Payout and Return of Capital: Dividends Plus Buybacks in One Ratio

A payout ratio divides what left the company for its shareholders by what the company earned for them. Give it profit attributable to owners, the dividend per share, the share count and any buyback, and four figures come back: the rupees distributed, the dividend payout ratio, the total payout ratio, and the yield. On Year 0 the last two read 52.17 and 4.00 per cent.

Two decisions sit under those four figures, and neither is a convention that can be looked up. The first is which profit line goes underneath. The second is what counts as having left the company. Decide either one wrongly and the arithmetic is still flawless while the answer is useless. The two decisions therefore deserve most of the room, and the division itself very little.

Every rupee below belongs to Sankalp Industrial Systems Limited, an invented manufacturer of industrial valves and precision castings, and to its five locked years of distribution history. Its five rows carry the three cases that keep turning up: a dividend that creeps upward, one year with a share buyback in it, and one year with an extra one-off payment on top. Between them those three produce every argument a payout ratio can start.

What goes in, and which profit line belongs underneath?

The shape is identical in a household and the words are easier, so start there. A salary of Rs 80,000 arrives in a month. Rs 30,000 goes out to the people the household supports: a parent in another town, a sibling in college. The rest stays, funds the fridge that finally has to be replaced, and cushions the month the scooter needs a gearbox. If somebody asks what share of the month went out, the answer is Rs 30,000 over Rs 80,000, and it is 37.50 per cent.

The household complicates in exactly the way a consolidated company is complicated. Suppose Rs 80,000 is not one salary but the pooled earnings of two brothers who run a shop together, and one of them holds only a quarter of it. The pooled figure is real and the shop earned it. The junior brother's quarter was never the household's to send, so the household cannot send it to the parent in another town. Dividing the Rs 30,000 by the pooled Rs 80,000 describes a household nobody lives in.

The junior brother's quarter is the whole of the denominator argument, and a consolidated company runs the same argument at scale. Sankalp Industrial Systems Limited holds 75.0 per cent of Sankalp Coatings Private Limited and fully consolidatesEvery rupee of a subsidiary's revenue, cost and profit is added into the parent's statements line by line, whatever slice of it the parent actually holds. it, so every rupee the subsidiary earns arrives in the group's profit after taxBottom line of the profit and loss account for the whole consolidated group, struck before anyone has asked whose it is.. At Year 0 that consolidated line is Rs 1,44,00,00,000. Sitting immediately beneath it is a deduction for minority interestOutside shareholders hold the rest of a part-held subsidiary. Their slice of its profit is stripped out before the group reports what belongs to its own shareholders. of Rs 6,00,00,000, and what remains, Rs 1,38,00,00,000, is profit attributable to owners.

Rs 6,00,00,000 is exactly one twenty-fourth of the consolidated figure. The minority deduction is small enough to look like a rounding line and large enough to move a payout ratio by more than two points, and that combination is what makes it dangerous. The rule that comes out of it is short: the denominator is the profit the parent's shareholders could in principle have been paid, and no rupee that belongs to somebody else can be in it.

The numerator has its own version of the same question, and it is the harder one. Cash leaves a company for its shareholders through two doors. Through the first it goes to everyone who holds a share, in proportion, as a dividend. Through the second the company buys its own shares back, and the cash goes to the people who chose to sell. Both doors move cash from the company to shareholders. Watching only the first door means that one particular year, whose true total came to Rs 1,05,65,00,000, gets reported at the dividend alone, namely Rs 45,65,00,000.

Output one: how many rupees actually walked out?

The first output is not a ratio at all, and skipping it is how people end up defending a percentage they cannot reconstruct. Rupees distributed is the dividend per share multiplied by the number of shares it was paid on. The Year 0 declaration by Sankalp Industrial Systems Limited ran to Rs 2.60 a share as the regular dividend, then Rs 1.00 a share as a special dividendDeclared once, described as one-off when it is declared, and carrying no promise at all about what next year will bring., on 20,00,00,000 shares. Rs 3.60 times 20,00,00,000 is Rs 72,00,00,000, of which Rs 52,00,00,000 arose from the regular declaration and Rs 20,00,00,000 the special.

The share count is the input people take least seriously and it moves. Two years earlier this company had 20,75,00,000 shares. Closing year minus 2, it retired 75,00,000 of them at Rs 80.00 apiece. The retirement cost Rs 60,00,00,000 and left 20,00,00,000 shares standing. Any calculation that carries last year's share count into this year's dividend is out by the whole of the difference, and here that difference is 3.75 per cent of the count.

Output two: what share of the year did shareholders get?

The rupees distributed divided by profit attributable to owners give the dividend payout ratio. Setting the Year 0 distribution of Rs 72,00,00,000 over Rs 1,38,00,00,000 of profit, the quotient prints as 52.17 per cent. The ratio reads the way the household reads: of every hundred rupees the company earned for its own shareholders in Year 0, a little over fifty two went straight back to them, and the rest stayed inside.

