Return of Capital: Handing a Rupee Back as One Use

Returning capital is the decision to hand a rupee back rather than deploy it, and it competes with every other use on the same test. A dividend returns the rupee itself. A share repurchase buys the company's own earnings at whatever the market price happens to be, so the price paid, and nothing else, fixes what that route returns.

Everything below rests on four things already settled. The five competing routes for a rupee, and the 14.2 per cent this business already earns on the capital it has, came from the opening guide in this sequence. The mechanism by which a repurchase retires shares was introduced there too, and is used here rather than explained again. The capital budget settled what is genuinely in contention for this period's money. And the figures themselves come from the record for Harivansh Packaging Limited, an invented manufacturer: 18.00 crore shares, net worth of Rs 1,650 crore, earnings before interest and tax (EBIT) of Rs 339 crore, cash of Rs 140 crore, and an illustrative share price of Rs 300/- carried at 28 August 2026.

What does the record not contain, and why say so first?

The record for Harivansh Packaging Limited publishes no distribution at all. There is no dividend, no payout ratio, no history of what has been handed back before, no split of the Rs 1,650 crore of net worth between share capital and accumulated profit, and no face value for the share.

So what is the Rs 7.78/- a share that appears below? The Rs 7.78/- is the whole Rs 140 crore cash balance divided by the 18.00 crore shares, worked out to see what handing the entire balance back would look like. That figure is an arithmetic consequence of a hypothetical act, not a distribution anybody declared, and the difference between those two things is the difference between teaching a decision and inventing a fact.

The absence turns out to suit the subject. The subject is not what a company pays out or why it settled on a level. The subject is the far narrower question of what happens to a set of numbers when a rupee walks out of the door, and that question needs a balance sheet and a share count, both of which the record does carry.

What does it actually mean to hand a rupee back?

Think of a shopkeeper who has had a good year and is holding Rs 4 lakh in the drawer. She can put a second counter in, she can clear the loan she took for the freezer, she can buy out the stall next door, or she can take the money home. Nobody would call the last of those generosity or discipline. Taking the money home is a statement, and a fairly blunt one: nothing she can do with that money inside the shop is worth more to her than the money itself.

A company handing capital back is making exactly that statement, at scale and in public. Harivansh Packaging Limited earns 14.2 per cent before tax on the capital employedNet worth plus borrowings, so the whole pool of money tied up in the business regardless of who supplied it. Built up alongside the statements and taken here as a given. it already has. When it hands a rupee back, it is saying that this particular rupee could not be put to work above that line, and so the shareholder should have it instead.

Returning capital is a claim about the opportunity set, not a virtue and not a failure. Whether a company should adopt a standing policy of returning capital, what a distribution signals to a market, how a sustainable level is chosen and whether it can be held through a bad year are all real questions and none of them is answered here. Those questions belong to the financing material, where a payout is treated as a policy. A payout is treated here as one competing use of one rupee, ranked against four others on the same test.

Every route on the list is judged by what it returns and what it forecloses. Returning capital gets no exemption from either question just because the money is going to the people who supplied it in the first place.

Which of the two forms moves the numbers, and which does not?

There are two ways to hand money back and they are far more alike than the argument around them suggests. A distribution takes cash out and leaves the share count exactly where it was. A repurchase takes the same cash out and retires some shares with it. The share count is the whole of the difference.

Most of the confusion starts in the profit ladder. Follow what happens to it in each case. Profit after tax is Rs 225 crore. Nothing about a distribution touches it: a payment out of a balance sheet is not a cost in a profit ladder, so Rs 225 crore stays Rs 225 crore. Nothing about a repurchase touches it either, for the same reason. In both cases the numerator of earnings per share is identical afterwards.

The denominator differs. Leave the count at 18.00 crore and earnings per share is Rs 225 crore over 18.00 crore, or Rs 12.50/-, exactly what it was. Retire shares and the same Rs 225 crore divides among fewer of them, so the figure rises. Both routes remove the same rupees from the same business, and the entire difference between them lies in how the remaining claims on an unchanged profit are arranged.

