As the IFRS 9 deadline draws closer, it is increasingly imperative that all affected lenders are in the process of switching to the new regulatory standard, due to the requirement to run the new models in parallel with existing processes. To help smaller lenders meet the regulation and larger lenders validate their models, Jaywing recently announced the launch of Horizon, our new IFRS 9 automated modelling software.
In this blog, we outline six key ways that Horizon can help you to solve IFRS 9.
1. Model development and execution can be undertaken within one environment
Horizon deals with all aspects of IFRS 9 impairment modelling using feeds of a lender's own data and macro-economic data. The model engine, developed using advanced machine learning techniques, produces granular forecasts of Probability of Default (PD), Exposure at Default (EAD) and Loss Given Default (LGD) components to calculate life time Expected Credit Losses (ECL) under different economic scenarios.
2. User-friendly interface allows users to adjust model parameters
Horizon offers lenders a user-friendly interface to adjust the models to their individual requirements while providing the necessary transparency for all components of the solution. More importantly, it affords lenders the opportunity to avoid lengthy and expensive projects or significant ongoing maintenance overheads while still meeting regulatory requirements by the deadline.
3. Deep industry knowledge and expertise embedded
With eleven UK lending organisations using Jaywing for their IFRS 9 implementations and countless previous IAS 39, Basel, Stress Testing and other banking regulation analytics engagements, our expertise in this field is peerless. Our data science credentials and ability to manage and corral big data have enabled us to develop an accessible product for lenders least likely to have access to quality skills and experience in-house or even through contractor resource.
4. Automated models developed using advanced data science that are constantly refreshed
The automated models are built based on your own data, coupled with a series of parameter settings giving you control over the way the models are constructed, including response to economic factors. Jaywing’s advanced machine learning algorithms then develop models that are intuitive (including exhibiting common-sense macro-economic sensitivities) and robust, according to those specifications. This allows a broad range of end-users to make use of their business knowledge and produce the necessary models, without requiring a background in statistics. Refreshing the models is a breeze, so it’s easy to ensure they are always up-to-date.
5. Just enough data
Horizon takes feeds of simple, aggregate performance data to produce its models, meaning that there is no requirement for bulky file processing or transfers. The inputs needed are similar to dynamic delinquency reports that are familiar to all credit risk professionals and so can be created with minimal adjustments to existing processes.
6. Supplemental data
IFRS 9 creates significant requirements for lenders in terms of the depth of history and volume of defaults available in historic data. Jaywing recognises that not all lenders have the data that is needed, and so work with data partners who can provide representative historic data for use with modelling in Horizon. This enables models to be produced on a robust historical dataset and then applied only to the lender’s own accounts to produce their provision numbers.