IFRS 9 ECL — Staging and the ECL Formula
In Lesson 1, you learned that IFRS 9 is the accounting pillar of modern banking — the framework that requires banks to provision for expected credit losses before those losses actually occur. In Lesson 2, you saw that the banking plugin's ifrs9-staging and ifrs9-ecl skills handle this calculation. Now you will build the staging model and the ECL formula from first principles, so you understand exactly what the skills are calculating and can verify their output.
The IFRS 9 Expected Credit Loss model is the single most important quantitative framework in global banking accounting. It governs how 140+ countries calculate provisions on every loan, bond, trade receivable, and financial guarantee on their balance sheets. Mastering staging and the ECL formula is not optional for any banking professional working outside the United States.
The Problem IFRS 9 Solves
Under IAS 39 (the predecessor standard), banks used an "incurred loss" model. A provision was recognised only when there was objective evidence that a loss had already occurred — the borrower had missed payments, gone bankrupt, or otherwise demonstrated an inability to repay. This created a fatal timing problem.
During the 2008 financial crisis, portfolios that would eventually lose billions were still being reported at full value because the losses had not yet been formally incurred. By the time objective evidence appeared, the losses were so large that banks could not absorb them. The incurred loss model was described by regulators as "too little, too late" — it delayed recognition of losses that were already economically inevitable.
IFRS 9, effective 1 January 2018, replaced this with a forward-looking model. Banks must now estimate expected losses over the future life of the instrument, not wait for evidence that losses have already happened. The mechanism for this forward-looking estimation is the three-stage model.
The Three-Stage Model
IFRS 9 classifies every financial asset into one of three stages based on the change in credit quality since the asset was first recognised (originated or purchased):
| Stage | Credit Quality | ECL Measurement | Interest Revenue Basis |
|---|---|---|---|
| Stage 1 | No significant deterioration since origination | 12-month ECL | Gross carrying amount |
| Stage 2 | Significant increase in credit risk (SICR) since origination, not yet defaulted | Lifetime ECL | Gross carrying amount |
| Stage 3 | Credit-impaired (defaulted) | Lifetime ECL | Net carrying amount (after ECL deduction) |
Three critical distinctions to internalise:
12-month ECL vs Lifetime ECL. Stage 1 assets require a provision equal to the expected credit loss over the next 12 months only. Stage 2 and 3 assets require a provision equal to the expected credit loss over the entire remaining life of the instrument. For a 20-year mortgage in Stage 2, the bank must estimate losses over all 20 remaining years — not just the next 12 months.
The staging cliff effect. When an asset migrates from Stage 1 to Stage 2, the provision typically increases by 5-10x because the measurement window expands from 12 months to the full remaining term. This cliff effect is one of the most significant features of IFRS 9 — a single rating downgrade can multiply a portfolio's provision by an order of magnitude.
Interest revenue in Stage 3. This is the distinction that catches many practitioners. In Stages 1 and 2, the bank recognises interest revenue based on the gross carrying amount (the full loan balance before deducting the ECL provision). In Stage 3, interest revenue is calculated on the net carrying amount (the loan balance minus the ECL provision). This reduces recognised income for credit-impaired assets.
The trigger that moves a loan from Stage 1 (12-month ECL) to Stage 2 (lifetime ECL) -- the single most consequential classification decision in IFRS 9.
If a borrower's PD doubles from 1% to 2% after origination, SICR is triggered. A $10 million mortgage moves from a $5,400 provision (12-month) to a $54,000 provision (lifetime) -- a 10x increase from one classification change.
SICR assessment drives provision volatility: one portfolio-wide rating downgrade can increase a bank's total ECL by hundreds of millions.
SICR: When Does an Asset Move to Stage 2?
The Significant Increase in Credit Risk (SICR) assessment is the most judgment-intensive part of IFRS 9. An asset moves from Stage 1 to Stage 2 when there has been a significant increase in the probability of default since initial recognition. IFRS 9 does not prescribe a single threshold — it requires banks to consider all reasonable and supportable information.
