# How to calculate bad debt provision under IFRS 9

If you have a large portfolio of trade receivables, then you face the same issue over and over again:

How to calculate bad debt provision to these receivables?

When I worked as an auditor, I used to discuss this issue with my colleagues very frequently.

Everyone of them agreed that yes, there is always some bad debt hidden among “healthy” receivables and it’s necessary to recognize some provision for that.

However, everyone had a different opinion on how to do it.

The most common approach was, to my surprise and disagreement, to create a provision in a few steps:

- Analyze receivables at the reporting date and sort them according to their aging structure
- Apply certain percentages of provision to the individual aging groups

Sounds easy, right?

In most cases, auditors applied something like 2% to trade receivables within maturity, 10% to trade receivables that were 1-30 days overdue… 100% to receivables more than 360 days overdue.

It always amazed me.

How the hell do you know that this particular company will suffer 10% credit loss on receivables that are 1-30 days overdue???

For me, it always seemed that these numbers were made out of thin air.

It was long time before IFRS 9 was adopted.

Now, luckily, IFRS 9 tells us how to create bad debt provision for trade receivables and how to get these percentages.

In this article, I’d like to explain this methodology and illustrate it on a simple example.

## What do the rules in IFRS 9 say?

IFRS 9 requires you to recognize the impairment of financial assets in the amount of expected credit loss.

In fact, there are 2 approaches for doing so:

*General approach*

In general approach, there are 3 stages of a financial asset and you should recognize the impairment loss depending on the stage of a financial asset in question.Thus, the impairment loss is either in the amount of a 12-month expected credit loss (ECL) or a lifetime expected credit loss (ECL).You can read more about the general approach here.There are a lot of implementation troubles and challenges, for example:

- How do you determine in which stage the financial asset is?
- How do you determine when the credit risk in some financial asset has significantly increased?
- How do you calculate 12-month ECL and lifetime ECL?
- How do you get and update your inputs into the ECL calculations?

Therefore, IFRS 9 permits an alternative for some type of financial assets:

*Simplified approach*

In simplified approach, you don’t have to determine the stage of a financial asset because the impairment loss is measured at lifetime ECL for all assets.This is great news because lots of troubles simply disappear.

However, let me warn you that the simplified approach is not for everybody and even if it’s simplified, you still need to make some calculations and effort.

## Who can apply simplified approach?

OK, that’s not the best question in the world, because everybody can apply simplified approach.

** Type of financial asset** is more important here.

**Have you already checked out the IFRS Kit ? It’s a full IFRS learning package with more than 40 hours of private video tutorials, more than 140 IFRS case studies solved in Excel, more than 180 pages of handouts and many bonuses included. If you take action today and subscribe to the IFRS Kit, you’ll get it at discount! Click here to check it out!**

*Special For You!*You ** have to apply **simplified approach for:

- Trade receivables WITHOUT significant financing component, and
- Contract assets under IFRS 15 WITHOUT significant financing component

For these two types of assets you have no choice – just apply simplified approach.

On top of that, you can** make a choice** for:

- Trade receivables WITH significant financing component,
- Contract assets under IFRS 15 WITH significant financing component, and
- Lease receivables (IAS 17 or IFRS 16)

For these three types of financial assets, you can apply either simplified approach or general approach.

## Can one entity apply both models?

Yes, of course – but ** not to the same type of financial asset**.

Take a bank, for example.

Banks usually provide lots of loans and under IFRS 9, they have to apply general models to calculate impairment loss for loans.

But occasionally, banks can have other financial assets, too.

For example, they may rent redundant offices and have lease receivables.

Or, they can provide advisory services and charge fees for that – thus they can have typical trade receivables.

For these types of assets, the same bank can apply simplified approach.

## How to apply simplified approach?

As written above, under simplified approach, you measure impairment loss as ** lifetime expected credit loss**.

IFRS 9 permits using a few practical expedients and one of them is a provision matrix.

What is a * provision matrix*?