Profit that stays inside becomes retained earningsProfit kept inside a business rather than handed out sits here, and it funds tomorrow's machines, stock and wages without anyone being asked for new money.. A rupee of profit is either handed over or it is not, so the payout ratio and the retention ratio always add to one. Whether a given split is the right split is a question about strategy, covered separately. The ratio measures the split and does not judge it.

The same symmetry set against the household is the cheapest sanity check available. Of every hundred rupees the salary brought in, thirty went to the people the household supports and seventy stayed. Nobody needs a second calculation to know the seventy: it is whatever the thirty was not. The same is true here. A dividend payout ratio of 52.17 per cent is a retention ratio of 47.83 per cent stated the other way round, and if the two figures under examination do not add to a hundred then one of them was computed on a different denominator from the other. The add-to-a-hundred check has caught more denominator mismatches than any amount of care over the arithmetic.

One warning about reading the ratio backwards. A payout ratio is not a statement about cash in the bank. Profit is an accounting figure and the dividend is a cash payment, and a company can perfectly well declare a payout ratio of 52.17 per cent in a year when its cash balance fell. The ratio compares a payment with an accounting result. The comparison is useful, and it still does not answer whether the payment was affordable out of the year's own cash.

Output three: what changes once the second door is counted?

Everything so far has watched one door. Year minus 2 is the year the second door opened. Profit attributable to owners was Rs 1,16,20,00,000. The dividend was Rs 2.20 a share on 20,75,00,000 shares, or Rs 45,65,00,000. On its own that is a dividend payout ratio of 39.29 per cent, sitting comfortably alongside the 37.50 and 38.46 per cent of the two years before it.

Then, at the end of that same year, the company spent Rs 60,00,00,000 buying 75,00,000 of its own shares at Rs 80.00. The buyback cash went to shareholders. The money reached a subset of them rather than all of them in proportion, and it went out under a different heading in the cash flow statement, but it left the company and it arrived with holders of the company's shares. Add it to the numerator and year minus 2 stops being an unremarkable year: Rs 1,05,65,00,000 against Rs 1,16,20,00,000 is a total payout ratio of 90.92 per cent.

Output four: why does the yield swap the denominator?

The fourth output looks like the others and is built on a different foundation. A dividend yield divides the dividend per share by the share price. At Rs 90.00 a share, the Year 0 total dividend of Rs 3.60 gives exactly 4.00 per cent, and the regular dividend of Rs 2.60 on its own gives 2.89 per cent.

Notice what just happened to the denominator: profit left and price arrived. A payout ratio asks what the company gave away out of what it earned, so both halves come off the company's own accounts. A yield asks what a share pays against what a share costs, so half of it comes from the accounts and half from the market. Two figures that both end in per cent and look alike in print are answering questions with different subjects.

The two yields on this company differ by more than a point, and the difference is entirely the special dividend. Any yield built on Rs 3.60 is a yield built partly on a payment the company described as one-off when it declared it. A difference of that size is not a reason to prefer either yield, but a reason to name which of the two is in use in the same sentence that quotes it.

What do five years of one company look like, row by row?

Here is the whole history in one table. Every figure in it was produced by the arithmetic above, and the two right-hand columns are identical in four rows out of five.

Two rows deserve a second look. Year minus 2 is the one already worked through: the two ratios separate by 51.64 points because a buyback sat in one numerator and not the other. Year 0 is the quieter one: both ratios read 52.17 per cent, no buyback is involved at all, and the jump from 38.10 to 52.17 comes entirely from the Rs 1.00 special dividend. The same company therefore produces two very different looking spikes for two entirely different reasons, and only reading the numerator settles which is which.

What does the whole calculation look like on a quiet year?

Year minus 1 is the row with nothing unusual in it: no buyback, no special payment, no change in the share count during the year. Year minus 1 is therefore the right row to work through by hand, and working one row by hand turns a table into something that can be defended in a meeting.

Collect the four inputs and hold every one of them in whole rupees. Profit attributable to owners is Rs 1,26,00,00,000. The dividend per share is Rs 2.40, all of it regular. The share count is 20,00,00,000, already reduced by the retirement that happened at the close of the year before. Buyback rupees for the year: none at all.

Output one is the multiplication and nothing more: Rs 2.40 on 20,00,00,000 shares is Rs 48,00,00,000. Output two divides that by the profit: Rs 48,00,00,000 over Rs 1,26,00,00,000 comes to 0.380952 and a fraction, printed as 38.10 per cent. Nothing was bought back, so output three adds nothing to the numerator, and the same division prints the same 38.10 per cent. Whenever the buyback input is zero the third output collapses onto the second. Four of the five rows in the table therefore carry identical figures in their last two columns.

