Dividend investing, and the arithmetic under the payout

A dividend is a transfer, not a return

A dividend is cash the company sends out of its own bank account and into yours. The company is worth that much less the moment it goes, and the share price is adjusted down for it on the first day the share trades without the right to that payout — the ex-date. The money did not appear from anywhere. It moved.

Take an illustrative holding, chosen to keep the arithmetic in sight rather than because it resembles anything: 100 shares at ₹400, so ₹40,000 of market value. The board declares ₹8 a share. On the ex-date the reference price is adjusted to ₹392, and your 100 shares are marked at ₹39,200. The ₹800 arrives separately. Add them and you are back at ₹40,000.

Nothing about your position improved. What changed is its composition — ₹800 of it is now cash sitting in a bank account instead of a claim on a company that was holding that cash on your behalf. A transfer, not a return, and the distinction is the whole of this article.

The cleanest demonstration is not in shares at all. When a mutual fund makes a payout under the income distribution cum capital withdrawal option — IDCW, the thing older material calls a dividend option — the scheme's NAV falls by the amount distributed per unit, because a fund's NAV is nothing but its assets divided by its units, and cash that has left is no longer an asset. Nobody argues about this in a fund. It is the identical arithmetic in a company, made visible because a fund's book is marked to market daily and a company's is not.

One honest qualification, because the rest of the article depends on it being stated precisely. The adjustment is mechanical: an exchange reduces the reference price for the payout and the share opens against that reference. What the price does next is not mechanical at all, and it is where most of the confusion lives. A share can go ex-₹8 and close ₹15 higher on the day for reasons that have nothing to do with the dividend, which makes it easy to conclude that the drop never happened. It happened. It was simply swamped.

The date machinery behind all of this — which date puts you on the register, which date you have to have bought by, and why buying on the record date is too late — belongs to corporate actions, and is not repeated here.

So why does anyone look at dividends at all?

Because the payment is not the point. The capacity to make it is, and the two are worth separating carefully.

Earnings are an opinion produced by applying accounting policy to judgements — the argument set out in PE and PB. A depreciation life, a capitalised cost, a revenue recognition call: each moves reported profit without a rupee going anywhere. A dividend cannot be produced that way. It requires a bank balance, and once the payment has cleared, the cash is gone in a way no policy choice can reverse.

So a payout that has been funded year after year, out of the business rather than out of borrowing, is a statement about cash generation that the profit line cannot make on its own. It is evidence, not the reward. The reward, if there is one, came from the business earning the money in the first place — the dividend only reports it.

There is a second thing a settled payout does, and it is about management rather than accounting. Cash inside a company has to be allocated by somebody, and it can be allocated into a diversification nobody asked for. A board that has committed to returning a share of it every year has narrowed its own options. Whether that constraint helps depends entirely on what the alternative use of the money would have earned, which is the trade-off in the next section.

Now the line this article will not cross. That a sustained payout is evidence about cash generation is a claim about mechanism, and it stands on its own. That companies which pay dividends have, as a group, delivered better outcomes for shareholders is a different kind of claim entirely — an empirical one, true or false of a specific market over a specific window, and answerable only with a named dataset and a stated period. The two get run together constantly, and the second is usually smuggled in behind the first. This article makes the first and does not make the second.

The rupee that was paid out is a rupee not reinvested

Every dividend has a cost, and it is the least discussed number in the whole subject: whatever the company would have earned on that money had it kept it.

A business that can put a rupee to work at a high rate of return has an expensive reason not to pay it out. A business that cannot has an expensive reason not to keep it, because retained cash earning very little still sits in the equity base and drags the return on it downwards — the mechanism behind return on equity and return on capital employed. Retention has a cost too, but it is an unannounced cost — nothing is declared, filed or paid out when a company simply keeps the money.

Which means a rising payout is not self-evidently a good sign or a bad one. It can be a mature business honestly returning cash it cannot deploy at an attractive rate. It can equally be a business out of ideas. The announcement looks identical in both cases, which is exactly why the payout on its own decides nothing.

This is also where the phrase “dividend investing” needs unpacking rather than endorsing. As a style it makes a specific bet: that a company's demonstrated willingness and ability to hand cash back is more informative about its future than the growth it might have bought with that cash. For that bet to pay, the payout must be funded from operations rather than reserves, it must survive a bad year, and the business must not have been sacrificing a better use of the money to fund it. Those are three testable conditions, not a verdict on the style.

Payout ratio, and the denominator it is measured against

The payout ratio is the dividend per share divided by earnings per share — the share of a year's reported profit that was handed out. Take an illustrative company earning ₹20 a share and paying ₹8: a payout ratio of 40%, so three-fifths of the year's profit stayed inside.

The ratio's use is in what happens when the two move apart. Suppose earnings fall to ₹5 a share and the board holds the payout at ₹8 rather than cut it. The ratio is now 160%, and the arithmetic of that is not subtle: the company paid out more than it earned, so the difference came from reserves built in earlier years or from borrowing. Paid from reserves or borrowing is a fact about this year that the dividend announcement itself does not mention.

None of which makes a payout above 100% wrong. A company with a genuinely one-off bad year and a large cash balance may reasonably choose to hold the payment steady rather than signal distress by cutting it. The point is narrower: the ratio tells you where the money came from, and that is a different question from whether the decision was sound.

Then the denominator problem, which is the same one that afflicts every ratio built on reported profit. Earnings are accrual figures, and a dividend is paid in cash. A company can report profit while collecting very little of it, if the sales sit in receivables or the year's gain came from selling an asset rather than from trading. The payout ratio measured against free cash flow asks the harder question — whether the cash to pay it was actually generated — and the two ratios can point in opposite directions for the same year.

