Common Challenges In Expected Credit Losses (ECL) Assessment

Expected Credit Losses Framework

Both US GAAP (ASC 326 - CECL model) and IFRS 9 require entities to estimate the expected credit losses (ECL) by incorporating historical experience, current conditions and forward-looking information. While IFRS 9 applies a staging framework, the CECL model considers the lifetime ECL from initial recognition. Despite the difference, the core principle aligns in estimating credit losses with a forward-looking approach.

Short-term trade receivables and provision matrix 

For trade receivables and similar short-term items without a significant financing component, IFRS 9 mandates a simplified approach, under which lifetime ECL is typically estimated using a provision matrix based on historical loss rates by aging category.

For longer-dated receivables or those with a significant financing component, applying the simplified approach is a policy choice and alternative methods may be considered. The CECL model also permits the use of similar aging-based loss rate methods for measuring lifetime ECL. 
 

Other financial exposures

For other financial exposures, entities applying the general approach under IFRS 9 or the CECL model typically use a probability-of-default (PD) methodology: 

ECL = PD x LGD x EAD

Where: 

• PD (Probability of Default): Likelihood of default over the relevant period
• LGD (Loss Given Default): the expected shortfall after considering recoveries, collateral and other credit enhancements
• EAD (Exposure at Default): the expected amount outstanding at the point of default 

While the formula is straightforward, estimating each of the three parameters can be complex. 

Challenges In ECL Computation

Ambiguous framework for credit status

ECL estimation requires clearly defined criteria to distinguish between performing assets, assets showing credit deterioration, and credit-impaired exposures, as probabilities are assigned across these states. Common challenges include:

• Different default or delinquency definitions may be applied across entities and for different purposes. Financial institutions, in particular, often encounter misalignment between internal credit risk management frameworks, regulatory reporting requirements and financial reporting definitions;
• Some non‑financial corporates may not formally define default or credit deterioration beyond extreme events such as bankruptcy and may treat recurring late payments as part of “normal” business practice. This weakens the disciplined identification of deteriorating or defaulted exposures;
• Certain products such as revolving credit facilities and on-demand loans may not have fixed repayment schedule or clearly defined contractual terms. Nevertheless, ECL must still be measured over the expected life of the exposure; and
• Recurring repayment delays introduce complex recording challenges due to their irregular nature and variable durations. This makes consistent tracking and categorisation difficult for historical loss rate analysis, leading to incomplete or inconsistent datasets that distort credit risk assessment and undermine the development and application of reliable provision matrices and benchmark probability of PD and LGD assumptions.

Data quality and availability of historical data 

High-quality historical data forms the foundation of reliable ECL calculations, particularly for methodologies that rely on historical credit statistics. However, entities often face significant structural limitations, including:

• Lack of detailed tracking of billing dates, partial settlements, overdue periods, and the timing of provisions which limits the accuracy of aging analyses;
• Newer businesses or specific markets frequently lack sufficient ageing statistics, making it difficult to construct meaningful provision matrices;
• Billing and accounting systems that embed survival bias, capturing only settled or overdue receivables and thereby skewing historical loss rates;
• Diverse counterparty profiles that complicate meaningful segmentation of exposures for robust provision matrix development; and
• Growing counterparty volumes that overwhelm spreadsheet-based processes, creating inefficiencies, control weaknesses, and heightened error risk. 

Market benchmarking difficulties 

Where internal data is insufficient, organisations often rely on market benchmarks to estimate default rates and recovery assumptions. While this reduces the need for detailed financial analysis of individual counterparties, several practical limitations remain:

• Large counterparty populations make consistent industry or risk segmentation difficult;
• Financial statements of counterparties are often unavailable for credit assessment;
• Credit ratings frequently rely on subjective judgements applied without consistent frameworks;
• Limited visibility over counterparties' other liabilities hampers assessment of repayment priority;
• Information on collateral or credit enhancements is typically unavailable; and
• Macroeconomic variables chosen for forward-looking adjustments may not be appropriate or sufficiently accurate. 

Practical Considerations

In summary, the challenges of ECL computation primarily stem from limitations in the availability and quality of internal credit statistics, the absence of clear and consistently applied frameworks for assessing counterparty credit status and the inherent subjectivity involved in qualitative assessments. 

To address these challenges, organisations should focus on stronger coordination across departments to maintain detailed and accurate billing records, setting up a unified credit assessment framework applicable across different reporting purposes, develop dedicated systems to support ECL data and analysis, and maintain comprehensive documentation to support qualitative judgements. These measures will significantly enhance the robustness and defensibility of ECL computations.