VittSphere ONE Calculators Blog CA Firm CA Prabhakar Kumar · FCA · ICAI 560762
Case law › Worked examples › Rs 1,42,00,000 written off in four parcels, disallowed as a provision, and added back to book profit as well

Rs 1,42,00,000 written off in four parcels, disallowed as a provision, and added back to book profit as well

The officer says I have not proved the debts went bad, that crediting a provision account is not a write-off, and that two of the parcels fail s.36(2) - which of those can I answer, and what is left if s.36(2) really does fail?

The Income-tax Act, 1961 was repealed on 1 April 2026. It still governs income earned up to 31 March 2026, and every proceeding about those years however late — assessment, reassessment, rectification, penalty, revision and appeal alike. Income earned from 1 April 2026 is governed by the Income-tax Act, 2025. What changed, and which Act governs your year →
A worked example, not advice on your case. The facts below are constructed to be typical, not real. Every legal step links to the authority behind it — follow those links before you rely on any of this, because no chartered accountant has yet signed this page off. Your facts will differ, and the difference is usually where the case is won or lost.

The situation

The client is a private limited company trading in industrial chemicals, assessed at a circle in Ahmedabad. Its AY 2023-24 return, filed on 31 October 2023, declared Rs 1,24,00,000 after writing off Rs 1,42,00,000 of debts in four parcels. Rs 62,00,000 is trade receivables on invoices of FY 2019-20 and FY 2020-21, all credited to the profit and loss account and offered to tax in those years. Rs 34,00,000 is an advance paid in FY 2021-22 to a supplier for solvent never delivered, since gone into liquidation. Rs 26,00,000 is the principal of a loan to a former distributor which carried interest that was offered to tax; the company is not a money-lender and its objects do not cover lending. Rs 20,00,000 is the unrecovered part of a Rs 50,00,000 invoice on which the customer paid Rs 30,00,000 and disputes the balance. In the accounts the whole sum was debited to the profit and loss account and credited to a provision for doubtful debts, and the balance sheet shows sundry debtors net of that provision; the individual ledgers were not closed. The order under s.143(3) dated 9 March 2026 disallows the whole claim on three grounds - that it is a provision and not a write-off, that no evidence shows any debt became bad, and that two parcels fail s.36(2) - adds Rs 18,00,000 under s.41(1) for old creditor balances, and adds the whole sum to book profit under clause (i) of Explanation 1 to s.115JB.

Before anything else

Split the Rs 1,42,00,000 into its four parcels on paper and run each one through s.36(2)(i) before drafting anything, because two of them pass and two do not, and they need different remedies pleaded in different grounds. A single ground asserting that the whole amount was written off invites a single order rejecting the whole amount, which is what happened at the assessment stage. The other half of the first move is to pull the signed accounts, not the draft, and check how the provision is actually presented on the face of the balance sheet - that presentation decides the officer's first ground on its own.

Working it through

8 steps. Each one shows the authorities it stands on.
  1. 1

    Test each parcel against s.36(2)(i) first, and be honest about which two fail.

    The library's concept page puts it plainly: the write-off is only half of it. The clause allows the deduction subject to the provisions of sub-section (2), and s.36(2)(i) sets the real precondition - the debt must have been taken into account in computing the income of the previous year or an earlier year, or it must represent money lent in the ordinary course of a business of banking or money-lending. The Rs 62,00,000 of trade receivables passes on the first limb, because the sales were credited and offered to tax, and the proof is the invoice, the ledger and the earlier year's computation together. The Rs 34,00,000 advance fails both limbs: an advance for goods never delivered was never taken into account in computing income, and the company lends nothing. The Rs 26,00,000 loan fails on the principal, because only the interest entered income and the principal is what has been written off.

    Careful here. This is precisely where the officer wins when the parcels are argued as one. Produce for the first parcel the invoice, the ledger extract and the computation of the earlier year showing the sale in income - the ledger on its own does not establish that the debt was taken into account in computing income, and that is the sentence the section uses. And do not concede the second and third parcels in the same breath as you identify the failure; they have an alternative route, and it has to be pleaded in the same appeal or it is gone.
  2. 2

    Answer the proof-of-badness ground by pointing out that the statute stopped requiring it in 1989.

