ECL Modelling
We build the expected credit loss models that determine your provisions, allowance for credit losses included, or independently validate the models you run.
IFRS
9
The provisions you report and the scenarios behind them, from the loss models through to the year end disclosure.
IFRS 9 Deliverables
Economic Scenario Forecasting
We generate the multiple scenarios and probability weights an expected credit loss calculation requires, including around your own central forecast.
Climate Risk Scenario Forecasting
We translate physical and transition pathways into the macroeconomic variables your provisioning models consume.
Automated Financial Disclosures
We generate the figures, notes and supporting evidence your auditors and supervisors require at year end directly from model output, so the year end pack rebuilds rather than being assembled by hand.
Monitoring Dashboards
We build live dashboards reporting current provisions, coverage and stage migration to the board.
Banking Case Studies
Banking
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.
Banking
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.
Frequently Asked Questions
-
Expected credit loss models are judged on whether a reviewer can reproduce the figure from the documentation, so ask any bidder to show you a model document rather than a methodology deck. At Gini we build the twelve-month and lifetime probability of default term structures, loss given default and exposure at default components, the staging framework, and the macroeconomic scenarios and weights behind them. We also validate models you already run.
-
Yes, and overlays are usually where a review earns its cost. Gini tests whether each post-model adjustment is still needed, whether it can be quantified from data now that the event it covered has run its course, and whether it double-counts something the model already captures. On staging we test the significant increase in credit risk trigger against subsequent default experience, since a threshold that never fires and one that fires constantly both fail for the same reason.
-
At Gini we calibrate the relationship between each economic variable and your own default and loss experience rather than importing an elasticity from elsewhere, then document the weighting judgement and show what the provision does under each scenario separately. The sensitivity of the reported figure to the weights is disclosed, because that is the first question an auditor asks and the one firms are least often ready for.
-
That is the standard Gini builds to. Models arrive as version-controlled code with tests, so the figure in the disclosure can be traced back to the data that produced it without anybody reverse engineering a spreadsheet. Every judgement is recorded with the evidence considered and the reason for the call, including the judgements we were least comfortable making.