In practice, banks use a combination of quantitative and qualitative triggers:
Quantitative SICR Triggers
| Trigger | Typical Threshold | Rationale |
|---|---|---|
| Rating downgrade | 2+ notches since origination | Internal rating systems measure credit quality; a significant downgrade signals deterioration |
| PD increase | PD doubles or increases by absolute threshold (e.g., +1.5%) | Direct measure of default probability change |
| Days past due | 30+ DPD (rebuttable presumption) | IFRS 9 includes a rebuttable presumption that SICR has occurred at 30 DPD |
Qualitative SICR Triggers
| Trigger | Example |
|---|---|
| Watchlist placement | Borrower placed on credit watchlist by the bank's risk management |
| Covenant breach | Financial covenant broken (e.g., debt service coverage ratio falls below minimum) |
| Industry stress | Borrower's industry experiences significant adverse conditions |
| Forbearance | Loan terms modified to provide relief to a struggling borrower |
The 30-Day Rebuttable Presumption
IFRS 9 includes a backstop: there is a rebuttable presumption that SICR has occurred when an asset is more than 30 days past due. "Rebuttable" means the bank can override this presumption if it has evidence that the past-due status does not indicate a significant increase in credit risk — for example, an administrative payment delay by a financially healthy borrower. But the default position is that 30+ DPD triggers Stage 2.
Stage 3: Credit-Impaired (Default)
An asset moves to Stage 3 when it meets the bank's definition of default. While each bank defines default in its own credit policy, Basel and IFRS 9 provide guidance on minimum criteria:
- 90+ days past due (rebuttable presumption of default)
- Borrower is assessed as unlikely to pay in full without realisation of collateral
- Debt restructuring with a loss to the lender
- Bankruptcy or insolvency proceedings
- Borrower deceased (for retail exposures) without sufficient estate coverage
Stage 3 assets require lifetime ECL measurement and the critical interest revenue change: interest is now calculated on the net carrying amount.
The ECL Formula
The Expected Credit Loss is calculated as the product of three components:
12-month ECL (Stage 1):
ECL = PD(12m) x LGD x EAD
Lifetime ECL (Stages 2 and 3):
ECL = Sum over each future period t of: PD(marginal, t) x LGD(t) x EAD(t) x DF(t)
Where:
- PD = Probability of Default (12-month for Stage 1, marginal for each future period in lifetime)
- LGD = Loss Given Default (percentage of exposure lost if default occurs)
- EAD = Exposure at Default (outstanding balance plus expected drawdowns)
- DF = Discount Factor (present value adjustment at the effective interest rate)
A Concrete Example: Stage 1 (12-Month ECL)
A bank holds a corporate term loan with these parameters:
| Parameter | Value | Source |
|---|---|---|
| PD (12-month) | 1.2% | Internal rating model, calibrated to point-in-time |
| LGD | 35% | Based on collateral coverage and recovery assumptions |
| EAD | $10,000,000 | Outstanding drawn balance (no undrawn commitment) |
12-month ECL = 0.012 x 0.35 x $10,000,000 = $42,000
The bank books a $42,000 provision against this loan. The loan remains at full value on the balance sheet ($10 million), but a $42,000 allowance reduces the net carrying amount to $9,958,000.
A Concrete Example: Stage 2 (Lifetime ECL)
The same loan experiences a SICR — the borrower's rating drops by 3 notches. The loan moves to Stage 2. Now the bank must calculate lifetime ECL over the remaining 5-year term:
| Year | Marginal PD | LGD | EAD | Discount Factor | Period ECL |
|---|---|---|---|---|---|
| 1 | 3.5% | 35% | $10,000,000 | 0.952 | $116,620 |
| 2 | 4.2% | 35% | $8,000,000 | 0.907 | $106,478 |
| 3 | 4.8% | 35% | $6,000,000 | 0.864 | $87,091 |
| 4 | 5.1% | 35% | $4,000,000 | 0.823 | $58,633 |
| 5 | 5.3% | 35% | $2,000,000 | 0.784 | $29,067 |
| Total | $397,889 |
Lifetime ECL = $397,889
The provision jumped from $42,000 (Stage 1) to $397,889 (Stage 2) — a 9.5x increase. This is the staging cliff effect. The loan's credit quality deteriorated, so the measurement window expanded from 12 months to the full remaining life. The economic loss expectation did not change by 9.5x — the measurement window did.
