TL:DR
IFRS 9 ECL (expected credit loss) is a forward-looking estimate of the losses a business expects on its financial assets before a default happens. You calculate it as PD × LGD × EAD, then sort each asset into one of three impairment stages. It replaced the incurred-loss method under IAS 39, so a provision now lands when credit risk rises, not after the money is already gone.
This IFRS 9 ECL overview covers the three stages, the UAE Central Bank rules, and the parts that trip up most finance teams once they move from reading the standard to actually running it. If you have been hunting for where to start, this is the shortcut.
Here is the position up front. Most of the pain in an IFRS 9 rollout is not the maths. It is the data. Clean up your payment history and staging rules and the calculation almost writes itself. The rest of this piece is about getting there.
Why this matters for UAE businesses now
The Central Bank of the UAE ran a Prudential Filter to phase in ECL provisions gradually from 2020 to 2025, cushioning bank capital from the first hit. That transition window has closed. Full IFRS 9 provisions now flow straight through, with no regulatory relief to soften them, so the numbers you report carry their full weight.
The IFRS 9 ECL formula, in one line
Three inputs drive every expected credit loss figure:
- Probability of Default (PD). How likely the customer is to stop paying.
- Loss Given Default (LGD). The slice of the exposure you actually lose if they do, after any recovery or collateral.
- Exposure at Default (EAD). How much they owe you at the moment of default.
Put them together and you get:
ECL = PD × LGD × EAD
What changes case to case is the horizon you apply that formula over, and that depends on which stage the asset sits in.
IFRS 9 ECL calculation example
Take a customer owing AED 200,000, with a 2% probability of default and a 40% loss given default. The maths is short:
| Exposure at Default (EAD) | AED 200,000 |
| Probability of Default (PD) | 2% |
| Loss Given Default (LGD) | 40% |
| ECL = PD × LGD × EAD | AED 1,600 |
That AED 1,600 is a Stage 1 figure, because it only covers losses expected in the next 12 months. Now suppose the same customer slips past 30 days late. The asset moves to Stage 2, and you re-run the exact same formula across the asset’s full remaining life instead of a single year. Same inputs, longer horizon, a bigger number.
Want more worked scenarios, staging transitions, and a full model walkthrough? Our ECL model IFRS 9 examples guide goes deep on that. This page stays focused on the stages, the UAE rules, and where rollouts actually get stuck.

The three stages of IFRS 9 ECL
The general approach sorts every financial asset into one of three buckets, based on how much its credit risk has moved since you first booked it. The stage sets the horizon; the horizon sets the size of the provision.
Stage 1: no significant increase in risk
The asset looks much like it did on day one. You provision for losses expected within 12 months, and interest income runs on the gross carrying amount. Most of a healthy book lives here, on lower provisions and routine monitoring.
Stage 2: risk has jumped significantly
Something changed. Payments are running late, the credit rating slipped, or the customer’s sector is under stress. Once credit risk has risen significantly since origination, the asset moves to Stage 2 and you provision for lifetime ECL, not just the next 12 months. Interest still runs on the gross amount. The trigger everyone argues about is what “significant” means, and that is a judgment call your policy has to pin down.
Stage 3: credit-impaired
There is now hard evidence of impairment. A covenant breach, a bankruptcy filing, a restructuring, or a market for the asset drying up. You still book lifetime ECL, but interest income switches to the net carrying amount, after the ECL is stripped out.
| Stage | Credit risk | ECL measured over | Interest on |
|---|---|---|---|
| Stage 1 | No significant increase | 12 months | Gross amount |
| Stage 2 | Significant increase | Lifetime | Gross amount |
| Stage 3 | Credit-impaired | Lifetime | Net amount |
Simplified vs general approach
Not every asset needs the three-stage machinery. IFRS 9 gives you two routes, and the one you pick depends on what you are provisioning.
The simplified approach applies to trade receivables and contract assets. You skip staging entirely and recognise lifetime ECL from the moment you book the asset, usually off a provision matrix built on historical loss rates and adjusted for what you expect ahead. Less admin, no significant-increase test to argue about.
The general approach is the full three-stage version, and it applies to loans, debt securities, and most other financial assets. More work, more judgment, but it gives you the risk detail a lender actually needs.
