Write-off ratio

The write-off ratio measures the value of loans removed from the balance sheet as uncollectible during a period, expressed as a percentage of the average gross loan portfolio. It is one of the core credit quality indicators for any lender, and the only one that records losses the institution has formally accepted rather than merely anticipated.

A write-off is an accounting action: the gross carrying amount of a loan is derecognised because there is no reasonable expectation of recovering it. Two points are frequently misunderstood:

• A write-off is not debt forgiveness. Unless the lender explicitly waives the debt, the legal claim survives derecognition and collection efforts may continue. Amounts subsequently collected are recorded as recoveries.
• A write-off is usually not a new loss. If the loan was already fully provisioned, the write-off simply removes the gross balance and the corresponding allowance together, with no further effect on profit. The loss was recognised when the provision was raised. Writing off an under-provisioned loan does hit the income statement — which is why a spike in write-offs alongside a stable provision expense is worth investigating.

Write-Off Ratio Formula

Write-off ratio = Value of loans written off during the period / Average gross loan portfolio × 100

Where the period is shorter than a year, annualise:

Annualised write-off ratio = (Write-offs in period / Average gross loan portfolio) × (12 / number of months) × 100

Average gross loan portfolio is normally the simple average of opening and closing balances. For portfolios that grow or contract sharply within the period, a 13-point monthly average gives a materially more accurate denominator and is the better practice.

Worked example

• Gross loan portfolio (1 January): 8,000,000
• Gross loan portfolio (31 December): 10,000,000
• Average gross loan portfolio: 9,000,000
• Loans written off during the year: 450,000
• Recoveries on previously written-off loans: 90,000

Write-off ratio = 450,000 / 9,000,000 × 100 = 5.0%

Gross vs Net: The Loan Loss Rate

The ratio above is the gross write-off ratio. Netting off recoveries gives the loan loss rate — sometimes called the net write-off ratio:

Loan loss rate = (Write-offs − Recoveries) / Average gross loan portfolio × 100

Using the same figures:

(450,000 − 90,000) / 9,000,000 × 100 = 4.0%

The gap between the two is a direct measure of collections effectiveness after write-off. An institution writing off 5% and recovering nothing is in a very different position from one writing off 5% and recovering a fifth of it, even though the gross ratio is identical.

Always state which version you are reporting. Gross and net write-off ratios are routinely conflated in management reporting and in published comparisons.

How Write-Offs Distort Portfolio at Risk

This is the most important practical point about the ratio, and the reason it should never be read alone.

Writing off a loan removes it from the gross loan portfolio. Because portfolio at risk (PAR) is calculated as overdue balances over gross portfolio, a write-off improves PAR twice: it removes the delinquent balance from the numerator and reduces the denominator.

Consider a portfolio of 10,000,000 with 800,000 in loans overdue more than 90 days:

• Before write-off: - Gross loan portfolio: 10,000,000

  • Overdue > 90 days: 800,000
  • PAR > 90: 8.0%

• After writing off the 800,000: - Gross loan portfolio: 9,200,000

  • Overdue > 90 days: 0
  • PAR > 90: 0.0%

Nothing about credit performance changed. No borrower repaid anything. An institution with a liberal write-off policy will always report better PAR than an identical institution with a conservative one.

The consequences:

• PAR and the write-off ratio must be read together. Falling PAR alongside a rising write-off ratio is portfolio cleaning, not improvement.
• PAR comparisons between institutions are meaningless without their write-off policies. Two lenders reporting PAR>30 of 4% may have entirely different underlying quality if one writes off at 180 days and the other at 360.
• A useful combined check is PAR plus the annualised write-off ratio, which is far less sensitive to policy differences than either figure alone.

