Tax-loss harvesting

It is late March. A reader looks at the portfolio and sees one holding deep in the red. A friend, or a broker's push notification, uses a tidy phrase: tax-loss harvesting. Book the loss, set it against this year's gains, pay less tax. So the reader sells - not because anything changed about the company, but because the calendar and the phrase arrived together.

The tax saving may be completely real. That is what makes this one tricky. The question is not whether harvesting works - under current rules it often does - but whether the trade still makes sense once the tax reason is set aside. A tax benefit is only worth having if the portfolio action underneath it survives on its own terms, after costs and after you have decided what you actually want to own tomorrow.

This chapter exists because a tax tactic can quietly become a trading excuse. Harvesting is one of the few moments in investing where the tax code appears to reward action - and action, near a deadline, is exactly the thing a careful reader should slow down.

The old saying is blunt: do not let the tax tail wag the investment dog. A loss is not created by selling; it already exists on the screen the moment the price fell. Selling only realises it - converts a paper loss into a booked one that the tax return can use. So the sell has two separate jobs riding on it at once. One job is the tax record. The other is a real change in what you own. Those two jobs must be read together, never one instead of the other.

Read the harvest as a cost decision as much as a tax decision. Every realising trade, and every re-entry behind it, drags brokerage, spread, fresh STT, and the turnover you are manufacturing. Every repeated cost is a claim on your result, whatever it is called - including a cost you invited in the name of saving tax. If the saving is ₹10,000 and the round-trip friction plus an unlucky overnight gap is ₹9,000, the harvest was a lot of motion for very little.

The protective habit is to write both reasons before trading: the tax reason, and the investment reason. If only the tax reason exists, the honest answer is usually to do nothing.

Start with the plumbing of the deduction, because that is where the real value sits - and where the rules matter most.

When you sell listed equity at a loss, you create a realised capital loss. Indian tax law then sorts that loss into a box by holding period. Sell within twelve months and it is a short-term capital loss (STCL); hold longer and it is a long-term capital loss (LTCL). The boxes are not interchangeable. Under current rules an STCL can be set off against both short-term and long-term gains, while an LTCL can be set off only against long-term gains. What you cannot set off this year you can generally carry forward for up to eight assessment years - but only if you file your return on time. Miss the filing deadline and the carry-forward can be lost even though the loss was real. illustrative

Now read the second job - the exposure. This is where investors quietly split into two situations. If the loss is in a holding you were going to exit anyway, the tax benefit is a clean bonus on a decision you had already made; there is no re-entry to worry about. But if you still want the position, harvesting means selling and buying back, and India - unlike the United States - has no general wash-sale rule for shares. You are legally allowed to sell at a loss and repurchase the same scrip immediately. That freedom is real, and it is also a trap, because "immediately" is never free: you cross the spread twice, pay fresh charges, and expose yourself to an overnight tax lot gap where the price can move before you are back in.

There is a boundary worth naming. Selling and rebuying purely to manufacture a loss - sometimes called bed-and-breakfasting - sits in a grey zone. A specific anti-avoidance rule (Section 94(7)) blocks losses engineered around dividend record dates, and the tax department can, in principle, look through wholly artificial round-trips. The clean version of harvesting is not a same-second round-trip to game a number; it is realising a loss you would recognise anyway, or accepting a genuine gap in exposure to get the deduction. The safest read follows the chain - the contract note, the tax-lot record, and the filed return that together prove the loss was real and correctly reported.

The point is not a formula. It is to see when the same tidy phrase - "harvest the loss" - is clean, incomplete, or misleading. Work it three ways. illustrative

The clean bonus. A reader holds a small-cap they have already decided to exit; the thesis broke months ago. It happens to be down ₹50,000, and the reader also has ₹1,20,000 of short-term gains this year. Selling does two useful things at once: it removes an exposure they no longer want, and the STCL sets off against the STCG, lowering the tax bill. There is no re-entry, so no round-trip cost. Here harvesting is genuinely close to free value - because the sell was justified on its own before tax entered the room.

The costly round-trip. Now the loss is in a holding the reader still believes in. To keep the exposure they must sell and rebuy. Two brokerages, two spreads, fresh STT, and an overnight gap all stand between the two legs. The read is a subtraction: estimate the tax saved, then subtract the full two-way friction and the risk that the price gaps up while you are out. If the saving comfortably beats the cost, and you are content to be briefly out of the position, it can still be worth doing - carefully. The misfire is treating the round-trip as costless because no wash-sale rule stops it.

The deadline stampede. It is 30 March and the reader sells three holdings at a loss in an afternoon, purely because the year is ending. Two of them were fine businesses they had no reason to leave. The tax return will show some set-off, but the portfolio now has holes the reader did not intend, opened at whatever prices the last two days offered. The repair is to separate the two clocks: the tax year has a deadline, but the investment decision does not have to be made in a panic to meet it. Decide what you want to own first; let the tax treatment follow.

Before "harvest it" becomes a reflex, a few slow questions are worth saying out loud.

Harvesting cannot rescue a weak thesis. A loss is a loss; booking it for tax does not undo the mistake that produced it, and the deduction is a fraction of the money, not a refund of it. The most it does is soften the tax on gains you made elsewhere. Reading a harvest as a way to "get something back" from a bad holding is a comfort story, not an economic one.

It also cannot replace a tax professional when the facts get complicated. The general shape here - classify the loss, match it to the right gain box, subtract the costs, respect the filing deadline - is enough for a straightforward equity portfolio. It is not enough when there is business income, intraday activity treated as speculative, multiple purchase lots at different dates, or cross-border holdings. In those cases the mechanism above tells you what to organise; it does not tell you the final number.

And it cannot remove market risk. The moment you sell to harvest a loss, you are out of the position at that price. Whether re-entering later helps or hurts depends on where the price goes next, which no tax rule controls. The deduction is certain; the exposure decision around it is not.

People get fooled when a tax tactic becomes a reason to trade. The deduction is real, so it feels like pure upside - and that feeling is exactly what hides the costs and the missing investment reason underneath.

The traps repeat across new investors:

1. Selling a good holding only to book a loss, then struggling to get the exposure back.
2. Treating the sell-and-rebuy as free because India has no wash-sale rule, while ignoring two spreads, fresh STT, and the overnight gap.
3. Rushing three trades on 30 March to beat a deadline that only the tax year, not the portfolio, actually has.
4. Assuming the loss can offset any gain, forgetting that a long-term loss offsets only long-term gains.
5. Booking the loss but filing the return late, and losing the carry-forward that was the whole point.