General Equilibrium Amplification and Crisis Vulnerability: Cross-Crisis Evidence from Global Banks

Battle Count: 76

The paper is primarily relevant to systemic risk management and macroprudential regulation rather than direct trading strategies. However, the amplification factor rankings and correlation dynamics findings could inform portfolio risk management, sector rotation strategies during crises, and tail-risk hedging decisions. The finding that exposure-based networks predict crisis outcomes while correlation-based networks do not has implications for how quantitative traders assess contagion risk. The crisis comparison (endogenous vs. exogenous shocks) could inform regime-switching trading strategies.

High complexity due to: (1) solving PDEs with spatial-network interaction terms, (2) computing Feynman-Kac path integrals numerically, (3) constructing bilateral exposure networks from confidential data, (4) implementing Lévy jump-diffusion with state-dependent intensity, (5) spectral decomposition of graph Laplacians, (6) channel decomposition requiring partitioning of diffusion paths. The theoretical framework requires strong mathematical background in stochastic processes, PDEs, and network theory. Empirical implementation requires access to bilateral exposure data and careful network estimation.

The paper provides detailed mathematical derivations (master equation, Feynman-Kac representation, amplification factor formula, channel decomposition) and specifies empirical methodology clearly. However, bilateral exposure data for the 16-bank network is not publicly available, and the specific network construction procedure for computing amplification factors is not fully detailed. Equity price data from standard financial databases is accessible but the exact sample selection criteria and network estimation parameters are not fully specified.