    The Supreme Court has held that for years after 1 April 1989 the assessee does not have to establish that the debt has in fact become irrecoverable: it is enough that it is written off as irrecoverable in the accounts. The officer's demand for legal notices, suits and recovery correspondence is therefore asking for something the clause no longer requires. What the same decision does leave with him is a narrower check - he can verify that the write-off was actually made in the books, which is why that matter was sent back rather than allowed outright, and it is the check he will fall back on once this ground is answered.

    Careful here. The decision removes the burden of proving badness. It does not remove the sub-section (2) precondition, and it does not stop an officer asking whether the debt existed at all - which is the real contest wherever a debtor is a related party or the sales themselves are doubted. Note also what it produced: a remission for verification of the books, not an allowance, so expect the same shape of relief here. Nothing in it helps with the officer's second and third grounds, and the reply should not treat it as answering the whole order.
    What this rests on
  3. 3

    Meet the provision-versus-write-off ground with the presentation in the signed accounts.

    The Supreme Court has held that debiting the profit and loss account and simultaneously reducing loans and advances or debtors on the asset side, so that the year-end figure is shown net of the provision, is an actual write-off, and that closing each individual debtor's account is not required. That is this balance sheet as the scenario describes it, and the ground is short: produce the schedule reconciling gross debtors, the provision and the net figure carried to the face of the accounts, with the auditor's note. The boundary on the other side is equally clear - the Supreme Court has held that a prudential provision for non-performing assets debited under regulatory directions must be added back in computing total income, because such directions govern presentation and not deductibility.

    Careful here. The presentation has to actually be net. If the signed balance sheet shows debtors gross with the provision sitting on the liabilities side or as a separate line, the decision does not reach the case and the officer's characterisation stands - check the signed accounts and, if the presentation is wrong, say so and argue the parcels on their own merits rather than staking the head on a fact that is not there. Remember the tail as well: s.41(4) charges a later recovery to tax in the year it is received, so a successful write-off now is not the end of the matter if the customer pays in 2028.
    What this rests on
  4. 4

    Plead the part write-off from the words of the clause, and the alternatives for the two parcels that fail.

    The clause allows the deduction of any bad debt or part thereof which is written off as irrecoverable in the accounts, so the Rs 20,00,000 written off out of a Rs 50,00,000 invoice is within its words and needs no separate justification - that is taken from the statutory text, because no decision in this collection deals with a part write-off. For the advance and the loan the route is not this clause at all. The Supreme Court has held that a loss springing from the way the business is carried on is deductible under the charging provision itself: it refused the claim there as a bad debt and refused it as expenditure, and then allowed it as a trading loss, because the profits of a business cannot be computed without taking account of losses incidental to it. The library's concept page says the same, that the allowance sections are not an exhaustive statement of what may be deducted.

    Careful here. The trading-loss route is strongest for the supplier advance, which was paid in the ordinary course of buying stock, and weakest for the loan to a former distributor, which the company had no business making. On that second one a Tribunal Bench has held that the principal advanced is a capital outflow and that neither the other-sources deduction nor the bad-debt route is open to a lender with no money-lending business - that entry is marked unverified in this library, which is weaker than anything else cited on this page, so use it as the shape of the department's argument rather than as authority, and do not let it be quoted at you without the marking being pointed out.
  5. 5

    Do not build anything on the bank and NBFC decisions, which are most of what this library holds on the section.

    The Supreme Court has held the rural-provision deduction distinct and independent of the deduction for a bad debt actually written off; a High Court has held that the provision created during the same year reduces the write-off claim; a Tribunal has read the 2013 Explanation as not reaching a clause (a) bank's non-rural debts; and the library's statutory position sets the proviso, the companion condition and that Explanation out together. Every one of them turns on an assessee to whom the provision clause applies. This client is a trading company: the proviso does not touch it, there is no provision account in its computation, and none of those entries decides anything about it. Their value on this file is negative - they tell you which arguments are unavailable, and which ones the officer may wrongly import.

