Start with the mechanism. The metaphor can wait. Suppose a business earns 30% on invested capital while its capital costs 12%. Every rupee it reinvests creates ₹2.5 of value, by the M3.04 value-driver logic. Now ask who else can see this. Everyone can. Competitors read the same filings you do, and capital is mobile, ambitious, and cheap for anyone with a credible plan. So new capacity enters. Prices get sharpened, customers get poached, talent gets hired away, and the 30% grinds toward 12%. In a competitive economy, the equilibrium return on capital is the cost of capital. Excess return is a disequilibrium, and markets abhor a disequilibrium.
The data says exactly this. Mauboussin's persistence studies (and McKinsey's, in Valuation) sort companies into ROIC quintiles and track them: the top quintile fades toward the middle over 5–10 years, the bottom quintile rises (bankruptcy and exit cull the worst), and the spread compresses. That is the classic fade. But the fade is not total: a minority of firms sit in the top quintile decade after decade. Those survivors are not lucky repeatedly; they own something structural that entry cannot copy or undercut. That something is the moat.
Advertisement
How is an economic moat defined?
So define it tightly. Below is the course's own working definition. We assembled it from the Morningstar and Mauboussin traditions, and it isn't quoted from either:
(our working definition) A moat is a structural characteristic of a business that allows it to sustain returns on invested capital above its cost of capital despite competitors' active, well-funded efforts to take those returns away.
Take it a phrase at a time:
- Structural: it belongs to the business, and the people don't own it. Management can dig a moat or fill one in. But if the advantage evaporates when a competent rival simply tries harder, it was never a moat. Morningstar's habit is worth copying here: rate the moat as if management were average.
- Sustain … above cost of capital: the moat's output is the persistent ROIC−WACC spread. No spread, no moat, whatever the story.
- Despite competitors' efforts: a high return no one has attacked yet is untested. Nobody has had to defend it. So don't ask "is anyone competing?" Ask "what happens to returns when they do?"
How do you recognise a moat, and why does it matter?
One corollary organizes everything that follows. The moat is the cause, and persistent excess returns are the evidence. You can never observe a moat directly. You observe two things: a plausible structural mechanism, and a numbers trail consistent with it. A claim without numbers is a story. Numbers without a mechanism are probably a cycle. You need both.
The valuation link (why you're really here). You've seen the relationship already: value = f(spread, reinvestment, duration). Take two companies with identical 25% ROICs and identical growth. If one's spread dies in 3 years and the other's survives 20, they're worth wildly different amounts. The moat is the duration variable. So moat analysis isn't a strategy elective. It's the third input to every DCF you'll ever build, and you'll soon meet its formal name, the Competitive Advantage Period (CAP). Unwind a reverse-DCF on a richly priced compounder and you'll usually find a market opinion about moat duration.
Your edge, if you have one, will often be a better-evidenced opinion on that same question.
What are the five sources of a moat?
The quick-test (memorize now, drill in Phase 8). Five sources, intangibles/brand · switching costs · network effects · cost advantages · efficient scale, each confirmed by high, stable-or-rising ROIC/ROCE across a full cycle. Beware the false moats: great product, hot management, high share, good execution. All of them erode. That's the ten-minute version, the one you'll run against a stopwatch.
Next comes the slow version, with the evidence.
A note on the taxonomy's pedigree. Pat Dorsey's The Little Book That Builds Wealth (2008) named four sources: intangible assets, switching costs, the network effect, and cost advantages. Morningstar's methodology later added efficient scale as a fifth. We teach all five, and the profession now treats "the five sources" as standard. In real companies they overlap. A rating agency is a license (intangible) and efficient scale. A payment network is a network effect and scale economics. Don't chase taxonomy purity. Name the primary mechanism, the one holding the moat up, because that tells you what evidence to check and what would kill it.
| Source | Mechanism in one line | Best single piece of evidence | Characteristic failure mode |
|---|---|---|---|
| Intangibles | Brand, patent, or license lets you charge more (or be chosen by default) for an equivalent product | Price increases stick: gross margin stable through input-cost cycles | Brand becomes merely familiar; patent cliffs; license repriced by the regulator |
| Switching costs | Leaving costs the customer more (money, time, risk) than staying | Retention/renewal ≥95%; revenue per client rises with tenure | Technology shift resets everyone to zero; a generational platform change |
| Network effects | Each additional user makes the service more valuable to every other user | Share and take-rate stable or rising simultaneously; liquidity metrics deepen | Multi-homing; the network caps out locally; regulation builds public rails (UPI) |
| Cost advantages | Structurally lower cost per unit at equal quality: scale, process, location/resource | Cost or price per unit visibly below peers' with equal-or-better ROIC | Peers replicate the process; scale advantage arbitraged by a bigger entrant; input cost shifts |
| Efficient scale | Market is only big enough for one or two efficient players; entry would destroy returns for all, so rational rivals stay out | Decades with no entry despite visible excess returns in a small market | Market grows large enough to invite entry; regulator forces access or caps prices |