Capital gains on property, and where the equity rules stop

What actually differs

A house and a share are both capital assets, and past that the answers diverge. The clock before a gain counts as long-term runs longer, no slice of the gain is free each year, the sale value can be overridden, and reinvesting the money can remove the charge outright.

The reason is structural rather than political. The flat rates everyone quotes are written for a named list of assets, not for capital gains at large: s.196 of the Income-tax Act 2025 (formerly s.111A) and s.198 (formerly s.112A) both address securities-transaction-tax-paid listed equity shares and units of an equity-oriented fund. A flat in Pune is neither. Property sits outside both, so the familiar figures simply do not reach it.

What follows is the comparison, in one place. Where a cell is blank it is blank deliberately — the 2025 Act renumbered and in places rewrote the property rules, and this library states no statutory figure it has not read in the new text. The last section says which figures are missing and why that is the safer answer.

Read down the right-hand column and the pattern is visible: every property rule exists because a house is an asset the state can see, value and register, and a share is not. That single fact generates the valuation floor, the buyer's withholding duty and the improvement account alike.

The clock is longer, and it may not start when you think

The holding period is not one rule with exceptions. Section 2(101) sets a threshold per class of asset, and the classes were drawn separately. Listed securities and equity-oriented fund units cross at 12 months; immovable property crosses later, and this article does not say how much later, because we have not read that figure in the 2025 Act, and a plausible number copied from a stale article is exactly the defect worth avoiding. The general mechanism is worked through in the capital gains explainer.

Two consequences follow that are specific to property. A resale inside the threshold is charged as a short-term gain, and a short-term gain on property is outside s.196 altogether — so the 20% figure an equity investor carries around describes nothing about it, in either direction. And because a property transaction takes months to complete, a sale agreed in one financial year and registered in the next can straddle the line by accident.

Inherited and gifted property behaves differently again. On the mechanism carried across from the repealed Act, you do not restart: the previous owner's cost becomes your cost, and their holding period counts as yours. A flat your father bought in 1998 and left to you last year is long-term in your hands from day one, and the gain is measured from his cost, not from its value on the day you inherited it. That second half is the part people are surprised by, and it usually makes the gain far larger than expected.

An under-construction flat has a genuine ambiguity at the front of the clock. The allotment letter, the builder-buyer agreement, possession and registration are four different dates, sometimes years apart, and which one starts the period has been litigated rather than settled by a figure. We are not going to tell you which date applies. What is worth knowing is that the question exists, that it can move a sale from one side of the line to the other, and that the documents establishing each date are worth keeping for that reason alone.

The improvement account, which shares do not have

A gain is not the sale price minus the purchase price. It is the sale consideration, less what it cost you to sell, less the cost of acquisition, less the cost of improvement. The third and fourth items are where property parts company with every listed security you have ever held.

Cost of acquisition is more than the number on the sale deed. Stamp duty, registration charges and the brokerage you paid on the purchase are part of what the asset cost you. So is the legal work. On the sale side, the brokerage, the legal fees and the charges incurred to effect the transfer come off the consideration before the gain is measured. Exactly where the boundary of “in connection with the transfer” falls is a question the 2025 Act rewording reopens; the shape of it is not controversial.

Cost of improvement is the genuinely different one, and it exists because you can spend money on a house. Over fifteen years an owner might add a floor, build a boundary wall, replace a roof, enclose a balcony. Each of those adds to what the asset cost, and therefore reduces the gain. There is no equivalent for a share: no amount of money spent by you makes your holding in a listed company larger.

The line that decides is capital against revenue. Money that creates something that was not there, or materially extends the life of what was, is improvement. Money that keeps the asset in the condition it was already in is not. Repainting, servicing the lift, replacing a leaking tap, the annual maintenance cheque to the society — none of that is improvement, however large the total. An extra room is.

Now the mistake, and it is the one that costs property owners the most rupees for the least reason. People keep the sale deed, because it is obviously important, and throw away the contractor invoices, because they are not. Fourteen years later the improvement account is real money and there is nothing to evidence it with. The arithmetic is unforgiving: an unevidenced improvement is a taxed gain, and the amount at stake is the whole of what you spent, not some fraction of it. The file to keep is dull and thin — dated invoices, the payment trail from your bank, the municipal approvals for anything structural.

The price on the deed is not necessarily the price

Here is the rule with no equity counterpart at all. Every state publishes a minimum value for each locality — the circle rate, the ready reckoner rate, the guidance value, depending on where you are — and uses it to charge stamp duty so that under-declaration does not cost the exchequer its duty.

