Recognising Stress Before It Spreads: ECL, ARC Sales and Future of Bank Bailouts in India

Summary: India’s banking regulatory framework is moving towards earlier recognition of financial stress, stronger provisioning and more timely resolution of distressed assets. The Reserve Bank of India’s 2015  Asset Quality Review exposed weaknesses in the recognition of non-performing assets and highlighted the consequences of delayed provisioning. Subsequent reforms, including the Insolvency and Bankruptcy Code, government recapitalisation and greater participation by  Asset Reconstruction Companies, sought to strengthen recovery and resolution mechanisms. The article examines how the RBI’s 2026  Expected Credit Loss framework, scheduled to become effective from 1 April 2027, represents a further transition from recognising losses after deterioration has occurred to estimating potential losses in advance. Under the framework described, banks classify exposures according to changes in credit risk and make corresponding provisions, potentially reducing unexpected balance-sheet shocks. The article also considers the increasing sale of stressed loans to ARCs, which may improve recovery and reduce banks’ direct exposure but can raise concerns about whether economic risk has genuinely transferred, particularly where banks retain security receipts. A related development is the RBI’s 2026 treatment of Specified Non-Financial Assets, including repossessed property and the reversal of certain unrealised interest income. Together, these regulatory measures are presented as instruments for improving transparency and limiting delayed recognition of losses. Nevertheless, their effectiveness depends on accurate credit-risk models, genuine asset resolution, adequate disclosure and institutional readiness. The reforms may reduce the likelihood of sudden banking distress without eliminating the possibility of future government intervention.

1. INTRODUCTION

The resilience of the banking industry is largely determined by the robustness of the financial system, which highlights the importance of timely identification and action on stuck assets as a critical issue in banking regulations. In India, the growth of Non-Performing Assets (“NPA”) has always been a major problem, especially for public sector banks, where insufficient handling of stress situations has led to major interventions by regulators and the government. The problem of under-declared NPAs gained prominence when the Reserve Bank of India’s (“RBI”) Asset Quality Review (“AQR”) in 2015 brought to light the flaws of a system that allowed failures to go unreported until the problems became worse and resolution harder and more expensive.

Consequently, efforts across regulatory frameworks were undertaken, resulting in improved identification, provisioning, and solvency of public sector banks via measures like  Insolvency and Bankruptcy Code 2016 (“IBC”) and a greater involvement of  Asset Reconstruction Companies (“ARC”). These developments prompt a larger inquiry into the developing nature of India’s response to troubled banks. Traditionally, banking crises have been met with ex-post interventions involving recapitalization and other types of public assistance. However, the modern situation appears to call for stronger regulatory measures, requiring potential losses to be detected earlier, proper provisioning is conducted, and the resolution process begins before banks face troubles.

Such transformations have gained more prominence in the year 2026 as the RBI’s Expected Credit Loss (“ECL”) framework signifies a new era in which banks enter a transition of moving away from the incurred-loss model of credit loss recognition. The new ECL framework will become operational from 1 April 2027 onwards and compels banks to make expectations of loss provisions. At the same time, banks are transferring their bad assets to ARCs in order to clear their books ahead of the new regime.

2. Evolution of Banking Stress Management in India

I. Pre-2015 Approach

Before 2015, distressed assets in India were characterised by delayed recognition and an over-reliance on restructuring options. The concept of regulatory forbearance allowed certain restructured loans to have the advantage of continued good asset classification and hence, the magnitude of the troubled status of the banks might have been underestimated. Later the RBI recognised that the restructuring scheme had been misused to minimise provisioning requirements rather than to resolve distressed assets properly.

This issue was especially important for Public Sector Banks (“PSB”), which had accumulated considerable vulnerabilities vis-à-vis distressed sectors. The result was the increased discrepancy between reported bank assets quality and real financial conditions. This practice showed that deferring the loss recognition made it harder to achieve the resolution and increased the need for external capital support.

