I am working out my house property income for the year. The flat was let for seven months and lay empty for five, I am still running the interest from before I got possession, and there is a second flat in joint names - how much of the loan actually comes off, and what is the officer going to attack?
A salaried assessee at Pune, computing his return for AY 2025-26, which is not yet filed. He owns two flats. The first, at Wakad, was booked under a builder's agreement of 12 May 2019 against a housing loan of Rs 62,00,000 sanctioned on 28 June 2019 and disbursed in construction-linked instalments; the possession letter is dated 20 August 2022 and the sale deed was registered on 4 February 2023. His own working from the lender's provisional statements puts the interest for the period to 31 March 2022 at Rs 7,25,000, one-fifth of which - Rs 1,45,000 - has been claimed in each of AY 2023-24 and AY 2024-25, so this year carries the third instalment. The lender's certificate for FY 2024-25 shows interest of Rs 4,38,600. The flat was let from 1 April 2024 at Rs 34,000 a month under a registered leave and licence agreement; the tenant left on 31 October 2024, so Rs 2,38,000 was received for seven months, and the flat stood empty from 1 November 2024 to 31 March 2025 with two brokers engaged and no tenant found. A separate letter provides furniture and appliances at Rs 6,000 a month, paid for the same seven months, and society maintenance of Rs 3,200 a month was billed to and paid by the tenant direct. The municipal tax demand for the year is Rs 11,400, of which Rs 7,600 was paid on 18 March 2025 and the balance of Rs 3,800 on 9 May 2025. The only municipal rateable value extract on the file is from 2019 and shows Rs 96,000 a year; a comparable flat in the same wing has been let since June 2024 at Rs 41,000 a month. The second flat, at Nashik, was bought in March 2021 for Rs 55,00,000 in the joint names of the assessee and his wife with no shares recited in the deed - Rs 18,00,000 from his account, Rs 11,00,000 from hers and a joint loan of Rs 26,00,000 serviced from his account, on which interest of Rs 1,92,400 ran in the year. His parents live in it and no rent is taken. What is missing from the file is most of the proof: the tenant's departure is recorded only in a message on a phone and there is no surrender letter; the broker listings are undated screenshots; the rateable value extract is six years old; the lender has issued a consolidated certificate for the year but no year-wise break-up of the interest before possession, so the Rs 7,25,000 is the borrower's own arithmetic; the furniture letter is unsigned by the tenant; and nothing on the file records who was to own what share of the Nashik flat.
Do not start with the interest certificate. Start with the seven months of letting and get it onto paper, because that single fact decides the five vacant months, and the five vacant months are worth more than the whole interest claim. If the flat is accepted as one that was let in the year, the annual value can be taken at the rent actually received and the computation runs off Rs 2,38,000; if it is not, the officer computes a notional figure for twelve months, and on the comparable sitting in the same wing that figure is roughly double the rent. The registered leave and licence agreement, the bank credits for the seven months and the rent offered in the return are what carry it, and they are already on the file. What is not on the file is the departure - a message on a phone is not a surrender letter - and that is the first thing to reconstruct, because the officer's easiest answer is that the flat was taken back for the family's own use, which puts it outside the vacancy clause altogether.
The charge is on the owner, and the library holds the Supreme Court's construction of that word for this head: the owner is the person entitled to receive the income from the property in his own right, so a buyer who has paid the consideration and holds possession is assessable even with no registered conveyance, and the corpus records that it was the assessee who argued the contrary there and lost. That settles the Wakad flat for the year of possession onwards and, more usefully here, it means the head does not shift merely because the deed came later than the possession letter. The Nashik flat is the harder one. The library's entry on co-ownership sets out that where the shares are definite and ascertainable each co-owner is taxed on his own share and an association-of-persons assessment is not open to the officer, and that the self-occupation relief is given to each co-owner in his own right. Where the deed is silent, the corpus records a Delhi High Court decision holding that signing the instrument of conveyance raises no presumption that the income is assessable in that person's hands and that the question is who in fact obtained the benefit of the property - a decision that reversed a Tribunal order which had presumed an equal split. Be honest about its limit: the Court set the order aside for want of any finding, it did not fix the smaller share the assessee claimed, so it establishes that the equal-shares presumption is wrong in law and nothing more. On this file the funding trail is the whole argument and it has never been documented, which is a job for now rather than for the appeal.