Most businesses run both. Simplified for the customer receivables, general for the lending book. It is rarely a single choice across the whole balance sheet.
Who Prima works with
Across industries and geographies — delivering measurable outcomes
🏦
Banks & Financial Institutions
IFRS 9 ECL models, credit risk, impairment methodology, regulatory reporting
🛡️
Insurance & Takaful Companies
IFRS 17 implementation, actuarial modelling, GMM/VFA/PAA, CSM calculations
🏢
Corporates & Multinationals
IFRS 15, 16, IAS 36 compliance, financial statements, internal controls
📈
Investment & Asset Managers
IFRS 9 classification, fair value (IFRS 13), derivative valuations, hedge accounting
🏗️
Real Estate & Construction
IFRS 16 lease accounting, IFRS 15 revenue recognition, project-based reporting
🏛️
Family Offices & Holding Groups
Governance frameworks, board advisory, risk management, portfolio evaluation
UAE regulatory requirements for IFRS 9
Central Bank of the UAE guidance
The Central Bank of the UAE sets specific expectations for how banks and businesses apply IFRS 9. The recurring themes are clean data, independent validation, and governance you can show an examiner. In practice that means:
- Regular back-testing of your assumptions against what actually happened
- Clear written documentation of every methodology you use
- An independent review layer, separate from the people who built the numbers
- Stress testing across several economic scenarios, not just the base case
Your provisioning also has to reflect UAE-specific conditions: GDP growth, oil price swings, the real estate cycle, and government spending. A generic global assumption set will not survive review here.
The Prudential Filter has ended
The five-year filter (2020 to 2025) let banks phase ECL in gradually and protect regulatory capital from a sudden shock. That relief is gone. Every provision now hits in full, so any business still treating IFRS 9 as a future problem is already behind. Increased provisions, tighter disclosure, and closer scrutiny are the new baseline, not a warning about what is coming.
Putting IFRS 9 into practice
Start with the data
Gather what the calculation actually needs: customer payment histories, credit scores and ratings, sector and location detail, collateral records, and macro forecasts. Then segment. UAE businesses often split by emirate, sector, and customer type, because a Dubai construction client and an Abu Dhabi retailer do not carry the same risk profile. Better segmentation, sharper ECL.
Automate the calculation
Manual spreadsheets hold up until your book grows, then they stop. Purpose-built IFRS 9 software runs the calculations, keeps an audit trail, and produces the reporting your auditor and regulator want to see. Look for something that handles AED, local macro inputs, and UAE reporting formats out of the box, rather than a generic engine you have to bend into shape.
Build in the forward-looking view
IFRS 9 requires macro forecasts inside the numbers: oil price paths, UAE GDP, interest rate expectations, the property market, infrastructure spend, and regional risk. UAE banks already fold oil and inflation into their provisioning so the output tracks real conditions rather than a static average. The trap here is double-counting, and it gets its own section below.
Where IFRS 9 gets stuck
Having run these across banks and corporates, the same three things cause most of the delay. None of them are the formula.
Thin or messy data
The calculation wants clean histories, recovery rates, and enough default observations to build a credible PD. Plenty of businesses simply do not have years of tidy records in one place. This is the part that actually eats the timeline. Where internal data runs short, external credit databases and sector benchmarks fill the gap, though you have to document the reasoning.
Defining “significant increase in credit risk”
The move from Stage 1 to Stage 2 hangs on the SICR test, and the standard leaves it open on purpose. You set the criteria. A 30-day-past-due backstop is common, alongside rating downgrades, covenant breaches, and sector stress signals. Set the bar too low and everything floods into lifetime ECL. Too high and you are recognising losses late, which is the exact failure IFRS 9 was built to kill.
Double-counting the forward-looking bit
Macro data is required, but it is easy to let the same risk show up twice, once in the PD and again in a management overlay. The fix is boring but it works: document each adjustment, define where it applies, validate against real outcomes, and apply it the same way every period.
ECL for manufacturing and financial-sector clients needs SICR criteria tuned to each industry’s own cycle, since a manufacturer’s stress signals look nothing like a bank’s.
