Paper: arXiv 2410.23275

Authors: Matteo Citterio, Marco D’Errico, Gabriele Visentin

Abstract

We introduce a novel Dynamic Graph Neural Network (DGNN) architecture for solving conditional $m$-steps ahead forecasting problems in temporal financial networks. The proposed DGNN is validated on simulated data from a temporal financial network model capturing stylized features of Interest Rate Swaps (IRSs) transaction networks, where financial entities trade swap contracts dynamically and the network topology evolves conditionally on a reference rate. The proposed model is able to produce accurate conditional forecasts of net variation margins up to a $21$-day horizon by leveraging conditional information under pre-determined stress test scenarios. Our work shows that the network dynamics can be successfully incorporated into stress-testing practices, thus providing regulators and policymakers with a crucial tool for systemic risk monitoring.

Complexity vs Empirical Score

  • Math Complexity: 5.5/10
  • Empirical Rigor: 6.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper presents an innovative application of Dynamic Graph Neural Networks (DGNN) to stress-testing and conditional forecasting of margin calls in temporal financial networks. While the empirical evaluation is conducted on simulated rather than real transaction-level OTC derivatives data, the theoretical framework and methodology are well-articulated and address a crucial gap in systemic risk monitoring.

Research Flowchart

  flowchart TD
    A[Research Goal: Forecast multi-step margin calls in financial networks under stress scenarios] --> B[Data Inputs: Dynamic IRS transaction networks & Reference Rate scenarios]
    B --> C[Methodology: Dynamic Graph Neural Network DGNN]
    C --> D[Computational Process: Model network topology evolution & conditional m-step ahead forecasting]
    D --> E[Key Outcomes: Accurate net variation margin forecasts up to 21-day horizon]
    E --> F[Impact: Enhanced systemic risk monitoring for regulators & stress-testing]

Cited References (131)