Simply said, it is a calculation of the impairment loss based on the * default rate percentage* applied to the

*.*

**group of financial assets**Here, we have 2 important elements:

- Group of financial assets
- Default rates

Let’s break it down.

### How to group the financial assets?

When you are using provision matrix for simplification, you still need to be as close to reality as possible.

Therefore, before applying any loss rates, you should group your financial assets first.

** Segment** them.

The reason is that all trade receivables do not necessarily share the same characteristics and therefore, it would not be reasonable to put them into the same pocket.

How to group them?

It depends on what factors affect the repayment of your receivables.

Maybe you noted that your retail customers (individuals) are less reliable and slower in payments than your business customers (companies).

Therefore, your segments or groups would naturally be retail customers and business customers.

Or, maybe you sell in a few geographical regions and you noted that customers from the capital city pay more reliably than customers in the rural areas (maybe it has something to do with unemployment rate…)

So, your segments would be customers from cities and customers from countryside.

I think you get the point – you should select the grouping of your trade receivables (or other financial assets in questions) depending on your circumstances.

Just a few suggestions for segmenting:

- By product type;
- By geographical region;
- By currency;
- By customer rating;
- By dealer type or sales channel;

etc.

The important point here is that the customers within one group should have ** the same or similar loss patterns**.

### How to get the default rates?

Remember – do NOT just trump the default rates up, just like auditors from the intro of this article.

You should really calculate them based on your own data.

IFRS 9 says that you should:

- Derive the default rates from your own historical credit loss experience; and
- Adjust them for forward-looking information.

#### Historical default rates

First, you need to * analyze the historical credit losses*.

How?

You should take the appropriate period of time and analyze which portion of trade receivables created during that period went default.

Just be careful when * selecting the appropriate period*.

It should not be too short in order to make sense and it also should not be to long because market changes quickly and long period might incorporate market effects that are no longer valid.

I recommend selecting one or two years.

Then you are going to * select the time buckets*, or periods when the receivables were paid.

Finally, you would calculate the * default rate for each bucket*.

No worries if this seems too unclear – you can find the illustrative example below.

#### Forward-looking information

Once you have your historical default rates, you need to adjust them by the forward-looking information.

What is it?

**Have you already checked out the IFRS Kit ? It’s a full IFRS learning package with more than 40 hours of private video tutorials, more than 140 IFRS case studies solved in Excel, more than 180 pages of handouts and many bonuses included. If you take action today and subscribe to the IFRS Kit, you’ll get it at discount! Click here to check it out!**

*Special For You!*They are all information that could affect the credit losses in the future, for example macroeconomic forecasts of unemployment, housing prices, etc.

You should adjust historical default rates for the information that is relevant for your financial assets.

For example, let’s say the telecom company has 2 segments of receivables:

- Retail customers, or individuals and for this group, unemployment rates are important factor affecting the payment rate.

If unemployment goes up, the credit quality of trade receivables to retail customers worsens. - Business customers: for this group, GDP (gross domestic product) and inflation rate are important factors in this particular country.

How to incorporate the forward-looking information?

When there is a * linear relationship *between the macroeconomic factor (i.e. unemployment rate) and the input (i.e. increase/decrease in collection of receivables), then the incorporation is quite simple.

In this case, you need to observe what effect has the change in the parameter on your default rates and make simple adjustment (see illustration below).

However, when the relationship is * not linear*, then the adjustment might require some modeling using either Monte Carlo simulation or other similar methods.

## Example: Impairment of trade receivables under IFRS 9

ABC wants to calculate the impairment loss of its trade receivables as of 31 December 20X1. ABC’s policy is to give 30 days for the repayment of receivables.

Note: This is an important point – 30 days credit period means that these receivables have NO significant financing component and therefore, you don’t have to worry about the present values.

The aging structure of trade receivables as of 31 December 20X1 is as follows:

Days after issuing invoice | Amount outstanding |
---|---|

Within maturity (0-30 days) | 800 |

31-60 days | 350 |

61-180 days | 280 |

180-360 days | 170 |

> 360 days | 100 |

Total |
1 700 |

ABC decided to apply the simplified approach in line with IFRS 9 and calculate impairment loss as lifetime expected credit loss.

