This study investigates the dynamic interdependence of global financial markets over time, focusing on identifying connection patterns among assets during periods of non-crisis and crisis, covering the period from 03/01/2000 to 06/23/2023. We employ Dynamic Bayesian Networks (DBN) to map the complex network of interactions between assets and reveal the topology of these connections. Our main findings indicate that over time, both in non-crisis periods and crisis periods, there is a tendency for financial assets to connect more strongly with assets located geographically nearby. This signals a process of "deglobalization" of financial markets, increasing considerations of trade-offs when analyzing risk and return asymmetries between different regionalized communities. Additionally, we observe that networks exhibit higher complexity in non-crisis periods compared to crisis moments. These conclusions substantially contribute to understanding the dynamics of global financial markets and emphasize the importance of considering factors related to regionalization and geographical interconnections when assessing risks and returns in the financial landscape. The implications of these findings are relevant for investors, portfolio managers, and policymakers, providing valuable insights for risk management and decision-making in an ever-evolving global financial environment.
Complete tool for submission, evaluation and publication of proceedings for scientific events.
