The IFRS 9 Requirement

IFRS 9 requires ECL measurements to reflect "reasonable and supportable information" about past events, current conditions, and forecasts of future economic conditions. This is a fundamental shift from the IAS 39 incurred loss model.

IFRS 9 Para 5.5.17

"An entity shall measure expected credit losses... in a way that reflects... reasonable and supportable information that is available without undue cost or effort at the reporting date about past events, current conditions and forecasts of future economic conditions."

Past Events
Historical loss experience and default data
Current Conditions
Current credit quality and economic state
Future Forecasts
Macroeconomic projections and scenarios

Scenario Design

IFRS 9 requires an unbiased, probability-weighted estimate of ECL. This is typically achieved using multiple macroeconomic scenarios.

Common Scenario Framework

Optimistic (Upside)
  • Strong economic growth
  • Low unemployment
  • Rising asset prices
  • Lower default rates
  • Typical weight: 15-20%
Base Case
  • Most likely outcome
  • Consensus economic forecasts
  • Moderate growth
  • Stable conditions
  • Typical weight: 50-60%
Pessimistic (Downside)
  • Economic recession
  • Rising unemployment
  • Falling asset prices
  • Higher default rates
  • Typical weight: 20-30%

Sample Scenario Variables

Variable Optimistic Base Pessimistic
GDP Growth (Year 1) +3.5% +2.0% -1.5%
GDP Growth (Year 2) +3.0% +2.2% +0.5%
Unemployment Rate 3.5% 4.5% 7.0%
House Price Change +8% +3% -15%
Interest Rates 4.0% 5.0% 3.0%
Scenario Probability 15% 55% 30%

Probability-Weighted Approach

The final ECL is calculated as the probability-weighted average of ECL under each scenario:

ECL = Σ (ECLscenario × Probabilityscenario)

Example Calculation

Scenario ECL Probability Weighted ECL
Optimistic $800 15% $120
Base $1,200 55% $660
Pessimistic $2,500 30% $750
Final Probability-Weighted ECL $1,530
Non-Linear Relationship

Note that the probability-weighted ECL ($1,530) is higher than the base case ECL ($1,200). This reflects the non-linear relationship between economic conditions and credit losses - downside risks typically impact ECL more than upside benefits reduce it.

Macroeconomic Models

Macro models translate economic forecasts into risk parameters (PD, LGD).

PD Adjustment Approaches

TTC to PIT Conversion

Convert through-the-cycle PD to point-in-time using:

  • Scalar adjustment factors
  • Z-factor or Merton-type models
  • Regression-based adjustments
Direct Macro Regression

Model default rates directly as function of macro variables:

ln(PD/(1-PD)) = α + β1GDP + β2UNEMP + ...

Key Macro Drivers by Portfolio

Portfolio Primary Driver Secondary Drivers
Residential Mortgages House Price Index Unemployment, Interest Rates
Consumer Credit Unemployment Real Income, Inflation
Corporate Loans GDP Growth Corporate Profits, Credit Spreads
Commercial Real Estate CRE Price Index Vacancy Rates, GDP

Management Overlays

Management overlays are adjustments to modeled ECL to capture information not reflected in the models.

Valid Reasons for Overlays
  • Emerging risks not in historical data
  • Sector-specific stress (e.g., COVID impact on hospitality)
  • Known model limitations
  • Recent credit events not yet in data
  • Government support measures
Governance Requirements
  • Clear documentation of rationale
  • Approval by appropriate committee
  • Regular review and expiry dates
  • Quantification methodology
  • Back-testing against outcomes
Best Practice

Overlays should be temporary adjustments with clear sunset provisions. Over time, the underlying models should be enhanced to incorporate the factors that necessitated the overlay.