Paper: arXiv 2402.08071

Authors: Sunday Akukodi Ugwu

Abstract

This mini-project models propagation of shocks, in time point, through links in connected banks. In particular, financial network of 100 banks out of which 15 are shocked to default (that is, 85.00% of the banks are solvent) is modelled using Erdos and Renyi network – directed, weighted and randomly generated network. Shocking some banks in a financial network implies removing their assets and redistributing their liabilities to other connected ones in the network. The banks are nodes and two ranges of probability values determine tendency of having a link between a pair of banks. Our major finding shows that the ranges of probability values and banks’ percentage solvency have positive correlation.

Complexity vs Empirical Score

  • Math Complexity: 6.5/10
  • Empirical Rigor: 2.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper introduces stochastic differential equations, graph theory, and network science models like Erdős–Rényi with probabilistic formulas, indicating advanced math complexity. However, it lacks empirical data, backtests, or implementation details, relying instead on a conceptual mini-project model without real-world validation.

Research Flowchart

  flowchart TD
  A["Research Goal: How does shock propagate in a financial network<br>and what factors affect bank solvency?"] --> B["Model Setup"]
  B --> C["Define Network: 100 Banks in Erdos-Renyi Network<br>Directed & Weighted Links"]
  C --> D["Input Data: 15 Banks Shocked to Default<br>Redistribute Liabilities to Connected Banks"]
  D --> E["Compute Processes: Simulate Shock Propagation<br>via Cascading Defaults"]
  E --> F["Key Findings: Positive Correlation Found<br>Higher Link Probability = Higher Bank Solvency"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes a contagion model on a directed, weighted Erdős‑Rényi random network of 100 banks to study shock propagation and its effect on solvency.

Source sentence
In particular, financial network of 100 banks out of which 15 are shocked to default (that is, 85.00% of the banks are solvent) is modelled using Erdos and Renyi network -- directed, weighted and randomly generated network.

Assumes 5

  • The network is generated as a uniform (Poisson) random directed graph where each link is created independently with probability p. model
    Source sentence
    They apply random networks which have arbitrary degree distributions, while assuming a uniform (Poisson) random network in which each directed link is generated with independent probability p. They explore how aggregate and idiosyncratic shocks – wiping out any bank’s external ∗Department of Mathematics, Michael Okpara University of Agriculture Umudike,
  • Shocking a bank removes its assets and redistributes its liabilities to connected banks. model
    Source sentence
    Shocking some banks in a financial network implies removing their assets and redistributing their liabilities to other connected ones in the network.
  • The system consists of exactly 100 banks, of which 85 % are initially solvent. data
    Source sentence
    4 Contagion on Financial Networks: Implementation Details and Results In this section we use 100 banks (only 85% are solvent) as nodes and 2 ranges of probability values with 15 and 20 equally spaced intervals which determine the rate of having a link between any 2 banks.
  • Probability values for links are examined using two ranges with 15 and 20 equally spaced intervals. model
    Source sentence
    4 Contagion on Financial Networks: Implementation Details and Results In this section we use 100 banks (only 85% are solvent) as nodes and 2 ranges of probability values with 15 and 20 equally spaced intervals which determine the rate of having a link between any 2 banks.
Show 1 more
  • Random iterations (10 and 20 runs) are sufficient to assess average solvency, and finer grid sampling does not affect results. statistical
    Source sentence
    Random Iterations (b) Summary Statistics of Probability of Interconnectedness of Banks and Average Percentage of Solvent Banks from 20 Random Iterations From the summary statistics of mean percentage of solvent banks and different ranges of the probability values (cf. Figures (8a), (8b), (9a) and (9b)), there are no statistical differences in the means of 15 and 20 equally spaced intervals of the probability values.

Data

  • Synthetic Erdős‑Rényi network of 100 banks synthetic unknown
    100 banks
    Source sentence
    4 Contagion on Financial Networks: Implementation Details and Results In this section we use 100 banks (only 85% are solvent) as nodes and 2 ranges of probability values with 15 and 20 equally spaced intervals which determine the rate of having a link between any 2 banks. Let us examine data of average number of the solvent banks and probability values with 15 and 20 equally spaced intervals.

Costs

Not applicable The paper reports no trading or execution result.