A common question: if the goal is forward-looking provisioning, why does Stage 1 use only 12-month ECL instead of lifetime ECL? The answer is practical: lifetime ECL for performing assets with 20-30 year terms would result in enormous Day 1 provisions that do not reflect genuine credit risk. The 12-month ECL is a compromise — forward-looking enough to capture near-term deterioration, but not so aggressive that it distorts the balance sheet for performing assets. The staging cliff from Stage 1 to Stage 2 is the mechanism that triggers full lifetime recognition when genuine deterioration is detected.
Stage Migration: What Moves a Loan Between Stages
Stage migration is not a one-way street. Loans can move in both directions:
| Migration | Trigger | Impact |
|---|---|---|
| Stage 1 to Stage 2 | SICR detected (rating downgrade, 30+ DPD, covenant breach) | ECL increases from 12-month to lifetime (5-10x typical) |
| Stage 2 to Stage 3 | Default (90+ DPD, unlikely to pay, restructuring with loss) | Interest revenue now on net carrying amount |
| Stage 2 to Stage 1 | SICR no longer exists (rating upgrade, sustained performance) | ECL decreases from lifetime to 12-month |
| Stage 3 to Stage 2 | Cure — borrower resumes payment, no longer credit-impaired | Interest revenue returns to gross carrying amount basis |
| Stage 3 to Stage 1 | Full cure — credit quality returns to origination level | Rare but possible for short-duration retail exposures |
The ifrs9-staging skill tracks these migrations and produces the stage migration table required for IFRS 9 disclosure — showing the opening stage, the closing stage, and the ECL impact of each migration.
Using the Banking Plugin
The ifrs9-staging and ifrs9-ecl skills handle both staging classification and ECL calculation. Here is what that looks like in practice.
Worked example. Sarah Chen, Head of Credit Risk at Meridian Commercial Bank, needs to stage a deteriorating corporate facility and calculate the ECL impact. She types:
Stage this facility and calculate ECL:
Borrower: Hartfield Engineering Ltd
Facility: GBP 10M term loan, 4 years remaining
Current rating: BB- (originated at BBB)
DPD: 0
Rating downgrade: 3 notches since origination
PD (12m): 1.2%, LGD: 35%, EAD: GBP 10,000,000
Jurisdiction: UK PRA
The agent routes through ifrs9-staging and ifrs9-ecl and responds:
GOVERNING STANDARD: IFRS 9 — UK PRA
DOMAIN: IFRS 9 ECL — Stage Assessment + ECL Calculation
JURISDICTION: United Kingdom — PRA Rulebook / UK CRR
STAGING RESULT
Facility: Hartfield Engineering Ltd — GBP 10M term loan
Stage: 2 (SICR detected)
SICR trigger: Rating downgrade of 3 notches since origination
(originated BBB, current BB-)
ECL CALCULATION
Measurement: Lifetime ECL (Stage 2)
PD term structure applied over 4-year remaining life
Lifetime ECL: GBP 397,889
Previous Stage 1 ECL (12-month): GBP 42,000
Provision increase: GBP 355,889 (+847%)
NOTE: The professional reviews the SICR trigger rationale
and the PD term structure assumptions; the agent calculated
the staging classification and lifetime ECL.
Sarah reviews the staging trigger (the 3-notch downgrade is a clear SICR indicator) and validates that the PD term structure the agent applied is consistent with her bank's internal rating model calibration. The next two lessons build the PD, LGD, and EAD components in depth (Lesson 4) and add macroeconomic scenario weighting (Lesson 5).
