Essay  ·  Return on Capital

A closer look at
return on capital.

Return on invested capital is the single most important number in corporate finance. Getting it right takes more than a glance at the financial statements.

By Marques Blank  ·  Eight-minute read

→  Good morning

Return on invested capital is the single most important number in corporate finance. It is also one of the easiest to get wrong. Both things are true at the same time, and the gap between them is where a lot of good and bad investment decisions get made.

Almost every framework in corporate finance comes back to the same question. Does the company earn more on the capital it deploys than the capital costs to raise? If the answer is yes and the gap is durable, value compounds. If the answer is no, value erodes no matter how fast revenue grows. ROIC is the number that settles the question.

The catch is that the ROIC you see in financial statements, and the one quoted in most research reports, often answers a slightly different question. It measures average returns on historical capital rather than marginal returns on new capital. It treats intangible investment as an expense instead of capital. It does not account for the mean reversion that academic researchers have documented for decades. The headline number is usually a starting point, not an answer.

A number behind every good business.

Michael Mauboussin and Dan Callahan, writing at Morgan Stanley's Counterpoint Global Insights in 2022, published the most practical framework for thinking about ROIC available outside the academic literature. Their empirical work across the Russell 3000 from 1990 to 2022 shows that enterprise value tracks cleanly with the spread between ROIC and the cost of capital. The market recognizes and prices high-ROIC businesses. The same paper is careful to note that ROIC itself regresses strongly toward the mean over time, which means identifying where a company's ROIC is heading next is usually more valuable than knowing where it is now.

The academic research supports the intuition. Robert Novy-Marx published a paper in 2013 in the Journal of Financial Economics titled “The other side of value: The gross profitability premium.” He showed that gross profits scaled by total assets, a close cousin of ROIC in the family of profitability ratios, predicts the cross-section of U.S. equity returns with roughly the same power as book-to-market, the classic value factor. The effect survived standard risk-factor controls. Companies with high profitability on the capital they deploy have, on average, rewarded investors beyond what conventional valuation metrics alone would predict.

Joel Greenblatt formalized a version of this idea for retail investors in his 2005 book The Little Book That Beats the Market. His “Magic Formula” combined high returns on capital with high earnings yield, in other words cheap valuations, as a stock-selection screen. Academic researchers replicated the approach with mixed results, but the core insight, that high-ROIC businesses purchased at reasonable valuations tend to outperform, held up in most samples. The framework is simple enough to explain in two sentences and robust enough to survive three decades of testing.

Three things the headline number leaves out.

The ROIC most investors calculate has three blind spots that make the reported number a poor guide to the underlying economics.

Intangibles

GAAP requires that research and development be expensed as incurred rather than capitalized as an asset. For a software company, a pharmaceutical company, or a consumer brand that invests heavily in marketing, the reported invested capital denominator is dramatically understated. A company that has spent ten billion dollars over a decade building a drug pipeline or a brand shows almost none of that spending on the balance sheet. The ROIC calculation divides operating profit by an invested capital figure that excludes the largest actual investment the company has made. The result is a ROIC that looks spectacular and bears little relationship to the economic return on the company's true capital base.

Off-balance-sheet obligations

Even after ASC 842 brought most leases onto the balance sheet, meaningful off-balance-sheet exposures remain in the form of unfunded pension obligations, certain supplier commitments, and structured financing. Aswath Damodaran has written extensively about how these items distort ROIC calculations for capital-intensive businesses. The reported number understates invested capital, overstates returns, and makes businesses that rent their capital look artificially superior to businesses that own it.

Average vs. marginal

A company's reported ROIC is an average across its entire installed base of historical capital. Some of that capital was deployed ten years ago in businesses that still earn attractive returns. Some of it was deployed last quarter in new projects whose returns have not yet materialized. The blended average tells you what the company has done. It does not tell you what the company is doing right now with the marginal dollar of capital. This is the most important of the three, and the hardest to correct for from published financials.

