IFRS 9 came into force on 1 January 2018. It is fair to say that the implementation projects have generally taken rather longer than planned, and the long-term impact remains unclear, but broadly speaking, banks can be said to have managed the process reasonably well. However, with implementation costs running into the millions of dollars, many banks are now looking to capitalize on their investment in IFRS 9 compliance and use it as an opportunity for the future. What, then, is coming next for these banks?
Options and opportunities
One obvious option is to optimise the credit risk cycle by using the results of IFRS 9 calculations to inform future thinking. By simulating ECL impacts, analysing model quality and appropriateness for different scenarios and modelling approaches, and selecting the best ones to implement, banks can now determine the best models to use within their credit risk cycle.
However, IFRS 9 implementation should also have resulted in more and better data availability, so it is worth considering how this might be used. The new infrastructure can be used to collect appropriate data to develop and calibrate a wide range of individual models. These can then be deployed in real-time to support a number of processes, including stress testing, asset and liability management, and capital calculations.
IFRS 9 has also changed the requirements for data architecture and data quality, providing opportunities as well as implementation challenges. This has resulted in an increase in input sources, deeper historical data needs, and requirements for new orchestration to incorporate new information such as cash flows and risk factors. In practice, IFRS 9 has generally led banks towards automated data supply from the beginning to the end of the process. This will follow other domains and lead to an interconnection with existing platforms of both risk and finance, which again opens up new avenues.
More and more banks are moving their activities into a flexible, cloud-based approach. This allows them to calculate complex models in a shorter time frame, and to understand and work with the model results, using the same resources for these very different tasks. This, however, is only the tip of the iceberg.
Artificial intelligence, machine learning and risk management
One topic that is becoming more and more relevant is the use of artificial intelligence and machine learning in risk management. We have already seen some early work using robotics and validation. The automatic building of models is still at a very early stage, and these projects remain highly dependent on individual human resources to provide suitable skills and intelligence, but there is considerable promise.
Risk management efficiency can be a competitive advantage. A reliable and integrated view for managerial purposes requires trust across the organisation through a proactive and systematic management of risk exposures. Join our upcoming webinar to discover the better path to Governance, Risk and Compliance
EY recently reported on a first experimental project in this area at an Italian bank. The bank uses automation and machine learning to manage time-consuming tasks like variable selection and loading. It also uses the AI system to select the best model, including a set of pre-charged criteria in the modelling. These system recommendations, of course, must be compared with existing models, but this leads to optimisation by checking a larger number of models in a much shorter time frame.
All these are examples of what is possible if you start to see compliance with new regulatory standards as an opportunity to re-examine your systems and processes, rather than a threat. Compliance can always be done as a minimum, but it seems sensible to capitalise on your investment and try to get more out.