How to treat reinvested dividends in your return calculation

Dividends are where return calculations quietly go wrong in two opposite directions. Leave them out and you understate your return — sometimes badly. Put them in the wrong place — as if a reinvested dividend were fresh money you contributed — and you overstate it, or you depress it, depending on the error. The correct treatment is simple once you see the underlying rule, and the rule is this: a reinvested dividend is internal, and the right benchmark is a total-return index.

This piece is a companion to our cornerstone on how investment return is actually calculated. The cornerstone establishes the methods; here we apply them to the one input that trips up careful people.

Price return is not your return

Start with the distinction that explains everything else. A price return measures only the change in price. A total return adds the income the holding paid — dividends, distributions, interest — on top of the price change. For anything that pays an income, those are different numbers, and the gap is the dividends.

This matters more than it sounds, because most of the index levels you see quoted are price indices. The JSE All Share level on the news, the S&P 500 number in the headline — those generally ignore dividends.[1] Your portfolio does not ignore them; it earns and (usually) reinvests them. So if you ever compare your dividend-earning portfolio to a headline price index, you are giving the index a handicap and crediting yourself, or your manager, with the market's dividends. Always compare against the total-return version.

A reinvested dividend is not a cash flow

Here is the rule that keeps your own calculation honest. When a dividend is paid and reinvested inside the portfolio, cash leaves as a dividend and immediately re-enters as a purchase of more units. Nothing crossed your account boundary. The net external flow is zero. So a reinvested dividend must not appear in your cash-flow column — not as a contribution, not as anything.

This has a clean consequence for time-weighted return: your TWR is identical whether or not dividends are reinvested. Reinvestment changes how many units you hold, not the percentage growth path of the money already invested. The percentages are the same; only the unit count differs.

For money-weighted return (XIRR), the same logic holds — a reinvested distribution is not an external flow. The one case where it does change your XIRR is when the dividend is taken in cash rather than reinvested: money you actually withdraw is a genuine external flow, a positive one, and it belongs in the column. Reinvested: invisible. Taken out: a withdrawal like any other.

A worked example you can check

The cleanest way to see the price-versus-total gap is a single share over a year.[2]

You buy one share at R100. Mid-year it pays a R3 dividend, reinvested at the then-price of R101 — buying 3 ÷ 101 = 0.029703 of a share. By year-end the price is R104.

• Price return = (104 − 100) / 100 = +4.0%

• Total return, dividend reinvested = (1.029703 shares × R104) / R100 − 1 = +7.09%

Ignoring the dividend understated the return by more than three points on a four-percent price move. (For comparison, if you had simply counted the R3 as cash rather than reinvesting it, the total return would be (104 + 3 − 100) / 100 = +7.0% — almost identical here, because it is one period and a small dividend, but the two diverge as dividends compound over longer horizons.)

That three-point gap is not a rounding detail. Over years, with a typical dividend yield compounding, the difference between price and total return is most of the reason "the market" has done better than the index level on the news suggests.

The two mistakes

Almost every dividend error is one of these two:

Treating a reinvested dividend as a new contribution. This double-counts the money — the dividend is already inside the portfolio's value, and entering it again as a contribution inflates the capital base, which depresses your apparent return (you look like you put in more to get the same result). It also corrupts any XIRR, because you have added a phantom external flow. A reinvested dividend is never a contribution.

Comparing a total-return portfolio to a price index. The mirror image — this inflates apparent performance, because your portfolio's dividends have nowhere to be matched on the index side. It is the single most common way a perfectly good fund is made to look like it is beating a benchmark it is merely keeping pace with. (This is also the trap behind judging whether a manager beat the market, covered in a companion piece.

One pointer we will not resolve here: in South Africa, dividends carry dividend withholding tax, and that changes your after-tax total return and the gross-versus-net index choice. The mechanics above are pre-tax; the tax layer is real and belongs to a dedicated treatment.

So what does this all mean?

Dividends are not an optional extra in a return calculation; for anything that pays an income, they are most of the gap between the number on the news and the number in your account. Keep two rules and you will not go wrong: a reinvested dividend is internal — it never enters your cash-flow column — so your time-weighted return is unchanged by reinvestment; and your benchmark is always the total-return index, never the price level. The +4.0% versus +7.09% on a single share is the whole story in miniature: same holding, same year, three points of difference, all of it dividends.