The RBI’s 2015 AQR marked a critical shift towards transparency by uncovering previously undisclosed stress, leading to the reclassification of several stressed accounts as NPAs and higher provisioning by banks. Consequently, gross NPAs of scheduled commercial banks rose sharply, highlighting the extent of the underlying stress. However, while the AQR addressed delayed recognition, mere recognition of stressed assets failed to resolve the underlying financial distress, necessitating effective resolution measures.

II. Post-2015 Regulatory Response

Post-AQR, Government recapitalisation was applied to improve the capital position of PSBs that were affected by the measurement of NPAs and provisioning thereof, while the IBC introduced a legally mandated mechanism for resolving insolvency within a time-frame. The RBI also enhanced mechanisms for the sale of stress assets to ARCs to enable banks to transfer distressed assets for better recovery and resolution. With these developments, the overall emphasis shifted from mere recognition of NPAs to resolution and recovery.

Although measures have been taken, a new regulatory philosophy is currently evolving, requiring banks to have mechanisms in place to identify and develop for possible losses before they become materialised as NPAs. This can be considered as an important step from retroactive regulatory action to a proactive approach.[1] The importance of this transformation can be seen in the 2026 regulatory changes introduced by the RBI, including the shift to Expected Credit Loss-based provisions.

III. The ECL Regime: A Shift Towards Pre-emptive Recognition

1. Existing NPA and Provisioning Framework

The pre-existing framework of the RBI required banks to categorize assets and create provisions against perceived losses. However, this approach was far from being proactive since, by the time the loss recognition process was in place, financial distress had already occurred, raising regulatory concerns that the losses would be recognized after the change in asset’s credit quality.

The ECL framework signifies an evolution towards pre-emptive loss accounting. It requires banks to determine, on every reporting date, whether there has been a significant increase in the credit risk of a financial asset or not. This classification happens in 3 stages, with Stage 1 typically being affixed a 12-month ECL while financial assets that witness a significant increase in credit risk shall shift to Stage 2, during which lifetime ECL will be applicable. Most crucially, ECL is being used in a parallel manner with the existing NPA classification system.

On 27th April 2026, the RBI implemented the  Reserve Bank of India (Commercial Banks  Asset Classification, Provisioning and Income Recognition) Directions (“Directions”), which will come into force on 1st April 2027. The Directions lays down a  three-step provisioning mechanism along with the essential protective threshold, which includes the requirement of the banks to set aside a minimum of 5% of the loan value if the loan is categorised as Stage 2 loans by the bank’s model for estimate excepted loss. Additionally, to reduce impact of loss incurred in terms of capital belonging to banks, RBI will now allow the loan to be recovered in a phased manner over a period of four financial years

ECL’s predominant rationale is to ensure that developing credit deteriorations make their way into bank financial statements at an early stage. By requiring provisions based on expected losses instead of just the losses already incurred by banks, the system focuses on minimizing the chances of sudden and severe provisioning shocks in case of the loans turning into Non-Performing Loan (“NPL”). According to the RBI, the new system will result in better credit-risk management practices, as well as more comparable and stable financial reporting by banks.

2. Impact on Banks

The transition process is expected to lead to enhanced provisioning requirements, especially for banks with significant exposure and experiencing early deterioration in credit quality. Thus, even if this transition leads to a reduction in profits and capital ratios at least in the short term, experts believe that the impact is manageable due to sufficient capital cushion. CRISIL, a major credit rating company, estimates that the cumulative projected impact on the Common Equity Tier 1 (“CET-1”) ratio could be in the form of as much as 120 basis points, though the impact may spread over four financial years.

ECL does not eliminate the possibility of bank failure or the need for government involvement. What it does change is the timing of loss recognition. Early loss recognition may encourage banks to increase and improve provisions, reassess risky positions, and intervene at the stage when problems are beginning to arise. Therefore, ECL may lower the risk of banks suffering unexpected losses and losing their capital stability, which would lead to the need for the public sector’s involvement. This advantage is conditional and depends mainly on the quality of banks’ credit risk models and their willingness to make accurate assessments of credit deterioration.

