Dividend vs Buyback: The Same Cash, Two Different Promises

Both send cash out. A dividend reaches every share in proportion, so a holder who does nothing is paid. A buyback reaches only the shares tendered, so a holder who does nothing is paid nothing and instead holds part of a company with fewer shares in it. A regular dividend also sets a level for next year. A buyback sets none. Sankalp Industrial Systems Limited, invented, has done both.

The shape of this is easier to feel than to read, so start on a street. Two neighbours run small businesses on the same lane, and each has a good year. The first one calls in all four of the people who put money into the shop when it opened, and hands each of them an envelope sized by how much they put in. Nobody had to ask, nobody had to be there, nobody had to decide anything. The envelopes were sized by the record of who put in what.

The second one does something else entirely. She puts up a notice saying that if any of the four wants out, she will buy their stake back at a stated price, and that the notice stands until the money she has set aside runs out. One of the four takes it. Three do not. At the end of the week the same quantity of cash has left the shop in both lanes, and yet almost nothing else about the two events is the same.

Ask the obvious question. Who is better off? The wrong instinct is to answer it, and the right instinct is to notice that the two shops did not do the same thing to the same people. In the first shop, four people were paid and four people still hold exactly the share of the shop they held on Monday. In the second, one person was paid and has gone, and three people were paid nothing at all and now hold a third each of a shop that used to be quartered between four. The cash that left is identical and the people it reached are not, and that single fact is the whole subject of this guide.

The finance version has bigger numbers and the same anatomy. The finance version runs on Sankalp Industrial Systems Limited, invented, a listed maker of industrial valves, precision castings and the aftermarket parts and service that go with them. The company has one subsidiary, Sankalp Coatings Private Limited, invented, and one associate holding in Aruna Tooling Private Limited, invented.

Both routes send cash out, so where exactly does the similarity stop?

The similarity stops almost immediately. A dividendCash paid to every shareholder in proportion to holdings. is a payment declared per share and made to whoever is on the register, in proportion, with no action required from the person receiving it. A buybackCash paid only to the shareholders who sell their shares to the company. is a purchase: the company goes into the market for its own shares, pays the people willing to sell, and the shares it buys stop existing. One is a distribution against a register. The other is a transaction with a counterparty.

Set five questions against each route and the pattern is easy to see. Does cash leave the company? Who receives it? What happens to the number of shares? What is committed for the year after? And where in the company's own record does it appear? On the first of those five the two answers are the same word. On the other four they are not merely different in degree; they are different in kind, and no amount of arithmetic converts one into the other.

Only the first of the five questions gets the same answer from both routes, and a comparison that stops at the first question has compared nothing. Everybody starts with the first question, and the warning is worth stating plainly. Cash left. Both are payout. Add them up. Everything after that sentence is where the two routes separate, and everything after that sentence is what a reader who stops there never sees.

One more thing before the detail. The size of the dividend and the size of the buyback do not decide the comparison. Sankalp Industrial Systems Limited, invented, distributed one quantity of cash as dividend in Year 0 and a different quantity as a buyback two years earlier, and their relative size is not what tells the two routes apart. A company can run a small dividend and an enormous buyback, or the reverse, and every structural difference below survives unchanged.

Who is standing on the other side of each transaction?

Everything else falls out of this question, so it is worth slowing down on. A dividend has the whole register on the other side of it. When Sankalp Industrial Systems Limited, invented, declared Rs 3.60 a share for Year 0, every one of its 20,00,00,000 shares was entitled to Rs 3.60, and the total leaving the company was Rs 72,00,00,000. A holder of one share received Rs 3.60. A holder of 4,000 shares received Rs 14,400.00. Neither of them had to lift a finger to get it, and neither of them could refuse it by inaction.

The word that carries this is pro rataIn proportion to holding, which is how a dividend reaches shareholders and a buyback does not.. Everybody receives in proportion, so nobody's slice of the company changes. A holder of one two-hundredth of the company on Monday holds one two-hundredth on Friday, with some cash in hand. A dividend leaves the shape of the shareholder base exactly as it found it. The property is strange and rather elegant once noticed.

A buyback has almost nobody on the other side of it. Sankalp Industrial Systems Limited, invented, bought 75,00,000 of its own shares at Rs 80.00 at the end of year minus 2. The cash leaving was Rs 60,00,00,000. There were 20,75,00,000 shares in issue at the time. So the cash reached the holders of 3.61 per cent of the shares and reached the holders of the other 96.39 per cent not at all. A buyback pays the people who leave and pays the people who stay nothing whatever.

