Profitability and Returns: ROE, ROIC, DuPont

DuPont: three ratios, one product

Return on equity is net income over equity, and it can be rewritten without changing the answer. Multiply by revenue over revenue and assets over assets and the ratio separates into three: net margin (net income over revenue, the profitability of each sale), asset turnover (revenue over assets, how hard the capital works), and the equity multiplier (assets over equity, how much of the balance sheet is financed by something other than equity). The revenue and the assets cancel, so the product is exactly ROE. That identity is the whole point. It means the same return can be produced by pricing power, by efficiency, or by borrowing — and the three have entirely different durability. A high margin can be competed away; a high turnover model can be out-executed; leverage does not go away when things get worse, it amplifies. When you see a return on equity above peers, the first question is which term is responsible, and the answer usually decides whether the premium is worth paying. There is a trap in the leverage term that catches almost everyone at least once. Buying back shares reduces equity, which raises the multiplier and raises ROE without the business doing anything at all. A company with a flat operating performance, a shrinking share count and a rising ratio can report improving returns for years — and its interest cover will be falling the whole time, which is the tell. Two roads to 12.2% — Company A: 6.91% margin × 0.83 turnover × 2.13 leverage = 12.20% · Company B: 3.41% margin × 2.00 turnover × 1.79 leverage = 12.20% · What separates them: A is levered and slow-turning; B is efficient and lightly levered ← · Which survives a demand shock: B — its model scales; A’s debt service does not Equity can be negative. Once it is, ROE stops meaning anything at all — the ratio flips sign and a loss can print as a positive return. Use ROIC and debt ratios instead when equity is thin.

ROIC, and the comparison that decides everything

Return on invested capital measures the operating return without regard to financing. The numerator is after-tax operating profit — EBIT multiplied by one minus the tax rate — and the denominator is the capital employed: debt plus equity less cash, which is the money the business actually has tied up in operations. Because debt and equity are both in the denominator and interest is left out of the numerator, the answer is the same whether the company borrowed or issued shares, which is what makes it comparable. The purpose of the number is a single comparison: ROIC against the cost of capital. When ROIC exceeds WACC, each additional dollar of capital invested earns more than the dollar costs, so growth creates value and reinvestment is a gift. When ROIC sits below WACC, growth destroys value — every new store, factory or acquisition makes the company worth less, and the most value-creating decision management can make is to stop investing and return the cash instead. A great many mature businesses live in the second case and behave as though they live in the first, which is why F16 is about capital allocation rather than about growth. Two caveats keep the metric honest. Invested capital should be measured on a normalized basis — goodwill from an expensive acquisition inflates the denominator and makes ROIC look worse than the operating reality, and leases need consistent treatment across the companies you compare. And ROIC is a ratio of a flow to a stock, so a company mid-way through a large capital programme will report a poor ROIC for a couple of years before the assets earn anything at all — which is a timing fact, not a verdict. A rough version of ROIC that requires nothing but the statements: EBIT over (total assets less current liabilities less cash). It is close enough for ranking companies, which is how it gets used in practice.

Which margin, and what each one isolates

Margin is not one number. **Gross margin** — revenue less cost of goods sold — is the pricing power of the product, isolated from everything spent to sell it. **Operating margin** then subtracts the cost of running the business: sales, administration, research. **Net margin** subtracts interest and tax as well, which is why it is the version that mixes financing into an operating story. The gaps between them are where the analysis lives. A company with a 60% gross margin and a 4% operating margin has a wonderful product and an expensive machine around it, and the question is whether the spending is building something or leaking. A company with a 22% gross margin and a 19% operating margin is a ruthless operator in a commodity — and its thin gross margin means a three-point move in input costs is the difference between a good year and a loss. Same operating margin, opposite risks. Two forces move margins, and separating them every time is the habit worth building. **Operating leverage** means revenue growing faster than fixed costs, so margins expand as the business scales — and the same mechanism runs in reverse, which is what makes a cyclical company’s margins collapse on a modest revenue decline. **Mix** means the margin change is arithmetic rather than performance: if a high-margin division shrinks and a low-margin one grows, the blended margin falls while every part of the business executes exactly as before. Revenue up, margins down, and nothing wrong — or revenue up, margins down, and everything wrong. Only the segment disclosure tells them apart. When comparing two companies, compare like margins: gross with gross, operating with operating. And compare them at the same point in the cycle, because a peak-year and a trough-year margin for the same business can differ by more than the gap between the two companies you are ranking.

