Beyond the Spreadsheet: A Better Way to Build Audit-Ready Macroeconomic Stress Scenarios

How KidbrookeONE turns macroeconomic forecasts into transparent, model-guided scenario workflows for credit risk and regulatory stress testing

Most risk teams building a stress testing framework arrive at the same point eventually: they already have a reasonable macroeconomic forecast, but no defensible way to turn it into the downside, central and upside scenarios their process actually needs. The starting baseline might come from an official national forecast, a board-approved set of assumptions, or the institution's own economics team. Wherever it comes from though, the harder task is turning that single baseline into coherent stress paths. A spreadsheet can move a policy rate or an inflation figure up or down by hand, but it rarely captures the historical relationships between variables, the uncertainty around each path, or the evidence trail an auditor will eventually ask for.

Where A Good Forecast Still Leaves A Gap

That is the gap KidbrookeONE's Economic Scenario Generator, or ESG, exists to close. The ESG already underpins pension forecasting, investment advice and portfolio analytics across the platform, running thousands of Monte Carlo simulations to produce the probability-based outcomes behind every percentile forecast it generates. Extending it into regulatory and credit risk stress testing meant adding a more granular macroeconomic layer, the central bank policy rate as a standalone variable, and a flexible inflation measure that each institution can align to its own internal definition, and richer GDP and unemployment models built to support structured stress work rather than generic market assumptions. Baselines typically run across a five-year horizon, a practical starting point built to extend as an institution's methodology matures.

Stress Severity And Scenario Weighting Are Two Different Decisions

The harder problem is conceptual rather than statistical, and it is where a surprising number of stress testing processes quietly lose their audit trail. Whether the context is IFRS 9 expected credit loss modelling, an internal capital adequacy process, or a regulatory stress test submission, institutions face the same structural question: how severe should each scenario be, and separately, how much weight should it carry? Those are distinct governance decisions. The scenario engine's job is calibrating severity, using a statistical framework grounded in the full distribution of simulated outcomes rather than an analyst's intuition about what "downside" ought to look like.

The weighting decision, or how much each scenario path contributes to the final ECL number, is a judgement call that sits with the institution's governance process. It might reasonably change from one reporting period to the next depending on the economic outlook or the institution's risk appetite at the time. What matters is keeping that decision separate from the question of how severe each scenario path actually is.

When the two get mixed together, problems follow. If an analyst adjusts the severity of a downside path to hit a particular ECL output, rather than building it from statistical principles and then letting governance decide how much weight it carries, the scenario becomes hard to explain and harder to defend. A regulator or auditor asking "why is this your downside?" deserves an answer grounded in the data, not one that quietly traces back to a target number someone had in mind.

Our approach keeps the two decisions cleanly apart. Scenario severity is set using percentile levels drawn from the simulated distribution, which means each path has a consistent, statistically grounded meaning from one period to the next, regardless of how market conditions or variable correlations shift over time. The governance process can then adjust its weighting on its own terms and timeline, without needing to revisit how the underlying scenarios were constructed.

From A Back-End Engine To An Analyst's Workspace

None of that modelling work matters if the people responsible for the process cannot actually use it. Historically, the Economic Scenario Generator (ESG) lived entirely behind the API: it produced numbers, but no one could open it up, question an assumption, or watch a scenario take shape. Alongside the modelling extensions, Kidbrooke built an interface that gives analysts a guided workflow instead of a black box. A user starts from their own baseline, locks whichever variables should be held as fixed assumptions, and lets the model project the rest forward using the statistical relationships built into the ESG, working through policy rate, inflation, GDP and unemployment in sequence. At every step, a single control sets how much the model's view shapes the path versus how much the analyst's own judgement does, and the projected path responds immediately so the analyst can see the effect of that choice rather than guess at it. The same workflow lets them interrogate the correlations driving a path and explore how one variable moves another, instead of taking the relationship on faith.

What Leaves The Building Has To Survive An Audit

That control is also what makes the output defensible. Each scenario set carries a visible record of which baseline it started from, which variables were anchored, how much model influence was applied at each step, and which percentile level was selected. Analysts can review fan charts, compare paths side by side, and export the finished scenarios as structured CSV files ready for direct ingestion into a downstream model, with dates, scenario names, variable values, sources and version information attached. When a risk committee or an external auditor asks why a particular stress path looks the way it does, the answer sits in the exported file rather than in someone's memory of a spreadsheet edit from three reporting cycles earlier. That trail matters as much for internal governance as it does for external scrutiny, since the committee signing off on the numbers needs the same clarity about how each path was built.

The Same Engine, A Different Argument

That same analytical engine that powers pension forecasting and investment advice on the platform is the one now stress-testing loan books and informing regulatory capital submissions. The variables change but the engine doesn't. Improvements made for one context, the policy rate modelling and inflation handling built for this work, strengthen every other ESG-driven forecast across the platform. That is how a modular platform should behave: each extension pays for itself more than once.

If your institution is wrestling with macroeconomic scenario construction, whether for credit risk, regulatory stress testing, or internal capital planning, we would be delighted to talk through what we have built and where it might fit your own process. KidbrookeONE is designed to be adopted piece by piece, so this kind of capability does not require touching everything else first.