Concept 1 of 61 min
IFRS 9: stages, expected credit loss and scenarios
Stage 1, 12 months
| PD | 3% a year |
|---|---|
| LGD | 50% |
| Exposure | 20,000,000 |
- Loss if default = 0.5 × 20,000,000 = 10,000,000
- ECL = 0.03 × 10,000,000 = 300,000
A performing loan carries a provision for 12 months of expected loss.
Stage 2, lifetime
| PD | 3% a year |
|---|---|
| Survival a year | 0.97 |
| Remaining life | 4 years |
| Loss if default | 10 million |
- Cumulative PD = 1 − 0.974 ≈ 0.1147
- ECL = 0.1147 × 10 = 1.147 million
1,147,000 against 300,000: moving to stage 2 multiplies the provision by nearly 4, the cliff effect.
- Stage 112-month expected credit loss
- Stage 2Significant increase in credit risk: lifetime loss
- Stage 3Credit-impaired: lifetime loss, interest on the net amount
Stage 3 credit-impaired loans keep lifetime loss, with interest on the net amount. ECL = PD × LGD × EAD weighted across scenarios, cumulative PD is 1 − (1 − annual PD)ⁿ, and cured loans can move back.
The jump from 12 months to the whole life can multiply a provision: the cliff effect. 30 days past due is a rebuttable presumption that credit risk has risen significantly.
Credit losses are asymmetric, so probability-weighting scenarios captures more than a single central case. Adding the annual PD year after year slightly overstates it, since a borrower can default only once.
After a restructuring and a probation period of regular payments, a loan can cure from stage 3 to stage 2 and then stage 1. Using the recovery rate instead of the loss rate is the classic slip.
- Keeping a loan in stage 1 until a payment is missed.
- Using the recovery rate instead of the loss rate.
- Provisioning on the central scenario alone.
The economic outlook worsens sharply. Walk me through what happens to a bank's IFRS 9 provisions.
A complete answer covers
- Point-in-time PDs rise as scenarios and their weights turn worse.
- Some loans show a significant increase in risk and move to stage 2.
- Moving from 12-month to lifetime loss makes provisions jump: the cliff effect.
- Collateral values and LGDs fall too, adding to the charge.
Also in our technical interviews, on camera, with a score.
Sources: IFRS 9, section 5.5 (impairment)
Concept 2 of 61 min
PD and LGD models: default, cycles, ranking and calibration
The why, the charts, the worked example and the interview questions are in Essentials, with every quiz.
See the plansSign inConcept 3 of 61 min
Collateral, haircuts and exposure at default
The why, the charts, the worked example and the interview questions are in Essentials, with every quiz.
See the plansSign inConcept 4 of 61 min
Portfolio quality: NPLs, concentration, early warning and forbearance
The why, the charts, the worked example and the interview questions are in Essentials, with every quiz.
See the plansSign inConcept 5 of 61 min
Pricing loans on risk and transferring it
The why, the charts, the worked example and the interview questions are in Essentials, with every quiz.
See the plansSign inConcept 6 of 61 min
AdvancedSame expected loss, very different tails
The why, the charts, the worked example and the interview questions are in Essentials, with every quiz.
See the plansSign inRecap
- ECL = PD × LGD × EAD; stage 1 takes 12 months of losses, stages 2 and 3 the whole life: the cliff effect.
- Default = more than 90 days past due or unlikely to pay; cumulative PD = 1 − (1 − annual PD)years.
- Collateral counts after its haircut; undrawn lines count through a conversion factor.
- Concentration hurts: large exposures are capped at 25% of Tier 1, connected clients are added up, Herfindahl = Σ shares².
- A loan rate covers funding, expected loss, the cost of capital and running costs; RAROC (risk-adjusted return on capital) = (revenue − costs − EL) ÷ capital.
25 questions on Credit risk
Timed, like the real one, drawn from a bank of 33. Every answer explained.