The library sets out three outcomes for a letting that carries furniture or services with it. If the letting of the building is separable from the furniture, the consideration is split and only the building portion stays under the house property head. If the two lettings are inseparable, the whole receipt goes to the residuary head. And if the letting is itself the business carried on, it is business income and neither applies, that question being decided on what the owner actually does and not on what his documents permit. The stakes here run the wrong way from the usual case: under the house property head this taxpayer gets the flat thirty per cent and the interest; under the residuary head he gets only what was actually spent, and the interest claim as he has computed it disappears. So the position to take is that the Rs 6,000 a month is part of the consideration for the flat and not a separate source. The corpus supports that in two ways. It holds a Bombay High Court decision stating the test as whether the service agreement could stand independently of the rent agreement, and holding that amenities which come with occupation rather than being separately supplied are rent by another name - and there the department was the party trying to push the charges out of the house property head and failed. And it holds the Calcutta decision that supplies the primary-object test: mere attachment of income to immovable property is not decisive, what matters is the primary object in exploiting it, and furnished accommodation let for a comprehensive monthly charge with no separate agreement for the fittings was held to be letting, not commercial exploitation. That decision went against the taxpayer in its own case, so it is being used here for its test and not for its result. The risk to name for the client is the middle outcome: an unsigned side letter for the fittings is the worst of both worlds, because it is evidence of a second arrangement without being evidence of a separable one.
This is the item worth the most money and it is the one the library answers best. The clause reduces the annual value to the rent actually received where the property is let, was vacant for the whole or any part of the year, and the actual rent falls short of the notional figure because of that vacancy. The corpus holds a Pune Bench decision of June 2024 that breaks the clause into exactly those three conditions and applies it to a property let for fifteen days and vacant for the rest of each year, deleting an addition built on a notional twelve-month figure; it also records the point that costs the department nothing to concede and much to fight, that an officer who has accepted the rent as income cannot at the same time say the property was never let. It distinguished the strict Andhra Pradesh High Court line rather than differing from it, which is the right posture here too, because that line is the answer the officer will give. The corpus also holds the Mumbai line allowing the clause where the letting was in an earlier year and the whole of the year under appeal was empty, and records that a coordinate bench line has carried it forward, with the qualification that no High Court has affirmed the reading. Now the point specific to this file. The strict High Court decision carries a second limit that is easy to miss and is usually fatal: the period for which a let-out property may remain vacant cannot exceed the period for which it had been let. Five months vacant against seven months let sits inside that limit, so on these facts the claim survives even on the reading least favourable to the taxpayer, and saying so in the computation is worth more than arguing the limit away. Two honest qualifications travel with it. The corpus records that the limit has never been adopted by any decision it holds, and that a 2024 Tribunal decision produced a result that cannot stand with it without anyone citing it. And the strict decision's principal follower, on unsold flats, is recorded as being under a grant of leave to appeal to the Supreme Court with the outcome unknown, so the strict line is itself unsettled.
There is a wider reading in the corpus under which the words describe the character in which the owner holds the property, so holding it for letting with genuine efforts to find a tenant is enough even without a completed letting. The library holds the Mumbai decision that adopted it - reproducing the construction from an earlier bench rather than laying it down itself - and its real contribution is evidentiary: where the same builder had found a tenant for the owner's other flat, letters to that builder could not be dismissed as fabricated, and demanding a despatch register or a stamped acknowledgment was an unreasonable standard of proof. The corpus holds a second Mumbai order on the same footing, allowing the claim where the flat departed from the sanctioned plan and could not be let until the defects were cured, with estate agents' letters accepted as evidence and an appellate authority's grading of the defects as minor rejected for want of material. Both are useful and neither should be the spine of this reply, for two reasons. The library records that the wider reading is contradicted at High Court level in two jurisdictions where the property was not let at all in the year, and records that even the Tribunal decision which follows the wider reading states the rule as requiring a letting in some earlier year. And this file does not need the wider reading, because there was an actual letting inside the year. What the broker material is actually for is the other half of the clause: every decision in this line attaches the condition that the vacancy was not the owner taking the property back for himself, and undated screenshots do not discharge it. Date them, get the brokers to confirm the mandate in writing, and get a letter from the tenant recording the date he left.
If the vacancy claim fails, the annual value is fixed under the notional limb and the comparable in the same wing is what the officer will reach for. The library sets out the steps: expected rent is the higher of municipal value and fair rent, capped by standard rent where rent control applies, and the gross annual value is the higher of that and the rent actually received. On how far the officer may go, the corpus holds the leading Bombay decision of 2014, recorded there as good law, which requires him to form and record a satisfaction that the bargain is unreasonable before he goes looking at all, to obtain comparables and test them for nature, area, measurement, location, floor, use, access and special advantages rather than adopting a locality rate straight, and to disclose the material to the assessee before fixing the rate; it also holds that the municipal rateable value is a safe guide that cannot be discarded in every case and that notional interest on an interest-free deposit cannot be added to rent. Against that, the corpus now also holds a Bombay decision of August 2025 upholding an annual value far above the declared licence fee, on the footing that the rateable value binds nobody and keeps its force only where the municipal determination is proximate to the year. Read that one for its limits: the library records that it was a plainly artificial letting where the stated fee merely covered the outgoings against a very large interest-free deposit, that notional interest was held impermissible under both limbs and survived only because it was not the sole factor, and - decisive here - that the assessee had put no rateable value before the officer at all and abandoned the two documents produced later. A six-year-old extract is the same failure in slower motion. Order a current one before the return goes in.
