Explainable Graph Neural Networks for Interbank Contagion Surveillance: A Regulatory-Aligned Framework for the U.S. Banking Sector

2026-04-14 · Source: Machine Learning · Field: Finance & Economics — Banking & Financial Services, FinTech & Digital Financial Services, Economic Analysis & Policy · Depth: Expert, quick

Summary

The Spatial-Temporal Graph Attention Network (ST-GAT) framework offers an explainable Graph Neural Network (GNN) solution for early detection of bank distress and macro-prudential surveillance within the U.S. interbank system. This framework models 8,103 FDIC-insured institutions across 58 quarterly snapshots from 2010Q1 to 2024Q2. It reconstructs bilateral exposures from public FDIC Call Reports using maximum entropy estimation to create a dynamic directed weighted graph. ST-GAT achieved an AUPRC of 0.939 +/- 0.010, outperforming other GNN architectures and nearly matching XGBoost's 0.944. Ablation studies show the BiLSTM temporal component adds +0.020 AUPRC, with temporal attention weights reflecting long-run structural vulnerability. Return on Assets (ROA) at 0.309 and Non-Performing Loan (NPL) Ratio at 0.252 were identified as key predictors, aligning with findings from the 2023 regional banking crisis.

Key takeaway

For financial regulators and risk analysts monitoring systemic stability, the ST-GAT framework provides a robust, explainable tool for early warning of bank distress. Its high AUPRC and identification of key predictors like ROA and NPL Ratio offer actionable insights, enabling more proactive and data-driven macro-prudential surveillance. You should consider integrating such explainable GNN approaches to enhance your current risk assessment methodologies.

Key insights

ST-GAT provides an explainable GNN for early bank distress detection and interbank contagion surveillance.

Principles

• Dynamic graph modeling enhances financial surveillance.
• Temporal components improve predictive accuracy.
• Explainability aids regulatory alignment.

Method

The ST-GAT framework uses maximum entropy estimation to reconstruct dynamic directed weighted graphs from FDIC Call Reports, then applies a BiLSTM temporal component and attention mechanisms for distress prediction.

In practice

• Utilize FDIC Call Reports for interbank network data.
• Incorporate temporal attention for financial time series.
• Prioritize ROA and NPL Ratio in risk models.

Topics

• Explainable GNNs
• Interbank Contagion
• Macro-prudential Surveillance
• Bank Distress Detection
• ST-GAT Framework

Best for: AI Scientist, Research Scientist, Policy Maker

Related on AIssential

• Unlocking the relational power of graph neural networks in fraud prevention — Graph Neural Networks enhance fraud detection by modeling relational data, uncovering networked fraud invisible to traditional tabular methods.
• The Silicon Soil — AI and real-time data can transform banking from land-dependent to a dynamic wealth orchestrator, mitigating systemic risk.
• AIMing for Standardised Explainability Evaluation in GNNs: A Framework and Case Study on Graph Kernel Networks — AIM provides a flexible framework for standardized, comprehensive explainability evaluation across diverse GNN models.
• Discovering Entity-Conditioned Lag Heterogeneity: A Lag-Gated Neural Audit Framework for Panel Time Series — AC-GATE provides a verifiable framework to expose entity-conditioned lag heterogeneity as a structured, inspectable model output.
• Bi-Level Chaotic Fusion Based Graph Convolutional Network for Stock Market Prediction Interval — Prediction intervals, not just point forecasts, are crucial for risk-aware financial market decision-making.
• Beyond Soft Masks: Hard-Perturbation Mixup Explainer for Robust GNN Explainability — HPME uses hard perturbations and structural mixup to generate robust, in-distribution GNN explanations by eliminating irrelevant information.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine Learning.