Overlays, or post-model adjustments (PMAs) is a term that is used to describe a spectrum of adjustments that are made outside the primary models. Banks often use post-model adjustments, where risks and uncertainties are not properly reflected in existing models. The objective of a post-implementation review is to assess shortcomings where models or data have limitations. A post-implementation review can identify improvements that can be made to a new requirement, to the standard-setting process or to the structure of an IFRS Accounting Standard. In-model adjustments and overlays exist in many forms and configurations.
What does Overlay Mean
Overlay refers to straightforward adjustments in order to correct known model errors or data deficiencies. They are subjective in nature. It sometimes refers to those adjustments which are capable of capturing risks and uncertainties that are not tracked by the models. It is usually applied at an aggregate level. These adjustments are also generally not expected to be material. Overlays may arise due to events for which there is insufficient time to reflect important model’s events, or granular adjustments. It may also arise during some exceptional economic conditions. Especially when the modelling and methodological approaches cannot be adapted at more granular levels to adequately reflect the exceptional conditions.
Types of Adjustments
When estimating ECL (Expected Credit Loss), banks make modelled and unmodelled adjustments to the outputs that are derived from the core models. These adjustments sometimes include both significant management judgements and estimation uncertainty. Some commonly used adjustments are:
Judgemental adjustments – these adjustments to the ECL estimate are made outside of the bank’s regular modelling process. It can change the amounts to reflect management judgements. Changes made to these assumptions underlying these judgemental adjustments can affect ECL within the next 12 months. Therefore, the pattern of utilization or release of these adjustments is likely to involve significant management judgment.
Post-model adjustments – PMAs are the adjustments to the ECL model output, which are usually calculated at a granular level. These are performed via modelled analysis and are allocated to provisions at a granular level. PMAs are incorporated in credit risk disclosures. These are calculated separately for each economic scenario.
Overlays – overlays are those adjustments made to the ECL model outputs that have been made outside the detailed ECL calculation and reporting process. Model adjustments which are made as a result of data limitations are not expected to be captured within the definition of judgemental adjustments. This is due to the lower level of judgement that they require.
Prerequisites for establishing Overlays
Overlays require internal control, governance to function properly coupled with transparent documentation. There are some pointers that should be kept in mind before establishing overlays. Some of them are mentioned below:
- What type of limitations are being addressed
- What kind of data was used, how was it consistent with similar data
- What plans have been laid out for overlay usage overtime, via model development/redevelopment, loan-level losses
- How will staging implications be addressed
- Has the overlay been reviewed, along with end-to-end ECL modelling process, to ensure that there is no double-counting
- What kind of inputs or review has the committee provided
These questions help in ensuring adequate processes to prevent challenges over time.
Need for Overlays
Due to Covid-19, many banks have expanded the frequency and scope of management overlays for impairment. The increased use of management overlays has highlighted gaps in governance and controls required to ensure ECL estimation is measured consistently across portfolios and over time. Leading banks offer a clear link between underlying model limitations and/or reasons for adjustment and each management overlay. Each of the model overlay is justified and subject to independent challenge. They also have clear-cut criteria mentioned for the overlay to reduce or change over time.
PMAs are a specific type of management overlays. These include targeted recalibration, along with additional mini-models or specific expert-based assumptions. PMAs offer a short-term fix to ECL calculations, whilst model recalibration and rebuild options remain limited. Policy makers are rightly pushing firms to remediate the root causes of PMAs, by enhancing models. Where this cannot be done in a timely manner, the monitoring, analysis and documentation of PMAs will retain a central role in the ECL control framework.
Top banks actively manage the portfolio of PMAs, keeping the scope, duration and conditions for unwinding management overlays under constant review. Decision-makers need to understand that PMAs do not remove the chances and requirements to improve models in the long term. Management overlays increase the reliability of ECL impairment levels in the short term.
On the other hand, over-reliance on PMAs in the long term creates costs and regulatory risks. These are neither sustainable, nor justifiable. Many top-tier banks are thus working to establish model development plans to enhance the granularity and sophistication of IFRS9 ECL model suites, to address model limitations uncovered by the crisis.
Conclusion
Traditional IFRS9 models sometimes miss out on reflecting actual credit risk expectations. This is a vital component for novel risks in a rapidly changing environment. Thus, the overlays have a role to play here. They are extremely useful tool when it comes to addressing emerging risks. Some in-model adjustments quantify the risk impacts by using scenarios and simulations. Nowadays, inflation risk and environmental risks are becoming crucial factors while determining expected credit loss. Effective use of in-model adjustments and evidence-based overlays help banks to quantify risk, and assess the right levels of capital to set aside in case of emergencies.
