This dissertation studies three distinct aspects of bank risk-taking from a novel portfolio perspective. First, I provide compelling evidence demonstrating that banks with high industry exposure adeptly adapt loan contract design to prevent potential adverse effects on loan portfolio value arising from the interaction between rival borrowers. Second, I stress the importance of banks’ pre-existing exposure and asset concentration in characterizing the risk-shifting incentives of banks enjoying government guarantees coverage. Last, I show that government guarantees coverage prompts bank risk-shifting at the intensive margin, and induces borrowers to leverage excessively, overinvest, and engage in low-quality projects. Altogether, I highlight the importance of pre-existing exposure in shaping banks’ risk management incentives and its connection to two key issues: the rise in bank concentration and the extent of government guarantees coverage. These findings carry noteworthy implications for bank lending behavior and loan contract design, ultimately impacting borrowers’ corporate policy and potentially giving rise to unintended policy consequences.