Incremental return on capital, and when it lies

Return on capital measures what the capital a company already has deployed is earning. The incremental return measures what each new unit earns, and it is the one that actually decides whether reinvestment creates value. It is also the one that most easily returns a number that looks good and means nothing.

Why the incremental figure is the one that matters

A company compounds value as long as it can reinvest at a return above its opportunity cost. The level of the return describes the past — it includes investments made twenty years ago; the incremental figure describes what is happening now, and it is what gets projected forward.

The link to growth is direct: the growth a company can finance out of its own results is the fraction of profit it reinvests, multiplied by the return at which it reinvests. If it grows faster than that, it is financing the difference with other people's money — suppliers' or customers' — or because it needs very little capital to grow. Both are real advantages, but it is worth knowing which of the two is at work.

When the number lies

The ratio is the change in after-tax operating profit over the change in invested capital. If both changes are negative, the resulting ratio is positive. A company that is shrinking — that sold divisions, recognised impairments and earned less — can publish a double-digit incremental return suggesting it compounds value, when it is doing the opposite.

It happened to us with a declining food manufacturer: after-tax operating profit fell, invested capital fell further through divestitures and impairments, and the ratio came out clearly positive. It would have painted the signal of a compounding business onto the page. That is why the rule is that if invested capital falls, the incremental return is not published: the row and the chart disappear, and the omission is declared. One fewer data point beats one wrong data point, and this error strikes exactly where the reader most needs the right signal.

A serial acquirer shows a structurally low return

When a company grew by buying other companies, its invested capital carries the price paid across decades of acquisitions, in the form of goodwill and intangibles. The return on that capital comes out low even when the operation is excellent, because the denominator includes what was paid above the book value of what was bought.

The mistake to avoid is chaining together that the return is below the bar, that the business is therefore mediocre, and that it therefore deserves a floor multiple. That chain produces absurdly low valuations of good companies. What is correct is to declare why — the numerator grows but the denominator carries accumulated goodwill — and to derive the multiple from the type of business rather than from the raw return.

And sometimes the denominator simply does not work

There are two ways to end up without usable invested capital, and both publish a plausible number. One is that the debt figure is stale and the calculation writes zero: a zero there does not read as missing data — it draws the quality bar at zero and asserts that the company generates no return. The other is that invested capital is negative, which is perfectly normal in a business with net cash and revenue collected in advance; dividing by a negative number returns a meaningless ratio.

In that case a defensible base is chosen — measuring on contributed capital, without netting out excess cash — and which one was used is declared, because the result depends on that choice.