As a practical expedient, ABC decided to use the provision matrix.

First, ABC needs to ** calculate historical default rates**.

In order to have sufficient historical data, ABC selected the period of 1 year from 1 January 20X0 to 31 December 20X0.

During this period, ABC generated sales of CU 20 000, all on credit.

Then, we can split the whole analysis process into a few steps.

#### Step 1: Analyze the collection of receivables by the time buckets

ABC needs to analyze when the receivables were paid and sort them out into table based on number of days from creation of invoice until the collection of the receivable:

When paid? | Paid amount | Paid amount (cumulative) | Unpaid amount |
---|---|---|---|

Within maturity (0-30 days) | 7 500 | 7 500 | 12 500 |

31-60 days | 6 800 | 14 300 | 5 700 |

61-180 days | 3 000 | 17 300 | 2 700 |

180-360 days | 2 200 | 19 500 | 500 |

> 360 days | 500 = write-off | 19 500 | 500 = write-off |

Total |
20 000 |
n/a |
n/a |

*Notes:*

- The amount of CU 500 in the column “Paid amount” for > 360 days represents in fact defaulted, unpaid amount.
- Paid amount cumulative is calculated as paid amount in certain time bucket plus paid amount in the previous bucket, i.e. cumulative paid amount in 31-60 days is calculated as 6 800+7 500. The exception is > 360 days – here, we can’t include CU 500 because it is not paid.
- Unpaid amount in the last column = total of 20 000 less cumulative paid amount.

#### Step 2: Calculate the historical loss rates

Then, ABC can calculate the historical default loss rates as the loss amount of CU 500 divided by the amount unpaid (outstanding) at the end of each time bucket:

When paid? | Unpaid amount | Loss | Default rate |
---|---|---|---|

Within maturity (0-30 days) | 20 000 | 500 | 2.5% |

31-60 days | 12 500 | 500 | 4.0% |

61-180 days | 5 700 | 500 | 8.8% |

180-360 days | 2 700 | 500 | 18.5% |

> 360 days | 500 | 500 | 100.0% |

*Note: Default rate = loss divided by the unpaid amount.*

Here you might note that data shifted a bit.

Unpaid amount for “within maturity” row amounting to CU 12 500 is now in the “31-60 days” row.

That’s OK because we are calculating amounts that fell into certain time bucket – that is, in the beginning of that bucket, not at its end.

So, in “within maturity” bucket, ABC created CU 20 000 of trade receivables; in “31-60 days” bucket, ABC created CU 12 500, etc.

Also, why did we apply the loss of CU 500 to all buckets?

The reason is that all receivables that were written off (CU 500) were in each stage over their life.

For example, all written off receivables amounting to CU 500 were current (within maturity), or within those CU 20 000 and therefore we can say that the loss generated during 20X0 (tested period) is 500/20 000.

The same applies for any other time bucket.

Now, we are not done yet.

We have only calculated the historical loss or default rates.

We still need to incorporate the forward-looking information.

#### Step 3: Incorporate forward-looking information

This is more difficult, but let me just outline one very simple approach.

Let’s say that ABC’s credit losses show almost linear relationship with unemployment rates.

Please note that “unemployment rate” is NOT a prescription for you – you should find your own macroeconomic factors that could affect your credit losses.

And, let’s say that the statistical office in ABC’s country assumes that unemployment rate will go up from 5% to 6% in 20X2.

ABC’s experience is that when unemployment rate increases by 1%, it triggers the increase in default losses by 10% (note – you should be able to prove that).

Therefore, ABC may reasonably assume that the loss of CU 500 can increase by 10% because of the increase in the unemployment rate – that is, to CU 550.

