To help rebuild the economy, financial institutions have an essential role to play in the supply of credit. However, there are many challenges to address to mitigate additional risks and uncertainty in the next few months.
Financial institutions that have exposures to hard-hit industries, including retail, hospitality, auto and shipping, can see the impact of COVID-19 on their provisions and allowance. A significant rise in the non-performing asset (NPA) ratio and credit cost is expected globally. This rise could be even more severe in the case of developing countries. Investors will become more risk-averse and increase a financial institution’s credit costs.
Both International Financial Reporting Standard (IFRS 9) and Current Expected Credit Loss (CECL) requires estimation of future expected credit losses (ECL) when calculating for provisions in investment portfolios, loan books, and trade receivables – essentially calling for predictive analytics. During this uncertain time, there are a few considerations institutions should keep in mind when calculating ECL.
Use of Traditional Credit Risk Models Vs. Advanced Models
Artificial intelligence (AI) and machine learning (ML) models are built on superior data quality, but it can take nine to 15 months for more meaningful and accurate data to become available. Given the current economic situation, AI and ML models might give less provision and allowances because currently, available data points do not describe the severity of COVID-19, which can create regulatory risk for financial institutes. Avoiding the risk of inaccurate IFRS 9 and CECL reporting entails the use of traditional credit risk models, including roll rate, vintage, and cash-flow. Along with the conventional credit risk model, using a qualitative overlay can capture information otherwise not included through the modeling.
Choosing Weights for ECL Scenarios and Revisiting Models
Increases in COVID-19 cases, with chances stay-at-home orders will be extended, will continue to slow down economic activities. Adverse economic scenarios will become the new normal in the current environment. Financial institutions can address this issue by using the appropriate weights for baseline, adverse, and severely adverse scenarios.
Revisiting and fine-tuning credit risk models in terms of assumptions, strengths, weaknesses, macroeconomic variable selections, portfolio behaviors, and business conditions is vital in order to get results that are closer to current economic conditions.
The CECL model works on recognition of expected credit losses over the entire lifetime of a loan, and this loss increases during severe economic conditions, such as the impacts from COVID-19. In the current situation, CECL loan losses will increase significantly and are further amplified by macroeconomic forecasts as models are not able to predict exact turning points in the business cycle. To address this challenge, firms must implement features such as mean reversion to the CECL model.
Staging & Segmentation Logic
Staging is of paramount importance in IFRS 9. Central banks across the globe have provided a moratorium option on all loans. A majority of customers across FICO scores have used this moratorium option. However, timelines for recognizing Non-performing Assets (NPAs) are being revised globally. Financial institutions cannot directly tag non-performing loans under stage 2 or stage 3. Stage classifications should be made based on qualitative analysis of the customer and the respective industry. In a continuously emerging situation, back-and-forth operations are required to perform stage determination. Existing staging logic needs revision to include new rule parameters, such as forbearance. Staging rule parameters and their trigger points will differ across industries.
Like Staging in IFRS 9, segmentation in CECL is paramount. Due to COVID-19, any earlier segmentation made is no longer valid. Firms must revisit the segmentation process by considering the longer-term impact on their portfolio. With the help of a stable, adaptable system, segmentation can be performed across products, customer types, and industries–the segment-wise appropriate model can be used to calculate the final CECL numbers.
Use of Internal Data vs. External or Supplemental Data
In the current scenario, financial institutions cannot rely on external data as meaningful and accurate. As noted above, external data points available today are not relevant and may take a while to capture. Instead, look for specific data inputs generated from daily operations (i.e., business liquidity and transactional data) to act as leading indicators. Firms can build an internal scoring model that solely depends on the type of expenses done by the customers, and this should differ across business sectors. To build such models, an integrated approach of account- and transaction-level data is required.
Addressing Compliance in the Days Ahead
Although most major financial institutes have implemented IFRS 9, there are a few geographies yet to meet their deadlines for implementation (nearly two or three years away). Institutions in these geographies can strategize their risk and finance activities to get the system ready for IFRS 9.
The Financial Accounting Standards Board (FASB) has delayed the implementation of CECL by a year to January 2022. This delay helps to reduce operational overhang in these difficult times and better serve clients and deposit holders. Even though CECL is deferred by a year, further changes in terms of phase-wise implementation are expected.
Solution Ideas from Oracle
To meet IFRS 9 and CECL requirements, financial institutions must incorporate a data-integrated approach with the ability to customize business rules as more changes with the moratorium, NPA recognition, and loan-restructuring are expected. Financial institutions working will Oracle have benefitted from integrated, risk-aware accounting and reporting environment that makes then quickly adapt to changing requirements, including those transpired by the pandemic.
Oracle’s IFRS 9 and CECL solutions integrate risk and finance, maintain data integrity, and deliver consistent and accurate reporting. It also provides transparent and rule-driven classification, stage classification, and preconfigured rules and methodologies for impairment-related calculations such as effective interest rate, effective interest spread, and ECL.
To learn more about IFRS 9 and CECL requirements, feel free to message me to explore more, or have a conversation.
Subscribe to our Blogs:
Oracle Financial Services Blogs: Sign up today