The Gap

40% → 15%

A company with 40% average ROIC may be deploying new capital at 15% and still look like a high-ROIC business for years before the average catches up. The stock tends to keep its premium multiple in the meantime. Investors who notice the gap early tend to do well. The ones who anchor on the headline tend not to.

High returns and the pull of the average.

The most overlooked feature of ROIC is that it mean-reverts. High-ROIC businesses tend to drift toward the cost of capital over time. Low-ROIC businesses tend to drift toward it from the other direction. The rate of reversion varies by industry and by firm, but the direction is consistent across most of the documented empirical evidence.

Eugene Fama and Kenneth French documented this pattern in their 2000 Journal of Business paper, “Forecasting profitability and earnings.” They found that cross-sectional variation in profitability is strongly mean-reverting, with roughly 38 percent of any deviation from the mean erased each year in their baseline estimate. A company with unusually high returns today will, on average, have returns meaningfully closer to the average within a handful of years. A company with unusually low returns will move in the other direction over a similar window. The rate varies. The direction does not. Fama and French also documented that reversion accelerates the further a firm is from the mean, which implies that the most extreme ROICs erode the fastest.

Louis Chan, Jason Karceski, and Josef Lakonishok extended the finding in 2003 in the Journal of Finance. Their paper, “The level and persistence of growth rates,” examined whether long-term earnings growth is predictable across a broad cross-section of U.S. stocks. The answer was largely no. There is little persistence in long-term earnings growth beyond chance, and valuation ratios and analyst forecasts had limited power to identify the firms that would sustain high growth. Both the numerator and the denominator of a ROIC projection, then, tend to be less durable than long-range models assume. Projections that extrapolate current high ROIC and current high growth into the future are, in the empirical record, almost always too optimistic.

The implication for valuation is direct. A discounted cash flow model that extrapolates current high ROIC indefinitely overestimates intrinsic value. A model that assumes reversion to the cost of capital on a reasonable timeline produces a valuation that usually tracks actual long-run outcomes more closely. The companies that resist mean reversion, the rare businesses that sustain high ROIC for two or three decades, are the true compounders. Identifying them requires understanding what prevents reversion in specific cases: network effects, switching costs, cost advantages, scale economies, intangible assets. The academic literature on competitive advantage has documented each of these, and the overlap with the ROIC persistence data is not accidental.

What the research actually recommends.

Stephen Penman and Francesco Reggiani took a related angle in their 2013 Review of Accounting Studies paper, “Returns to buying earnings and book value: Accounting for growth and risk.” They showed that earnings yield and book-to-price, when read together rather than in isolation, predict future returns in a way consistent with rational compensation for the risk that expected earnings growth will not actually be realized. The practical lesson for ROIC work is that valuation metrics carrying information about both current profitability and the risk to future growth tell you more than any single headline multiple read on its own.

The practical framework that emerges from the combined literature has four components.

i.

Calculate ROIC with capitalized R&D and adjusted for operating leases, not as reported.

ii.

Distinguish marginal ROIC on recent capital deployment from average ROIC on historical capital.

iii.

Apply a reasonable mean-reversion schedule when projecting forward.

iv.

Reserve premium valuations for the narrow set of businesses whose structural advantages justify persistent high returns.

Most sell-side research skips these adjustments. Most buy-side screens use headline ROIC as reported. The gap between what the research recommends and what the industry actually does is part of why the Novy-Marx profitability premium has held up for as long as it has. The information is public and the framework is well understood. Applying it consistently is the hard part.

In Closing

ROIC is still the right variable to care about. The number on your financial data terminal, with R&D expensed and invested capital mismeasured, is where the analysis starts, not where it ends.

The gap between reported ROIC and economic ROIC is where investors with patience earn their edge. Most of the market is working off the headline.

Tomorrow the memo turns to the related measurement problem on the capital return side: stock-based compensation as hidden share issuance that silently offsets buybacks, and the math that shows why most headline buyback yields are overstated.

Marques

Chanhassen  ·  22.04.26

For informational purposes only  ·  Not investment advice