Price Return, Total Return and Yield Are Three Different Numbers

Three definitions, one holding, one year

Buy a share at $100. A year later it trades at $108 and it has paid $3 in dividends. Capital-gains yield is (108 − 100) ÷ 100 = 8.00 percent. Dividend yield is 3 ÷ 100 = 3.00 percent. Total return is (108 − 100 + 3) ÷ 100 = 11.00 percent, which is exactly the sum because both are measured against the same starting price. Nothing more complicated is going on over a single period, and the whole confusion in the subject comes from what happens when you string periods together.

Over many periods they stop adding. Price growth compounds on the price, reinvested income compounds on the share count, and the two multiply rather than sum: a 5 percent price growth with a 2.5 percent reinvested yield gives a total return of 1.05 × 1.025 − 1 = 7.5625 percent a year, not 7.50. The difference looks like rounding and is not: over thirty years the extra 0.0625 point is worth 1.7 percent of the final position. This is the same multiplicative structure our companion article on dividend reinvestment decomposes into a share-count effect and a price effect — it is worth reading alongside this one rather than repeating here.

The number in the news is a price index, and the gap grows

Almost every headline index level is a price index: it tracks constituent prices and drops the dividends on the floor. Total-return versions exist for all of them and are what fund benchmarks actually use, but they are not the number read out on the evening news. The consequence is arithmetic. Take $10,000 at 5 percent annual price growth with a 2.5 percent yield reinvested. After ten years the price measure says 1.629 times your money and total return says 2.085 — the price measure has already missed 42.0 percent of the gain. After twenty years it is 2.653 against 4.348, missing 50.6 percent. After thirty, 4.322 against 9.066: $43,219 against $90,656, and 58.8 percent of the gain invisible. After forty years the price index reports less than a third of what actually happened.

There is one prominent exception, and it is worth knowing which side of it you live on. The German DAX is quoted by default as a performance index, meaning dividends are treated as reinvested inside the index level itself; a separate price version (the DAX Kursindex) exists but is not the headline. Almost every other national headline index — and the widely quoted American, French, Spanish, Portuguese and Italian benchmarks among them — is quoted as a price index with a total-return variant published alongside. That single methodological difference makes long-horizon comparisons between headline levels across countries meaningless unless you first check which construction each one uses. Compare total return with total return, or price with price, and never one against the other.

The ordering between two assets can invert

This is the practical reason the distinction is not pedantry. Asset A grows 7 percent a year in price and yields 1 percent. Asset B grows 4 percent and yields 4 percent. On price return A wins by three full points a year, which over twenty-five years turns $10,000 into $54,274 against B's $26,658 — A looks twice as good and any price chart will say so. On total return the ranking flips: A compounds at 1.07 × 1.01 − 1 = 8.07 percent and B at 1.04 × 1.04 − 1 = 8.16 percent, so after the same twenty-five years A is worth $69,603 and B $71,067. B wins, by $1,464, having looked like the obvious loser throughout.

Notice how narrow the margin is and how wide the price gap was. Two points of yield bought back three points of price growth and a little more, which tells you something useful about how to read a chart: a price line is not a return line, and the distance between two price lines carries no information about which holding paid its owner more. It also tells you that the comparison is fragile — change B's yield to 3.9 percent and A wins again. Anyone claiming a general result about dividend-paying versus growth assets from this arithmetic alone is over-reading it. The honest claim is narrower and more useful: the ranking depends entirely on which of the three numbers you look at, so state which one you mean.

Total return only exists if the income is actually reinvested

A total-return figure is not a description of the asset; it is a description of the asset plus a policy. Every published total-return series assumes each distribution is reinvested at the price prevailing on the day it is paid, with no tax withheld and no dealing cost. A holder who spends the income earns the price return and nothing more — that is the whole 4.322 versus 9.066 difference over thirty years. A holder who reinvests but pays tax on the way, or pays a fee per purchase, or holds cash for a week before buying, lands somewhere in between and closer to the lower figure than intuition suggests, because the shortfall compounds. Our companion article on dividend reinvestment quantifies the tax drag term by term rather than as a single haircut, which is the right way to see it.

One further point that separates the arithmetic from the story: a dividend is not new money. On the ex-dividend date the price falls by roughly the amount paid, so a distribution moves value out of the price and into your account rather than adding it. That is why total return is the only one of the three numbers that is conserved across the payment — capital-gains yield falls by the dividend and dividend yield rises by it, and the sum is unchanged. Any argument that a high payout is free income is an argument about the accounting, not about the money.

Yield on cost is a personal history, not a yield

A yield is income divided by a price, and everything turns on which price. Buy at $100 a share paying $3, and both the market yield and your own are 3.00 percent. Let the dividend and the price both grow 6 percent a year. After ten years the dividend is $5.37 and the price is $179.08 — the market yield is still 3.00 percent because both numerators and denominators grew together, but the dividend measured against your original $100 is 5.37 percent. After twenty years it is 9.62 percent; after thirty, 17.23 percent. That figure is real in the sense that the cash is real, and useless in the sense that it says nothing about the share.

