Turning a monthly monitoring pack into a one-parameter run
A UK bank asked us to build the monitoring pack for its LGD models and get it running each cycle.
Recalculating actuals against expected by hand each month left the pack arriving too late for anyone to act on it.
What We Provided
A defensible read on whether the models still work
The book runs through the models, realised recoveries are calculated over the period, and stability and discrimination are tested alongside actuals against expected. Every headline number comes from a repeatable calculation.
A pack its readers already knew how to read
Content was settled in Excel, where it is quick to argue about, then ported to Power BI with the reporting team without losing the structure its readers recognised.
A monthly cycle that costs a parameter, not a week
The scripts were productionised so that a single run-month parameter regenerates the outputs, taking manual assembly out of the monthly cycle along with most of the room for error.
When something moves, knowing where to look
Sub-model analysis behind the pack localises a drop, and each component is graded by how far it can move the output, so a flag arrives with a shortlist attached.
The Results
The pack regenerates from a single run-month parameter, so the numbers arrive early in the review window. Analyst effort goes into the exceptions it flags rather than into assembling it, and the diagnostics point to the component to check first.
Other Case Studies
Independent validation of a Pillar 2A concentration add-on
A European bank needed to know that the capital it holds against concentrated lending would stand up to scrutiny, so it asked us for an independent view of the model behind the number.
Redeveloping every IFRS 9 model across two portfolios
A UK lender needed every model in its IFRS 9 expected credit loss suite rebuilt, at the point when the team that had built them was no longer there.