The income-tax computation then borrows that figure. If the consideration stated in the deed is below the stamp-duty value, the computation substitutes the stamp-duty value as your sale consideration, subject to a tolerance band that allows a modest shortfall to pass. You are taxed on a price nobody paid. The band exists because published rates are coarse and a genuine sale can land slightly under one; its size is a figure we have not read in the 2025 Act, so this paragraph does not give one.

The mechanism runs on both sides of the table, and that is the part people miss. Where the buyer acquires property for less than the stamp-duty value, the shortfall can be charged to the buyer as income — on the reasoning that they received something worth more than they paid. So a sale written far enough below the published value can produce a charge on the seller, who is taxed as though they had received the higher figure, and a second charge on the buyer, who is taxed on the difference. The same gap, taxed twice, in two different hands — each side with its own tolerance before its own provision bites.

That is the failure mode worth naming, because it is not a penalty and does not feel like one at the time. It is two separate provisions doing their jobs on the same transaction, and nobody involved needs to have done anything deliberate for it to bite. A distress sale, a disputed title, an unfashionable floor in an otherwise expensive building — each can genuinely trade below the published rate.

Where the value is genuinely below the published rate, the mechanism does not simply ignore it: the seller can ask for a referral to a valuation officer, whose figure then drives the computation. That route takes time and evidence, and it runs after the fact. The published value for a locality, by contrast, is public and can be read before a price is agreed — it sets a floor under the tax whatever the parties settle on, which is the whole reason it is worth reading first rather than at the point of filing. One complication sits on top of that: where an agreement and the final conveyance fall in different years, the two dates can carry different published values, and which of them fixes the figure — and on what condition — is not stated here.

The reliefs that equity has no version of

Now the direction the difference runs the other way. Equity gets an annual threshold of ₹1.25 lakh under s.198 and nothing else; property gets no annual threshold and, instead, a set of reliefs that turn on what you do with the money afterwards.

They come in three shapes. Roll the gain into another residential house within a window measured from the transfer. Put it into specified bonds within a shorter window, accepting a lock-in and a modest coupon. Or, where the deadline to reinvest falls after the date your return is due, park the money in a designated capital gains account so the intention is on record, and spend it from there. Each carries its own qualifying window, its own cap, and a lock-in on the new asset during which selling it claws the relief back. None of those figures appears here, for the reason given throughout: the 2025 Act renumbered and rewrote this material and we have not read the new text.

What matters more than the figures is what kind of provision this is. An equity threshold is automatic — you do nothing and it applies. A reinvestment relief is conditional on future conduct, and the condition binds for years after the sale. That is a different bargain, and it has a price.

The price is flexibility. A relief that requires buying another house converts a liquid outcome into an illiquid one; the tax saved is real and so is the fact that you now own a second property you may not want, which cannot itself be sold inside the lock-in without the relief unwinding. Bonds cost you less freedom and less money at once: the lock-in is finite, the coupon is modest, and the capital sits idle in the meantime. Neither is free, and an owner who compares only the tax saved has priced one side of the trade.

The named failure mode here is the clock. The window runs from the transfer, not from the day the money reaches you. A buyer who pays the last instalment nine months late has not extended your deadline by nine months, and the relief does not care whose fault the delay was. The holding account exists precisely for the gap between the two, and it has to be opened before the return is due — which means the decision belongs in the month of the sale, not in the month you get round to filing.

Which regime you are in, and why these reliefs survive it

Almost every article on property tax written before 2026 quietly assumes the old regime. Most readers are not in it, because the new regime is the default under s.202 of the 2025 Act — you are in it unless you opt out. Under it the deductions that shaped a generation of property decisions are gone.

The last row is the point of this section, and it is the least obvious thing in the article. Section 202(2) switches off deductions from total income — the Chapter VIII list, which is where s.123, s.126, s.129 and s.134 all live — along with interest on a self-occupied house. The reinvestment reliefs are not deductions from income at all. They sit inside the capital-gains computation and change what the gain is before any of that machinery starts.

So the two are not switched by the same lever. A reader who moved to the default regime lost the deduction that made a home loan feel cheap, and did not lose the relief for rolling a sale into another house. That is this article's central claim, and it is a reading of the section rather than a settled point: it holds only so long as s.202(2) does not name these reliefs, which is the thing to check in the section text before it decides anything. The wider comparison is its own article, and the home-loan side is worked through in the home loan guide.