Thus, the calculation of loss (default) rates adjusted by forward-looking information is as follows:

When paid? | Unpaid amount | Loss | Default rate |
---|---|---|---|

Within maturity (0-30 days) | 20 000 | 550 | 2.75% |

31-60 days | 12 500 | 550 | 4.40% |

61-180 days | 5 700 | 550 | 9.60% |

180-360 days | 2 700 | 550 | 20.40% |

#### Step 4: Apply the loss rates to the current trade receivables portfolio

And finally, coming to the end of this exercise, let’s apply these loss rates to actual portfolio of trade receivables as of 31 December 20X1:

Days after issuing invoice | Amounts outstanding | Loss rate | Expected credit loss |
---|---|---|---|

Within maturity (0-30 days) | 800 | 2.75% | 22.0 |

31-60 days | 350 | 4.40% | 15.4 |

61-180 days | 280 | 9.60% | 26.9 |

180-360 days | 170 | 20.40% | 34.7 |

> 360 days | 100 | 100.00% | 100 |

Total |
1 700 |
n/a |
199.0 |

Done.

ABC can recognize the impairment loss on trade receivables as

- Debit P/L Impairment loss on trade receivables: CU 199
- Credit Trade receivables – adjustment account: CU 199

Any questions? Please let me know in the comments below this article. Thank you!

##### Tags In

JOIN OUR FREE NEWSLETTER AND GET

report "Top 7 IFRS Mistakes" + free IFRS mini-course

Please check your inbox to confirm your subscription.

## 56 Comments

### Leave a Reply Cancel reply

### Recent Comments

### Categories

- About IFRS (12)
- Accounting estimates (IAS 8) (3)
- Consolidation and Groups (14)
- Employees (6)
- FAQ (1)
- Financial Instruments (33)
- Financial Statements (14)
- Foreign currency (7)
- How To (17)
- IFRS Accounting (60)
- IFRS Summaries (25)
- IFRS videos (36)
- Impairment of assets (5)
- Income Tax (7)
- Intangible assets (6)
- Inventories (10)
- Leases (12)
- Not just IFRS (7)
- Podcast (14)
- PPE (IAS 16 and related) (23)
- Provisions and Contingencies (4)
- Revenue recognition (10)
- Sectors&Industries (1)
- Uncategorized (2)
- US GAAP (2)

Excellent article on PBD.Thanq Silvia.

It is very useful overview of IFRS 9

Dear Silvia,

As a financing company, our loans receivable are of the same product type, i.e.,consumer loans. The terms of the loan are not more than 2 years, aging structures are the same as your examples, collections are analyzed by the loan system by time buckets, historical loss rates can be calculated, and forward-looking information is not hard to identify. With these data, can we apply the simplified approach in calculating the loan loss provision instead of the general approach?

Dear Dante,

I dont think you can use the simplified approach in this regards as these are interest bearing financial assets….. kindly refer to the General approach……my opinion

Dear Oladele,

Thank you very much for your opinion.

Hi Dante,

I agree with Oladele in this point. Consumer loans are loans – they were created as a result of providing financing to the clients, not as a result of selling goods/rendering services on credit. Therefore as such, they can’t be considered trade receivables. Anyway, you can still apply portfolio approach to these loans – just the percentages will be different. S.

Hi Silvia,

Thank you very much for your reply.

Really happy for all your efforts in this IFRS issues.

Please continue sharing your great knowledge in this regard.

Thank you, mussie 🙂

Dear Silvia,

Thank you so much for making IFRS so easy for us.

I have a question regarding individual receivables. What I mean is, if we take a debtor’s card and want to check the receivable by the end of a period, how can we extract to portion of the total impairment (and hence accrued provision).

Also its interesting, how do we treat the provision when the receivables get paid after some time (fully, of partially).

Thank you in advance!

Hi Natalia,

you can always exclude individual debtors from the group assessment – as a part of your segmentation strategy. Then you can assess the debtors individually and apply the above principles to the remaining receivables assessed in the group. If the receivable is paid, then there’s no impairment anymore, is there? 🙂

What I mean is after the impairment has occurred as a result of applying the provision matrix, and in the future periods some of the receivables (impaired) are collected. Probably we recognize profit, is it correct?

