Since the financial crisis of 2007-2009, quantitative models of systemic risk have experimented a revolution. One of the main causes resides in the failure of the models responsible of predicting and providing insights about the crisis, as well as to the lack of explanations about the interlinks between the real economy and the financial system. It is now clear that economic and financial systems are complex by nature, that many of the underlying processes shaping them are unobserved, and that there is a necessity of inferential models to estimate those processes. With the ongoing flows of data, new methodologies are needed to support traditional methods that worked well with simpler data and smaller samples. The relevance of the interconnections between economic and financial agents cannot be now neglected, and therefore, methodologies for analyzing networked data are penetrating rapidly into the econometric toolkit. As the systems of study are dynamic and time-evolving, there is a natural space for time series methods to join.
This thesis aims to contribute to this novel stream of literature that merges time series econometrics, networks and machine learning models to generate inferences, dynamic analyses and forecasts on financial stability, macroeconomics and monetary economics, the intercept between the three, and their now undeniable interconnections.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados