Applications of Network Analyses to Systemic Risk in Financial Systems and to Macroeconomic Growth and Volatility

This thesis contributes to the applications of network analysis to the areas of macro-prudential policy and granular macroeconomics for GDP growth and volatility. Following the main introduction of the thesis, Chapter 2 investigates the properties of the global banking system flows, as a cross-border banking system, given by the BIS consolidated banking statistics. It contributes to the literature in two ways. First, by extending the systemic risk analysis in Markose et al (2017) to quantify the implied loss in case of failure of the systemically most important banking system. For this, I use the Eigen-pair method of Markose-Giansante with the maximum eigen-value yielding the systemic risk index and the right and left eigenvector centralities providing measures, respectively, for systemic importance and systemic vulnerability of banking systems. Second, by filling in major data gaps in the within country sectoral flow of funds in the BIS data, and analysing the sectoral cross-border flows (non-financial sectors across and within countries). In Chapter 3, a new and innovative approach based on the Ghosh inverse is used to quantify the falling in GDP growth given an increase in the financial sector share of gross operating profits to the detriment of other sectors of the economy. The final chapter builds on the Carvalho-Gabaix-Acemoglu approach of granular macroeconomics. It innovates by analysing the impact of sectoral final demand shocks on GDP volatility given the centrality of the sectors. This is compared with the Carvalho-Gabaix-Acemoglu approach of supply side productivity shocks. Both approaches show the growth of the financial sector centrality as a major contributor to GDP volatility