Dividend vs Buyback India: Tax, Signal & EPS

The cleanest tell a board ever gives you is not the buyback it announces. It's the one it talks up and then quietly drops.

In July 2025, a small, profitable IT-services firm — Ksolves India — told its own shareholders, on the record, that it was keeping dividends low on purpose to leave room for a buyback. Eighteen months later it had bought back nothing, said "No plan" to the next person who asked, and gone back to paying fat dividends. Nothing illegal, nothing even unusual. But it's the whole subject of this article in one company: the choice between a dividend and a buyback is treated by most investors as a tax footnote, when it is actually one of the loudest signals you get about how management thinks — about its own stock, and about whether it has run out of better ideas for the money.

(Ksolves and GHCL appear below as illustrations of how to read a cash-return decision — not as a view on either stock.)

One question ties the whole thing together: does the cash-return policy agree with what management said it would do with the money?

How does a company actually return cash to shareholders?

Reinvestment and debt repayment keep cash inside the business. Returning cash means deliberately sending it out the door, and there are only two formal routes in India: a dividend, or a buyback.

A dividend is a direct cash payment, declared per share, paid to everyone on the register on a record date. It is visible, periodic, and — once a company establishes a rhythm — psychologically sticky. Boards hate cutting dividends because the market reads a cut as distress.

A buyback is the company using cash to purchase its own shares from the market and extinguish them. The shares disappear; the count shrinks. In India a buyback happens one of two ways. A tender offer invites all shareholders to sell a proportion of their holding back to the company at a fixed price, usually at a premium to the market — retail investors get a reserved quota under SEBI rules. An open-market buyback has the company buying shares on the exchange over a window, at prevailing prices, with no guarantee any particular shareholder participates.

Same goal, two very different instruments. One pays everybody the same rupee. The other shrinks the pie so each remaining slice is bigger. The difference matters for tax, for the per-share maths, and — most of all — for what the decision tells you. Start with the dividend.

What does a dividend actually signal — and how is it taxed now?

A dividend's signal is commitment. When a board declares a 30% payout ratio and repeats it year after year, it is telling you the cash is recurring and dependable enough that management is willing to be judged on maintaining it. That reluctance to cut is the whole point — a stable dividend is a costly signal, because the cost of breaking it is reputational pain.

The tax picture flipped in Budget 2020. Until then, the company paid Dividend Distribution Tax (DDT) before the dividend ever reached you. As of FY21, DDT was abolished, and dividends are taxed in the shareholder's own hands at the slab rate. A company also deducts TDS of 10% (u/s 194) once annual dividends to a shareholder exceed ₹5,000.

That single change made dividends tax-inefficient for anyone in a high slab. If you're at the top marginal rate, a sizeable chunk of every dividend rupee is gone the moment it lands. A retiree in a low slab barely feels it; a high-income investor compounding a portfolio feels it a lot. So the same dividend is a different deal depending on who holds the share — worth remembering before you chase a fat payout. (For why a high headline yield is often a warning rather than a gift, see the dividend yield trap.)

Note what a dividend does not do: it doesn't change the share count, so it doesn't flatter per-share metrics. It is the honest, boring way to return cash. The buyback is where the optics get interesting.

How does a buyback lift EPS and ROE without the business improving?

A buyback removes shares from existence. Two ratios respond automatically.

EPS is profit divided by share count. Shrink the denominator and EPS rises even if profit is flat — no new customer, no new product, just fewer slices. ROE is profit divided by shareholders' equity, and paying cash out to buy shares reduces the equity base, so ROE rises too. Both improvements are arithmetic, not operational.

Here is the mechanical effect, side by side. Take a company earning ₹100 crore on 10 crore shares, with ₹500 crore of equity, deploying ₹50 crore either as a dividend or as a buyback at ₹50 a share (1 crore shares bought back).

The dividend lifts ROE (smaller equity base) but leaves EPS untouched. The buyback lifts both. Notice nothing in the business changed — same ₹100 crore of profit throughout. So a per-share number that "improved" can mean the company grew, or it can mean the company simply shrank its own denominator while standing still. That is exactly why a rising ROE driven by buybacks is not the same as a rising ROE earned by the business, and why you should always ask what's driving the trend before you applaud it. We unpack that distinction at length in ROCE vs ROE: what the gap reveals.

A buyback can genuinely create value — but only at the right price, which is the next question.

When does a buyback create value, and when does it destroy it?

The everyday version: imagine four friends own a pizzeria equally. One wants out. If the remaining three buy his quarter-share for less than it's truly worth, they each end up richer per head. If they overpay to get rid of him, they each end up poorer. A buyback is exactly that, at scale and in public.

