top of page
ClsoeIT Logo

IFRS9 Provisions Plausibility Engine

New IFRS 9 methodology

In early 2018, a new methodology for the calculation of provisions for Expected Credit Losses (or ECL) under IFRS9 for the banking sector came into force. Our team was tasked with implementing a solution for the retail credit exposures of Raiffeisen Bank International (RBI). The calculation is used to check the plausibility of results provided by each RBI subsidiary bank. It also serves as a reference implementation in the initial phase of the IFRS9 project and in later production deployment.

"We decided to go into cooperation with CloseIT after several introductory sessions, where they convinced us both in terms of the proposed technical solution and in terms of understanding our business needs. We appreciated the overall perspective and ability to design individual components so that they later fit into a working suite without interfering with the design process." Deyan Ivanov, Head of Retail Risk Analytics & Methodology at Raiffeisen Bank International.

Our solution for all RBI banks

In addition to the rigorous implementation of IFRS 9 ECL calculation, the use of open technologies and the application of standards that provide support and future scalability were important criteria. For an international bank with a large number of clients, the satisfying speed of calculation on relatively common modern hardware is expected.

One of the challenges was the diversity of the risk models environment within each RBI subsidiary bank. Typically, several tools are used to develop IFRS 9 macroeconomic models, such as SAS Data Miner, SAS Base, as well as Python or R. To consolidate the results of the whole group, macroeconomic overlay factors must be applied consistently across the portfolio. 

"We were looking for a solution that adapts to the diversity of the entire RBI group. The result is a unique integration that unites the whole Group." Deyan Ivanov, Head of Retail Risk Analytics & Methodology at Raiffeisen Bank International.

This solution consists of developing a language that describes the mathematical functions of macroeconomic models. This language is easy for humans to read, but models can also be exported from external systems. For this language, we have created a compiler and runtime environment to calculate macroeconomic overlay factors by defined scenarios. The factors are stored in a central database for subsequent use in ECL calculation. Thanks to this solution, it is possible to unify individual application portfolios and work with them together - this proved important in the subsequent implementation of stress tests.

Functional prototype in 3 months

Tool architecture is designed for ever-changing demands and new challenges. Individual, independent components communicate with each other through an established interface, so they can be easily modified or exchanged to accommodate, for example, changes in source data, risk parameter format, etc.

Within only three months, we progressed from the start of development through the first, minimally viable product, complete architectural design, user verification, solution implementation, and ending with a function prototype. The regular two-week review, where progress in development was presented, including release to a customer server with consultations with a team of RBI experts, was the basis for success.

"This was a clear example of agility. CloseIT was presenting a steady development in regular iterations as the tool was progressively taking shape. The speed of development was impressive. From zero to functional IFRS 9 ECL prototype in 3 months!" Deyan Ivanov, Head of Retail Risk Analytics & Methodology at Raiffeisen Bank International.

Part of the solution beyond the technical specification was the creation of a modern user interface in the form of a web application. Here, the user can easily check the calculation status of each portfolio using clear progress bars, track their results, manage calculation parameters, and create, or modify, macroeconomic models for each portfolio. The user can modify the model parameters or models themselves and compare their outputs in real time. In order to implement various what-if analyses, we have added the ability to simply save the entire configuration to the user's disk and reload it in a different environment - similar to saving a position in a video game :-) 


"The final result is, in terms of UX and graphics, perfect, very intuitive, and meets the high requirements of current standards." Deyan Ivanov, Head of Retail Risk Analytics & Methodology at Raiffeisen Bank International.

IFRS9 impairment calculation overview in Model Governance Suite

CCF Macro Models in Model Governace Suite

IFRS9 impairment calculation event log in Model Governance Suite

IFRS9 impairment calculation maintenance in Model Governance Suite

PD Macro Models - Bosnia and Herzegovina subsidiary visualization in Model Governance Suite

PD Macro Models - Bosnia and Herzegovina subsidiary graphs in Model Governance Suite

The result? Reliable software

Raiffeisen Bank International has received a reliable tool in a timely manner, the use of which allows the verification of the implementation of new calculation methodologies in individual banks. The software was immediately installed and integrated on client servers where authorized users can access it. The tool is now being used extensively.

"Since the beginning of the collaboration, I was surprised by the involvement of CloseIT in the whole problem. They focused not only on software or custom design solutions, but they tried to incorporate all our requirements into the overall solution, especially from the business point of view."

Deyan Ivanov, Head of Retail Risk Methodology & Validation at Raiffeisen Bank International

Related Posts

See All

MGS integration with antivirus

One of the MGS features is to manage model-related files and documents. Of course, common and other model non-related files can be uploaded also to the public folder, users' private folder, or shared

MGS to Databricks connector

Analytics teams in banks and insurance companies use platforms like Databricks for data science, ML/AI model development, model testing, and more. These same companies also have teams such as Senior R


bottom of page