This study focuses on the bailout of systemic risk. Under the net liability clearing mechanism, we propose two bailout models under the one-layer lending network and two-layer network with lending and share cross-holding, respectively. The bailout models aim at maximizing the payment of the financial system while controling the bailout budget. We initially introduce an upper bound constraint on bailout money in the models to save the government’s bailout cost. The two models are rewritten as 0–1 mixed integer programming problems. By sufficiently utilizing the structural properties of reformulated models, we develop pre-processing rules for fixing some 0–1 variables, and then propose two new branch and bound type algorithms whose maximum iteration number equals to the number of banks in the system. This ensures that our algorithms are computational efficient and applicable for large-scale problems. We carry out empirical tests on the Chinese banking system and provide some helpful policy guidance to the government, which demonstrate the effectiveness and practicality of our models and solution methods.
© 2008-2026 Fundación Dialnet · Todos los derechos reservados