Tax slabs tax slices, not people

What a slab rate is charged on

A slab rate applies only to the income inside that slab, not to your whole income. Move into a higher band and the extra rate touches the rupees above the threshold and nothing below it. A raise therefore always leaves you with more money after tax than before, under any slab ladder.

The belief that it works the other way is the most durable misconception in Indian personal finance, and it costs real money: people decline overtime, delay a promotion conversation, or ask for a component to be paid next year, in order to avoid a penalty that does not exist. It is worth killing with arithmetic rather than reassurance, because reassurance is what everyone has already been offered.

Start with the ladder itself. Under the default regime — the new one, which the Income-tax Act 2025 applies at s.202 unless you actively opt out — the bands run nil up to ₹4 lakh, 5% to ₹8 lakh, 10% to ₹12 lakh, 15% to ₹16 lakh, 20% to ₹20 lakh, 25% to ₹24 lakh, and 30% above that. Read that as a list of thresholds with a rate attached to the stretch above each, not as a list of categories of taxpayer.

The distinction matters because the language encourages the wrong picture. Saying somebody “is in the 20% bracket” sounds like a description of the person. It is a description of their last rupee. The ladder taxes slices, not people — and the same person is simultaneously in every band below their top one.

One preliminary, because it decides which number you feed into the ladder. Slabs are applied to total income, which for a salaried reader is pay after the standard deduction of ₹75,000 under the new regime and ₹50,000 under the old. That deduction is automatic; it is already inside both bills in any comparison, and it is not something you claim.

The raise, worked out to the rupee

Take somebody whose total income is ₹15.8 lakh and who receives a ₹60,000 raise, carrying them to ₹16.4 lakh — across the threshold where the rate steps from 15% to 20% on the ladder quoted above. If the myth were true, the whole ₹16.4 lakh would be charged at 20%, giving a bill of ₹3.28 lakh against the ₹1.17 lakh their payslip already shows. The raise would have cost them more than three times what it paid — which is exactly the fear, stated in rupees.

Here is what the ladder actually does with the ₹16.4 lakh. Each row charges its own rate on its own slice, and the slices are added.

Before the raise, at ₹15.8 lakh, the first three rows are identical and the fourth charges 15% on ₹3.8 lakh rather than on ₹4 lakh — ₹57,000 instead of ₹60,000, and there is no fifth row. The bill is ₹1.17 lakh.

So the raise of ₹60,000 raised the tax by ₹11,000, not ₹2.11 lakh. Take-home rose by ₹49,000. Only the ₹40,000 of the raise that sat above ₹16 lakh met the 20% rate at all; the ₹20,000 below it met 15% — the same rate their top rupee was already facing before the raise arrived. Both figures are tax before surcharge and cess, which sit further down the calculation and do not change the shape of this.

Those band edges are the ones quoted in the previous section, and a Finance Act can move them. If it does, the mechanism above survives untouched and every rupee figure in it does not — so recompute, do not extrapolate.

Marginal and average are two different numbers

The example produced two percentages, and confusing them is what keeps the myth alive. Both are correct; they answer different questions.

The marginal rate is what the next rupee costs. For the taxpayer above it is 20%, because that is the band their top rupee sits in. The average rate is what the whole income cost: ₹1.28 lakh on ₹16.4 lakh, which is 7.8%. The gap between 20% and 7.8% is not an error or a loophole. It is the arithmetic consequence of the lower bands being charged at lower rates on the way up.

Almost every conversational claim about tax quotes the marginal rate and means the average one. “I pay 30% tax” is a sentence about somebody's last rupee, and it is wrong about their bill by a wide margin. Nobody in India has ever paid their top slab rate on their whole income under a slab ladder, because the ladder makes it structurally impossible.

The marginal rate is still the number that matters for a decision, which is the part worth keeping. It tells you what an extra rupee earned is worth to you after tax, and what a rupee of deduction saves you. The average rate tells you what happened. Neither substitutes for the other, and the useful habit is knowing which one a given sentence needs.

Where a small raise really does cost more

Now the part that the reassuring version of this article leaves out, and the reason the myth has survived so long: the folk belief is not baseless. It is attached to the wrong step. There is a point in the Indian calculation where a few thousand rupees of extra income produces a change in tax out of all proportion to it. It is not in the slab ladder. It is at the rebate.

A rebate is subtracted from the tax, not from the income. Under the default regime s.156(2) gives a rebate of up to ₹60,000 where total income does not exceed ₹12 lakh. Note the shape of that condition. It is not a taper that fades out as income rises. It is a step, not a slope — the rebate is available in full below the ceiling and not at all above it.

Follow what that does to two taxpayers who are ₹10,000 apart, using the same ladder as before.

₹10,000 of extra income, and the bill goes from nothing to ₹61,500 before cess. That is the effect people think slabs have, sitting two steps further down the calculation where almost nobody looks for it. The slab ladder cannot do this to you. A threshold in the rebate can, because it switches something off rather than tapering it.

Two qualifications, and the first is important enough that the arithmetic above should be treated as the unrelieved case. A tax system cannot really defend an outcome where earning ₹10,000 more leaves you ₹51,500 worse off, and the standard repair is a marginal relief provision that caps the additional tax at the additional income. Whether one attaches to this rebate threshold under the 2025 Act is a specific question worth confirming for the year you are filing, and this article does not answer it. What survives either answer is the shape: with relief, the bill climbs rupee-for-rupee across a stretch instead of jumping; without it, it jumps. Neither is a slope.

The second qualification is that this is a mechanism, not a plan. Knowing where a threshold sits tells you why a bill moved the way it did, and where a rule of thumb about marginal rates stops applying. It says nothing about what to do with an offer of extra income, which is a question about the offer.