Invested capital is a definition, not a fact

Return on invested capital is the most useful return metric in this subject and it is not a single number that a company reports. Both halves are choices. On the numerator you have to decide whether to use net operating profit after tax, operating profit, or earnings, and whether to add back amortisation of acquired intangibles. On the denominator you have to decide what counts as capital: total debt plus equity, or only the operating assets, or operating assets **excluding goodwill**, and whether to average the balance over the year or take the closing figure. Those choices are not academic, because goodwill is exactly where the difference concentrates. A business bought at a large premium has a small return on *total* invested capital, since the purchase price is in the denominator, and a large return on capital **excluding goodwill** — which measures the economics of the business rather than the price the acquirer paid. Both questions are legitimate, and they answer different ones: the first is what the current owner is earning on the money actually committed, and the second is whether the business itself is a good business. Quoting one number as “the” ROIC without saying which question it answers is how two analysts reach opposite conclusions about the same company using the same filings. The comparison that matters survives the definitional noise if you are consistent. ROIC against the cost of capital decides whether growth creates value, and it has to be computed the same way across every company you compare, with the same treatment of leases, excess cash and goodwill. A more robust version for a business with a heavy acquisition history is to look at the return on incremental capital — the change in operating profit divided by the change in invested capital over several years — because it asks the forward-looking question directly and is much harder to flatter with an accounting choice. • Decide the numerator and the denominator, and say which question you are answering. • Including goodwill answers “what is the owner earning”; excluding it answers “is the business good”. • Leases, excess cash and averaging all move the denominator materially. • Return on incremental capital is harder to game and better suited to acquisitive companies. Whichever definition you pick, hold it for every company in the comparison and for every year of one company’s history. ROIC is only meaningful as a series or a comparison; a single figure with an unstated denominator is decoration.

The return on the next dollar

Return on invested capital is an average. It describes what the capital already inside the business is earning, and it is a valuable thing to know — but it is not the number that decides whether the next dollar of capital should be spent. That number is the return on the **incremental** capital, and the two can differ enormously within the same company. The incremental measure is simple to state and awkward to compute: the change in after-tax operating profit over a period divided by the change in invested capital over the same period. Both terms are noisy, so it is read over a multi-year window rather than a single year, and it is read alongside the cash-flow statement, because invested capital also moves for reasons that have nothing to do with investment — a disposal, an impairment, a lease-accounting change, a currency translation. The reason the distinction matters is arithmetic. Value is created when a company reinvests at a return above its cost of capital, and it is destroyed when it reinvests below it, no matter how fast revenue grows. A company earning a handsome average return on capital but expanding into a division with mediocre economics will show a falling incremental return long before the average gives any hint — the average is anchored by the legacy business, which is exactly the part that will not grow. The mirror case is a company with a plain average number that is redeploying capital into its best activity; the average lags while the increment leads. This is why the two numbers should always be read as a pair. High average, high incremental is a compounding machine. High average, low incremental is a business that is running out of good places to put money, and growth in it is worth less than the headline suggests. Low average, high incremental is a turnaround that has found something. Low on both is a business that should be returning capital rather than deploying it — and the capital-allocation question in this subject, buybacks versus dividends versus acquisitions, is really the question of what to do when the incremental return falls below the hurdle. There is an honest limit to how far a few years of data can take this. The incremental figure is dominated by measurement noise at short horizons, so a single year’s reading is not evidence of anything; a five- to seven-year trend is worth acting on. Where the two measures disagree, the increment usually deserves the benefit of the doubt, because it is the one that describes the company you are buying rather than the one that existed when the capital was first committed. • Average return on capital describes the past; incremental return describes the next dollar. • Value is created only when the incremental return clears the cost of capital. • Capital moves for reasons other than investment, so read the increment with the cash-flow statement. • Read the pair together: their disagreement is the most informative thing on the page. A company that will not disclose segment-level capital is telling you something about how the average is being held up. Where the returns are reported only in aggregate, the increment is the place to look for the number the aggregate conceals.

Sources

• The DuPont decomposition of return on equityStandard financial-statement analysis; Penman, "Financial Statement Analysis and Security Valuation"
• Return on invested capital and its relationship to economic profitMcKinsey, "Valuation"; Damodaran, "Investment Valuation"
• The equivalence of ROIC above WACC and value creationModigliani & Miller (1958, 1963); standard corporate-finance texts
• Why leverage flatters returns in an upcycle and destroys them in a downturnStandard leverage arithmetic; CFA Institute FSA curriculum