Expected Credit Losses on Trade Receivables under IFRS 9: How to Build a Provision Matrix Step by Step

What is the simplified approach, and when is it mandatory?

The completed version of IFRS 9 issued in 2014 added impairment requirements based on expected credit losses, effective for annual periods beginning on or after 1 January 2018. ^[1]

For trade receivables and contract assets arising under IFRS 15 that do not contain a significant financing component (including where the IFRS 15 one-year practical expedient is applied), the loss allowance is always measured at an amount equal to lifetime expected credit losses. ^[2] In other words, the simplified approach is not optional here. For receivables that do contain a significant financing component, and for lease receivables, the entity may elect the same approach as an accounting policy, and the policy for each category can be chosen independently of the other. ^[2]

The practical benefit of the simplified approach is that it spares the entity from tracking significant increases in credit risk for each receivable, because the loss is measured over the lifetime from the outset.

What must the estimate reflect?

IFRS 9 requires expected credit losses to reflect three elements: an unbiased, probability-weighted amount determined by evaluating a range of possible outcomes; the time value of money; and reasonable and supportable information about past events, current conditions and forecasts of future economic conditions, available without undue cost or effort. ^[2]

The standard does not require every possible scenario to be identified, but it does require the possibility that a credit loss occurs, and the possibility that none occurs, to be considered, even if the probability of a loss is very low. ^[2] That is why receivables not yet past due are not automatically given a zero rate.

Reasonable and supportable information that reporting teams typically use includes:

• Internal: ageing reports, actual collection history for each band, write-off decisions and recoveries after write-off.
• About customers: changes in payment terms, requests to reschedule debts and breaches of credit limits.
• External: activity indicators for the sectors in which customers are concentrated, and economic forecasts published by official bodies.

A costly search for every piece of information is not required, as the standard asks for what is available without undue cost or effort. ^[2] But the data the organisation already holds must feed into the estimate.

How do you build a provision matrix?

The matrix is applied to a portfolio of receivables, with an expected loss rate for each ageing band, including the not-past-due band. It uses historical observed default rates over the expected life of the receivables, then adjusts those rates for macroeconomic factors that may affect future defaults. ^[3] The IATA Industry Accounting Working Group guidance refers to Illustrative Example 12 in IFRS 9 as one way of applying a provision matrix. ^[3]

In practice, work through these steps:

1. Extract historical data: receivables ageing at the end of each past period, what was collected, what was written off, and what was recovered after write-off.
2. Calculate a loss rate for each band: the proportion of receivables in that band that ended as an actual loss.
3. Separate what is not a credit loss: service disputes, price adjustments and discounts are dealt with apart from credit losses. ^[3]
4. Adjust for the future: if current or expected conditions are worse or better than in the historical data period, adjust the rates and document the rationale.
5. Update regularly: refresh historical rates and review forward-looking estimates periodically. ^[3]

How do you segment the portfolio?

A single rate for the whole portfolio is valid only if customers’ risks are similar. Segmentation usually follows the characteristics that genuinely change the loss pattern:

• Customer type: government bodies, companies and individuals.
• Payment terms: cash, 30-day credit and 90-day credit.
• Security: receivables covered by a letter of guarantee or credit insurance versus uncovered receivables.
• Sector or region: where their economic conditions differ.

If you assess some large or credit-impaired receivables individually, remove them from the group to which the matrix is applied so that the loss is not counted twice.

Hypothetical example: a matrix for one customer group

A hypothetical organisation provides services to companies on 30-day terms. After analysing its historical data and adjusting it for current conditions, it arrives at the following rates for its corporate customer group:

Total receivables are AED 135,000 and the required allowance is AED 4,600. If the existing allowance in the books is AED 3,000, an increase of AED 1,600 is recognised in profit or loss. If the existing balance is higher than required, the difference is reversed.

Suppose, in the same example, that the historical rate for the over-90-days band was 32%, and the organisation raised it to 40% because several of its customers operate in a sector whose activity has recently declined and whose payments have slowed. That is the forward-looking adjustment the standard requires, and it must be documented: which indicator was used, why this size of adjustment was chosen, and its effect on the allowance. Had the rate stayed at 32%, the allowance for this band would have been AED 1,600 instead of 2,000.

The rates here are illustrative inputs only. An actual estimate needs the organisation’s own data, with documentation of how each rate was derived and the rationale for each forward-looking adjustment.

Write-offs and recoveries: how do they feed the matrix?

The gross carrying amount of an asset is reduced directly when there is no reasonable expectation of recovering all or part of it, and the write-off is a derecognition event. ^[2] Give your organisation a written policy stating when that applies, such as when collection procedures have been exhausted or a bankruptcy ruling has been issued.

Write-off records and subsequent recoveries are the raw material for historical rates. If recoveries after write-off are not recorded, the matrix will overstate losses. If write-offs are delayed for years after hope of collection has gone, balances in the oldest band swell and the rate calculation is distorted. So tie write-off decisions to written criteria, and keep a log linking each written-off amount to its ageing band at the time of write-off.

Errors that recur when preparing the matrix

• Confusing late payment with loss: a receivable delayed by a customer’s long approval cycle may never become a loss, so test this with data rather than impressions.
• Including disputes in the rate: a dispute over service quality or price is not a credit loss. ^[3]
• Freezing the rates: using rates from previous years without updating or adjusting them for current conditions. ^[3]
• Neglecting the not-past-due band: giving it a zero rate despite the requirement to consider the possibility of loss. ^[2]

Application in Gulf organisations

Receivables portfolios in the Gulf often combine government clients, private companies and individuals, with different payment and loss patterns. So place government clients in a group of their own, and test whether their delays reflect genuine credit risk or a procedural cycle. If your data shows that receivables in this group are eventually collected despite delays, do not load them with the same rates as private companies, although they should still carry a rate of their own.

Compare the estimate with what was actually collected at the end of each period, and discuss material changes with the person responsible for financial reporting. Confirm which accounting framework your organisation must follow, whether IFRS as issued or as adopted by the regulator in your country, and the disclosure requirements that apply.

Conclusion

The simplified approach makes lifetime loss measurement mandatory for the trade receivables it covers, and the matrix is a practical way to apply it. Segment the portfolio by risk, give each ageing band a rate grounded in your own data and adjusted for current conditions, and update it at every reporting date.