5. ARC Sales and Pre-emptive Balance-Sheet Clean-up

I. Role of ARCs in stressed-asset resolution

ARCs occupy a slightly awkward position in the Indian financial system. They exist to take on what banks would rather not carry, yet they were built by the same regulator i.e., RBI, that supervises the banks selling to them. Set up under the Securitisation and Reconstruction of Financial Assets and Enforcement of Security Interest Act 2002  (“SARFAESI Act”), ARCs buy NPLs from banks and other lenders, generally paying part in cash and part through security receipts, and then pursue recovery. More than two decades on, the sector has moved from being a marginal recovery channel to a routine balance-sheet tool that lenders reach for well before a loan has exhausted every other avenue of resolution.

II. Why banks are accelerating asset sales

ARCs are reported to have acquired stressed debt worth more than ₹2 lakh crore during FY 2025-26, made up of roughly ₹1.5 lakh crore in corporate exposures and close to ₹50,000 crore in retail assets, with retail acquisitions alone rising by around ₹54,727 crore over the previous year. Part of this shift traces back to a 2022 regulatory change that allowed ARCs to acquire special mention accounts, meaning loans that are overdue but have not yet crossed into non-performing status. That single change let banks sell earlier in the distress cycle rather than waiting for full NPA classification, and low-vintage acquisitions of this kind are said to have grown from about 14 per cent of new purchases in FY 2024 to roughly 30 per cent in FY 2025. The recovery rate on security receipts issued against these purchases has also been projected to reach 75-80 per cent by FY 2026, a marked improvement on the low twenties recorded only a few years earlier.

The connection between this trend and the incoming provisioning regime is not incidental. Analysts flagged the link years in advance, arguing that the anticipated move to ECL provisioning would itself become a driver of ARC business, since banks facing forward-looking, stage-based provisioning have a strong reason to sell an asset before it migrates into a heavier provisioning bucket. Under the finalised framework, i.e., the April 2026 Directions provisions must be built on an estimate of losses over the life of an exposure once its credit risk has increased materially, rather than only once a default has actually occurred. Selling a weakening account to an ARC before it is reclassified into a higher stage therefore does two things at once: it removes the exposure from the balance sheet, and it avoids the additional provisioning charge that would otherwise follow the reclassification.

In 2015, the entire sector was hit badly and experienced a sudden rise of bad loans making it difficult for banks to make profits after having realized the issue. The current situation is of a different nature although both situations have almost similar nature of instincts, i.e., banks don’t want to report bad loans because it would make it impossible for them to make profit. The 2015 market anticipation process was corrective and retrospective while the ongoing process of ARC sales in the light of enacted 2026 Directions is more anticipatory. It indicates that the banking sector is now prepared to changes that the regulators are likely to introduce in the future in terms of new rules that could potentially affect banks’ operations.

III. Are ARC sales genuine resolution or regulatory arbitrage?

When a bank sells a stressed loan to an ARC largely against security receipts rather than cash, and then continues to hold a meaningful share of those receipts on its own books, the loan has changed its accounting label without necessarily changing hands in any economic sense. Historically, banks have subscribed to a large majority of the security receipts issued against their own sales, which raises the question of how much risk transfer is actually taking place. There are also longstanding concerns, raised at various points by the Central Vigilance Commission (“CVC”), about the fees ARCs charge and about how well banks actually track recoveries once an asset has left their books. None of this means the sales are worthless. Even where a bank retains exposure through security receipts, it typically no longer has to hold a distressed loan at par on its own balance sheet, and a dedicated recovery specialist may well do a better job of extracting value than the original lender could have. But it does mean that a rising volume of ARC sales should not, on its own, be read as proof that stress is disappearing from the system; it may simply be moving into a different, less visible corner of it.

