Growth, Reinvestment and Moats

Growth is a decision, not a destiny

Growth = reinvestment rate × return on invested capital. The reinvestment rate is the share of earnings put back into the business rather than distributed; the return on those assets is F6’s ROIC. Multiplied, they give the sustainable growth rate from internal funds, and the identity is worth internalising because it says growth is a choice about capital rather than a property of an industry. It also explains why two companies with identical growth rates can be worth very different amounts. Forty percent growth produced by reinvesting everything at a 25% return is a compounder. Forty percent growth produced by reinvesting everything at a 5% return is a business buying revenue with capital — and the tell is that the second company’s free cash flow is negative for years while its revenue rises, and its returns on capital fall as it gets bigger. Growth always shows up in the statements; whether it shows up in value depends on the return. There is a third term the identity hides: the duration of the excess return. A return of 25% that persists four years is worth far less than the same 25% that persists twenty, and the only thing that makes it persist is a structure that keeps competitors from bidding the return away. That structure is what the word moat means, and it is the subject of the second half of this lesson. The identity, on two companies — Company A: reinvestment 60%, ROIC 25%: Growth 15%, and each dollar adds $0.16 of value · Company B: reinvestment 60%, ROIC 7%: Growth 4.2%, and each dollar destroys $0.02 · Company A at the same growth as B: A can grow 4.2% with a 17% reinvestment rate and distribute the rest · What that means: The high-return business pays you while it grows; the low-return one charges you for it ← Acquisition growth hides here. Paying a full price for a competitor’s earnings is deployment of capital at whatever return the purchase price implies — often below the cost of capital, and always invisible in a revenue chart.

A moat is a structure, and there are only a few kinds

A moat is not being good at what you do. It is a reason a well-funded, competent competitor cannot take your returns away. The discipline in that sentence matters, because almost every company claims durable advantages and the return history of most of them shows the advantage was a period, not a structure. The framework that has survived best proposes five sources, and a claimed advantage that is not one of them is usually a temporary gap. Supply-side scale economies: the incumbent’s cost per unit is lower because it is bigger, and it can profitably cut prices to a level a new entrant cannot match while it builds. This is a real moat where fixed costs dominate, and it is why so many industrial markets are oligopolies. Demand-side scale, or customer captivity: switching is expensive, slow or risky — entrenched software, a bank relationship, an aircraft part that requires re-certification. Here the moat is the cost of changing, not the price. Then there are three that get claimed far more often than they exist. Brands are only a moat where they produce a pricing premium a competitor cannot replicate by spending more on marketing, which is true for a few consumer categories and false for most. Regulation is a moat while it lasts, and it is a political asset that can be repriced. And the most common claim of all — a network effect — needs the network to become more valuable to each user as it grows, not merely larger; a network that is expensive to join is customer captivity instead. The test that separates a moat from a good run: if a competitor with unlimited capital entered tomorrow, what specifically would stop them? If the answer is a named mechanism — plant economics, switching cost, a licence, a certified part — you are describing a structure. If the answer is an adjective, you are describing a mood.

How a moat shows up in the numbers

A durable advantage leaves fingerprints. Returns on capital stay above the cost of capital for a decade or longer, and they do not fade toward the average the way most profitability does. Gross margins are stable while competitors enter, because the incumbent is not forced to discount. And capital intensity stays low relative to growth: a business with a real demand-side moat grows revenue without spending proportionally more on assets to defend it. The inverse fingerprints are just as informative. Falling returns on capital as the company grows is the signature of a strategy without a moat: the easy customers were taken first, and each additional dollar of growth requires deeper discounting or heavier assets. Rising receivables days or falling gross margins right after a competitor announces capacity are the same signal arriving early. One warning belongs with the concept. A moat is measured against a specific competitor set, so it is only as stable as that set. Newspapers had an impregnable local monopoly on classified advertising until the internet made a different market the relevant one; retail banks had a switching-cost moat until mobile banking lowered it. The moat did not disappear because a direct competitor beat it; it disappeared because the definition of the competitor changed, which is why the right question about a moat is not whether it is strong but what could make it irrelevant. A high return on capital with high reinvestment is a compounder. A high return on capital with no reinvestment opportunity is a cash machine. Confusing the two is how investors overpay for the second, and how they underpay for the first.