Read the bottom three rows of both panels before anything else. Net worth lands at Rs 1,510 crore on both routes. Capital employed lands at Rs 2,250 crore on both. The return on that capital lands at 15.1 per cent on both. A reader who believes the two forms are fundamentally different acts has to explain why three of the six lines are identical to the rupee.

What does a repurchase actually buy?

When a company repurchases its own shares, it is buying something specific: a slice of its own future earnings. And like anything else bought, what it costs decides what it was worth buying.

The arithmetic is short. Rs 225 crore of profit sits behind a market capitalisation of Rs 5,400 crore at the illustrative Rs 300/- share price. Spend a rupee acquiring that stream and the rupee buys Rs 225 crore over Rs 5,400 crore of it, or 4.17 per cent. That 4.17 per cent is the earnings yieldEarnings expressed as a percentage of the price paid for them, so the inverse of a price to earnings multiple. Constructed in the valuation material and used here as a given., and it is what the repurchase returned.

Now change one thing and only one thing. Suppose the market price on the day had been an illustrative Rs 200/- instead. The same Rs 140 crore, the same Rs 225 crore of profit, the same plant, the same customers, the same order book. The rupees now buy Rs 225 crore of profit against a market capitalisation of Rs 3,600 crore, a yield of 6.25 per cent. At an illustrative Rs 450/- they buy 2.78 per cent.

Sit with the spread for a moment. The best of those three bars is more than twice the worst, and the company did not do anything to earn the difference. The company did not improve a margin, win a customer, cut a cost or change a plan. The company simply happened to be transacting on a day when the market had one view rather than another.

The same repurchase is a different decision at a different price. No other route on the list has its return set outside the company. Growth spending returns whatever the project returns. Repaying borrowing returns the contracted rate that stops being paid. An acquisition returns whatever the target earns against what was paid for it, and the price there is at least negotiated. A repurchase alone has its return handed to it by a market on a particular morning.

A repurchase can therefore be the strongest use of a rupee available and the weakest use available, in the same company, in the same year, on the same set of accounts, depending only on when it happened.

What does a distribution return to the company that pays it?

Nothing. The one-word answer is the complete one, and it is not a trick.

A repurchase at least acquires something. Rs 140 crore buys 0.4667 crore shares at Rs 300/-, and those shares carry earnings, so there is a yield to compute even if it turns out to be a poor one. A distribution acquires nothing whatever. The rupee leaves at par with itself, and the company gets back exactly no claim on any future stream.

The answer sounds like a criticism and is the opposite of one. Because a distribution returns nothing, the comparison it invites is beautifully clean. There is no yield to argue about, no basis to reconcile, no multiple to defend. There is only the rupee against what that rupee would have earned had it stayed where it was. On this record that is 10.6 per cent after tax.

The construction of that 10.6 per cent matters more than the number. Notice how it was built. The figure is EBIT of Rs 339 crore taxed at this company's effective tax rateThe tax charge actually borne divided by profit before tax, so a measured outcome rather than a headline rate. Here it is Rs 75 crore on Rs 300 crore, or 25.0 per cent, and it is this invented company's own figure. of 25.0 per cent, giving Rs 254.25 crore, set against capital employed of Rs 2,390 crore. The 10.6 per cent is not 14.2 per cent multiplied by three quarters. Multiplying a number already rounded for display only rounds it again.

Because it buys nothing, a distribution is the cleanest test on the entire list: it is a direct, unhedged admission that the internal return was not worth having on that rupee. There is nowhere for the argument to hide. A weak acquisition can be defended with a synergy nobody has measured. A weak project can be defended with a second phase nobody has funded. A distribution has no such story available to it. Boards therefore find a distribution harder to announce than they find spending the same money badly.

Why does the hurdle rise when the capital leaves?