  1. Graph Neural Networks: Foundations, Frontiers, and Applications (2022)
  2. The graph neural network model (2008)
  3. Responsible graph neural networks (2023)
  4. Indices, graphs and null models: analyzing bipartite ecological networks (2009)
  5. Global alignment of protein–protein interaction networks by graph matching methods (2009)
  6. Neural message passing for quantum chemistry (2017)
  7. Towards a realistic track reconstruction algorithm based on graph neural networks for the HL-LHC (2021)
  8. Graph learning for combinatorial optimization: a survey of state-of-the-art (2021)
and 123 more
  1. Graph neural network for traffic forecasting: A survey (2022)
  2. A review of graph neural networks and their applications in power systems (2021)
  3. Liaisons dangereuses: Increasing connectivity, risk sharing, and systemic risk (2012)
  4. Wavelets on graphs via spectral graph theory (2011)
  5. Gaussian Markov random fields: theory and applications (2005)
  6. Exploring network structure, dynamics, and function using NetworkX (2008)
  7. A global geometric framework for nonlinear dimensionality reduction (2000)
  8. Graph embedding for pattern analysis (2012)
  9. Deepwalk: Online learning of social representations (2014)
  10. Semi-supervised classification with graph convolutional networks (2016) — link
  11. Graph attention networks (2017) — link
  12. Inductive representation learning on large graphs (2017)
  13. Graph convolutional neural networks for web-scale recommender systems (2018)
  14. Cluster-gcn: An efficient algorithm for training deep and large graph convolutional networks (2019)
  15. Approximation by superpositions of a sigmoidal function (1989)
  16. The reduction of a graph to canonical form and the algebra which appears therein (1968)
  17. How powerful are graph neural networks? (2018) — link
  18. Forecasting the future of artificial intelligence with machine learning-based link prediction in an exponentially growing knowledge network (2023)
  19. Fast graph representation learning with PyTorch Geometric (2019) — link
  20. Pytorch geometric temporal: Spatiotemporal signal processing with neural machine learning models (2021)
  21. Graph neural networks in recommender systems: a survey (2022)
  22. A graph neural network method for distributed anomaly detection in IoT (2021)
  23. Friends and neighbors on the web (2003)
  24. Variational graph auto-encoders (2016) — link
  25. Link prediction based on graph neural networks (2018)
  26. Evolvegcn: Evolving graph convolutional networks for dynamic graphs (2020)
  27. Recurrent event network: Autoregressive structure inference over temporal knowledge graphs (2019) — link
  28. Predicting dynamic embedding trajectory in temporal interaction networks (2019)
  29. On the evolution of random graphs (1960)
  30. Random graphs (1998)
  31. Structural diversity and homophily: A study across more than one hundred big networks (2017)
  32. Netgan: Generating graphs via random walks (2018)
  33. Graphvae: Towards generation of small graphs using variational autoencoders (2018)
  34. A survey on deep graph generation: Methods and applications (2022)
  35. A systematic survey on deep generative models for graph generation (2022)
  36. Generative adversarial nets (2014)
  37. MolGAN: An implicit generative model for small molecular graphs (2018) — link
  38. Graph convolutional policy network for goal-directed molecular graph generation (2018)
  39. Wasserstein generative adversarial networks (2017)
  40. Graphrnn: Generating realistic graphs with deep auto-regressive models (2018)
  41. node2vec: Scalable feature learning for networks (2016)
  42. Emergence of scaling in random networks (1999)
  43. Hands-On Graph Neural Networks Using Python: Practical techniques and architectures for building powerful graph and deep learning apps with PyTorch (2023)
  44. Flexible, high performance convolutional neural networks for image classification (2011)
  45. Principal component analysis (1987)
  46. Long short-term memory (1997)
  47. A guide to temporal networks (2016)
  48. Connectivity and inference problems for temporal networks (2000)
  49. An introduction to the theory of point processes: volume II: general theory and structure (2007)
  50. Point processes (1980)
  51. Hawkes processes in finance (2015)
  52. Spectra of some self-exciting and mutually exciting point processes (1971)
  53. Foundations and modeling of dynamic networks using dynamic graph neural networks: A survey (2021)
  54. GC-LSTM: Graph convolution embedded LSTM for dynamic network link prediction (2022)
  55. Graph representation learning (2020)
  56. Rethinking the financial network (2013)
  57. Mathematical modeling of systemic risk (2012)
  58. Interbank clearing in financial networks with multiple maturities (2019)
  59. Dynamic clearing and contagion in financial networks (2024)
  60. A continuous-time model of financial clearing (2020)
  61. Bilateral credit valuation adjustment for large credit derivatives portfolios (2014)
  62. The price of complexity in financial networks (2016)
  63. Compressing over-the-counter markets (2021)
  64. Systemic risk and central clearing counterparty design (2015)
  65. Optimal control of interbank contagion under complete information (2014)
  66. Clearing payments in dynamic financial networks (2023)
  67. Control of dynamic financial networks (2022)
  68. Representation learning for dynamic graphs: A survey (2020)
  69. Foundations and modeling of dynamic networks using dynamic graph neural networks: A survey (2021)
  70. A Comprehensive Survey of Dynamic Graph Neural Networks: Models, Frameworks, Benchmarks, Experiments and Challenges (2024) — link
  71. A deep learning approach to dynamic interbank network link prediction (2022)
  72. Dynamic interbank network analysis using latent space models (2020)
  73. Network valuation in financial systems (2020)
  74. DebtRank and the network of leverage (2016)
  75. Simulation methods to assess the danger of contagion in interbank markets (2011)
  76. Default cascades: When does risk diversification increase stability? (2012)
  77. Systemic risk in financial networks (2007)
  78. A dynamic network model of interbank lending—systemic risk and liquidity provisioning (2020)
  79. Control of interbank contagion under partial information (2015)
  80. Financial network and systemic risk—a dynamic model (2021)
  81. Dynamic default contagion in heterogeneous interbank systems (2021)
  82. Endogenous Distress Contagion in a Dynamic Interbank Model (2022) — link
  83. Bank runs, deposit insurance, and liquidity (2000)
  84. Systemic risk and stability in financial networks (2015)
  85. How likely is contagion in financial networks? (2015)
  86. Failure and rescue in an interbank network (2013)
  87. Stability analysis of financial contagion due to overlapping portfolios (2014)
  88. Liquidity risk and contagion (2005)
  89. Systemic liquidity shortages and interbank network structures (2013)
  90. Liquidity hoarding, network externalities, and interbank market collapse (2010)
  91. Contagion in financial networks (2010)
  92. Financial networks, cross holdings, and limited liability (2009)
  93. Structured sequence modeling with graph convolutional recurrent networks (2018)
  94. Term-structure models: A graduate course (2009)
  95. An introduction to the theory of point processes: volume I: elementary theory and methods (2003)
  96. Interest rate models: theory and practice (2001)
  97. Brownian motion and stochastic calculus (2014)
  98. Predictive temporal embedding of dynamic graphs (2019)
  99. Convolutional LSTM network: A machine learning approach for precipitation nowcasting (2015)
  100. Empirical evaluation of gated recurrent neural networks on sequence modeling (2014) — link
  101. Real-valued (medical) time series generation with recurrent conditional gans (2017) — link
  102. An attentional multi-scale co-evolving model for dynamic link prediction (2023)
  103. Options, futures, and other derivatives (2016)
  104. Interest rate models-theory and practice: with smile, inflation and credit (2006)
  105. Pathways towards instability in financial networks (2017)
  106. Will a large complex system be stable? (1972)
  107. Risk and global economic architecture: Why full financial integration may be undesirable (2010)
  108. Systemic risk in banking ecosystems (2011)
  109. Payvae: A generative model for financial transactions (2021)
  110. Dynamic term structure models for SOFR futures (2021)
  111. Term structure modeling of SOFR: Evaluating the importance of scheduled jumps (2023)
  112. A theory of the term structure of interest rates (2005)
  113. Monte Carlo methods in financial engineering (2004)
  114. Two singular diffusion problems (1951)
  115. Simulating a multi-factor term structure model over relatively long discrete time periods (1996)
  116. Exact simulation of point processes with stochastic intensities (2011)
  117. Monte Carlo algorithms for default timing problems (2011)
  118. SciPy 1.0: fundamental algorithms for scientific computing in Python (2020)
  119. Risk and valuation of collateralized debt obligations (2001)
  120. D 'e monstration simplifi 'e e d’un th 'e or `e me de Knight (1971)
  121. Adam: A method for stochastic optimization (2014) — link
  122. Systemic risk in financial systems (2001)
  123. Debtrank: Too central to fail? financial networks, the fed and systemic risk (2012)

0 of 131 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.