IFRS 9 as a standard has brought in transformation in the way banks and financial institutions have been operating . The fundamental change has been in terms of ex-ante recognition and disclosure of the risk as opposed to ex-post disclosures under IAS 39. It has forced financial institutions to be more prudent and transparent in the way they deal with public money for managing their businesses and focus on quality of assets in an otherwise marketing focused asset expansion regime.The adoption has come with its own set of challenges the biggest one being collection, storage, retrieval and transformation of internal and external data which forms the backbone of the risk models. As an outcome, the IFRS 9 implementation has for majority of the banks resulted in increased provisions.
We have advised multiple banks in managing their increased provisions due to implementation of ECL calculation framework under IFRS 9. While in a traditional credit context, the focus is on identification of non performing assets beyond 60 and 90 days, the game has changed in the IFRS 9 context. It is too late to identify and cure these assets as provisions will typically be higher with assets being classified as stage 2 and stage 3. Provision calculations alongside stage deterioration is prohibitive.
Our experience suggest that it is critical for the banks to the do the following to cope with this challenge:
– Quick classification of assets at early stage (i.e. pre- delinquency). We have our proprietary models for early warning signs on consumer portfolio, which has generally not been an industry norm
– Optimizing efforts by identifying self curers. This ensures more collection capacity is available for high risk customers
– Identification of high risk customers who have higher chances of straight rolling into higher delinquency buckets. Our proprietary machine learning methods have relied on non traditional variables that focus on customer spend data, utility consumption and payment, customer’s banking behaviour beyond historical delinquency. These have helped in not only identifying the high risk customers, but a segmentation based on reason for riskiness helps the collectors in better resolution
Our engagement with banks in Middle East region suggests that a fairly protective environment based on salary transfers and restrictive criteria for expatriates needs to be enhanced so as to enable the banks to acquire better customers on one hand and also keep a tab on the provision numbers on the other. Techniques relying on machine learning and big data can help in reducing provisions by more than 10%
To know more about our experience, feel free to contact us at ag@sutra-management.com
Download full article in pdf here.
About the authors
Abhishek Gupta
Abhishek is Managing director at Sutra Management Consultancies. He has expertise in setting up analytics and CRM capabilities for various organizations. He has provided strategy advisory on risk and customer management for banks, insurance, Telecom and governments.
Shravan Shrikrishna Potnis
Shravan is a director at Sutra. Shravan has experience of over 17 years in the area of Risk advisory and big data management for banking, government institutions and insurance clients
Saurabh Assat
Saurabh is a senior project manager at Sutra. Saurabh has worked on over 40 engagements entailing risk management, IFRS9 and big data management engagements