The gap grows without limit, because the numerator compounds and the denominator is frozen at whatever you happened to pay. That makes yield on cost useless for the only thing yields are for, which is comparison: a 17 percent yield on cost and a 3 percent yield on market value describe the same share, and only the second can be set against an alternative you might buy today. It also makes it a poor reason to hold. The decision to keep a position is a decision to buy it at today's price with today's proceeds, and the price you paid in 1998 has no claim on that. Report yield on cost as what it is — a satisfying record of a past decision — and do the comparison on market yield and total return.

A falling price raises the yield, which is a warning and not a bargain

Dividend yield has the price in the denominator, so it moves inversely to the price with no news about the company required. Hold the dividend at $3 and let the price fall from $100 to $80, and the yield rises from 3.00 to 3.75 percent. Let it fall to $60 and the yield is 5.00 percent. Let it fall to $40 and the yield is 7.50 percent. That last figure is a very attractive-looking number produced entirely by a 60 percent loss. Compute the year's total return alongside it and the picture reverses: from $100 down to $60 with $3 paid, the total return is (60 − 100 + 3) ÷ 100 = −37.00 percent. The screen shows a 5 percent yield and the holder is down more than a third.

This is why screening on yield alone selects for exactly the wrong thing. A high yield is either compensation for a risk the market has already priced, or an artefact of a price that fell for a reason, or a payout the company will not sustain — and in the third case the yield is quoting a dividend that no longer exists. The forward-looking version of the question is whether the payout is covered by earnings and cash flow, which the payout ratio addresses and the yield does not. Read the yield as a ratio between two numbers you should look at separately, not as a rate of return, and pair it with total return before drawing any conclusion at all.

Frequently asked questions

What is the difference between capital-gains yield and total return?

Capital-gains yield counts only the price change: (ending price − starting price) ÷ starting price. Total return adds the income received over the same period, still divided by the starting price. On a share bought at $100 that ends at $108 having paid $3, capital-gains yield is 8.00 percent and total return 11.00 percent. Over one period the two differ exactly by the dividend yield; over many periods they differ by more, because reinvested income compounds on top of price growth — 5 percent price growth with a 2.5 percent reinvested yield compounds at 7.5625 percent a year, not 7.50.

Does the index I see on the news include dividends?

Usually not. Almost every headline index level is a price index, tracking constituent prices with distributions discarded; a total-return version is published alongside and is what fund benchmarks actually use. The German DAX is the well-known exception, quoted by default as a performance index with dividends treated as reinvested, and with a separate price version. This matters more than it sounds: at 5 percent price growth and a 2.5 percent yield, the price measure misses 42.0 percent of the gain after ten years and 58.8 percent after thirty. Before comparing two index levels across countries or across decades, check which construction each one uses.

Why does a share's yield go up when its price falls?

Because the price is the denominator. With a $3 dividend, a price of $100 gives 3.00 percent, $80 gives 3.75 percent, $60 gives 5.00 percent and $40 gives 7.50 percent — the company has done nothing and the yield has more than doubled. The same move is a disaster on total return: falling from $100 to $60 while paying $3 is (60 − 100 + 3) ÷ 100 = −37.00 percent for the year. This is why screening on yield alone tends to select falling prices and unsustainable payouts. Check whether the dividend is covered by earnings and cash flow before treating a high yield as an opportunity.

Is yield on cost a useful number?

It is a real number and a poor decision tool. Yield on cost divides today's dividend by the price you paid, so as the dividend grows and the purchase price stays frozen, it rises without limit: a share bought at $100 yielding 3.00 percent, with dividend and price both growing 6 percent a year, shows 5.37 percent after ten years, 9.62 after twenty and 17.23 after thirty — while the market yield stays at 3.00 percent throughout. Only the market yield can be compared with anything you might buy instead. Deciding to hold is economically the same as deciding to buy at today's price, and what you paid years ago has no bearing on that.

Can a lower-growth asset beat a higher-growth one?

Yes, when the income makes up the gap. An asset growing 7 percent a year in price with a 1 percent yield compounds at 8.07 percent in total; one growing 4 percent with a 4 percent yield compounds at 8.16 percent. Over twenty-five years, $10,000 becomes $54,274 and $26,658 respectively on price alone — the first looks twice as good — but $69,603 and $71,067 on total return, so the second wins by $1,464. The margin is narrow and the example is not a general result: drop the second asset's yield to 3.9 percent and the first wins again. The transferable point is that a price chart cannot rank two holdings by what they earned.

Do I actually get the published total return?

Only under the assumptions it is built on. Published total-return series reinvest every distribution at the price on the payment date, with no tax and no dealing costs. Spend the income and you get the price return alone — the whole 4.322 versus 9.066 difference over thirty years in our example. Reinvest but pay tax on each distribution, or a fee per purchase, or leave the cash idle for a fortnight, and you land between the two, closer to the lower figure than intuition suggests, because each shortfall compounds for the rest of the holding period. Our companion piece on dividend reinvestment works the tax drag through term by term.