RBI Expected Credit Loss (ECL) Framework

Indian Banks Spent 30 Years Looking in the Rearview Mirror. The RBI Just Changed That Forever. There is a date that will quietly define Indian banking for the next decade. It is not the day a major bank failed or the day the RBI bailed one out. It is April 27, 2026, when the Reserve Bank of India issued 14 final directions that fundamentally rewire how every major bank in India must think about the loans on its books.

The regulation is called the Expected Credit Loss (ECL) Framework, formally titled the RBI (Scheduled Commercial Banks - Asset Classification, Provisioning and Income Recognition) Directions, 2026. It comes into force on April 1, 2027, and it does something Indian banking has never formally required before: it asks every bank to stop looking at what has already gone wrong and start estimating what might go wrong next.

This is not a technical accounting change. It is a philosophical shift in how risk is managed, how capital is held, and how early warning systems function inside the backbone of India's economy. When the system works the way its designers intend, it makes bank failures less sudden, bank balance sheets more transparent, and the entire credit cycle less violent.

"This proposed change is more cultural, as it compels banks to anticipate risk rather than react to it, strengthening resilience and aligning India's financial ecosystem with global standards."

To understand why this reform matters, you need to understand what it replaces. India's old provisioning system, formally called the IRACP framework, worked like a doctor who refuses to diagnose illness until the patient collapses. It waited for a loan to go visibly bad before requiring a bank to set aside money to cover the loss.

The system classified loans by how many days overdue a borrower was, and applied fixed provisioning rates accordingly. A borrower rated AAA and a borrower rated C received essentially similar treatment, until visible default signs emerged. There was no mechanism to distinguish worsening credit quality early, and no forward-looking lens.

The consequences were predictable and severe. During India's banking crisis between 2015 and 2020, provisioning surged suddenly and sharply. Banks that appeared adequately capitalised one quarter were reporting massive provision requirements the next. Credit growth collapsed. Capital adequacy ratios fell at exactly the moment they needed to be strong. The system amplified the crisis rather than cushioning it.

Globally, this problem was solved much earlier. The IFRS 9 standard, adopted in 2018, required entities worldwide to use forward-looking credit loss models. India applied IFRS 9 logic to NBFCs and corporates through Ind AS 109. Banks, the most systemically important part of India's credit infrastructure, had been waiting. For nine years, India's banks operated under rules that the rest of the world had replaced. That wait is now over.

The concept behind ECL is simple even if the implementation is complex. Instead of waiting for a loss to happen, a bank estimates three things for every loan it holds and multiplies them together.

ECL = PD x LGD x EAD. That formula, run across every loan in a bank's portfolio, produces the expected credit loss that the bank must provision for. The revolutionary part is that all three inputs must incorporate forward-looking macroeconomic information, not just historical loss data. If GDP growth is forecast to slow, if unemployment is expected to rise, if a sector is under structural stress, those expectations must feed directly into the PD, LGD, and EAD estimates.

Banks must also run scenario analyses across multiple macroeconomic pathways, weight them by probability, and provision accordingly. This is why ECL is genuinely predictive rather than merely descriptive. It forces a bank to think systematically about the future before the future arrives.

The ECL framework organises every financial asset a bank holds into one of three stages based on how its credit quality has changed since the bank first made the loan. The concept anchoring all three is called Significant Increase in Credit Risk (SICR).

The most important of the three stages for systemic stability is Stage 2, the early warning stage. The old IRACP system had no equivalent. A borrower went from performing to NPA with minimal intermediate signalling. Under ECL, Stage 2 is the framework's radar screen, requiring banks to build lifetime provisions well before a loan technically defaults. This is where the real reform lives.

One critical clarification: the RBI has retained its existing NPA classification norms alongside the ECL framework. ECL governs provisioning; NPA designation and its regulatory consequences, such as restrictions on credit to NPA borrowers, continue under existing rules. This dual structure means ECL is a provisioning reform, not a complete NPA replacement.

Every regulatory change eventually comes down to a number that investors need to put in their models. For ECL, that number is 120 basis points. According to CRISIL Ratings, the transition to ECL could have a gross impact of up to 170 basis points on the Common Equity Tier-1 (CET-1) ratio of most banks. After factoring in provisions already made, the net impact is estimated at up to 120 basis points.

To put that in context: most large Indian banks currently maintain CET-1 ratios well above the regulatory minimum of 8% (ICAAP typically targets 11-13%). A 120 bps hit is meaningful but manageable, especially given the four-year glide path. The RBI has allowed banks to spread the one-time provisioning impact on profitability and capital across FY2027 to FY2031.

