Judging Moat Quality Without a Valuation Model

Two businesses land on the same screener on the same afternoon. One trades near 45 times earnings, charges an annual membership fee that almost nobody strictly needs to pay, and still sees more than nine in ten members renew it anyway. The other trades at 8 times earnings, has missed same-store-sales estimates for six straight quarters, and every ratio on the page says it's a bargain. Feed both into a discounted cash flow model using the same discount rate and the same terminal growth assumption, and the spreadsheet will often say the second business is the better buy. The spreadsheet isn't lying. It's just answering a question nobody actually asked it.

What would it take to displace either business from where it sits today? That's the real question, and no valuation model — however carefully built — was designed to answer it. A DCF takes moat quality as an input, usually buried inside a growth-rate assumption or a margin forecast several tabs deep in the workbook. It doesn't independently test whether that input deserves to be there. Confuse the two jobs — judging the moat and judging the price — and you'll misprice both, consistently, in the same direction.

This is the single most common analytical error among otherwise careful investors: treating a low multiple as evidence of quality and a high one as evidence of its absence. A multiple describes what the market currently believes about a business. On its own, it says almost nothing about whether that belief is correct. Separating those two questions — is the moat real, and is the current price fair given that — is what this post is actually about. The second question belongs to fair value work. The first one is the one a spreadsheet can't do for you.

Why the Valuation Model Was Never Built for This Job

A discounted cash flow model is a machine for converting assumptions into a number. Feed it a growth rate, a margin trajectory, a discount rate, and a terminal multiple, and it will return a fair value estimate with total confidence — regardless of whether the underlying assumptions reflect a business with a genuine structural advantage or one running on fumes. The model doesn't know the difference. It just does the arithmetic on whatever you tell it.

This isn't a flaw in DCF methodology. It's a description of what the tool is for. A valuation model exists to translate a view of the future into a present value, given that you already hold a view. It was never built to generate that view in the first place, and treating its output as a substitute for competitive analysis gets the order of operations backwards. You don't discover whether a moat is durable by discounting cash flows. You decide whether cash flows are worth discounting by first establishing whether the moat is durable — and that step happens before the spreadsheet ever opens.

The practical failure mode shows up constantly in stock screeners. A screen built around low P/E, low EV/EBITDA, or high free cash flow yield will reliably surface businesses whose moats are eroding, because a market that's correctly pricing in structural decline produces exactly those numbers. It will just as reliably miss businesses whose durable competitive advantage commands a persistent premium, because 'expensive relative to trailing earnings' and 'overvalued' are not the same claim, and a mechanical screen can't tell them apart. Cheap and hollow can look identical to expensive and justified. So the screen is the wrong tool for the first question, even when it's the right tool for the second one.

What Moat Quality Looks Like Before Any Multiple Is Applied

Moat quality is legible without a model. It shows up in customer behavior, in pricing power exercised without volume loss, in the friction a competitor would face trying to rebuild the position from scratch. None of that requires a discount rate. No spreadsheet required.

Costco: An Advantage Wide Enough to Explain the Premium

Costco has traded at a persistent premium to the broader retail sector for well over a decade, and the multiple alone tells you nothing about whether that premium is earned. The renewal data does. In its fiscal 2024 10-K (fiscal year ended September 1, 2024), Costco reported a membership renewal rate of 92.9% in the U.S. and Canada and 90.5% worldwide, across a base that had grown to nearly 137 million cardholders. Membership fee income rose 5% that year to roughly $4.8 billion — a figure worth sitting with, because it arrived alongside total operating income of about $9.3 billion against net sales of $249.6 billion, an operating margin of well under 4%. Costco runs its merchandise business close to breakeven and earns the bulk of its profit from a fee that members pay year after year for the right to keep shopping there.

That's the structural tell. A business that can charge an annual toll for access to itself, hold that toll's renewal rate above 90% through recessions and inflation spikes alike, and still keep merchandise prices aggressive enough that the toll feels worth paying — that's a wide moat by any reasonable definition, built on switching costs and scale-driven cost advantage reinforcing each other. The 45-times-earnings multiple from the opening example is a fair description of what a market that has actually done this analysis is willing to pay. Whether it's the right price today is a separate question, and one this post deliberately isn't answering. That's fair value territory, not moat-quality territory.

