A study on the impact of strategic behaviour in financial networks

A financial system comprises a set of institutions, such as banks, that engage in financial transactions. The interconnections showing the liabilities among the banks are represented by a network and can be highly complex. Hence, the utility (well-being) of a bank depends on the entire network and not just on the well-being of its immediate borrowers/debtors. For example, the possible bankruptcy of a bank and the corresponding damage to its immediate lenders/creditors resulting from the bank's failure to repay, can be propagated through the financial network by causing the creditors' (and other banks', in sequence) inability to repay their debts, thus having a global effect. In this thesis, we study the global impact induced by a sequence of individual strategic operations in the realm of financial networks. In particular, we investigated a range of financial operations including cash injection, debt removal, debt transfer, and priority-proportional payments, as adopted by banks. We study each operation in both centralized and decentralized manner. In the centralized context where the financial authority has the power to control the behavior of each bank as desired, we aim to determine if there exists an algorithm that efficiently computes the optimal combination of these operations to achieve desired systemic objectives. In cases where no efficient algorithm exists, we provide hardness results on the computation aspects and turn our attention to the development of approximate algorithms. In the game-theoretic (decentralized) setting, we consider each bank as a utility-maximizing agent that can be strategic about the corresponding operations. Regarding the games in financial networks, we study the existence and quality of equilibria, as well as the computational complexity of equilibrium-related problems. Overall, our findings contribute to a good understanding of the impact of strategic behavior on financial networks.