Return on invested capital

The numerator is the easy half: operating profit after the tax the business would pay on it, before interest, so that it belongs to lenders and owners together. The denominator is where every analyst makes choices. Financing side: equity plus debt, minus the cash the business does not need to operate, because idle cash is not working in the business and would drag the return down for a reason that has nothing to do with pumps. For Brightwell that is 70 plus 30 minus 20, so 80 million. Operating side: the assets that make the product, factories, stock, money owed by customers, minus the operating liabilities suppliers effectively lend for free, 110 minus 30, the same 80.

Each choice moves the answer. Call all of Brightwell's cash "needed" and the denominator is 100, the return 12.0 percent instead of 15.0. Leave out leased equipment the business cannot run without and the return looks better than the economics. The professional habit is to write the denominator down line by line, so that a reader who disagrees can change one line, not argue about the whole number.

The cost of capital here is a constructed 9 percent, a judgment about the return investors could expect elsewhere for Brightwell's risk. How to derive it (CAPM, betas) is outside these ten sessions; the intuition that money has a price over time is in Economics: money, inflation and interest ↗.

Suppose 20 million of Brightwell's 80 is goodwill from buying a valve maker a few years ago (constructed). Goodwill is the premium paid over the value of the assets acquired. Two respectable views:

Goodwill in: 12.0 / 80 = 15.0 percent. This is the return on everything the owners actually paid, acquisition premiums included. It judges management as allocators of capital: if they overpaid for the valve maker, this number shows it.

Goodwill out: 12.0 / 60 = 20.0 percent. This is the return on the operating assets the business runs on today. It judges the underlying economics: what another dollar invested in the existing operation is likely to earn, since organic growth does not require paying anyone's acquisition premium again.

Neither is the correct one. They answer different questions, and the gap between them, 5 points here, is itself information about how much of the business was bought rather than built. Report both, and say which one a valuation leans on.

The value formula on the figure is a growing perpetuity of the cash owners actually receive: next year's after-tax operating profit, times the share not reinvested, divided by the discount rate minus growth. The share reinvested is growth divided by the return on the reinvestment. All Brightwell numbers below are constructed arithmetic.

Standing still: no growth, nothing reinvested, 12.0 / 0.09 = 133.3. Reinvest at 15 percent: 3 percent growth needs 20 percent of profit reinvested, owners keep 9.6, value 9.6 / 0.06 = 160.0. Growth added 26.7. Reinvest at 6 percent: the same 3 percent growth needs 50 percent reinvested, owners keep 6.0, value 6.0 / 0.06 = 100.0. Growth subtracted 33.3. Reinvest at exactly 9 percent: 33 percent reinvested, value 133.3, identical to standing still.

That last line is the whole session in one number. Growth at a return equal to the cost of capital adds nothing; above it, growth is worth paying for; below it, a growth forecast lowers the value. A model that grows earnings without asking what the growth costs has silently assumed the first kind of business.

Read Buffett's 1989 letter, the section on mistakes of the first twenty-five years, at berkshirehathaway.com. Two of the mistakes he names are this Part in prose. Buying a cheap textile business, Berkshire itself, and the habit he calls "cigar butt" investing, buying poor businesses because they are cheap, both fail for the same reason: the price was low, but the capital went on earning little. And he writes that good managers cannot fix bad economics that skilled management does not rescue a business whose economics are poor.

In the language of tonight: management can move a return a little, but if the business earns below its cost of capital on what it reinvests, a better manager mostly makes it grow faster in the wrong direction. How long a business can keep earning above its cost of capital is the valuation input called a moat, and the subject of Analyst session 8; the economics of why some businesses can is in Economics: games and moats ↗.

Depreciation is a rough stand-in for the wear on existing assets; capital spending above it suggests the asset base is growing, below it suggests it is shrinking or being replaced more cheaply. Over the last ten fiscal years the ratio of total capex to total D&A was 0.57 for Caterpillar, 0.82 for Exxon Mobil, 0.57 for Johnson & Johnson and 1.19 for Procter & Gamble. Read naively, only P&G put back more physical capital than wore out.

Now the honest caveats, one per filer. Caterpillar: its capex tag leaves out equipment its finance arm buys to lease, while its D&A includes depreciation on that fleet, so 0.57 understates what it spends. Exxon Mobil: fiscal 2020's D&A of 46,009 million included large write-downs, which inflate the denominator. Johnson & Johnson: much of its D&A is amortisation of acquired intangibles; its reinvestment shows up as acquisitions and research spending, neither of which is capex. And for all four: capex below D&A does not prove the business stopped reinvesting. It proves only that this one line was smaller than that one line.

That is why the signal is crude and why it cannot replace a return on capital. It tells you where to look in the filing, not what you will find.

Damodaran's terminal-value slides set two rules for the forever stage of a valuation. The stable growth rate cannot exceed the growth rate of the economy, though it can be lower, and the riskfree rate is a simple proxy for nominal economic growth. And the reinvestment in stable growth equals the stable growth rate divided by the stable-period return on capital (Damodaran, A., valuation class slides, terminal value session, NYU Stern, 2023, read at source).

Here is the arithmetic, labelled as arithmetic. A terminal growth rate of 2.5 percent at a 10 percent return on capital requires reinvesting 2.5 / 10 = 25 percent of earnings, every year, forever. A model that assumes 2.5 percent terminal growth while handing all of earnings to owners has assumed an infinite return on capital, which no business earns. Analyst session 4 uses 2.5 percent terminal growth on owner earnings; this rule is the reason to ask, there, what reinvestment that number quietly assumes.

A useful corollary from Part 2: if the stable-period return on capital equals the cost of capital, the terminal growth rate stops mattering to value at all. Whether a particular business deserves a stable return above its cost of capital is the moat question of Analyst session 8.