6. The Repossessed-Asset Amendment: Regulatory Consistency or a Separate Development?

I. The 2026 amendment

The RBI issued its final Amendment on Specified Non-Financial Assets (“SNFA”) on 16 July 2026, following draft directions circulated on 5 May 2026, and set the effective date at 1 October 2026. The  Amendment is applicable to immovable property acquired by a lender in satisfaction of its claims, with the non-banking assets already recognised under  Section 9 of the Banking Regulation Act 1949. An asset can only be treated as an SNFA once its title has actually passed to the lender and once the exposure it relates to has already been classified as non-performing, with other recovery routes shown to be unworkable. Taking mere symbolic or physical possession under Section 13(4) SARFAESI Act is not enough to meet that de-recognition test. The Amendment also states that disposal must generally happen within seven years and the acquisition is on a non-recourse basis. The asset cannot be sold back to the original borrower or a related party as defined under Section 29A of IBC, and the change was issued alongside consequential Amendments to the existing Directions on resolution of stressed assets and on income recognition, asset classification and provisioning.

II. Treatment of uncollected interest after collateral repossession

As per the Amendment, any interest or other charges that had accrued on an exposure cannot be recognised as income before the underlying SNFA is formally acquired. Where a lender had already booked such income on an account that later becomes an SNFA, and that income remained outstanding on 30 September 2026, it must be reversed through the profit and loss account by 30 September 2027, to the extent it was never actually realised. Simply put, a bank that had been silently accumulating interest on a non-performing loan headed towards foreclosure, interest the debtor was never going to repay, will now be obliged to reverse that entry and take the write off in its earnings rather than leaving it unrecorded as if it were an active account receivable.

III. Anti-evergreening rationale

Evergreening entails the practice of leveraging accrued interest on the previous bad loans by making new loans in the favour of the struggling borrower to give the impression that a NPL has become performing on paper. Hence, by not enabling banks to postpone the recognition of their losses, RBI has also tackled evergreening through alternative funding schemes. Now banks require to immediately account for any loans as cash-based once the default has been launched and to undo anything that might previously have been accounted for as accruals.

7. Conclusion and Way forward

The ECL framework, the rise in ARC sales, and the repossession regulation, all three regulatory tools are aligned and lead to the same outcome. ECL framework mandates banks to acknowledge losses before defaults occur. ARC provides banks with an opportunity to remove the risk of bad debt from the balance sheet before that debt is actually written off.  Lastly, the regulation of SNFA makes impossible for the bank to take possession of collateral and continue reporting unrealized imputed interest as revenues. Each of the three approaches does not resolve the issue independently, but together they constrain the bank from postponing the acknowledgement of a loss that it is aware of. All these developments create an environment where the distance between a loan and a bank’s admission of a problem becomes increasingly narrower. At the same time, it should be noted that these changes relate to a change in timing and transparency, but they do not serve as a guarantee against future bailouts.

Although these reforms do not make a future government bailout of a bank or group of banks impossible, they do increase the chances of it happening at a less surprising time. To strengthen the argument that this transition is not simply a cosmetic one, there are a few specific actions that can be taken. First, the requirements for disclosure and monitoring of security receipts should be tightened so that the regulators and market participants can determine the extent to which risk has been transferred from the selling bank instead of simply being renamed. Second, the deadlines for the implementation of expected credit loss should be adjusted according to institutional readiness with public sector banks and smaller banks receiving actual help in building appropriate data and modelling infrastructure, rather than just receiving an extended timeline. Finally, the criteria for Stage 2 provisioning should allow enough flexibility to separate a case of an institution that is temporarily under stress from one that is experiencing real decline, so that financing these segments of the economy will not be choked off when they need it the most.

Referance

[1] Friederike Niepmann and Tim Schmidt-Eisenlohr, ‘Bank Bailouts, International Linkages and Cooperation’.

****