India’s expected-loss rules will make bank judgement more visible

The RBI’s expected-credit-loss framework will bring more of a bank’s judgement about future losses into its accounts. The final directions issued on 27 April take effect on 1 April 2027, with the previous framework remaining operative until then. Earlier recognition can make an emerging problem more visible while the bank still has choices about pricing, capital and loan management. It also puts more weight on assumptions about borrowers, collateral and the economy. The rule will work best when investors and supervisors can understand why provisions changed and whether the same evidence would produce a similar result across comparable portfolios. RBI’s 2026 directions on asset classification and provisioning · Business Standard on the final provisioning framework

The horizon changes as risk changes

Under the new framework, a loan can require a provision before a missed payment makes it visibly impaired. The directions distinguish 12-month expected credit losses from lifetime expected credit losses and require the longer horizon when credit risk has increased significantly, subject to their terms. Twelve-month expected loss is the portion of lifetime losses associated with defaults possible over the next twelve months. It should not be read as a measure of cash shortfalls incurred only within one year. The transition between horizons can have a material accounting effect even before a default occurs. RBI’s 2026 directions on asset classification and provisioning

Suppose an illustrative ₹100 crore portfolio has a two per cent probability of defaults associated with the first horizon and loss severity of 40 per cent. The simplified product is ₹0.8 crore. If a later deterioration triggers a lifetime probability of eight per cent with the same assumed severity and exposure, the illustrative amount is ₹3.2 crore. Actual calculations require expected timing, scenarios, exposure at default and discounting, among other requirements. This example isolates the effect of the horizon. It shows why a change in the bank’s assessment can affect profit and capital before collections fall by a similar amount. RBI’s 2026 directions on asset classification and provisioning

Forward-looking means the scenarios have to be explained

A lender’s estimate can change because borrowers have become weaker, because the lender changed the weight assigned to an adverse scenario, or because the composition of the portfolio changed. These are different mechanisms. The directions require multiple scenarios and probability weights determined using evidence and judgement as appropriate. A disclosure that only reports the net increase in provisions leaves outsiders unable to tell whether the loan book deteriorated or the bank became more cautious about the same loans. Reconcile the effects of new loans, migrations between risk categories, recoveries, write-offs and model changes where feasible. RBI’s 2026 directions on asset classification and provisioning

The effect on credit can arrive before the accounting deadline. Banks may improve data, tighten lending standards or change pricing while preparing for transition. That can improve risk recognition. It may also make it harder for borrowers whose recent performance looks weak despite a viable business. A rise in rates or decline in new lending should be analysed alongside evidence on applications, borrower quality and the bank’s funding costs. A shortfall in data can be costly if it causes lenders to treat similar risks differently or to rely on conservative assumptions that are difficult to test.

The case for earlier recognition and its limits

The strongest argument for the change is that a lender should account for credible signs of losses before they are fully realised. Earlier recognition gives owners, regulators and bank managers more time to respond. The strongest objection is that several banks can raise provisions at once when the outlook worsens, potentially reducing willingness to lend at a difficult moment. Avoiding that effect altogether would require giving up some information about risk. The relevant question is how the banking system manages the recognised risk with capital, pricing and borrower support, and whether the models distinguish a temporary shock from lasting deterioration. RBI’s 2026 directions on asset classification and provisioning · Business Standard on the final provisioning framework

The transition arrangements deserve separate attention. A bank can face a change in reported retained earnings because it applies the new methodology to an existing book, even if its borrowers’ circumstances did not change that day. The RBI sets out how transition impacts are measured and how eligible amounts may be phased for regulatory capital under the specified schedule. Read opening effects alongside subsequent provisions on new or deteriorating exposures. Comparing accounting profit before and after the change without that reconciliation could confuse a change in measurement with a change in performance. RBI’s 2026 directions on asset classification and provisioning

One useful exercise would hold the loan portfolio fixed and show how a plausible change in the economic scenarios alters the provision. A second would hold the scenarios fixed and show the effect of borrowers moving between risk categories. The first isolates sensitivity to the model and the outlook. The second gives more information about deterioration in the existing book. Neither comparison predicts realised losses, but both help readers establish what changed. A lender that publishes the reason for a large adjustment is easier to assess than one that reports the adjustment as an isolated figure. RBI’s 2026 directions on asset classification and provisioning

There is a distributional issue in model error. A lender may have much better data on larger firms and salaried households than on smaller businesses with volatile receipts. That can create larger uncertainty and conservative assumptions for the second group. Track how often decisions are corrected when additional information becomes available and whether estimates improve as the lender collects reliable evidence.

What to look for in the first disclosures

Compare reported risk migrations, provision coverage, changes in scenario assumptions and realised recoveries over time. Look for differences between otherwise similar exposures, and consider whether the underlying borrowers or the methodology explain them. Earlier warning would be valuable if it improves the bank’s response without causing unjustified differences in access to credit. Repeated large model adjustments that cannot be reconciled with portfolio performance would weaken confidence. The most informative first report will show how the bank reached its estimates and where the uncertainty remains. RBI’s 2026 directions on asset classification and provisioning · Business Standard on the final provisioning framework