The library is specific about the order and about the condition. Municipal taxes come off the gross annual value first, but only those actually paid by the owner during the year - not arrears outstanding and not taxes borne by the tenant - and that step belongs to the annual value provision rather than to the deduction section, so it survives whichever regime applies. On these numbers that means Rs 7,600 this year and Rs 3,800 next year, and the payment of 9 May 2025 cannot be pulled back by the fact that the return is still unfiled. It is worth telling the client plainly that this is a diary problem and not a legal one: paying the whole demand before 31 March would have made the whole of it deductible. Then the flat thirty per cent of what is left, which the library reasons is not in the list of deductions the default regime removes and is available whether or not anything was spent - and which it points out is the deduction practitioners drop the moment they see the default regime. If the officer enhances the annual value, the thirty per cent goes up with it, and the corpus says in terms that he will not compute that for you.
For a property that is let there is no maximum on the interest deduction, and the corpus holds a 2024 Bangalore Bench order saying exactly that, with rent actually received in the year as the finding it rests on. Know what that order is and is not. The library records that it cites no authority at all, that it decides the point on findings of fact - a certificate from the lending bank, the same claim allowed in earlier years, and a finding that the loan had not been taken afresh for renovation - and that it is recorded as unverified for later treatment. It is therefore a good template for how to prove the claim rather than a proposition to lean on: file the lender's certificate, show the rent, and show the claim has been consistent. On the pre-construction interest, the library's entry on the interest deduction sets out the mechanism - interest for the period before the year of acquisition or construction is allowed in five equal instalments beginning with the year of acquisition - and it makes the practical point that the five years run from the year of acquisition and not from the year the taxpayer remembers to start them. That is the exposure on this file. The Rs 7,25,000 is the borrower's own arithmetic off provisional statements, and the year the instalments start turns on when the flat was acquired, which on the possession-and-payment reasoning above is 2022-23 and not the year of registration. Get a year-wise certificate from the lender before the third instalment is claimed, because the same figure has to survive two more years and an officer who disturbs the base disturbs all five. For the Nashik flat the position is different: the library records the ceiling for a self-occupied property as Rs 2,00,000 where the capital was borrowed on or after 1 April 1999 for acquisition or construction, the acquisition or construction was completed within five years of the end of the financial year of borrowing and the lender certifies the purpose, and Rs 30,000 where any of those fails - and it records that the Rs 2,00,000 is an aggregate across both houses treated as self-occupied and not a limit per house. The library is candid that those figures rest on secondary material and on the 2025 Act text rather than on a primary reproduction of the current section, so quote them from the certificate and the statute, not from the page.
The regime decides two of the figures above. The library reasons from the statutory list itself that the default regime removes the interest deduction only in respect of the house whose annual value is nil as self-occupied, so the Nashik interest goes and the Wakad interest stays; that the flat thirty per cent is not in the list at all; and that what the regime really takes away is the set off of a house property loss against any other head. That last point is where the money is on this file, because if the vacancy claim succeeds the annual value is Rs 2,38,000 and the interest is several times it, so the loss is real and under the default regime it shelters nothing. The library records that it could not establish whether such a loss may still be carried forward within the head, so do not tell the client it can. The forward-looking point is the interest going into the cost of acquisition on a sale. The corpus holds a Tribunal decision that allowed it, on the reasoning that the deduction under the house property head and the capital gains computation sit under different heads and neither excludes the other - and it holds a Delhi High Court decision from the 1922 Act on the same instinct, that what a buyer laid out to acquire an asset is its cost, including interest on money borrowed to buy it, though not the running costs of holding it. Both must now be given with the qualification the library carries: the Tribunal decision is recorded as superseded by amendment, the capital gains provision having been amended with effect from AY 2024-25 to exclude interest already claimed under the interest deduction from the cost of acquisition or improvement, and the Karnataka High Court was against the claim even before that. So the claim is dead for interest that has been deducted, and the only live version is interest never claimed at all.
A return computed this way is usually accepted, and where it is picked up the vacancy claim is the strongest item on the file because there was a real letting inside the year and the five months of vacancy are shorter than the seven months of letting, so the claim survives even the strictest High Court reading. What decides it is not law but paper: a registered agreement and seven months of bank credits against a departure recorded only on a phone. The realistic risk is not that the annual value is taken at a notional twelve-month figure - that is hard for an officer who has accepted the rent as income - but that he taxes the seven months' rent and refuses the vacancy for the five, which is the outer-limit argument standing on its head and has to be met with the clause's own words. The furniture letter is the item most likely to cause avoidable damage, because it invites a head argument the taxpayer does not want and cannot win cheaply; signing it or dropping it are both better than leaving it as it is. On the interest, the current year's claim on the let flat is rarely disturbed once the lender's certificate is on record, while the pre-construction instalment is the item that gets picked apart three years later when nobody can rebuild the base figure. The Nashik flat is the sleeper: an equal split is what an officer will assume, and there is nothing on the file to displace it.