A buyback below intrinsic value transfers wealth to the continuing shareholders — spend ₹50 to retire something worth ₹70, and everyone who stays is better off. Above intrinsic value it does the reverse: ₹70 spent to retire something worth ₹50, quietly destroying value for the people who hold on, while flattering EPS in the press release. The maths looks the same on the EPS line either way. That's the trap.

So a buyback embeds a claim: management believes its own shares are undervalued. That claim is testable. Was the stock actually cheap on cash-flow terms when the buyback ran, or near a high? Did management buy aggressively when the price fell, or only when it was already expensive? Boards that buy back near peaks, or use buybacks mainly to offset dilution from generous stock grants, are not returning value so much as managing the optics.

There's also a hard legal constraint worth knowing. Under the Companies Act and SEBI rules, a buyback may only be funded from a company's reserves and surplus — never from borrowed funds — and a company that has defaulted on its debt, deposits, or declared dividends isn't eligible to run one. So while a dividend can be quietly propped up with borrowing (a red flag we return to below), a buyback that gets done is at least cash-backed by retained earnings by law.

The tax angle used to tilt this decision. For years, buybacks were taxed at the company level — a buyback distribution tax of roughly 20% plus surcharge and cess — which made the proceeds tax-free in your hands and gave buybacks an edge over dividends for high-slab investors. As of FY25, after the change effective 1 October 2024, buyback proceeds are taxed in the shareholder's hands, treated like a deemed dividend at the slab rate. That largely neutralised the old arbitrage. Which forces a sharper question.

The said-vs-did test: one buyback that happened, one that didn't

Two real boards, both small-cap, both cash-generative, both standing at the same fork — return cash by dividend, or by buyback. One did exactly what it said. The other said one thing and did the opposite. The gap between them is the whole lesson.

Ksolves India — the buyback that was talked up, then dropped. On the Q1 FY26 call (21 July 2025), a shareholder asked promoter-MD Ratan Srivastava directly whether, with the stock down roughly 50% from its peak, a buyback now made more sense than dividends. His answer: "my future planning is that definitely buyback in future and by keeping this dividend on the lower side" (Ksolves Q1 FY26 concall). Read that twice. He was explaining the low dividend — the first interim dividend that year was just ₹1 a share — as a deliberate choice to conserve cash for a buyback.

Then it didn't happen. By the Q3 FY26 call, another investor asked whether — given the stock had fallen further — a buyback or promoter buying might "instil some confidence in the market." The same promoter's full reply: "No plan. I think we have given good numbers for the confidence. If it is not giving the confidence, then what I can do" (Ksolves Q3 FY26 concall). And the cash that had supposedly been husbanded for a buyback? It went straight back out as dividends — a third interim dividend of ₹5 a share that quarter, taking the FY26 total to ₹11 (Ksolves Q3 FY26 concall). By the Q4 FY26 call (30 April 2026), management confirmed the policy was now a plain 40–60% dividend payout, "as we are not looking for any acquisitions or inorganic growth currently" (Ksolves Q4 FY26 concall). The buyback was never mentioned again. Inve's Promise Tracker logs that buyback commitment as missed — it was guided, then went silent, then explicitly withdrawn.

Here's the part that makes it worth your attention rather than your outrage. There was nothing dishonest about any single statement. The business had genuinely wobbled — Ksolves' quarterly net profit fell from ₹10.3 crore in Q3 FY25 to ₹5.9 crore in Q4 FY25, and its operating margin slid from 37% to 26% over the same window (Inve data, FY25). A board that guides toward a buyback "depending on the situation" and then meets a worse situation by hoarding flexibility is behaving rationally. But an investor who heard "definitely buyback" in July 2025 and sized a position around it was reading a forecast as a commitment. The words were soft; the takeaway shouldn't have been.

GHCL — the buyback that was guided and delivered. Now the contrast. In Q2 FY26 the GHCL board announced its third buyback, for ₹300 crore. One quarter later, on the Q3 FY26 call, the managing director didn't hedge: "I am happy to report that we successfully completed our shareholders' buyback program of Rs. 300 crores. This reflects our strong cash generation ability and our confidence in the company's future" (GHCL Q3 FY26 concall). And it was cash-backed in the boring, lawful way — GHCL carried roughly ₹3,460 crore of reserves against just ₹83 crore of borrowings at the time (Inve data, FY26), so this was retained earnings going out, not debt dressed up as a return. On the same call the CFO totted it up plainly: ₹300 crore of buyback plus ₹115 crore of dividends meant "we have rewarded our shareholders by distributing Rs. 415 crores during the 9-month period of this financial year, which is 116% of PAT for the nine months period" (GHCL Q3 FY26 concall). Inve's Promise Tracker logs that one as achieved.

