Title:Optimal Systemic Risk Bailout: A PGO Approach Based on Neural Network

Abstract:The bailout strategy is crucial to cushion the massive loss caused by systemic risk in the financial system. There is no closed-form formulation of the optimal bailout problem, making solving it difficult. In this paper, we regard the issue of the optimal bailout (capital injection) as a black-box optimization problem, where the black box is characterized as a fixed-point system that follows the E-N framework for measuring the systemic risk of the financial system. We propose the so-called ``Prediction-Gradient-Optimization'' (PGO) framework to solve it, where the ``Prediction'' means that the objective function without a closed-form is approximated and predicted by a neural network, the ``Gradient'' is calculated based on the former approximation, and the ``Optimization'' procedure is further implemented within a gradient projection algorithm to solve the problem. Comprehensive numerical simulations demonstrate that the proposed approach is promising for systemic risk management.