Almost nobody predicts the movement that follows, and anyone who reads ratios for a living needs to.

Return on capital employed is EBIT over capital employed. Hand Rs 140 crore back and the numerator does not move at all: the machines still run, the customers still buy, EBIT stays at Rs 339 crore. The cash was part of the pool the business had tied up in it, so the denominator moves by the full Rs 140 crore, from Rs 2,390 crore to Rs 2,250 crore.

Divide it out. EBIT of Rs 339 crore against capital employed of Rs 2,390 crore gives 14.2 per cent, and the identical EBIT against a pool of Rs 2,250 crore gives 15.1 per cent. The business got measurably better on the headline efficiency ratio by doing precisely nothing to the business.

Now put that next to the way the rest of this sequence works. Every competing use of a rupee is judged against the return the company already earns, and that line has been 14.2 per cent throughout. The 14.2 per cent line is not a fixed property of the business. Shrink the capital and the ratio rises, so any company can lift the hurdle its own uses are judged against.

The household version of this is uncomfortably familiar. A person carrying Rs 5 lakh of idle savings alongside a small business earning good returns has a mediocre average return on everything they hold. Take the Rs 5 lakh out and spend it on a holiday and the average return on what remains improves. Nothing about the business changed and nobody got richer. The measurement simply stopped including the lazy part.

The practical warning follows directly. If a company hands capital back and then reports an improved return on capital employed, those two facts are one fact reported twice, not evidence that the decision worked. Anyone assessing whether a return of capital was well judged has to compare against the ratio as it stood before, and say out loud that the denominator moved.

What happens when both routes run on the same Rs 140 crore?

Harivansh Packaging Limited has Rs 140 crore of cash sitting on its balance sheet. The share price on the record is an illustrative Rs 300/- carried at 28 August 2026. Take that cash and run each route to the end.

Route A, the whole balance handed out

Rs 140 crore over 18.00 crore shares is Rs 7.78/- a share. Neither the Rs 225 crore of profit nor the 18.00 crore count moved, so earnings per share stays at Rs 12.50/-. Net worth falls from Rs 1,650 crore to Rs 1,510 crore and capital employed from Rs 2,390 crore to Rs 2,250 crore, so return on capital employed rises from 14.2 per cent to 15.1 per cent on the same EBIT of Rs 339 crore.

Route B, the same balance spent on its own shares

Rs 140 crore at Rs 300/- buys 0.4667 crore shares, taking the count from 18.00 crore to 17.53 crore. Earnings per share rises from Rs 12.50/- to Rs 12.83/-, a gain of 2.7 per cent. Capital employed falls by the same Rs 140 crore, so return on capital employed rises to the same 15.1 per cent.

Then look at the per-share book figure. Almost no report carries it. Net worth per shareShareholders' funds divided by the number of shares. The number is taken from the balance sheet and carries no claim about what a share is worth in a market. was Rs 1,650 crore over 18.00 crore, or Rs 91.67/-. After the repurchase it is Rs 1,510 crore over 17.53 crore, or Rs 86.12/-. The figure fell for a completely legible reason. The shares were bought at Rs 300/- against a book figure of Rs 91.67/-, or 3.27 times it.

One row in that table complicates the story rather than tidying it. Net worth per share falls on both routes, not only on the repurchase. Handing out Rs 140 crore against an unchanged 18.00 crore shares gives Rs 1,510 crore over 18.00 crore, or Rs 83.89/-, lower than the Rs 86.12/- the repurchase leaves. The repurchase preserves more book value per share than the distribution does, and the repurchase is also the route people accuse of destroying book value. The accusation is made that loosely.

The honest statement is narrower and more useful. A repurchase lowers net worth per share only when the price paid exceeds the book figure, and lifts it when the price is below. Check the boundary yourself. Buying at exactly Rs 91.67/- would retire 1.5273 crore shares, leaving Rs 1,510 crore over 16.4727 crore shares, or Rs 91.67/- again, unmoved to the paisa. Rs 300/- is well above that boundary, so the figure falls. The rule is a condition with a number attached, not a general property of repurchases.