Write-Off Policy

Every lender needs a documented write-off policy specifying:

• The trigger. Most commonly an aging threshold — for example, all loans more than 180 or 365 days past due. Some policies add case-specific triggers such as borrower death, business closure, absconding, or exhaustion of legal remedy.
• Whether write-off is automatic or case-by-case. Automatic aging-based write-off is more consistent and less open to manipulation; case-by-case allows judgement but requires strong controls.
• Approval authority. Write-offs are typically approved at management committee or board level, above defined thresholds.
• Treatment of accrued interest. Whether the written-off amount includes accrued and unpaid interest or principal only. This materially affects the ratio and must be applied consistently.
• Post-write-off collection. Whether and for how long recovery efforts continue, and how recoveries are recognised.
• Memorandum records. Written-off accounts should remain traceable off balance sheet, including for credit bureau reporting.

Conservative vs aggressive policy

• Reported PAR: Higher under late write-offs (e.g. 360+ days) vs lower under early write-offs (e.g. 90–180 days).
• Write-off ratio: Lower and lumpier under late write-offs vs higher and smoother under early write-offs.
• Balance sheet realism: Late write-offs carry assets with little recovery prospect, whereas early write-offs keep the balance sheet cleaner.
• Recovery rate on written-off loans: Lower when written off late (collection remedies exhausted) vs higher when written off early (collection still live).
• Risk of misuse: Late write-offs risk delaying loss recognition, while early write-offs risk masking underlying delinquency.

Neither is inherently correct. What matters is that the policy is documented, consistently applied, disclosed, and not changed opportunistically. A change in write-off policy makes ratios non-comparable across the change date, and should be disclosed whenever it occurs.

Note also that many regulators prescribe mandatory write-off timelines, which may override internal policy.

Relationship to IFRS 9 and Provisioning

Under IFRS 9, write-off occurs when there is no reasonable expectation of recovery, in whole or in part. It is a derecognition event rather than a measurement judgement — distinct from the expected credit loss allowance, which is an estimate carried against loans still on the books.

The sequence in a well-run portfolio is: a loan deteriorates and migrates to stage 2 and then stage 3, the ECL allowance builds toward the expected loss, and by the time write-off occurs the allowance largely covers the balance being removed. When that sequence works, write-off has minimal income statement impact.

Signals that it is not working:

• Write-offs materially exceeding the existing allowance, producing a charge at write-off. This indicates provisioning has been running behind actual loss experience.
• A very high allowance relative to eventual write-offs, indicating over-provisioning or delayed write-off of loans already fully provided for.
• Provision coverage ratio — allowance divided by non-performing loans — is the standard cross-check.

Common Measurement Pitfalls

Failing to annualise. A quarterly write-off ratio quoted without annualisation looks four times better than the equivalent annual figure. This is one of the most frequent reporting errors.

Denominator inconsistency. Opening balance, closing balance, simple average and monthly average all give different answers. On a fast-growing portfolio the difference is large — using the closing balance flatters the ratio significantly.

Lumpiness. Where write-offs are approved once or twice a year at board level rather than processed continuously, monthly and quarterly ratios become erratic and trend analysis breaks down. Rolling twelve-month figures are more readable.

Mixing principal and interest inconsistently between periods, or between the numerator and the policy definition.

Ignoring restructured loans. Rescheduling a distressed loan into a new contract can postpone or avoid write-off entirely while the underlying exposure remains impaired. A low write-off ratio alongside a high volume of restructuring warrants scrutiny.

Comparing across institutions without policy alignment. As above, the write-off ratio is only comparable between lenders using similar triggers and timing.

How to Use the Ratio in Practice

• Trend it, on a rolling twelve-month basis, on a fixed policy. Direction matters more than level.
• Read it beside PAR, provision coverage and restructuring volume. Any one of the four can be improved in isolation without any real change in credit quality; together they are much harder to distort.
• Break it down by vintage. Write-offs by origination cohort show which lending decisions produced the losses, rather than which period happened to process the paperwork. This is the single most useful cut for credit policy purposes.
• Segment by product, branch, loan officer, sector and loan cycle. First-cycle borrowers should produce a materially higher write-off rate than repeat borrowers; if they do not, either screening or the progressive ladder is miscalibrated.
• Track the recovery rate separately — recoveries as a percentage of amounts previously written off — as a direct measure of post-write-off collections performance.
• Compare against portfolio yield. The write-off ratio is a cost of doing business; what matters commercially is whether pricing covers it with margin remaining after operating and funding costs.