Analysis of terminal risks and interbank correlations in Iranian listed banks: A hybrid model approach of deep learning and Gaussian processes)

Azimi, Sepideh; Kianpoor, Saeed; Homaioonfar, Ali Reza

Analysis of terminal risks and interbank correlations in Iranian listed banks: A hybrid model approach of deep learning and Gaussian processes)

Document Type : Origional Article

Authors

Sepideh Azimi ¹ ^¶

Saeed Kianpoor ²

Ali Reza Homaioonfar ³

¹ Instructor, Department of Basic Science, Faculty of Basic and General Studies, Technical and Vocational University(TVU), Tehran,Iran.

² Assistant Professor, Department of Management, Payame Noor University, Tehran, Iran

³ Instructor, Department of Social Sciences, Payame Noor University, Tehran, Iran

Abstract

Background and Objective: This research was conducted to measure tail risks and interbank correlations within the Iranian banking system and to identify the root causes of systemic fragility. The focus is on the extended period following the global financial crisis (from 2008 to 2024), during which the banking system has been under severe pressure from structural factors such as fixed provisional profit payments to depositors, widespread overdrafts from the central bank, severe balance sheet imbalances, and macroeconomic shocks. The primary objective is to assess the current state of systemic vulnerability and forecast probable future paths using advanced hybrid modeling approaches.
Methodology: This study employs a novel hybrid model integrating deep learning, Gaussian processes, Time-Varying Parameter Vector Autoregression (TVP-VAR) models, interbank network analysis, and extensive Monte Carlo simulations. Quarterly data on twelve key banking variables (including overdrafts, non-performing loans, capital adequacy ratio, liquidity, etc.) alongside macroeconomic variables (inflation, exchange rate, economic growth, etc.) from 2008 to 2024 were analyzed. A systemic vulnerability index and the magnitude of risk transmission through various channels (particularly the trust channel) were calculated and compared with conventional methods.
Results and Findings: The results indicate that the Iranian banking system entered a stable critical and chaotic regime at the beginning of 2008 and has remained in this state until the end of 2024. The systemic vulnerability index surpassed 0.96 in 2024, signifying a highly fragile condition nearing a critical point. Shocks to the system operate in a highly asymmetric manner, with negative shocks being approximately ten times stronger than positive ones and tending to be nearly permanent. Risk transmission among banks occurs almost entirely (close to 100%) through the trust channel and hidden correlations. The root cause of this fragility is the continued policy of paying fixed provisional profits to depositors (despite real resource-use imbalances) and the widespread overdrafting of banks from the central bank, which has created a vicious cycle of liquidity expansion and steadily increasing systemic risk. Simulations estimate the probability of a systemic collapse by the end of 2026, assuming the current trend continues, at over 87%. Conversely, immediate and decisive structural reforms, including the complete elimination of fixed provisional profit payments, the dissolution or merger of insolvent banks, and a serious overhaul of corporate governance, could steer the banking system toward a stable and self-reinforcing regime. The forecasting accuracy of the proposed model is significantly higher than that of traditional methods.

Keywords

Tail Risk

Interbank Correlation

Systemic Risk

Hybrid Deep Learning Model

Iranian Banking System

Subjects

Sustainable Regional Development

· Abdurraffi, N., & Banna, H. (2025). AI-Enhanced Risk-Return Optimization in Islamic and Conventional Banking Portfolios. EkBis: Jurnal Ekonomi dan Bisnis, 9(1). https://doi.org/10.14421/ekbis.2025.9.1.2414

· Dahliani, Y., El Wafa, F., & Supendi. (2025). The Paradox of Risk and Sharia Compliance: Analyzing Islamic Bank Strategies in Navigating Global Economic Uncertainty. International Journal for Science Review, 2(5). https://doi.org/10.71364/ijfsr.v2i5.28

· Damianou, A. C. (2015). Deep Gaussian processes and variational propagation of uncertainty (Doctoral dissertation, University of Sheffield).