Yes, of course, that would be reversal of impairment loss. Anyway, you would do these calculation at each reporting date and you are just booking the change in provision (either increase or decrease) in comparison with the previously recognized amount. S.

Thank you Silvia. Your web site is the best finding for me in IFRS issues so far!!

Sincerely speaking, you are amazing..

Well done Silvia for putting theory into practice.

Thanks as always for the supper document. Regarding LGD and PD do you have any documents or workings like the above. How to determine PD and LGD is very important to get in to the loan provision. kindly do share if you have one.

Hi Gadissa,

thank you! Well, there are some materials available on the net if you look for some Big 4 guidance. Anyway, I can write an article in the future about more practicalities. Take care!

Hi Sylvia ,

I am from Canada. I am from the Real Estate industry dealing mainly in Residential leases. I found your article so insightful and easy to follow in the application of the calculation of the impairment provision for trade receivables. To my surprise the auditors in the current year audit that was just completed a few days back did not question our method of bad debts provision that we have been using since.

Certainly a lot of learning for me here that I will pass on to my staff.

Thank you very much for sharing.

Hi Ravi,

great, I am very happy that this article has a practical sense for you. In relation to auditors – well, sometimes they just accept whatever clients give them, as soon as it looks elaborate enough… I don’t want to be too critical, because I was the auditor, too! Take care! S.

This a powerful yet simple illustration! Even the ‘idiot on the street’ would have understood easily, I guess. Thank you!

Hi Charles,

OMG, I don’t think my readers are idiots on the street 🙂 Thank you! 🙂

Hi Silvia,

As usual, you are at your best. we need more practical oriented articles like this.

Live long.

Aaaa, thank you 🙂 Maybe if you can give me a hint what other practical topics you would like me to cover, that would be fantastic! S.

Dear Silvia,

This is really a very good Article explaining and giving further guidance on the application IFRS 9 on setting Bad Debt with illustrative examples. It will assist Finance and Accounting teams. I appreciate your relentless efforts,

Thanks, Shimellis 🙂

Silvia consistently amazes me. Your simplicity is incredible. Thanks for this. Keep up the good work. Cheers.

Hi! silvia TANZANIA We are not back,Thank you very much keep it up you made my day

Hi Silvia,

Marvellous job . I could only say that your work is great.

Thanks for sharing.

Great article as always Silvia ! This is so informative, gonna share that with my colleagues asap ! 🙂

Amazing work. Very informative. Keep the good work going.

Thank you for the absolutely amazing article, Silvia

A short question: can the simplified approach be applied to calculate the ECL for NPLs?

Hi Ivaila,

thank you! For non-performing loans, you still need to apply general approach, but as the loan is non-performing, it would be in stage 3 in most cases and you would calculate impairment loss equal to life-time ECL – which is very similar to simplified approach. S.

Thank you Silvia. Appreciated!

I have checked this article and illustrated example, However it doesnt talk on the use of collaterals to arrive at Impairement charges, I thot i heared somewea IFRS 9 provide such oppurtunity, ”To include collaterals on computation of Loan Loss Provision”

Hello Silvia,

I am Fredrick from Ghana, that was a very useful article. But if i can ask, what happens to forward-Looking when from experience the macroeconomic factors those not have a liner relationship with the historic loss rate?.

As I wrote above, you have to do some modelling and simulations, it’s not very easy to describe it in the article. Best, S.

I have been reading multiple articles on the topic and yours was really the most clear and understandable explanation that I have come across. I always felt that the simplified approach was not simple. Thank you for simplifying it. 🙂

Many Thanks!

Dear Silvia,

First of all, I would like to thank you for your interesting and practical article.

I have just one question for you: Can we still compute the default rates if there are no written off receivables during the analyzed period? Thank you!

Hi Serghei,

excellent question. Let me tell you this: the credit loss is simply the present value of difference between the amount of the receivable per contract and the amount expected to receive. The amount expected to receive is what matters here. In the above example, it was clear that we expect to receive everything less write-off. The same applies to any other situation – you don’t have to take write-offs into account, instead you should take the difference between the contractual amount and expected receipt. Of course, if you expect to receive everything (and there’s no significant financing component), then you have no credit loss. Anyway, for large portfolios, I would perform analysis in more years separately to see how the loss rates develop. S.

