Transitioning to ECL framework: A move towards expected loss-based norms

Recently, the Reserve Bank of India (RBI) announced that Indian commercial banks will have to start implementing the expected credit loss (ECL)-based provisioning norms from April 1, 2027, according to the International Financial Reporting Standards (IFRS). There will also be a transition path (till March 31, 2031) to smoothen the one-time impact of higher provisioning, if any, on their existing books (indicating prudential floors). The ECL norms will replace the ‘incurred loss’ model with an ‘expected credit loss’ model. Under the new norm, banks will have to adopt a forward-looking approach considering past events, current conditions and forecast information. This will provide a timely and adequate accounting treatment of loan loss provisions. Given the strong balance sheet as well as capital position of banks currently, this is the right time to introduce more risk-sensitive ECL-based provisions. 

An important consideration in the IFRS 9 Expected Credit Loss model is to make provisions before the occurrence of a loss event. This is for the first time Indian banks are going to depend on their internal models for estimating regulatory provisioning requirements. Banks will thus need to internally model the key elements of their credit risk-related losses, viz., probability of default (PD), loss given default (LGD), and exposure at default (EAD), and thereby derive more risk-sensitive measures for expected credit loss and credit risk capital. ECL is the product of EAD, PD and LGD with a discount factor. PD is the “likelihood” that a borrower will not be able to meet the contractual obligations within a specified time period, usually one year. LGD is defined as the realised losses after recovery by the lender when a borrower defaults [LGD = 1- (Expected Recovery/Exposure at Default)]. Cash flows are discounted back to the point of default based on appropriate discounting rates. Exposure at default tracks the current outstanding as well as the expected drawn portion from the total unavailed loan limits. 

The forward-looking ECL-based allowance (or provision) is applied at origination and for all subsequent reporting periods to financial assets till derecognition. Under IFRS 9, banks need to classify their assets into three stages. The good performing assets with low credit risk can be categorised as stage 1, for which the 12-month PD needs to be estimated for the ECL. For assets with a significant increase in credit risk (due to downgradation of rating by two notches, or more than 30 days past due), they need to be categorised as stage 2 for which lifetime PD has to be used for estimation of ECL. The non-performing assets (overdue more than 90 days) will be treated as stage 3, and the provision amount will be purely dependent on the loss given default rate (as PD will be 100 per cent). 

Expected loss is the losses anticipated on a credit exposure or credit portfolio due to defaults expected to occur during the normal course of business. It is the current internal estimate of probable future credit loss. The forward-looking estimates can be obtained from historically derived numbers as a good starting point. However, for advanced approaches, forward-looking PD adjustments need to be made by incorporating macroeconomic factors into PD estimation. 

Figure 1 presents overall corporate PD trends as predicted by the leading credit rating agency over a 21-year period. The average PD over a 21-year period according to the estimate is at 2.71 per cent. Interestingly, the average PD estimate is significantly lower than rating wise portfolio and clearly shows that banks with a large diversified portfolio will stand to gain over the longer term instead of short term. Though the capital charge is likely to increase in the short term, the benefits of a diversified portfolio, differentiated pricing will allow the banks to be even more proactive in terms of risk assessment. Also, as can be seen from the graph, overall corporate default rates clearly move with the economic condition and gross domestic product (GDP) growth rate. Banks need to link the PD estimates with common macroeconomic factors to derive more accurate forward-looking scenarios and this will ensure banks are always better prepared and frontloaded in their actions in this uncertain global environment. 

Besides PD, the Loss Given Default Rate (LGD) based on the loan recovery experience of banks also plays a crucial role in determining the expected credit loss. We have also projected the LGD trend obtained from various recovery channels (Debt Recovery Tribunal, Securitisation and Reconstruction of Financial Assets and Enforcement of Securities Interest Act, Insolvency and Bankruptcy Code) for commercial loans. The average LGD over the same 21-year period is around 77 per cent. In fact, corporate LGD values are relatively higher compared to other markets (in the US, it is 45 per cent, Singapore, 25 per cent), and LGD may have cyclicality. As LGD rate changes with the macroeconomic cycle and it is linked to the default rate (PD), banks will be more proactively making loss provisions and containing their future NPAs. 

At the portfolio level, banks need to segment their credit portfolios (product-wise, industry-wise, risk-wise, etc.) and derive multiple PDs, LGDs, and EADs. By linking their credit planning with these credit risk drivers, commercial banks will be able to derive portfolio diversification benefits, which will enable them to sustain their credit growth as well as manage capital. Transition to the ECL-based provisioning norm is expected to bring greater solvency and soundness in the banking system. 

 

Arindam Bandyopadhyay is Professor, National Institute of Bank Management; and Soumya Kanti Ghosh is Member, 16th Finance Commission and Economic Advisory Council to the Prime Minister, and Group Chief Economic Advisor, State Bank of India. Views are personal