The timing is also fortunate. India's banks enter this transition in their strongest asset-quality position in a decade. Net NPA ratios have fallen below 1% for most major banks, and capital buffers remain above regulatory requirements. CareEdge Ratings assessed Indian banks as well-positioned to absorb the ECL transition without systemic stress.

Not all banks will experience this transition equally. The impact varies dramatically based on portfolio mix, existing provision cover, data quality, and model readiness. Here is an honest read of the landscape.

The ECL directions did not arrive alone. On the same day, April 27, 2026, the RBI also issued the Commercial Banks - Capital Charge for Credit Risk (Standardised Approach) Directions, 2026, operationalizing the Basel III final reforms for Indian banks, also effective April 1, 2027.

This companion framework changes how much capital banks must hold against different types of credit risk. Where the ECL framework changes provisioning, Basel III changes capital requirements. Together, they represent a twin overhaul of the entire credit risk architecture.

The RBI has provided a four-year transition period from FY2027 to FY2031. Banks may spread the one-time provisioning impact across this window. This is not just financial relief. It is also an acknowledgment that the infrastructure required for ECL is substantial and cannot be assembled overnight.

The RBI has been consultative across a three-year journey: a January 2023 discussion paper, an October 2025 draft consultation, and the April 27, 2026 final directions. That careful, iterative approach reflects both the reform's complexity and its stakes. Banks knew this was coming. The ones that have been preparing since 2023 are already building model inventories, cleaning data, and training staff. The ones still treating ECL as a 2027 problem are already behind.

The honest answer to where India stands: it is nine years behind the global frontier, but it is catching up correctly. IFRS 9 went live globally in 2018. India is implementing its prudential equivalent in 2027. That gap matters, but it also means India benefits from nearly a decade of evidence on what works and what does not in ECL implementation across other markets.

India's ECL framework has one distinctive feature that sets it apart from pure IFRS 9: the mandatory prudential floors for Stage 1 and Stage 2 exposures. IFRS 9 is principles-based; banks have wide discretion in model design. The RBI, characteristically prudent, has imposed minimum provisioning backstops to prevent under-provisioning during the early years of model development. This is a uniquely Indian safeguard and a sensible one.

The ECL framework is a banking regulation, but its effects radiate far beyond the risk departments of scheduled commercial banks. Here is what it means across the stakeholder spectrum.

It is a new provisioning system for Indian banks issued on April 27, 2026 and effective April 1, 2027. Under ECL, banks must estimate future credit losses on every loan they hold using three inputs: Probability of Default (PD), Loss Given Default (LGD), and Exposure at Default (EAD), incorporating forward-looking macroeconomic forecasts. It replaces the old incurred-loss model that only recognised losses after a loan became bad.

Banks with higher proportions of unsecured lending, such as personal loans, credit cards, and microfinance, face the largest provisioning increase. Nuvama analysts specifically named IndusInd Bank, RBL Bank, and IDFC First Bank as higher-impact institutions. Well-provisioned PSU banks like SBI and private banks with secured portfolios like HDFC Bank and Kotak are better positioned. CRISIL estimates the net CET-1 impact at up to 120 basis points for most banks.

Stage 1 covers performing loans with no significant credit deterioration since origination. Banks provision for 12-month expected losses only. Stage 2 covers loans showing a significant increase in credit risk, requiring lifetime expected loss provisions. Stage 3 covers credit-impaired or NPA-equivalent assets, also requiring lifetime ECL. Stage 2 is the critical early warning mechanism the old system lacked entirely.

No. The RBI has retained its existing NPA classification and regulatory consequences alongside the ECL framework. ECL governs how much banks must provision; NPA designation and its associated regulatory restrictions continue under existing rules. The two systems operate in parallel.

India's ECL framework closely mirrors IFRS 9, the global standard adopted across the EU, UK, Singapore, and many other markets since 2018. The key India-specific addition is mandatory prudential floors for Stage 1 and Stage 2 provisioning, which prevent under-provisioning during early model development. IFRS 9 is principles-based with more bank discretion; India's version has built-in regulatory backstops.

Sources: RBI, CRISIL Ratings, CareEdge Ratings, Global Market Intelligence, Nuvama Analysts via Business Standard, IFRS Foundation, Grant Thornton ECL framework analysis, KPMG India, S&P Global, Banking Regulation: RBI ECL Framework