Probability of default (PD)

PD is usually the most important of the three ECL parameters, because small changes in it move the allowance most. This guide explains the two PD horizons, the difference between point-in-time and through-the-cycle PDs, how PDs are estimated, and works through a point-in-time adjustment.

By Awais Jameel, Chartered Accountant. Reviewed by Muhammad Bilal, Chartered Accountant. 3 minute read.

Short answer

Probability of default (PD) is the likelihood that a borrower will default over a given period. IFRS 9 uses two horizons: 12-month PD for stage 1 assets and lifetime PD for stages 2 and 3. PDs must be point-in-time, reflecting current conditions and reasonable and supportable forecasts, unlike the through-the-cycle PDs used for regulatory capital. Lenders estimate PDs from their own default history by rating grade or segment, from external ratings, or from scoring models, and adjust them for the economic outlook.

At a glance

Measures
Likelihood of default over a period
Stage 1
12-month PD
Stages 2 and 3
Lifetime PD (100% in stage 3)
IFRS 9 needs
Point-in-time, forward-looking PDs
Regulatory capital
Through-the-cycle PDs
Excel
PD term structure model
Probability of default (PD)Measures: Likelihood of default over a period; Stage 1: 12-month PD; Stages 2 and 3: Lifetime PD (100% in stage 3); IFRS 9 needs: Point-in-time, forward-looking PDs; Regulatory capital: Through-the-cycle PDs; Excel: PD term structure model.KEY FACTS AT A GLANCEProbability of default (PD)MeasuresLikelihood of defaultover a periodStage 112-month PDStages 2 and 3Lifetime PD (100% instage 3)IFRS 9 needsPoint-in-time,forward-looking PDsRegulatory capitalThrough-the-cycle PDsExcelPD term structure modelTax BakersProbability of default (PD)Measures: Likelihood of default over a period; Stage 1: 12-month PD; Stages 2 and 3: Lifetime PD (100% in stage 3); IFRS 9 needs: Point-in-time, forward-looking PDs; Regulatory capital: Through-the-cycle PDs; Excel: PD term structure model.KEY FACTS AT A GLANCEProbability of default (PD)MeasuresLikelihood of default over a periodStage 112-month PDStages 2 and 3Lifetime PD (100% in stage 3)IFRS 9 needsPoint-in-time, forward-looking PDsRegulatory capitalThrough-the-cycle PDsExcelPD term structure modelTax Bakers
Key facts at a glance, as set out in this guide.

What is probability of default?

The chance that a borrower fails to meet its obligations within a set period, defined using the lender's definition of default, which IFRS 9 presumes to be no later than 90 days past due. A 12-month PD of 2% means that, of 100 similar borrowers, two are expected to default in the next year, though which two is unknown. PD is one of three parameters in ECL = PD x LGD x EAD; see how to calculate ECL.

What are 12-month PD and lifetime PD?

12-month PD is the probability of default within the next 12 months, used for stage 1 assets. Lifetime PD is the probability of default at any point over the asset's remaining life, used for stage 2 assets and, at 100%, for stage 3. Lifetime PD is built year by year from marginal PDs; see lifetime PD and the term structure.

Point-in-time vs through-the-cycle probability of default

Point-in-time vs through-the-cycle PDPoint-in-time vs through-the-cycle PDTOPICPoint-in-timeThrough-the-cycleReflects current conditionsYesAveraged outIncludes economic forecastsRequiredNot requiredMoves with the economyStronglyLittleUsed for IFRS 9 ECLRequiredNot allowed aloneUsed for regulatory capitalNot usuallyUsually
IFRS 9 needs PDs that move with the economy.

A through-the-cycle PD is a long-run average that changes little as the economy moves. A point-in-time PD rises in downturns and falls in good times. IFRS 9 requires the second, so lenders whose regulatory models produce through-the-cycle PDs must convert them before using them in ECL.

An example: adjusting a PD for the economy

A lender's long-run average PD for a segment is 2.0%, observed when unemployment averaged 4.5%. Its history shows that each percentage point of extra unemployment raises PDs by about 25%. The forecast for next year is unemployment of 6.0%.

Forecast unemploymentScalarPoint-in-time PD
3.5%0.75x1.50%
4.5% (long-run average)1.00x2.00%
6.0% (forecast)1.375x2.75%
7.5%1.75x3.50%

The scalar is 1 + 25% x (6.0 - 4.5) = 1.375, so the point-in-time PD is 2.75%. This simple scalar is transparent and easy to audit; larger lenders use regression models linking default rates to several economic variables, or the Vasicek formula. The Point in time sheet of the PD term structure model (Excel) runs the scalar.

How are PDs estimated?

  • Internal default history: observed default rates by rating grade, score band or segment over many years, the most common source for banks.
  • Scorecards and rating models: statistical models using borrower data, such as financial ratios for companies or credit bureau scores for individuals.
  • External ratings: for bonds, banks and large corporates, rating agency default studies; see PD from external ratings.
  • Market data: PDs implied by credit default swap spreads or bond prices, adjusted for the risk premium they contain.

How do companies outside banking estimate PD?

A company with an intercompany loan or a deposit rarely has default data of its own. It usually maps the borrower's credit quality to a rating grade, using the borrower's own rating, its financial ratios, or a parent's rating, and takes the PD for that grade from a published default study. Trade receivables usually skip PD altogether and use a provision matrix of loss rates instead.

What are common PD mistakes?

  • Using regulatory through-the-cycle PDs in ECL without a point-in-time adjustment for current conditions.
  • Treating a 12-month PD as if it applied unchanged to every year of a loan's life.
  • Estimating PDs from too short a history, which may contain no downturn and so understate risk; back-testing catches this.

See also significant increase in credit risk, which compares lifetime PDs at origination and today.

Need help applying the standards?

Our Chartered Accountants help finance teams and students apply IFRS and US GAAP to real transactions.

Book a call Chat on WhatsApp

Questions people ask

What is probability of default in IFRS 9?

The likelihood that a borrower defaults over a given period: 12 months for stage 1 assets and the remaining lifetime for stages 2 and 3.

What is the difference between point-in-time and through-the-cycle PD?

A point-in-time PD reflects current conditions and forecasts; a through-the-cycle PD is a long-run average. IFRS 9 requires point-in-time PDs.

How do banks estimate probability of default?

From internal default history by rating grade or segment, scorecards, external ratings and market data, adjusted for the economic outlook.

Can regulatory PDs be used for IFRS 9?

Only after adjustment, because regulatory PDs are usually through-the-cycle and may include conservative floors and margins.

Sources

Every fee, date and rule on this page was taken from these official and primary sources.

  1. IFRS Foundation: IFRS 9 Financial Instruments

Rules and fees change. If you are reading this long after October 4, 2026, confirm the figures with the source before you rely on them.

ECLLifetime PD and the PD term structureECLPD from external credit ratingsECLHow to calculate ECL, step by step

More in ECL

Previous in this topicCredit-impaired assets and stage 3TopicAll 40 guides in ECL

This guide is general information. It is not tax or legal advice for your situation.