Two boards, two buyback sentences spoken on a call. One became a press release and a completed programme; the other became "No plan." The only way to tell them apart in advance was to stop trusting the announcement and start tracking the follow-through.

So which is better — and what is the choice really telling you?

Here's the non-obvious part. Now that the October 2024 change has flattened most of the tax difference, the dividend-versus-buyback choice is barely a tax decision anymore. It's a signal about how management thinks about its own stock and its own cash.

A buyback is management implicitly making two claims at once: that the shares are undervalued, and that it has no better use for the cash than buying itself. Both are testable against the concall. Ksolves is the cautionary version of the first claim — the promoter said the stock looked cheap and a buyback was coming, then declined to act when it got cheaper still. The second claim has its own trap: did the same management spend the prior year guiding aggressive growth capex — new plants, new geographies — and then pivot to buying back stock? If so, one of those two statements isn't true. You can't simultaneously be capital-starved for growth and so cash-rich that the best available return is your own equity.

Read the cash-return policy and the growth guidance together; the moment they contradict, the decision exposes what management actually believes versus what it says.

The contradictions to hunt for: a big buyback running while management still guides heavy reinvestment, or a dividend maintained out of borrowed money rather than free cash flow. A payout funded by debt is not a sign of strength; it's a board protecting a signal it can no longer afford. To tell a real, cash-backed payout from a cosmetic one, trace it through the cash flow statement — profit can be massaged; the cash that funds a dividend cannot. GHCL passes that test out loud: ₹415 crore distributed against ₹3,460 crore of reserves and almost no debt. A board borrowing to maintain a payout would fail it.

How do you check capital-allocation guidance against what management actually did?

This is where the talk-versus-do gap lives, and it's a behaviour you see across the market, not just in one company. Capital-allocation lines are the ones that quietly drift: the payout target softens, the buyback-if-cheap line goes silent, the cash gets redirected — and four quarters on, nobody circles back. Inve's tracking bears this out. Across more than 13,000 management commitments logged over 1,500+ listed Indian companies (Inve data, as of 2026-06-12 — a read on how managements communicate now, across roughly the last two years of calls, not a lifetime verdict), only about 55% of the commitments that have come due were delivered as stated; more than 900 were never mentioned again on any later call, and 35% of companies (534 of 1,519) are carrying at least one such commitment that quietly went silent. Capital-allocation lines are exactly the class that drifts. "We'll maintain a 30% payout," "we'll buy back if the stock stays here," "definitely buyback in future" — these are easy to say on a call and easy to walk back four quarters later when nobody's checking. Ksolves isn't an outlier; it's a clean specimen of the general pattern.

So check. When a board states a payout policy or floats a buyback, the useful move isn't to record the headline — it's to pull the last several quarters of concall transcripts and see whether the prior cash-return guidance survived contact with reality. Did the maintained payout hold when profit dipped? Did the buyback-if-cheap line turn into an actual buyback, or evaporate once the stock recovered? Inve's Promise Tracker logs each such commitment and marks it kept, missed, or quietly dropped, so you read the delivery record — not the press release — alongside the policy. The KPI screener lets you line that up against the payout ratio and ROE trend across a watchlist, which is the only practical way to do it for ten or fifteen holdings, quarter after quarter, without living inside the transcripts yourself.

The discipline is simple to state and rare to practise: judge a cash-return decision by the company's track record of doing what it said, not by the elegance of the latest slide.

Where this read can be wrong. Tracking the follow-through is necessary, but it isn't a verdict on the management. A board that drops a buyback because the business genuinely deteriorated — as Ksolves' margins did — may be acting more responsibly than one that ploughs ahead on autopilot; flexibility is sometimes the right call, and a single dropped commitment over a two-year window isn't a character indictment. The transcript record many of these tools rest on is only about two years deep, which is enough to spot a pattern of one-way guidance but not enough to judge a management's lifetime. And a completed buyback isn't automatically value-accretive — GHCL did exactly what it said, but whether ₹300 crore retired shares below intrinsic value depends on the price it paid, which is a separate question from whether it kept its word. Keeping one's word and making a good capital-allocation decision are two different tests; this article is about the first. For more on separating credible guidance from theatre, see which management guidance to trust.

So before you applaud the next buyback announcement, ask the owner's question: five years from now, do I want this board's word to have meant something — and does its record of doing what it said give me any reason to believe it will?

Frequently asked questions

Related reading

• Asset Turnover Ratio: Formula, DuPont & ROE Link
• Capital Work-in-Progress (CWIP): A Balance-Sheet Red Flag
• Cash Conversion Cycle Explained: DSO, DIO, DPO