Understanding Exposure at Default (EAD) in Risk Management

Why Exposure at Default (EAD) matters for your IFRS 9 ECL work

For banks, finance companies and corporates with credit exposures, EAD determines the monetary base to which Probability of Default (PD) and Loss Given Default (LGD) are applied — and thereby drives the Expected Credit Loss. An understated EAD produces understated provisions and potential misstatements in financial statements; an overstated EAD reduces reported profitability and may distort pricing and capital allocation.

Beyond accounting, EAD is central to model governance, reporting to the board and audit committees, and to disclosures under IFRS 7 Disclosures — all areas where stakeholders expect robust methodology, traceability and sensitivity analysis. Well‑documented EAD approaches also improve Risk Committee Reports and help manage the broader Impact of ECL on business decisions.

Core concept: definition, components and clear examples

What is EAD?

Exposure at Default is the expected monetary exposure to a counterparty at the time of default. It includes on‑balance sheet exposures (outstanding balances) and the expected portion of off‑balance sheet items (undrawn facilities, guarantees, letters of credit) that are likely to be drawn or called at default.

Key components

• On‑balance sheet outstanding (principal + accrued interest)
• Undrawn commitments multiplied by a Credit Conversion Factor (CCF)
• Contingent liabilities and collateral netting effects
• Forward‑looking behaviour (prepayment, drawdown patterns)

Simple numeric example

Loan A: outstanding balance = 100,000; undrawn committed facility = 50,000. If empirical analysis suggests a CCF of 40% for this product, EAD = 100,000 + (0.40 × 50,000) = 120,000. If LGD = 45% and 12‑month PD = 2% (for a stage 1 exposure), the 12‑month ECL ≈ EAD × PD × LGD = 120,000 × 0.02 × 0.45 = 1,080.

EAD under Three‑Stage Classification and ECL Methodology

IFRS 9’s Three‑Stage Classification affects the horizon used for PD and therefore how EAD is combined in the ECL equation. For Stage 1 you typically use 12‑month PD but modelled EAD may still use expected lifetime drawdowns for certain products; for Stage 2 and Stage 3 the entire lifetime PD is used. The ECL Methodology must therefore document whether EADs are modelled on a 12‑month or lifetime basis and why.

For a wider introduction to the ECL framework and how PD, LGD and EAD interact, see our Introduction to ECL.

Practical use cases and scenarios for your team

1. Retail credit card portfolio

Cards have large undrawn balances. Typical approach: estimate a product‑level CCF from 24 months historical drawdown behaviour, segmented by delinquency bucket. Example: outstanding balances average $200m; undrawn commitments $150m; CCF 60% → EAD = 200 + 0.6×150 = $290m.

2. Corporate revolving facility

For large corporates, drawdown behavior shifts with liquidity cycles. Use scenario modelling (base, adverse, severe) with weighted CCFs. Produce a stress table: base CCF 30%, adverse 45%, severe 60%; weight outcomes by scenario probability to obtain forward‑looking EAD.

3. Term loans with committed refinancing

Term loans often have minimal off‑balance exposures; focus shifts to prepayment and accrual inclusion (interest in arrears). Document whether accrued interest is included in EAD or treated separately for provisioning.

4. Non‑financial corporates and intercompany lending

If your company is non‑bank, see specific guidance for treatment of trade receivables and leases in ECL for non-financial companies. EAD modelling for receivables often uses invoice ageing cohorts and historic drawdown is less relevant.

Impact on decisions, profitability and reporting

Changes in EAD, all else equal, scale the ECL charge linearly. For example, a 10% increase in portfolio EAD moves provisions up by 10% (assuming unchanged PD and LGD). That can materially affect quarterly profit and regulatory capital ratios.

Accounting impact on profitability and capital

Properly modelled EAD avoids earnings volatility from ad‑hoc adjustments. Transparency on methodology reduces audit adjustments and supports management’s forward planning. Read more on how provisioning changes feed into capital and earnings in our piece about the Impact of ECL.

Reporting and disclosures

IFRS 7 Disclosures require entities to explain ECL methodology, assumptions and sensitivity. Practical reports to the Risk Committee should include EAD drivers, CCF ranges, scenario tables, and reconciliation of opening vs closing EAD by segment. For corporate reporting and investor relations, coordinate with the team responsible for Disclosures & investors to ensure consistency between internal risk reports and external disclosures.

Also ensure consistency with any public ECL disclosures and the internal pack used for the audit committee.

Common mistakes and how to avoid them

• Ignoring drawdown patterns: Using a blunt average CCF can misstate EAD. Avoid by segmenting by product, tenure and delinquency.
• Stale data: Historical CCFs without adjustment for macro changes produce bias. Use forward‑looking overlays tied to macro scenarios.
• Poor segmentation: Aggregating heterogeneous products (e.g., credit cards and mortgages) will hide risk drivers — model separately.
• Lack of reconciliation: Not reconciling modelled EAD to accounting balances invites audit queries. Maintain a reconciliation table for each period.
• Weak governance: No model validation or version control. Implement formal Risk Model Governance with independent validation and periodic backtesting.
• Disclosure gaps: Missing or inconsistent narrative in the notes (IFRS 7 Disclosures) leads to regulatory questions — coordinate with finance and disclosures teams and reference our ECL disclosure guidance in internal packs such as ECL disclosure.

Practical, actionable tips and a checklist

1. Segment exposures: By product, origination vintage, geography, and delinquency. Example segments: mortgages, term loans, credit cards, overdrafts.
2. Choose method for off‑balance items: Empirical CCF, behavioural CCF (from customer lifecycle modelling), or stress scenarios. Document rationale and sample sizes.
3. Include accruals and fees: Confirm whether accrued interest and fees are part of EAD under accounting policy; include them consistently.
4. Forward‑looking overlays: Link CCFs to macro indicators (GDP, unemployment) and produce at least 3 scenarios with weights.
5. Validation and backtesting: Backtest predicted drawdowns vs actuals over rolling windows (e.g., 24 months) and log p‑values and performance metrics in validation reports.
6. Document governance: Keep an auditable repository of model versions, assumptions, validation reports and approvals under your Risk Model Governance framework.
7. Reconciliation and controls: Produce a reconciliation each reporting period: modelled EAD -> general ledger exposure -> disclosure table.
8. Automate reporting: Use reproducible pipelines so EAD calculations, sensitivity runs and disclosure tables are generated quickly for Risk Committee Reports and auditors.

KPIs / Success metrics

• Coverage ratio (Provisions / EAD) by product and portfolio
• CCF by product and vintage — mean and 95% confidence interval
• % of total exposure with modelled EAD vs proxy
• Backtest accuracy: predicted drawdown vs actual (MSE or MAE)
• Time to produce EAD & ECL report (target: ≤ 5 business days after period end)
• Number of audit or regulator queries related to EAD per year (target: 0–1)
• Volatility of ECL as a % of provisions across scenarios

FAQ

Reference pillar article

This article is part of a cluster that complements our pillar guide The Ultimate Guide: The basic equation for calculating ECL – explanation of PD, LGD, and EAD, how the formula is applied in practice, and a simple illustrative example. For the fundamental equation and a worked example tying EAD to PD and LGD, refer to that guide and the broader material on Expected credit losses (ECL.

When operationalising EAD you will rely on robust inputs and governance: ensure your teams use consistent ECL data definitions, align with the ECL disclosure schedule, and refer back to the Introduction to ECL if stakeholders need foundational context.