· Ding, W., Li, W., Zhang, Z., Wan, C., Duan, J., & Lu, S. (2023). Time-Varying Gaussian Markov Random Fields Learning for Multivariate Time Series Clustering. IEEE Transactions on Knowledge and Data Engineering. https://doi.org/10.1109/TKDE.2022.3232331

· Elsinger, H., Lehar, A., & Summer, M. (2006). Using Market Information for Banking System Risk Assessment. International Journal of Central Banking, 2(3), 1–32.

· Frey, R., & Hledik, J. (2014). Correlation and Contagion as Sources of Systemic Risk. SSRN. https://doi.org/10.2139/ssrn.2541733

· Jiang, S., Wang, J., Dong, R., Li, X., & Zhang, Y. (2023). Systemic Risk with Multi-Channel Risk Contagion in the Interbank Market. Sustainability, 15(3), 2727. https://doi.org/10.3390/su15032727

· Li, C. (2025). Research on financial risk prediction and management models based on big data analysis. International Journal of High Speed Electronics and Systems, 12(4), 2441007. https://doi.org/10.1142/s0129156425410075

· Li, G., Jing, Z., Feng, Y., & Li, J. (2023). Drivers of risk correlation among financial institutions: A study based on a textual risk disclosure perspective. Economic Modelling, 125, 106468. https://doi.org/10.1016/j.econmod.2023.106468

· Liu, F., Mao, T., Wang, R., & Wei, L. (2022). Inf-convolution, optimal allocations, and model uncertainty for tail risk measures. Mathematics of Operations Research, 47(4), 2757–2782. https://doi.org/10.1287/moor.2021.1217

· Nakhli, S., & Rafei, M. (2025). How can optimal monetary policies reduce oil sanctions’ impacts? Evidence from Iran. Economic and Political Studies. https://doi.org/10.1080/20954816.2024.2312759

· Paul, M., Nagwovuma, M., Nansamba, B., Hellen, N., Jingo, D., & Marvin, G. (2024). Trustworthy Deep Learning Techniques for Credit Risk Assessment. 2024 5th International Conference on Electronics and Sustainable Communication Systems (ICESC). https://doi.org/10.1109/ICESC60852.2024.10689741

· Peykani, P., Sargolzaei, M., Oprean-Stan, C., Kamyabfar, H., & Reghabi, A. (2025). The effect of macroeconomic shocks on non-performing loans and credit risk in the Iranian banking system using time-varying parameter vector autoregressions. PLOS One. https://doi.org/10.1371/journal.pone.0329587

· Rahman, M. R., Misra, A., & Tiwari, A. (2024). Interbank systemic risk network in an emerging economy. Review of Accounting and Finance. https://doi.org/10.1108/raf-07-2023-0206

· Rasmussen, C. E., & Williams, C. K. I. (2005). Gaussian Processes for Machine Learning. MIT Press. https://doi.org/10.7551/mitpress/3206.001.0001

· Wang, Z., Zhao, Q., & Qiu, L. (2022). Multi-Dimensional Factor Correlation, Multiple Interbank Network Contagion, and Conditional VaR of Banks. Frontiers in Physics, 10, 895603. https://doi.org/10.3389/fphy.2022.895603

· Xie, C., Liu, Y., Zhang, J., Wang, H., Li, M., & Chen, Y. (2016). The Stability of Interbank Market Network: A Perspective on Contagion and Risk Sharing. Advances in Mathematical Physics, 2016, 1297832. https://doi.org/10.1155/2016/1297832

· Yao, Y. (2025). Applications of Bayesian inference in financial econometrics: A review. Economics and Management Innovation, 8(1), 45–60. https://doi.org/10.71222/pf0sg388

· Zedda, S., Patanè, M., & Miggiano, L. (2018). The Role of Correlation in Systemic Risk: Mechanisms, Effects, and Policy Implications. Social Science Research Network (SSRN). https://doi.org/10.2139/ssrn.3252363

· Zedda, S., Patane, M., & Miggiano, L. (2021). The Role of Correlation in Systemic Risk: Mechanisms, Effects, and Policy Implications. In Financial Risk Management and Modeling (pp. 123–145). Springer, Cham. https://doi.org/10.1007/978-3-030-78965-7_58