IFRS 9 vs IAS 39: what actually changed
The old standard, IAS 39, waited. A loss was recognised only once a loss event had happened, so provisions arrived late and often too small. Regulators called it “too little, too late” after 2008, and that criticism is why IFRS 9 exists.
| Aspect | IAS 39 | IFRS 9 |
|---|---|---|
| Recognition trigger | Incurred loss events | Expected future losses |
| Time horizon | Historical focus | Forward-looking |
| Loss calculation | Point-in-time | Probability-weighted scenarios |
| Staging | Binary, impaired or not | Three-stage progression |
Regional banks including Emirates NBD and National Bank of Kuwait already apply the full IFRS 9 approach, with corporate tax effects folded into their provisioning from 2023, per their published 2023 financial statements. The practical effect: you book a provision on day one, on assets performing perfectly well, purely because some small probability of loss exists. That surprises people the first time it hits the P&L.

Implementation checklist for UAE businesses
A realistic rollout runs about ten months end to end, though a clean data environment can shorten that and a messy one will stretch it well past.
Assessment (Months 1 to 3)
- Inventory every financial asset that falls under IFRS 9, then check what data you actually hold on each
- Compare your current process against the standard and mark every gap in systems, skills, and documentation
- Set a timeline and budget you can defend, and line up the training or external help the gaps demand
Build and test (Months 3 to 8)
- Design your ECL methodologies and either build or buy the tooling to run them
- Stand up the governance and control framework, with documentation written as you go, not bolted on after
- Test against historical data, check the output makes sense, refine, then get independent validation where it is required
Go live and monitor (Month 8 onward)
- Deploy for live reporting, train the operational team, and watch the first cycle closely for surprises
- Wire in the IFRS 9 disclosure requirements from the first reporting date
- From there it is ongoing: back-test regularly, update for regulatory change, and keep tightening as you learn what your book really does

Prima Consulting
Need expert guidance on this topic? Our advisory team is available for a free initial consultation.
No obligation · Responds within 1 business day · GCC, Europe & APAC
IFRS 9 ECL: common questions
How do you calculate expected credit loss under IFRS 9?
Multiply three inputs: probability of default, loss given default, and exposure at default. A 2% PD, 40% LGD and AED 200,000 exposure gives an ECL of AED 1,600. Trade receivables usually skip the staging and use a simplified lifetime estimate off a provision matrix.
What are the three stages of IFRS 9 ECL?
Stage 1 covers assets with no significant rise in credit risk, measured on 12-month ECL. Second covers assets where risk has jumped significantly, measured on lifetime ECL. Last stage covers credit-impaired assets, still lifetime ECL, but with interest calculated on the net carrying amount.
When does an asset move from Stage 1 to Stage 2?
When credit risk has increased significantly since first recognition. Common triggers are payments more than 30 days late, a credit rating downgrade, a covenant breach, or clear sector-wide stress. Your own SICR policy sets the exact thresholds.
What is the simplified approach under IFRS 9?
A shortcut for trade receivables and contract assets. You recognise lifetime ECL from day one and skip the three-stage test entirely, usually using a provision matrix built on historical loss rates adjusted for what you expect ahead.
Does the UAE Prudential Filter still apply?
No. The filter phased ECL provisions in from 2020 to 2025 and that window has closed. Provisions now flow through in full, with no transitional relief, so UAE businesses feel the complete impact on their statements.
The shift from waiting on losses to forecasting them changes how you read your own book. Provisions stop being a year-end surprise and become an early signal. If you are building your first IFRS 9 calculation, or your auditor has flagged the staging, that is the work Prima’s risk and financial advisory team does every day. Start with a free consultation and we will tell you where your real gaps are, before the auditor does.
Free Consultation
Ready to discuss your IFRS 9 ECL rollout?
Prima’s advisors work with organisations across the GCC, Europe, and Asia-Pacific. First conversation is always free — no pitch, just expertise.
Author
-

Ibrahim Ahmed Zahidie, FCA, brings 18+ years of technical depth across IFRS financial reporting, regulatory risk frameworks, and business transformation in the banking sector. His experience spans KPMG and UBL, with a practice focus on IFRS implementation, disclosure optimisation, sustainable finance reporting, and digital compliance strategies for regulated institutions operating in Saudi Arabia, the UAE, Ireland, and European markets.