Here is the mistake worth naming, because it is the common one. A single year's payout ratio is read as a statement about policy, when it is mostly a statement about that year's denominator. A ratio that jumped because the dividend rose and a ratio that jumped because earnings collapsed look the same in a screener column and are not remotely the same event. The only way to tell them apart is to look at the two series separately — the dividend per share year by year, and the earnings per share year by year — rather than at the ratio they produce.

The yield trap is a fact about division

Dividend yield is the annual dividend per share divided by the current share price. So it has two inputs, and it rises when either the top goes up or the bottom comes down. That sentence contains the entire trap.

Continue the illustration. A ₹8 dividend on a ₹400 share is a yield of 2%. If the board raises the payout to ₹12 and the price is unchanged, the yield becomes 3% — the numerator moved, and something happened at the company.

Now hold the dividend at ₹8 and let the price fall to ₹160. The yield is 5%. It more than doubled, the company declared exactly the same rupees as before, and the move came entirely from the denominator. Whatever caused the price to fall by three-fifths is not addressed by the yield rising; it is the reason the yield rose.

There is a second mechanism stacked on top, and it is the one that catches people sorting a screen. A trailing yield divides last year's declared dividend by today's price. When a business deteriorates, the price responds within days and the dividend responds at the next declaration, which can be many months later. In that gap the trailing yield is computed from a payout that may already be gone. A stale numerator, a live price — and a screen sorted by yield puts exactly those cases at the top, because that is what sorting by a ratio with a falling denominator does.

So “high yield” is not a category of company. It is a position in a sorted list, and shares arrive at the top of that list by two routes that have nothing in common. Separating them costs one extra step: look at what the dividend per share itself has done over several years, in rupees, before looking at the yield the ratio reports.

And the reframe worth carrying away: a yield is a ratio, not a payment. What lands in your account is the rupees, and the rupees do not change when the price does.

Three things a company can do with a rupee of profit

Setting the alternatives side by side makes the transfer visible in a way the dividend on its own does not. Take an illustrative company with 10 crore shares at ₹400, so ₹4,000 crore of market value, deciding what to do with ₹40 crore.

Read the fourth row against the fifth. The dividend leaves the per-share value ₹4 lower and puts ₹4 in your pocket; the buyback leaves the per-share value at ₹400 and puts nothing in your pocket, while making each surviving share a slightly larger slice of a slightly smaller company. The same ₹40 crore left the building both times.

Note what the buyback row does not say. It does not say the remaining shares became more valuable — at a purchase price of ₹400 the arithmetic gives exactly ₹400 after, and any claim beyond that is a judgement about whether ₹400 was the right price to pay, which is a different argument requiring different evidence.

Total return, and the dividend you can make yourself

Because the price is adjusted down for every payout, a price chart of a company that has paid for twenty years understates what holding it did. The payments are not on the chart; the drops they caused are. Adding the dividends back gives total return, and the gap between the two is exactly the cash that was handed over on the way.

The same distinction runs through index reporting. A headline index level is normally a price index, while the total return version of the same index assumes distributions were reinvested. Comparing a fund's performance against a price index credits the fund with the dividends and the index with none, which is a comparison of two different things. The correct benchmark is the total return version, and which one a factsheet used is worth checking rather than assuming.

Now the part that most changes how the subject reads. If a dividend is simply cash moved out of your holding and into your bank account, you can perform the same operation yourself. Return to the 100 shares at ₹400. Sell two of them for ₹800 and you hold 98 shares worth ₹39,200 plus ₹800 in cash — ₹40,000, arranged exactly as the ₹8 dividend arranged it.

The positions are identical in composition. What differs is not the money but the control: who chose the timing. A dividend arrives when the board decides and in the amount the board decides; a sale happens when you decide, in the amount you decide, with dealing costs attached and with a tax treatment that is not the same as the dividend's — the reason the two are worth reading side by side in dividend tax. A mutual fund investor faces precisely the same choice between an IDCW payout and a systematic withdrawal, for identical reasons.

This is also where the strongest emotional pull in the subject lives, and it is worth saying plainly. Spending a dividend feels like living off the income while leaving the capital untouched; selling two shares feels like eating into the capital. The arithmetic says the two did the same thing. Treating one rupee differently from another because of which pocket it arrived in is mental accounting.

The label is not only a criticism, though, and this is the part usually left off. An investor who will spend a dividend but will not sell a share may hold through a fall that would otherwise have shaken them out. That is a judgement about behaviour rather than a mechanical claim, and it leaves the arithmetic exactly where it was: the two transactions still do the same thing to the position.

What to look at instead of the yield column

Everything above points at the same practical move: the yield is the end of a calculation, and the interesting objects are its inputs. The dividend per share, in rupees, year by year. The earnings and the free cash flow that funded it, separately. What the price did, and when, relative to the last declaration.

FNOTrader's Stocks app screens across roughly 2,390 stocks and 17 NSE sector and size indices and filters on the reported fundamentals rather than only on the ratios computed from them, so a screen can be built on the series you actually meant instead of on the ratio a data vendor happened to publish. How a screen quietly answers a different question from the one you asked is the subject of the screener guide.

What no tool does is turn a payout into a decision. It can put the same calculation across every company so two numbers are comparable, and it can show the inputs next to the output. Reading them remains yours.

FNOTrader is not a SEBI-registered investment adviser or research analyst, does not recommend shares, and nothing here is a view on any company, any yield or any payout policy.

Common questions

Continue reading

More in Stocks & Screening · App: Stocks · Definitions: glossary · Free tools: calculators · All: every article