Modeling structure and credit risk of the economy: a multilayer bank-firm network approach

Battle Count: 50

The paper is primarily focused on macroprudential systemic risk assessment and regulatory stress testing rather than direct trading strategies. However, the identification of systemically important firms, sectoral vulnerability rankings, and contagion pathways could inform sector rotation strategies, credit risk pricing, and tail-risk hedging. The multilayer network structure could be used to build correlation models between sectors and financial institutions for portfolio risk management. The framework is more relevant to institutional risk managers and regulators than to high-frequency or algorithmic trading.

The framework involves multiple interconnected components: (1) three distinct network reconstruction methods (DCGM, ECAPM, SCGM/IOGM) with parameter calibration; (2) ESRI dynamics with essential/non-essential input constraints and iterative convergence; (3) FSRI credit loss computation; (4) DebtRank propagation with nonlinear default probability; (5) ensemble averaging over 100 network realizations; (6) statistical analysis (OLS, quantile regression). Requires understanding of network science, maximum entropy methods, input-output economics, and financial contagion models. The GitHub repository provides code but the conceptual complexity is high.

Code is publicly available on GitHub (https://github.com/mnlknt/bank-firm_multilayer_shocks) with sample data. However, the proprietary dataset (BankFocus, AIDA from Bureau van Dijk) requires commercial subscription. ISTAT Input-Output tables are publicly available. The methodology is fully described with equations, parameter settings, and filtering criteria. Results are averaged over 100 network realizations.