What the repurchase bought, stated beside what it cost

Earnings of Rs 12.50/- for a price of Rs 300/- is a yield of 4.17 per cent. The Rs 225 crore is profit after tax, so that yield is an after-tax figure. Set it against the 10.6 per cent after tax the existing capital earns, and against the 14.2 per cent it earns before. The rupees left a pool returning 10.6 per cent after tax and bought a stream returning 4.17 per cent after tax.

The comparison is the whole finding, and its limits matter. The comparison does not say the repurchase was a mistake. The comparison names what was given up and what was acquired, on one basis, in one line, and anyone can then see the two quantities beside each other. Whether Harivansh Packaging Limited should have done it depends on things no published figure carries.

Would the lift have appeared at any price at all?

People habitually mash two ideas together, and one question separates them. Earnings per share rose 2.7 per cent. Is that rise evidence that Rs 300/- was a sensible price?

Test it directly by paying more. At an illustrative Rs 600/- the Rs 140 crore retires 0.2333 crore shares, leaving 17.7667 crore, so earnings per share becomes Rs 12.66/-, a lift of 1.3 per cent. At an illustrative Rs 1,200/-, four times the recorded price, it retires 0.1167 crore shares and earnings per share is still Rs 12.58/-, a lift of 0.7 per cent. At an illustrative Rs 200/- it would be Rs 13.01/-.

The lift never turns negative, at any price tested. A rise in earnings per share appears at every price on this record, so it cannot possibly be evidence that the price paid was worth paying.

Do not mistake the result for a law. The result holds here for a specific and checkable reason. This record attributes no part of the Rs 225 crore of profit to the Rs 140 crore of cash, so when the cash leaves, no earnings leave with it. Change that assumption and the result changes.

So the honest version of the claim is conditional and it names its condition. On a company holding genuinely idle cash, a repurchase lifts earnings per share at any price, and the lift is therefore worthless as evidence. On a company whose cash was earning something real, the lift becomes informative, and the threshold it has to clear is exactly the earnings yield at the price paid. Either way, the number that carries the information is the yield, and the number everybody quotes is the lift.

What is gone once the rupee is gone?

Rs 140 crore is not a large number against a Rs 3,180 crore business, and that is precisely why it needs sizing properly rather than eyeballing.

Measure it in time. The business generates Rs 342 crore in a period, being earnings before interest, tax, depreciation and amortisation (EBITDA) of Rs 477 crore with Rs 60 crore of finance cost and Rs 75 crore of tax taken out of it. Rs 140 crore against Rs 342 crore is 4.9 months of everything the business produces. Handing it back is therefore not a rounding decision; it is close to five months of output leaving in one act.

Measure it in alternatives instead, and the number bites harder. The acquisition of Sundarban Polymers Private Limited is funded by exactly Rs 140 crore of this company's own cash alongside Rs 1,000 crore of new borrowing. Not approximately. Exactly. The cash that would fund a return of capital and the cash that would fund the acquisition are the same rupees. The two routes are mutually exclusiveTwo courses that cannot both be taken, because taking either consumes the same resource the other needs. The phrase is used strictly here: the same identified rupees are required by both. down to the last rupee rather than merely different.

Then there is the direction of the door. People describe returning capital as reversible, on the reasoning that a company short of money can always raise more. The reasoning is true, and it answers a different question.

Rs 140 crore of retained cash went out on a day whose price is on the record. A new issue of securities would come back, sold to whoever will buy them, at a price a market sets on some future morning, carrying whatever costs and conditions attach to it then. The two transactions are related in intent and unrelated in terms. Calling a return of capital reversible describes a possibility rather than an undo.