The effect on the shape of the shareholder base is the mirror image of the dividend. Everybody who stayed still holds the same number of shares, and there are now fewer shares in total, so everybody who stayed holds a slightly larger fraction of the company. The holders who stayed were not paid in cash. They were paid, if the word fits at all, in proportion. Whether the larger fraction is worth anything depends on the value of the company, and a share count cannot settle the value.

What does a shareholder who does nothing actually receive?

Readers resist this difference harder than any other in the subject, and the calculator below puts a control on it. Under a dividend, the shareholder who does nothing receives cash. Under a buyback, the shareholder who does nothing receives no cash at all. Not a smaller amount. Not a delayed amount. Nothing.

The reason people resist it is that they have heard buybacks described as returning cash to shareholders, and that description is true of the company and false of most of the shareholders. The company did return the cash; it went out of the door and does not come back. The cash went to the specific people who agreed to hand over their shares in exchange, and it was those people, not the company, who decided the cash would reach them.

There is one case where the two routes land in exactly the same place, and it is worth knowing because it marks the boundary of the difference precisely. Suppose the company splits a fixed quantity of cash between the two routes, and suppose a holder tenders exactly their proportional share into the buyback: a holder of one per cent of the company offers one per cent of the shares being bought. Then whatever the split, the rupees reaching that holder are identical. The dividend lost by having less paid out is exactly the sale proceeds gained by tendering. The two routes are the same transaction for a holder who tenders in proportion, and completely different for a holder who does nothing. The interesting question is therefore never about the routes and always about the holder.

The proportional tender dissolves a lot of confusion. Nothing about a buyback is designed to disadvantage the holder who stays. The holder who stays has simply not participated in a transaction that was open to them, and has ended up with a bigger slice instead of some cash. Whether that suits them depends on facts about them, not about the company.

What happens to the share count, and what does that change?

The share countThe number of shares in issue, which a buyback reduces permanently and a dividend leaves alone. is the single arithmetic thing only a buyback touches. A dividend of any size, in any year, on any scale, changes the count by exactly nothing. Rs 72,00,00,000 went out of Sankalp Industrial Systems Limited, invented, in Year 0 and the count was 20,00,00,000 before it and 20,00,00,000 after it. The buyback, by contrast, took 75,00,000 shares out of existence and moved the count from 20,75,00,000 to 20,00,00,000 in one step.

Every figure that has the share count in its denominator therefore moves when a buyback happens and does not move when a dividend happens. Earnings per share is the obvious one. Divide the same profit by a smaller count and the quotient is larger, and the size of the effect from the count alone is easy to state exactly: 20,75,00,000 over 20,00,00,000 is exactly 1.0375, a lift of exactly 3.75 per cent. The 3.75 per cent is the count effect and nothing else. Every reader can do that arithmetic in their head, and that is precisely why it needs a warning attached.

The warning is this. The Rs 60,00,00,000 the company spent was not free. The money existed, was doing something, and could have gone on doing it. Once what that cash was itself earning is netted off, the honest effect on earnings per share is lower than 3.75 per cent, and the full arithmetic of it, together with the return at which the whole effect vanishes, is covered separately. The count effect is the subject here.

And there is a larger warning that outranks that one. Earnings per share is not value. A buyback can lift earnings per share and leave the company worth less than it was, and a company that spends cash it needed in order to move a per share figure has done something arithmetically successful and financially foolish. The count effect is a fact about division. A fact about division is not a finding about value.

What does each route commit the company to next year?

Now the largest of the differences, and the one that shows up least in the arithmetic. A regular dividend is a commitmentThe level a company will be measured against next year, which a regular dividend sets and a buyback does not.. Not a legal one, in the ordinary case, and not one a company can be sued over. A practical one: once a company has declared Rs 2.60 a share, the number Rs 2.60 becomes the reference point against which the following year's declaration is read. A declaration of Rs 2.80 has raised it. A declaration of Rs 2.60 has held it. A declaration of Rs 2.40 has cut it, whatever the accompanying words say.

The five years of Sankalp Industrial Systems Limited, invented, make a textbook shape. The regular dividend went Rs 1.80, then Rs 2.00, then Rs 2.20, then Rs 2.40, then Rs 2.60. Five years, five rises, never a pause and never a step back. A company that has produced that sequence has, without writing a word of it down anywhere, told everybody what a normal year looks like, and it will now be read against that pattern until it breaks it.