Sears: Cheap on Every Screen, Structurally Hollow

Sears Holdings spent much of the decade before its October 2018 Chapter 11 filing trading at multiples that looked, on paper, like a value investor's dream — low price-to-book, low price-to-sales, a real estate portfolio bulls argued was worth more than the entire market capitalization. None of those numbers were fabricated. Not even unreasonable as arithmetic. What they missed was that the retail moat had already dissolved: store traffic was declining structurally, not cyclically, and the 'value' in the real estate required either a sale-leaseback or a liquidation to realize, neither of which does anything for a shareholder banking on the operating business recovering.

This is the no-moat case, and it's worth naming explicitly because it's the mirror image of Costco. A screen looking for statistically cheap stocks would have flagged Sears for years leading up to the filing. A moat-quality assessment — asking what kept customers coming back, and finding the honest answer was 'increasingly, nothing' — would have pointed the opposite direction much earlier. The valuation looked attractive precisely because the market had, correctly, priced in the erosion. Cheap wasn't a signal of undervaluation. It was the market doing its job.

Where the Valuation-Model Habit Misleads You Specifically

Three places this goes wrong most often when investors work through names.

The first is treating 'expensive' as synonymous with 'overvalued.' A business with genuine pricing power and a growing renewal base can sustain a premium multiple for a very long time, and investors who screen it out on P/E alone are filtering for cheapness rather than for quality. The second is the mirror error: treating 'statistically cheap' as evidence of a bargain rather than as a market signal that the underlying business has already weakened. Both mistakes come from the same source — asking a valuation ratio to answer a question about durability, which it was never built to answer.

The third failure is subtler, and it's the one I trust myself least on: building a DCF first, then reverse-engineering a moat story to justify whatever growth rate makes the output attractive. It's an easy trap. The model rewards optimism with a bigger number, and a bigger number feels like confirmation rather than what it usually is — a symptom. The discipline that actually works is sequencing the two tasks so the moat conclusion gets reached independently, before a single growth assumption is typed into a cell. Otherwise the 'analysis' is motivated reasoning wearing a spreadsheet as a costume.

Judge the Moat First, Then Call the Price

None of this argues against valuation work. Same discipline, two separate jobs. It argues for doing them in the right order and refusing to let either one stand in for the other. Ask the structural question first: what would it take for a well-capitalized competitor to replicate this position, and how has the business actually behaved under stress in the past. Answer that using renewal rates, pricing behavior exercised without volume loss, customer concentration, and the plain evidence of whether growth requires proportional capital or comes nearly free. Only then does it make sense to open a valuation model and ask what that position is worth at today's price — a question that belongs, in more depth than this post attempts, to the discipline James Whitfield covers when he walks through what a business is actually worth.

Get the sequence backwards, and the multiple starts doing work it was never suited for — standing in as a proxy for quality it can't actually measure. Get it right, and a high multiple on a durable business and a low multiple on a decaying one stop looking like opposite ends of a single 'value' spectrum. They start looking like exactly what they are: two markets, doing their jobs correctly, pricing two very different competitive realities.

This framing deserves some pushback, because the two-step process makes this sound cleaner than it usually plays out. Moat judgments and price judgments inform each other more than a strict ordering implies — a business trading at an unusually rich multiple for a decade is itself weak evidence that the market has already done a version of this work for you. But treating that consensus as a substitute for doing the structural analysis yourself is exactly the shortcut this post is arguing against.

Key Takeaways

• Valuation multiples describe current market belief about a business, not competitive durability. A DCF takes moat quality as an input; it doesn't independently verify it.
• Costco's fiscal 2024 10-K shows the structural tell of a wide moat: 92.9% U.S./Canada renewal (90.5% worldwide) on nearly 137 million cardholders, funding the bulk of roughly $9.3 billion in operating income against a merchandise business run near breakeven.
• Sears Holdings traded at statistically cheap multiples for years before its October 2018 bankruptcy filing — cheap wasn't mispricing, it was the market correctly reflecting a moat that had already eroded.
• Judge the competitive position first, independent of price. Only then is it useful to ask what that position is worth — a separate discipline with its own tools.