IFRS 9: Shorten your reporting cycle by leveraging R in SQL Server

IFRS 9 is one of the biggest endeavors of recent years for financial institutions. Using programming languages such as R or Python directly into the database space opens new perspectives on how your data is processed for regulatory purposes.
accentureinsights-ifrs16-header-01.jpg

Rooted in the 2008 financial crisis, IFRS 9 (International Financial Reporting Standard 9) remediates historic issues by setting a new unifying accounting language for financial assets. Effective since January 2018, it replaces IAS 39, the previous accounting reporting standard, and aims to provide clearer classification, simplified regulation and increased transparency and consistency on financial instruments, especially with regards to the recognition cycle of Expected Credit Losses.

IFRS 9 implementation consists of three phases:

  1. Classification and measurement: to define a new classification methodology for financial assets and liabilities;
  2. Impairments: to improve financial reporting regarding provisions for Expected Credit Losses, taking into account credit quality, forward looking macro-economic information and point-in-time estimation;
  3. Hedge accounting: to reflect an accounting treatment more in line with risk activities.

Three stages of loan

IFRS 9 defines three possible stages for a loan: Stage 1, which recognizes a 12-month Expected Credit Loss for all loans; Stage 2, which recognizes a Lifetime Expected Credit Loss for loans experiencing a significant increase in risk since their initial recognition; and Stage 3, which recognizes a Lifetime Expected Loss for loans experiencing a loss event. 

This new approach and IFRS 9 requirements with regards to implementation, such as including macro-economic drivers, implies the development of new impairment models. More concretely, these new models must be plugged into legacy data systems because their outputs are needed to source the new IFRS 9 disclosure reports.

Architecting for new impairment models 

From an architectural implementation perspective, one could consider deploying the new impairment models and related reports via a fully integrated suite such as SQL Server, hence covering the data, the modelling and the reporting requirements of IFRS 9.

Newer versions of database systems enable running R scripts ‘in-database’, meaning that the server and its resources can now carry out complex operations such as Expected Credit Loss calculations without having to move delicate data outside its data environment. Consequently, all the calculations from the ECL models are then readily available upstream for any internal and external reporting needs of firm related to the new regulation.

R is an open-source programming language and statistical computing environment with an established community in the academic world and which is now making its way to the corporate world. Until recently, R was mainly used as a prototype tool. However, it offers a lot of advantages: a strong knowledge base, pre-built and publicly available complex statistical libraries, big data and distributed computing capabilities, among others...

Seamless integration and faster reporting 

From a development and deployment standpoint, R integration in SQL Server is pretty seamless: R code is shared and versioned through the usual project code management tool; and any R script, when ready and tested, can be deployed in and run from a SQL stored procedure.

With regards to IFRS 9, using R in SQL Server means impairment models can be run ‘in-database’. The calculated Expected Credit Losses can thus be seen as any other data step, pushing data forward into a dedicated database designed for reporting purposes. Subsequent reports are then built using traditional reporting tools such as SQL Server Reporting Services or Power BI. Ultimately, having a fully automated data pipe, with no external step specific to the modelling, the firm's IFRS 9 reporting cycle is streamlined and shortened. 

Eventually, this will also facilitate the modelling process from prototyping and validation to deployment because data scientists can carry out both parts of the workflow. In a world where quantitative analytics is becoming a key differentiator for success, mastering these technologies can mean a competitive edge.

Want to know more about how Accenture can help your company manage the challenges and opportunities of IFRS 9? Don’t hesitate to contact us for a chat!

If you missed our first two episodes of our 'Data at the Service of Regulatory Reporting' series. It's all here!

Authors: Charles Marsily Nesrine Besbes