Set that against the same company's Rs 1.00 special dividend in Year 0. A special dividend is declared once and labelled as a one-off precisely so that it does not enter the sequence. Nobody who receives Rs 1.00 of special dividend expects Rs 1.00 of special dividend next year, and that is not a happy accident; it is the entire reason the board structured it as a special rather than raising the regular from Rs 2.60 to Rs 3.60. The label on a distribution is doing work that the amount of it cannot do: it says whether this is a level or an event.

A buyback sits in the same category as the special dividend on this one criterion. A buyback commits the company to nothing at all. The transaction happens, the shares go, and the following year begins with no expectation attached. Freedom from commitment is not an argument for the route. The freedom is a structural fact about it, and one of the two or three things actually driving the choice a board makes between the routes when it has cash and has to decide how to send it out.

Which is why the question a board is really answering is not about this year at all. The real question is whether the company expects to have this cash again next year. If the answer is yes, a rise in the regular dividend says so and can be sustained. If the answer is no, saying so through a rise in the regular dividend creates an expectation the company will have to disappoint, and two structures exist that avoid the problem entirely.

Why did the same declared dividend become cheaper to pay?

This is the effect almost every comparison of the two routes leaves out, and it is the one place where a one-off transaction quietly rewrites a repeating one. A dividend is declared per share and paid on every share. Shrink the number of shares and the same declaration costs less money. The declaration has not changed. The recurring costWhat a declared dividend per share costs in total each year, which falls when the share count falls. of honouring it has.

Work it on the invented company. In year minus 1, the year straight after the buyback, Sankalp Industrial Systems Limited, invented, declared Rs 2.40 a share and paid it on 20,00,00,000 shares, at a cost of Rs 48,00,00,000. Had the buyback never happened there would have been 20,75,00,000 shares, and the identical Rs 2.40 declaration would have cost Rs 49,80,00,000. The buyback made the same declared dividend Rs 1,80,00,000 a year cheaper to pay, and it will go on being cheaper for as long as the company pays a dividend at all.

Notice the shape of that last row, because it is the part that surprises people. The saving is not a fixed rupee amount fading into the background. The saving is the declared rate multiplied by the shares that no longer exist, so every time the company raises its dividend the saving from the old buyback gets larger too. Across the two years since the transaction it comes to Rs 1,80,00,000 and then Rs 1,95,00,000, a total of Rs 3,75,00,000 of dividend the company did not have to find.

Now the discipline that has to travel with that, and it is exactly the same discipline as with the count effect. A lower recurring cost is not a gain. The company paid Rs 60,00,00,000 up front to get it. Whether Rs 60,00,00,000 was a sensible price for 75,00,000 of its own shares is a valuation question and not a cash flow one, and it turns on what those shares were worth. Something that reduces a future outflow by paying a present one has not created anything by itself. The buyback changed the timing and the shape of the outflows.

Where does each one show up in the company's own record?

A dividend is easy to find because it announces itself. A dividend appears as a rate declared per share, as a total distributed in rupees, and in the statement of cash flows as a payment to shareholders. Three places, all of them labelled with the word dividend, all of them stated in a form a reader can compare across years without doing anything clever. A person who wants to know what Sankalp Industrial Systems Limited, invented, paid out in Year 0 finds Rs 3.60 a share and Rs 72,00,00,000 without effort.

A buyback is easy to miss because it announces itself in two places that a reader watching dividends is not looking at. A buyback appears in the share count, moving it from 20,75,00,000 to 20,00,00,000, and in the cash flows as an outflow for the purchase of the company's own shares. A buyback is not a dividend, so it never appears on the dividend line, at any point, in any form.

A reader tracking distributions through the dividend line alone will record year minus 2 as a year in which Rs 45,65,00,000 went out, when the true figure was Rs 1,05,65,00,000, and the entire missing Rs 60,00,00,000 is the buyback. The gap is not a small error of measurement. The missing Rs 60,00,00,000 is more than half of what actually left the company, invisible because it left through a different door.

There is a household version of this that makes it stick. Consider someone tracking what a shop pays out by watching only its salary account. Every month the same figures appear, and the conclusion is that the shop is steady. The one thing the watcher cannot see is that in one of those months the shop also bought out a partner. The salary account was perfectly accurate. The salary account was simply not the whole of what left.

How is a year with a buyback compared against a year without one?

