How Simulation Can Strengthen Credit Risk Modelling

Risk teams developing stress-testing frameworks often begin with a credible macroeconomic baseline, such as a national forecast, a board-approved outlook or projections from an internal economics team. The challenge is turning that baseline into downside, central and upside scenarios in a way that preserves relationships between variables, reflects uncertainty and can be clearly explained to auditors and regulators.

Kidbrooke’s Economic Scenario Generator is designed to support that process through stochastic modelling and Monte Carlo simulation. The engine, already used across the company’s WealthTech platform for risk assessment and cash-flow modelling, generates thousands of possible economic paths to produce structured and explainable scenario outputs.

The tool has now been extended for regulatory and credit risk stress testing, with more detailed modelling of macroeconomic variables. These include a separately modelled policy rate, configurable inflation measures, and broader projections for gross domestic product and unemployment. Baseline scenarios are typically set over a five-year period, although the framework can be extended as modelling requirements develop.

Separating scenario severity from weighting

A central feature of the approach is the separation of scenario severity from scenario weighting. This is particularly relevant for institutions using IFRS 9 expected credit loss models, which require probability-weighted, forward-looking estimates of potential losses.

Rather than adjusting a downside scenario manually to achieve a desired result, Kidbrooke’s framework links severity to percentile levels within simulated distributions. The probability assigned to each scenario remains a separate governance decision, allowing institutions to change weightings without altering the methodology used to construct the scenarios.

This distinction can help improve transparency and reduce the risk of scenarios becoming difficult to justify during model validation, audit or regulatory review.

Combining modelling with expert judgement

The system also includes a guided workflow that allows analysts to begin with their own baseline assumptions, fix selected variables and project the remaining factors using the model’s statistical relationships. Variables such as policy rates, inflation, GDP and unemployment can be developed in sequence, while users can adjust the balance between model-generated outputs and expert judgement.

Analysts are also able to inspect the correlations behind each projection, rather than treating the model as a black box. Each scenario set creates a record of the original baseline, fixed assumptions, level of model influence and selected percentile, with structured outputs available for use in downstream risk and validation models.

For operational research practitioners, the development highlights the role of simulation, scenario analysis and structured judgement in financial decision-making. It also reflects a wider challenge across risk modelling: how to combine analytical rigour with transparency, governance and practical usability when decisions must be made under uncertainty.


References:

https://fintech.global/2026/07/16/why-stress-testing-scenarios-fail-audits-and-kidbrookes-fix/

https://kidbrooke.com/solutions/scenario-projection-engine

Operational Research News

Operational Research News

Stay Ahead with the Latest in Operational Research – Dive In Now!

Read more

Join The Operational Research Society

Join The Operational Research Society

Join Today! As a member of the Operational Research Society the door opens to a world of opportunities and resources that can propel your career forward.

Read more

Sign Up As A Volunteer

Sign Up As A Volunteer

Become an operational research volunteer, support charities, and help promote analytics and decision science in your community.

Read more