Banking & market infrastructure
Tracing contagion through a counterparty network
Where an aggregate stress test reports a loss, a network simulation reports the route — and the point at which the response stops being linear.
Illustrative example
The question
A stress scenario produces a system-wide loss estimate. Which institutions transmit it, in what order, and where does a marginal effect become a non-linear one?
Approach
01 · Specify
Institutions represented as heterogeneous agents with balance-sheet state, exposure links, funding constraints and an explicit deleveraging rule. The rule set is extracted from supervisory documentation into a reviewable specification before any code is written.
02 · Calibrate
Exposure structure and balance-sheet distributions fitted to published aggregates using approximate Bayesian computation, so the simulated population reproduces observed concentration rather than assuming it.
03 · Validate
Sensitivity sweeps across the deleveraging threshold and the network density parameter, plus plausibility checks against the historical episodes the scenario is modelled on.
Output
- A transmission map: which nodes propagate the shock and in what sequence
- The threshold region where the aggregate response becomes non-linear
- A sensitivity boundary showing which assumptions the conclusion depends on
Illustrative application. No client engagement, institution or result is described.