A household feels this instantly. Withdrawing Rs 2 lakh from a deposit to redo the kitchen is not undone by the possibility of borrowing Rs 2 lakh next year. The borrowing could indeed be done. The borrowing would be a different arrangement, at a rate nobody has quoted today, and the household would be the one paying it.

How does a lender read exactly the same act?

Everything above has been read from inside the company. A company that returns capital has to survive the reaction as well as the arithmetic. Change seat and the same Rs 140 crore looks different.

A lender assessing Harivansh Packaging Limited does not look at return on capital employed first. The lender looks at net debtBorrowings less cash held, so what would still be owed if every rupee on hand went to lenders tomorrow morning. The construction of net debt belongs with the statements. against EBITDA, the ratio a covenant usually attaches to. Today borrowings are Rs 740 crore and cash is Rs 140 crore, so net debt is Rs 600 crore and the ratio against EBITDA of Rs 477 crore is 1.26 times.

Return the whole Rs 140 crore and the cash goes to nil while the borrowings stay exactly where they were. Net debt becomes Rs 740 crore. Against the same Rs 477 crore of EBITDA that is 1.55 times. The identical act that lifts return on capital employed from 14.2 per cent to 15.1 per cent also pushes leverage from 1.26 times to 1.55 times, and both movements come from the same disappearing Rs 140 crore.

A leverage figure without a stated basis is worth very little. The 1.26 times and the 1.55 times are both struck on this company's own borrowings against this company's own EBITDA of Rs 477 crore, the standalone basis. Mixing that numerator with a combined denominator is a known way to understate leverage.

An equity analyst reading the same announcement does a third thing again, and it takes about a minute. Take the rupees committed. Take the price. Divide the earnings by the market capitalisation at that price to get the yield bought. Put that yield beside the after-tax return on capital employed. On this record those two numbers are 4.17 per cent and 10.6 per cent, and the analyst now has a question worth asking management rather than a press release worth repeating.

None of this needs translating to a domestic scale. A household with Rs 3 lakh set aside and a Rs 12 lakh home loan can hand the Rs 3 lakh to a relative, and their average return on what they still hold might well improve. Their cushion, meanwhile, is gone, and the bank looking at them next year sees the second fact rather than the first.

What is settled somewhere else, and by whom?

Three questions a reader will reasonably have after all of the above are not answered here, and it is better to say who answers them than to guess.

The first is policy. The signal a distribution sends to a market, how a level is chosen, whether it can be sustained through a weak period, and what happens to a share price when a company that has always paid stops paying all belong to the financing material, where a payout is treated as a standing policy rather than as one use of one rupee.

The second is procedure. The acts a listed company must perform, resolve, disclose and file in order to distribute cash or to repurchase its own shares are set by authorities, and they move. The third is tax, in the hands of whoever receives the money. Tax also moves, and it changes the answer for the recipient without changing a single number in the company's own arithmetic.

Free reservesThe part of accumulated profit that a company is permitted to draw on for acts such as a repurchase, as distinct from amounts locked away for a stated purpose. The Companies Act sets what qualifies, and mca.gov.in carries the wording. are what such an act is actually funded out of in a company's books, and the record for this business publishes no split of net worth at all. Nor does it publish a face valueThe nominal amount attached to a share when it is issued, carried in the share capital line of the balance sheet. A face value is a book convention rather than a measure of worth, and Indian distributions are often announced as a percentage of it., and a percentage of face value is therefore an announcement this record cannot support.

Which two questions keep getting answered as though they were one?

One confusion sits under everything above, and it is worth naming as sharply as possible. When a repurchase happens, two entirely separate things occur at once and get reported as one.

The first is a division effect. The same Rs 225 crore of profit divides among 17.53 crore shares rather than 18.00 crore, so the per-share figure rises. The rise is arithmetic on a denominator and it says nothing about anything.

The second is a return. Rs 140 crore bought earnings at a yield of 4.17 per cent, and that is what the money actually did. The division effect and the return are answers to two different questions, and only the second says what the money did.