Jurisdiction Variant: Pakistan (SBP)
The State Bank of Pakistan (SBP) mandated IFRS 9 adoption with a staggered timeline: effective 1 January 2023 for large banks with assets of PKR 500 billion or more, and 1 January 2024 for all other banks and microfinance banks (the original timeline under BPRD Circular No. 04 of 2019 was revised by BPRD Circular No. 3 of 2022). Pakistani banks follow the same three-stage model described above, but SBP introduced a local modification: a mandatory floor of 0.5% for 12-month PD on any performing Stage 1 exposure, regardless of the bank's internal model output. This prevents banks from understating ECL on newly originated assets. Additionally, SBP requires Pakistani banks to use a minimum of three macroeconomic scenarios (base, optimistic, pessimistic) weighted by probability, and the regulator publishes suggested macroeconomic variable paths during periods of economic stress. Banks reporting to SBP must reconcile their IFRS 9 provisions against SBP's Prudential Regulations for Corporate/Commercial Banking (PR-R8), which may produce a higher regulatory provision floor than the IFRS 9 model output. The banking plugin's pakistan-sbp jurisdiction overlay captures these local modifications.
Try With AI
Use these prompts in Claude or your preferred AI assistant to explore this lesson's concepts.
Prompt 1: Staging Classification
I have a portfolio of 6 commercial loans. Classify each
into IFRS 9 Stage 1, 2, or 3. For each Stage 2 or Stage 3
classification, state the SICR trigger or default criterion.
Facility A: Current, 0 DPD, rating unchanged since origination
Facility B: Current, 0 DPD, rating downgraded 3 notches
Facility C: 45 DPD, rating downgraded 1 notch
Facility D: 95 DPD, borrower filed for bankruptcy
Facility E: Current, 0 DPD, placed on credit watchlist
Facility F: Was 60 DPD, now current for 12 months, rating
restored to origination level
For Facility F, explain whether it can return to Stage 1
and what evidence would be required.
What you are learning: Staging classification is the first step in every ECL calculation. By classifying facilities yourself, you develop the judgment the ifrs9-staging skill encodes — and you can verify whether the skill is staging correctly when you use it on real data.
Prompt 2: The Staging Cliff Effect
A bank holds a 20-year residential mortgage with these
parameters:
- PD (12-month): 0.8%
- LGD: 15%
- EAD: $450,000
Calculate the 12-month ECL for Stage 1.
Now assume the borrower's PD increases and the mortgage
moves to Stage 2. Using these simplified lifetime
parameters:
- Average annual marginal PD: 1.8% per year
- LGD: 15% (constant)
- EAD declining linearly from $450,000 to $0 over 20 years
- Ignore discounting for simplicity
Calculate the approximate lifetime ECL.
How many times larger is the lifetime ECL compared to
the 12-month ECL? Explain why this cliff effect matters
for bank financial reporting.
What you are learning: The staging cliff effect is the most consequential feature of IFRS 9 for bank earnings. A portfolio migrating from Stage 1 to Stage 2 can produce provision increases of 5-10x, directly reducing reported profit. Understanding this cliff helps you interpret bank financial statements and evaluate whether staging decisions are reasonable.
Prompt 3: IFRS 9 vs IAS 39
Explain to a non-technical board member why IFRS 9's
forward-looking ECL model is better than IAS 39's
incurred loss model. Use the 2008 financial crisis
as your primary example.
Structure your explanation as:
1. How IAS 39 worked (incurred loss — wait for evidence)
2. What went wrong in 2008 (losses existed but were
not recognised)
3. How IFRS 9 fixes this (forward-looking ECL, three stages)
4. The trade-off: IFRS 9 provisions are higher in good
times, but losses are never a surprise
Keep the explanation under 300 words and avoid jargon
that a non-accountant would not understand.
What you are learning: The ability to explain IFRS 9's rationale to non-technical stakeholders is a professional skill. Board members, audit committee chairs, and senior management need to understand why provisions change, and the IAS 39 vs IFRS 9 comparison is the clearest way to explain the "why" behind the numbers.
Flashcards Study Aid
Continue to Lesson 4: PD, LGD, and EAD — Building the ECL Components →