Systemic Financial Risk Model
A public flagship research build focused on network contagion, stress propagation, and the hidden fragility of connected financial systems.
Project overview
This is the flagship public research build on the site. The goal is to model how shocks move through connected financial institutions and why local weakness often matters less than network structure once pressure rises. In plain language: a system can look healthy in calm periods and still be dangerously fragile if funding, exposures, and incentives line up in the wrong way.
The current page is intentionally honest about its stage. It includes a methodology sketch, illustrative synthetic sample data, and a stress chart designed to show the shape of the problem. It is not yet presented as a finished empirical paper. That is deliberate. A credible public project gets stronger when it shows its scaffolding clearly instead of pretending to be fully complete on day one.
Important: the downloadable CSV on this page uses illustrative synthetic data. It is included to show structure and scenario logic, not to make fake empirical claims. Replace it with real public-source inputs before presenting results as research findings.
Core research question
How does a localized shock become system-wide fragility once institutions are connected through short-term funding dependence, correlated exposures, and network centrality?
Working model logic
- Represent institutions as nodes in a connected network rather than as isolated balance sheets.
- Track characteristics that make stress more dangerous: capital ratios, dependence on short-term funding, and network centrality.
- Run multiple scenarios instead of one static “risk score.”
- Compare how the same institution behaves in calm conditions versus funding shocks, asset shocks, and combined shocks.
- Use interpretable outputs so the project stays legible to non-specialists while remaining analytically serious.
Illustrative scenario chart
The chart above is a first-pass demonstration: institutions with high short-term funding dependence and high network centrality show much sharper deterioration in the combined-shock scenario.
Downloads
Next milestones
- Replace synthetic data with real public-source institutions and documented provenance.
- Push the notebook, code, and visualization logic to GitHub.
- Add a cleaner explanation of network topology assumptions and propagation mechanisms.
- Write a summary article that journalists and recruiters can understand in under five minutes.
- Publish a stable external version on LinkedIn, Medium/Substack, or SSRN that links back to this canonical page.