify SIFIs and manage systemic risk within the Danish financial system. We find that the Danish SIFI framework is very simplistic in the identification of SIFIs, as it does not fully capture effects of interaction and interconnectedness, hence interdependence of institutions, which is the very core of systemic risk. The overall conclusion is that the SIFI-framework is a step in the right direction, but that it needs to be further developed. Although the factors discussed in this thesis are far from trivial to implement, it is factors that should be taken into considerations, when doing so.