    Careful here. The same is true of the other regulated-entity entries here - the finance company that could not deduct a provision beyond the statutory percentage, the general insurer's case under the First Schedule, and the statutory position on interest on bad or doubtful debts of a bank or finance company. Three of those are marked no later treatment found. If any of them is cited against a trading company, the answer is the same in one line: they construe the provision clause and the special interest section, not the write-off clause. That answer is worth putting in the grounds pre-emptively, because the assessment order already reasons as though a provision account had to be exhausted first.
  6. 6

    Argue the book-profit add-back on the same characterisation you are running below, and expect it to be the weakest head.

    A Full Bench of a High Court has held that after the retrospective insertion of clause (i) with effect from 1 April 2001, a provision for diminution in the value of any asset - which is what a provision for doubtful debts is - must be added back to book profit, and that the earlier line decided under the general clause no longer governs. The difficulty on this file is that the argument on the normal computation is that the entry is a write-off and not a provision, and the book-profit argument has to be the same argument: the accounts carry debtors net, so no provision for diminution survives on the balance sheet to be added back. Say it once and identically in both places, because the two grounds will be read together.

    Careful here. Nothing in this collection decides whether a write-off presented net of a provision account escapes clause (i). The Supreme Court decision on what counts as an actual write-off is on the normal computation, and the Full Bench is about a provision that remained a provision, so the bridge between them is an argument and should be pleaded as one. Check first whether the head is worth a ground at all: the book-profit charge matters only if it exceeds the normal tax for the year, and a company that has opted for the concessional rate is outside the book-profit section altogether.
  7. 7

    Treat the s.41(1) addition on the creditor balances as a separate contest with a separate burden.

    The library's concept page sets out two conditions, not descriptions: the officer must establish that an allowance or deduction was made in an earlier assessment in respect of the very loss, expenditure or trading liability he is now taxing, and that the assessee has since obtained an amount in respect of it or a benefit by way of remission or cessation of it. Both have to be shown for each of the balances making up the Rs 18,00,000, which means the officer has to go creditor by creditor to the earlier year's purchase and the earlier year's deduction. An old balance still carried as payable in the accounts, without more, is not a cessation, and the ground should ask him to identify the year of the deduction and the event of remission for each item.

    Careful here. The company's own treatment decides how hard this is. If any of the balances were written back to the profit and loss account in the accounts, the second condition becomes much easier for the officer and the contest narrows to the first. If they still sit as payable, he has to find the cessation, and confirmations or subsequent payments are the answer. Keep this head in its own ground: it is the mirror image of the bad debt heads and merging them invites the officer's argument that the company writes off what it is owed and does not write back what it owes.
  8. 8

    Draft one ground per parcel, with its own statutory footing on its face.

    The discipline is procedural and it costs nothing. Four parcels, each with the sub-section (2) footing stated for it; the alternative under the charging provision pleaded expressly for the two that fail; the accounting characterisation as its own ground; the book-profit addition as its own ground; the creditor addition as its own ground. The reason to be strict about it is what happens if a point is not dealt with: a High Court has held that the Tribunal's rectification power is a power to amend and not to review, and that the mistake must be obvious and patent rather than one to be established by a long process of reasoning on a point on which two views are possible. A parcel folded into another parcel's ground and then passed over in the order is hard to recover afterwards.

    Careful here. That decision is about the Tribunal's rectification power and not about the first appellate authority, so the connection drawn here is a practical one rather than a holding - a composite ground invites a composite order, and the remedy for a composite order is limited. Keep the concession on the two failing parcels inside their own grounds too, so that the concession cannot be read as extending to the trading-loss alternative pleaded in the same breath.
    What this rests on

Where this usually lands

The trade receivables parcel is usually allowed, and the accounting ground usually falls with it once the net presentation is shown - together that is Rs 62,00,000 of the Rs 1,42,00,000. The supplier advance goes as a trading loss more often than not where the liquidation is documented and the payment was in the ordinary course of buying stock. The distributor loan is usually lost, and is the parcel on which the department's argument is strongest. The part write-off is rarely contested once it is pleaded from the words of the clause. The book-profit head is genuinely uncertain and is the one most likely to be decided against the client even where the main claim succeeds. Expect at least one parcel to be remitted for verification of the books rather than allowed outright, which adds a year.

What to do

What this library could not tell you

Written down rather than papered over. These are points where the argument needed authority we do not hold, so the study stops short instead of guessing.

Every authority used above

17 entries. Nothing in this study cites anything outside the library.