Hello Silvia, thank you for such a wonderful article. You said on ‘how to group the financial assets’ that not all trade receivables share the same characteristics and therefore should not be placed into the same bucket. And on grouping them it depends on the factors affecting the repayment of those receivables for example ‘Geographical region’ in such a case does it mean the rates/percentages for customers from cities and customers from countryside within the same business will be different or a single average rate/percentage should be used?

Hi Emmanuel,

if your groups or segments of receivables have different characteristics affecting the collection of cash, then you should perform the same analysis for each group separately and it means that yes, they will have different loss default percentages. S.

Very amazing, simplified, and wonderful , but here you considered the amount of CU 500 >360 Days is written-off, suppose if i have A/R balance over 360 days and i don’t have bad debts ( written-off), shall i consider that balance over 360 days as the written-off and shall i divide it to the total credit sales over the year to get the credit loss ?

Hi Salah, well, please see my response to Serghei above – with some minor modification, I would write you the same. S.

Very neatly explained and simple to understand. Awesome work.

Hi Silvia,

i’m working at telecom company and we have huge trade receivables account with other operators, also banking interest rate is 10 % for one year. and based on historical data they paid their debt after 21 months from issued invoices, so that my question is about net present value (NPV), do i calculate NPV for current receivable balance as credit loss and illustrate in balance sheet or not? if yes with which rate? what is the IFRS 9 opinion abot net present value and net future value?

Thanks to you Maam Sylvia and thanks to your website. It’s so user friendly.. More power to you.. 🙂

Silvia madam made me to remember her name whenever I hear about IFRS. It’s like she made me to think IFRS is International Financial Reporting by Silvia(IFRS). Hahaha..

Thank you so much madam.

LOL 🙂

Dear Silvia,

For trade receivables with credit terms of almost 90 days. wouldn’t the historical trends and current situation enough to calculated ECL? Forward looking can be done if information is available without undue cost or effort. But for servicing business or trading business an effort and cost will be for incorporating forward looking factors, especially if modeling those factors would be more suitable.

Please give us your thoughts in this regard

Good Morning Dear Silvia, I am trying to apply your above mentioned Example: Impairment of trade receivables under IFRS 9. where i am not in a position to arrive the unpaid amount, if i take a cut of date 31 Dec 2017. For ex: total Revenue 4million for the period Jan to Dec 2017 and the corresponding receipts 3million and the balance 1 million is not yet due for collection, the same is due in subsequent period and received. now do i need to take all the receipts including for the period years to arrive unpaid amount? Please advise, and thank in advance for your Help.

Hi Neelakantam,

yes, of course. You should not trace total receipts of cash in 2017 – you should trace the receipts of the receivables generated in 2017. It means that sure, you have to take subsequent repayment of 1 mil. into account. S.

HI Silvia,

Wonderful practical insights, which my auditors hold the same view! Just wanted to ask further, for the forward looking rate, can it be a different rate for each reporting period, e.g one period Inflation rate and the other Interest rates? Reason is to minimise unnecessary large fluctuations in ECL’s to be provide for each reporting period – meaning to say taking the lowest possible rate every quarter that impacts the business and receivables. Thnxs!

Dear Mark,

No, because as I wrote above, you should identify the factor affecting the behaviour of your receivables and I doubt it would change period to period. I mean – it can sometimes change, but the reason for the change should be justified with the change of various factors in environment really affecting your receivables, and NOT by the “avoiding large fluctuations in ECL”. S.

Superb article ! Really like it. I have a question, when there are multiple macroeconomic factors so what should be the approach

Aamir, in this case, you would need to apply some simulation techniques like described above. S.

Many thanks , dear Silvia.

A have a question: If in our analysis for previous receivables there were no unpaid/written off debt, what can we do for determining the percentages. Can it be zero for all of aging analysis??