Finds 5

  • The ranges of probability values and banks' percentage solvency have a positive correlation. statistical
    Source sentence
    Our major finding shows that the ranges of probability values and banks' percentage solvency have positive correlation. The range of probability values and banks’ percentage solvency have positive correlation.
  • Increasing network connectivity reduces the impact of shocks because the shock is shared among many banks. theoretical
    Source sentence
    Effects of such shocks are reduced if the banks are highly connected since all banks in the network share a little bit of the shocks; the shocks’ impacts become minimal.
  • No statistical differences were observed between using 15 versus 20 equally spaced probability intervals in terms of average solvency. statistical
    Source sentence
    Random Iterations (b) Summary Statistics of Probability of Interconnectedness of Banks and Average Percentage of Solvent Banks from 20 Random Iterations From the summary statistics of mean percentage of solvent banks and different ranges of the probability values (cf. Figures (8a), (8b), (9a) and (9b)), there are no statistical differences in the means of 15 and 20 equally spaced intervals of the probability values.
  • The fineness of the probability‑value grid and the number of contagion trial samples do not affect the evaluated percentage solvency or default. statistical
    Source sentence
    Hence, we deduce that the number of grid points – fineness of the intervals of the probability values, has no effect on evaluation of percentage solvency or default of the banks. The number of sampling of the contagion trials does not affect the average contagion saturation.
  • Complete network structures can withstand economic crises by evenly redistributing shock impacts, resulting in no subsequent financial contagion. theoretical
    Source sentence
    Their findings reveal that complete network structure has capacity of withstanding economic crises because impact of such shocks to the system can be evenly redistributed among other banks. No financial contagion results afterwards.

Checked by others

Contradicted by

  • Systemic risk measured by systems resiliency to initial shocks (2023)
    On the question: How does increasing network connectivity (density p and size n) affect the impact of financial shocks?
    Compare the findings

    This paper: Increasing network connectivity reduces the impact of shocks because the shock is shared among many banks.

    That paper: Increasing network density (parameter p) and size (n) systematically reduces the allowed maximal uniform shock for all parameter values.

    A finds higher p and n lower the maximal uniform shock tolerated, while B finds higher connectivity reduces shock impact; the conclusions conflict.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.

Cited References (44)

  1. Financial contagion (2000)
  2. Recent developments of constructing adjacency matrix in network analysis (2014)
  3. What is network science? (2013)
  4. Analyzing social networks (2018)
  5. Graph Theory, Fifth Edition (2017)
  6. Resilience to contagion in financial networks (2016)
  7. Contagion in financial networks (2010)
  8. Solvency distress contagion risk: network structure, bank heterogeneity and systemic resilience (2021)
and 36 more
  1. The" no significant difference" phenomenon: A literature review (2002)
  2. The Cox-Ross-Rubinstein Option Pricing Model
  3. Fundamentals of actuarial mathematics (2014)
  4. Coherent measures of risk (1998)
  5. Introduction to the mathematics of finance: from risk management to options pricing (2004)
  6. Estimation of geometric Brownian motion parameters for oil price analysis (2017)
  7. Probability: theory and examples (2019)
  8. Transport in Nigeria https://en.wikipedia.org/wiki/ (2023)
  9. Life Insurance Mathematics with exercises contributed by Samuel H: Cox (1997)
  10. Stochastische Finanzmärkte: Quantitative Finance Lecture Notes, Summer Semester 2022 TU Chemnitz (2022)
  11. Calculating and comparing security returns is harder than you think: A comparison between logarithmic and simple returns (2015)
  12. Brownian motion (1980)
  13. Stochastic calculus and financial applications (2001)
  14. Predicting real growth using the yield curve (1996)
  15. Geometric Brownian motion, option pricing, and simulation: Some spreadsheet-based exercises in financial modeling (2012)
  16. Paul Wilmott on quantitative finance (2013)
  17. On the evolution of random graphs (1960)
  18. Random graphs (1959)
  19. On the strength of connectedness of a random graph (1961)
  20. Network Science (2016)
  21. Likelihood-based model selection for stochastic block models (2017)
  22. Stochastic blockmodels: First steps (1983)
  23. The structure and function of complex networks (2003)
  24. Properties of Random Networks (2022)
  25. Graph clustering (2007)
  26. Clusters, networks, and innovation (2005)
  27. Erd "o s-R 'e nyi Poissonized (2022)
  28. Random walks on graphs (1993)
  29. Spreading on Networks (2019)
  30. Diffusion and random walks on graphs (2015)
  31. Infinite-degree-corrected stochastic block model (2014)
  32. Fast random graph generation (2011)
  33. Erdős–Rényi model https://en.wikipedia.org/wiki/ (2023)
  34. Random walk models in biology (2008)
  35. The multiplex nature of global financial contagions (2020)
  36. Financial contagion: the viral threat to the wealth of nations (2011)

0 of 44 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.