The nil-tax return that still owes tax

The rebate has a second property that decides real bills and is missing from almost every explanation of it. The rebate does not shelter special-rate income. Section 156(3) bars it from being set against income charged at its own flat rate, and the clearest case of that is long-term capital gains on listed equity under s.198.

Take somebody whose total income is ₹10 lakh of salary-based income, plus long-term gains of ₹1.5 lakh on equity fund units. The slab tax on the ₹10 lakh comes to ₹40,000 by the ladder above, which the rebate wipes out entirely. Their slab liability is nil.

The gains are dealt with separately. Only the amount above ₹1.25 lakh of such gains in the year is charged, so ₹25,000 of the ₹1.5 lakh meets the s.198 rate of 12.5%. That is ₹3,125 — and the rebate cannot touch it. A taxpayer who reads “no tax up to ₹12 lakh” and concludes there is nothing to pay has a bill, a return to file, and often a surprise.

The same logic runs through short-term gains under s.196 at 20%. It runs the other way for anything charged at slab rates: units of a debt-oriented scheme bought on or after 1 April 2023 fall under s.76 and are taxed at slab rates, with the gain always treated as short-term, which means they climb the ladder with your salary and are covered by the rebate exactly as salary would be. Two mutual fund redemptions of identical size can therefore produce completely different answers, and the deciding fact is what the scheme holds rather than what it is called. The mechanics of each are set out in capital gains tax.

There is a further open question here worth naming rather than glossing: whether an unexhausted basic exemption can be set against gains of this kind. It has a real answer and this article does not have it, so treat the ₹3,125 above as the arithmetic of the rate, not as a complete computation of anyone's liability.

Which ladder are you even standing on

The new regime's, unless you actively opted out of it — the Income-tax Act 2025 makes it the default at s.202, and it is the ladder quoted here. Every figure above belongs to one specific rate table, and quoting slab arithmetic without saying which is how most articles on this subject go wrong.

The old regime still exists, has its own thresholds and its own rates, and is now the one you have to ask for — so the default slab table is the right starting assumption for most readers rather than an alternative to check.

The regimes differ in a second way that changes which ladder is even reachable. Under the default regime there is no s.123 deduction — the provision everyone still calls 80C, capped at ₹1.5 lakh — no s.126 for health insurance premium, no s.129 for education-loan interest, no deduction for interest on the house you live in, and no house rent allowance exemption, which is available in the old regime only. Those deductions are what let somebody in the old regime land on a lower rung than their gross pay suggests.

Which of the two produces less tax is arithmetic on your own figures rather than a general fact, and it turns on one quantity: the deductions you would actually claim. The comparison is worked through in the old regime against the new, and where slabs sit in the eight-step sequence that produces a final bill is in income tax basics.

One transition note, because two statutes are live at once. Income for the year to 31 March 2026 is still governed by the repealed Income-tax Act 1961, so a return being filed for that year runs on its slab table and its section numbers. Income from 1 April 2026 falls under the 2025 Act. The bands quoted here are the current ones, and the ones on an older return are not wrong — they belong to a different year.

Four mistakes this misunderstanding produces

Each of these is specific, common, and traceable to confusing a slab with a threshold.

1. Refusing income to stay in a band. Declining overtime, a bonus or a promotion because of the slab above is protecting against a penalty that the ladder cannot impose. The only threshold in the calculation that can justify the worry is the rebate, and it sits at one specific income level rather than at every band edge.
2. Quoting a marginal rate as a bill. “I lose a third of my salary to tax” describes the top rupee, not the salary. The average rate is always lower, and knowing your own is what makes any comparison of two offers or two regimes honest.
3. Buying a deduction to change bands. A deduction is worth its size multiplied by your marginal rate, which is a discount rather than a return — except close to the rebate threshold, where the same rupee can be worth far more. Away from that threshold the arithmetic is ordinary, and the instrument still has to be worth owning on its own terms. What the eligible ones actually are is set out in tax-saving investments.
4. Assuming a nil slab bill means nil tax. The rebate is worth nothing above its ceiling and worth nothing against special-rate income below it, so equity gains can leave a return with tax payable and a filing obligation even when the salary computation came to zero.

The trade-off underneath all four is the same one. A slab ladder gives you a bill that rises smoothly with income, which is why no band edge is worth managing around; the thresholds bolted on to it — the rebate, and surcharge further up — buy relief for people below them at the price of a discontinuity for people beside them. Smoothness and targeting cannot both be had, and Indian income tax has chosen targeting at the rebate ceiling and again further up at surcharge. Those thresholds, not the band edges, are the only places in the whole calculation where an extra rupee is worth a second look.

What the arithmetic needs from you

Nothing above requires a calculator you do not have. It requires two inputs the reader has to supply: the current rate table, and an honest figure for total income, including the parts that arrive from outside a payslip.

That second input is where most self-computations go wrong, because capital gains are invisible to every deductor and are the item people reconstruct worst from memory. Working out whether a fund sale fell short-term or long-term needs the acquisition date and the redemption date; working out the gain needs the per-unit value on both. FNOTrader's Mutual Funds app runs on the full published AMFI history of daily per-unit scheme values — the net asset value, or NAV — around 34 million rows of it, and reports invested amount against value, maximum drawdown, and the annualised return on cashflows landing on irregular dates, which is XIRR. That is the transaction picture a gain computation takes as its input, not the computation itself.

FNOTrader is not a chartered accountant, a tax practitioner or a SEBI-registered investment adviser, and nothing here is tax advice or a recommendation about any product or regime. The section numbers are the Income-tax Act 2025 as it stands; the bands, the rebate and the rates on top of them change with each Finance Act, so take them from the current table rather than from any article — this one included.

Common questions

Continue reading

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