Two further regime interactions are worth flagging without asserting. The rebate under the default regime is up to ₹60,000 where total income does not exceed ₹12 lakh, and it cannot shelter special-rate income such as s.198 long-term gains. And surcharge on capital gains is capped at 15%, which is why a single large sale does not necessarily drag the gain into the top surcharge band. Both of those are established for equity gains; whether either reaches a gain on property specifically is something we have not confirmed, and neither belongs in your own arithmetic until it is.

Two names on the deed, two separate computations

A jointly held property does not produce one gain that somebody pays tax on. It produces a gain per co-owner, in proportion to the share each of them owns, and each co-owner computes and reports their own.

The consequences run further than the split. Each co-owner has their own reinvestment decision to make on their own share, and their own deadline for making it. If one of them buys another house and the other does not, only the first share is sheltered; the relief is not a property-level fact that the family shares. Each also brings their own other income to the computation, so the same rupee of gain can be charged differently in two hands.

Here is the mistake. Families routinely assume the gain belongs to whoever paid for the house — usually one earner — while the deed carries two names because a lender wanted a co-applicant, or because adding a spouse seemed prudent. The deed share is the starting point, not the funding. Where the two diverge, the question becomes who beneficially owned the asset, and the clubbing provisions can pull a share back to the person who funded it. That is contested territory and the outcome turns on facts, so we state the shape and not the answer.

What follows from the shape is practical and unglamorous. The document that settles a funding dispute years later is the bank trail from the original purchase — who paid the builder, from which account, and who serviced the loan. That trail is easy to keep at the time and impossible to reconstruct at the point of sale. The same is true of a share held with a joint holder, but the sums involved make property the case where it actually matters. One thing that trail is not is a nomination. A nominee is the person an asset is released to on death, which is a different question from who owned it — and therefore settles nothing about whose gain a later sale produces. The two get confused constantly, and on a property they are rarely even the same person.

One sale, one year, and a buyer who owes the paperwork

A property sale is lumpy in a way an equity portfolio is not. You can trim a shareholding across two financial years to spread a gain; you cannot sell two-thirds of a flat. One transaction can be the largest taxable event of a lifetime, and it lands whole, in a single year.

That collides with the fact that tax is not payable once a year. It is payable in instalments through the year, and a large gain arising in a later quarter raises a timing question with a real cost attached. Whether such a gain is picked up from the instalment following the transfer, or attracts interest for the instalments already gone, is not settled here — the mechanism itself is set out in advance tax and TDS.

The other structural difference is that somebody else has a duty in your transaction. On a purchase of immovable property above a stated consideration, the buyer deducts tax before paying the seller and deposits it against the seller's account. Nothing comparable happens when you sell shares on an exchange. The rate, the threshold and the provision are not stated here; the consequence is.

The consequence is that a private individual, usually buying a house for the first time, is now a withholding agent. If the buyer deducts and never files the statement, the money has left the seller's price and does not appear as credit against the seller's tax. The seller pays twice and then spends months chasing a stranger for a correction. The check that catches it is mechanical and takes minutes — the deduction either shows up in the seller's own tax credit statement after completion or it does not, and the moment to find out is while the buyer is still reachable and still motivated to fix it, not at the point of filing.

Why the blanks are safer than an old number

Read back and you will find a set of blanks: the holding-period figure for property, the long-term rate, whether indexation survives in any form, the tolerance band under the valuation floor, the caps and windows on all three reinvestment reliefs, and the buyer's withholding rate and threshold.

Those are not omissions of nerve. The Income-tax Act 1961 was repealed on 1 April 2026 and replaced by the Income-tax Act 2025; the rates barely moved and the section numbers all did. Every figure this article does state has been read in the new text, and every one it does not has been left out rather than carried over on the assumption that it survived. A stale rate stated confidently is worse than a blank, because the reader cannot tell it is stale — and most of what is currently written about property tax was written before April 2026 and does not say so.

The section numbers cited above — s.196, s.198, s.2(101), s.202, s.123 with Schedule XV, s.126, s.129, s.11 with Schedule III Sl. No. 11 and s.134 — are 2025 Act numbers. If you are filing for FY 2025-26, meaning April 2025 to March 2026, that year is still assessed under the repealed Act and carries the old numbering, which is why both appear.

What the article does give you is the part that outlasts a Finance Act: property runs a longer clock, has no annual free slice, is measured against a published valuation floor, carries an improvement account you must evidence, offers reliefs that are conditional on what you do next, and splits by ownership share when there are two names on the deed. Every rate in the country can change and all six of those remain true. FNOTrader is not a tax adviser and none of the above is advice on your position.

Common questions

Continue reading

More in Tax Planning · App: Mutual Funds · Definitions: glossary · Free tools: calculators · All: every article