Output four is the one that cannot be finished, and the reason is worth more than the answer would have been. A yield needs a share price, and the only price locked for this company is the Rs 90.00 at Year 0. Nothing in a set of accounts supplies a price for year minus 1. Three of the four outputs are built entirely out of the company's own statements; the fourth needs a figure that lives outside them. Where that figure is not available, the honest move is to leave the output blank rather than to borrow a price from a different year and quote the result as though it belonged.

The same walk on year minus 3 takes about twenty seconds. Profit attributable to owners Rs 1,07,90,00,000. Rs 2.00 a share on 20,75,00,000 shares gives Rs 41,50,00,000. The division gives 38.46 per cent, with no buyback to add, so the total payout ratio is the same 38.46. The retirement had not happened yet, so the share count in that year is the larger one. A reader who reached for the current 20,00,00,000 would have computed Rs 40,00,00,000 of dividend and reported 37.07 per cent. The share count belongs to the year, not to the reader's most recent memory of it.

What does the run look like when it is drawn instead?

Step back from the individual figures and look at the outline. Practitioners draw payout as a line before they quote it as a number: a line makes the difference between a drift and an event visible in about a second, and a table does not.

The dividend line is close to flat with one step at the far end; the total line is the same flat run with a single tower rising out of the middle of it. Those two shapes carry different information. A gentle drift upward in the dividend line is a series of separate decisions, each one small, each one taken in a different year. A tower is one decision, taken once, that happens to be enormous. Averaging the two shapes together would describe a company that has never existed.

Add the five years up: is that the average of the five ratios?

Now ask a different question. Not what happened in a year, but what happened across the run. Foot the distributed column of the table and the five dividends come to Rs 2,44,50,00,000. Foot the profit column and it reaches Rs 5,87,70,00,000. One division on those two totals gives a five year dividend payout of 41.60 per cent. Set the single buyback of Rs 60,00,00,000 on top of the numerator, lifting it to Rs 3,04,50,00,000, and the five year total payout reads 51.81 per cent against the identical denominator.

The aggregate 41.60 per cent is not the average of 37.50, 38.46, 39.29, 38.10 and 52.17, and the difference is not a rounding artefact. The average of those five is 41.10 per cent, exactly half a point lower. The aggregate divides one total by another total, so a year that earned more contributes more to both halves and therefore weighs more. An average of ratios treats every year as one vote regardless of size. One vote each would be right for five separate companies and is wrong for one company across five years of growth.

What will these four outputs never show?

Four outputs is a small number, and the things left out matter more than the things left in. The set does not show what a buyback did to earnings per shareA year's profit divided by the number of shares in issue gives what a single share earned over that year.. The question has a naive answer and an honest one that differ, and it is worked through where buybacks are covered. The set does not show which of the yield and the ratio belongs in a given sentence. The four outputs do not show what a change in any of these figures means, and that is where signallingReading a change in what a company pays as a message about what its managers expect, rather than reading it as arithmetic. lives. And it says nothing whatever about whether any of these numbers is good.

Which of the two ratios to compute is decided by the question, not by the company, and the calculator returns both rather than choosing. A question about what shareholders received in cash from an entitlement they all shared is answered by the dividend ratio. A question about what the company handed back in total, through every door, is answered by the total ratio. A question about what one share paid against what one share cost is answered by neither ratio, and by the yield.

Which figures here are exact, and which are rounded?

One rule governs precision in every figure here, and a reader who checks the arithmetic without knowing it will find a discrepancy that is not one. Every rupee amount here is exact, and every percentage is the exact division rounded once, at the end, to two decimals. Rs 45,65,00,000 is not an approximation of anything: it is 20,75,00,000 shares multiplied by Rs 2.20, to the rupee. The 39.29 per cent beside it is 39.285714 and so on, cut to two decimals for display.

The consequence shows up the moment two displayed percentages are subtracted. Year minus 2 reads 39.29 and 90.92, and subtracting those two printed figures gives 51.63 points. The same subtraction on the underlying divisions gives 51.64. Neither is a defect and neither is more correct in general: the first is the difference between two rounded numbers and the second is the rounded difference between two exact ones. The 51.64 figure is the one to quote, computed on the exact values and rounded once.

The same discipline explains why the five year figure is 41.60 rather than 41.61 or 41.6. Rs 2,44,50,00,000 divided by Rs 5,87,70,00,000 is 0.41602858 and a little more, or 41.60 per cent to two decimals. Nothing in that chain was rounded before the final step, and the two trailing digits are only there because the third decimal would be inventing precision the inputs do not carry.

Sankalp Industrial Systems Limited and Sankalp Coatings Private Limited are invented.
Educational material. Not advice on any investment, tax, budget or market position.