Carefully, and by refusing to let one number stand for the answer. Here are the five years of Sankalp Industrial Systems Limited, invented, laid out with the two routes kept in separate columns and one further column that no summary ever carries: what a holder of 100 shares who did nothing at all actually received in cash that year.

Reading the last column down and then the buyback column sets out the whole difficulty of comparing these years. Year minus 2 is by a very wide margin the biggest year in the table by cash returned. Year minus 2 is also the year in which the holder of 100 shares who did nothing received Rs 220.00, the third smallest figure in the column and entirely unremarkable. Both of those statements are true at once, and no single figure can carry both.

The only honest comparison across a buyback year keeps three numbers apart that a summary wants to merge: the regular dividend, the special dividend and the buyback. Merged, they give a total that is arithmetically correct and useless for almost every question anybody actually asks. Kept apart, every question becomes answerable, including the ones about who was paid and what was committed.

Two more mechanical cautions when a buyback sits in the middle of a series. First, the share count is not the same on both sides of it, so a per share figure from year minus 3 and a per share figure from year minus 1 are not two readings of the same instrument. State the count alongside any per share figure that crosses the transaction. Second, Year 0 in this table contains a Rs 1.00 special dividend inside its Rs 3.60, so the Rs 3.60 is not the fifth term of the Rs 1.80 to Rs 2.60 sequence. The regular sequence at Year 0 is Rs 2.60. The Rs 3.60 is the regular sequence plus an event.

What do the two routes have in common that readers keep forgetting?

With the two routes separated, it is worth being precise about what they share. The shared parts are real, and getting them wrong produces its own confusion. Three things.

First, both are cash actually leaving. Neither is an accounting entry, neither is a promise for later, and in both cases the rupees are gone from the company on the day they go. A company that has done either has less cash than it had, and whatever it was going to do with that cash it can no longer do. Rs 72,00,00,000 out as dividend and Rs 60,00,00,000 out as buyback are equally out.

Second, both come out of the same pool and compete with everything else that pool funds. Every rupee that goes out by either route is a rupee not reinvested in the business, not used to pay down borrowing and not held against a bad year. The decision about how much to send out is made before the decision about which route to use, and the second decision does not change the first.

Third, and people find this hardest to accept, under a stated and quite restrictive set of conditions the choice of route does not change the value of the company at all. The argument belongs to Miller and Modigliani, whose 1961 paper in the Journal of Business set it out, and the conditions it needs and the ones this invented company breaks are worked through separately. The route becomes consequential only where the world departs from those conditions, and naming which condition is doing the work in a particular case is the whole of the analysis.

Tax is not on that list of shared things. How each route is taxed, in whose hands, and at what point, is set by law and it changes. A printed rate would eventually be wrong in a way a reader could not detect, and it would be exactly the kind of wrong that somebody acts on. A rate belongs to the current text of the law, so the treatment appears below as a category with an authority attached and no number.

What does a lender, an analyst or a household do with the difference?

A lender reads the two routes as two different kinds of claim on future cash. A regular dividend that has risen five years running is, for practical purposes, a standing annual outflow that the company will fight hard to protect, so a lender sizing headroom treats it as close to fixed and asks what is left after it. A buyback is a discretionary event that happened and is over, so it comes out of one year's cash and out of the balance sheet and does not sit in next year's model at all. The same rupee leaving produces two completely different lines in a lender's projection depending on which door it went through.

An analyst reads them as a data problem before reading them as anything else. The denominator changed at one point and did not change gradually, so any per share series that crosses the buyback has a break in it. Any measure of cash returned that comes off the dividend line alone is missing the buyback entirely. So the first job is mechanical: pull the regular dividend, the special dividend and the buyback into three separate rows for every year, write the share count next to each row, and only then compute anything. The second job is different and more interesting: read what the pattern of the regular dividend says about what the board expects.

A household with shares in a listed company reads it as the difference between something that happens to them and something they have to decide about. A dividend arrives. There is no form, no timing, no decision, and it is the same rupees per share for the person holding four shares and the person holding four lakh. A buyback asks a question and the answer defaults to no if nothing is done. Doing nothing is a decision in a buyback and is not a decision in a dividend, and that asymmetry is the practical difference most people actually experience. Whether to act on such an offer is a question about a person's own position.

A board reads all three of those readings back at once, and the choice between the routes is rarely about arithmetic. The choice is about whether the cash is expected to recur, about what the company wants to have committed to itself twelve months from now, and about who the company wants to have on its register at the end of it. None of those is a calculation, and none of them is settled by whichever route makes a per share figure look larger.