How long a moat lasts, and the two questions that test it

The empirical literature on returns on capital is unkind to the idea of a permanent advantage. Across large samples, high returns on invested capital **decay toward the cost of capital** over periods of roughly five to fifteen years, and the decay is faster the higher the return is to start with. That is what competition is supposed to do, and it means a valuation that assumes today’s excess return persists for thirty years is assuming something about competitive dynamics rather than something in the financial statements. The practical form of the assumption is the forecast horizon: the explicit years where you let the return stay elevated are your estimate of how long the structure holds, and they should be a number you chose deliberately rather than the default of a model. Two questions do most of the work of testing durability, and neither requires a valuation. The first is **what would a well-funded, competent competitor do with five years and unlimited capital?** If the answer is that they could replicate the product and the customers would switch, there is no structure — there is a head start. If the answer is that they would need to build something that can only be built once (a network with both sides already on it, a distribution system in a country with limited routes, a brand that took decades of advertising to build), there is a structure. The second is **what does the company itself do with the cash the advantage produces?** A moat that generates cash which is reinvested at high returns compounds; the same cash paid out is still a good business, but it is a bond with equity risk rather than a compounding machine. That second question is where moat analysis meets the growth identity in this lesson, and where the most common analytical failure sits. Businesses are often described as having a moat when what they have is a **high return on existing capital and nowhere to put new capital** — a wonderful asset to own and a poor one to compound, because growth can only come from acquisitions or from reinvestment at lower returns. The distinction between a moat that can be extended and one that is merely present decides whether the value of the business is mostly its current earnings or mostly the reinvestment of them, and those two cases deserve very different multiples for the same return on capital. • Returns on capital decay toward the cost of capital over roughly five to fifteen years in large samples. • The forecast horizon is your explicit estimate of how long the advantage lasts. • Test durability by asking what a well-funded competitor could replicate in five years. • Separate a moat that can be extended from one that is merely present — the second is a bond with equity risk.

The capital cycle: how a moat is competed away

A moat is easier to understand as a process than as a list of sources, because the thing that destroys most advantages is not a rival’s brilliance but the ordinary behaviour of capital. The mechanism runs as follows. A business earns a return on capital well above its cost, which attracts capital — from competitors building capacity, from adjacent firms entering, and from customers deciding to do it themselves. Capacity takes years to build, so it arrives after the demand that justified it, the industry ends up with a surplus, and prices fall until the marginal producer earns nothing; then capital leaves, the surplus clears, and the cycle begins again. The **capital cycle** is the reason a high return on capital is a temporary condition in most industries unless something is actively blocking the flow, and it is the frame in which a moat should be judged: not “does this business earn a lot today”, but “what stops the next dollar of capital from doing to it what capital did to the last industry that earned a lot”. Only a few things block that flow, and they are narrower than the list of claimed advantages. A structural cost advantage — scale that cannot be replicated, a resource that only exists in one place, a location that cannot be duplicated — blocks it because the entrant’s economics are worse from day one. A switching cost blocks it because the customer’s cost of leaving exceeds the price difference, so the entrant has to pay the customer to move. A regulatory licence blocks it because the entrant is not permitted. A network blocks it because the entrant’s product is worth less to each user until the users are already there, which is the definition of a trap. Everything else — a good brand, superior execution, a first-mover head start — is a lead rather than a moat, and leads are what capital closes. The test that separates them is one sentence: who is the marginal new entrant, and what would it cost them to be as good as this company? If the answer is “they have already tried and it was unprofitable”, the advantage is real; if it is “nobody has bothered yet”, the question is still open. What kills a moat, once it exists, tends to arrive through one of four doors. Technology changes the customer’s alternative, which is the mechanism the previous read described as the competitor set changing, and the rate matters: a moat that assumed customers needed a physical intermediary is worth nothing when the intermediary becomes a website. Regulation removes it from outside — success invites scrutiny, and an advantage that shows up as a visible price to consumers is the kind that gets legislated away. Capital inflows erode it from inside, which is the cycle above running at the industry’s pace. And customer concentration ends it abruptly, because a moat built around serving one client is not the company’s moat at all, it is the client’s convenience. The empirical backdrop is worth keeping alongside those doors: measured returns on capital mean-revert, and the fraction of companies that sustain a high return on capital for two decades is small and falls as the horizon lengthens. That is the prior any forecast of a durable advantage should beat, and when a model assumes twenty years of above-cost returns, the burden is on the analyst to name which door is locked. • Capital flows toward high returns, builds capacity, and the surplus competes the return back down. • Only scale, switching costs, licences and true network effects actually block that flow. • The test: who is the marginal entrant, and what would it cost them to match this business? • The four doors: technology changing the alternative, regulation, capital inflows, customer concentration. • Returns on capital mean-revert, so a durable advantage has to beat a documented prior. The capital cycle also gives a second, cheaper way to look for the next broken moat: watch where the industry is spending. Collective capex in an industry earning high returns is the surplus arriving, four years before the prices fall — which is the same signal the sector lesson reads as the commodity price’s best predictor.

Sources

• The five sources of durable advantage, and why most claimed moats are not oneBruce Greenwald, "Competition Demystified" (2005)
• Returns on capital and the persistence of profitabilityStandard competitive-analysis literature; Porter, "Competitive Advantage"
• The tension between competitive markets and persistent excess returnsModigliani & Miller, and the empirical literature on profitability persistence