D. R. Horton: A Wide Moat Built on Endurance, Priced for Perfection
D. R. Horton: A Wide Moat Built on Endurance, Priced for Perfection
D. R. Horton is the largest homebuilder in the United States, and it presents an instructive exercise: a company that passes three of our four gates cleanly, and fails the fourth for the most classic of reasons. The business is knowable, the moat is real and widening, and management does nearly everything right with capital. The price simply refuses to cooperate. Here is how the judgment unfolds, gate by gate.
Gate 1: Circle of Competence
The business model can be stated in a paragraph, which is itself a good sign. Horton builds and sells entry-level homes at scale across 126 markets in 36 states. Rather than owning land outright, it controls lots through option contracts and through Forestar, its affiliated land developer — an asset-light, “lot-light” structure. Its subsidiaries, DHI Mortgage and title services, attach to 81% of closed homes, extending earnings per transaction, and mortgages are sold to third parties shortly after origination, so interest-rate risk is not held for long. Most of the profit comes from home sale margins, and the source of those margins is scale: purchasing power and fast construction turns.
Information access passes without difficulty. SEC filings, earnings releases, and housing starts and sales statistics are sufficient to reconstruct the business and its competitive position.
Predictability passes conditionally, and the condition matters. The structural case — that the U.S. will likely remain undersupplied in housing a decade from now, and that Horton will likely remain the largest builder focused on entry-level homes — is an analytical argument, and a reasonable one. But any specific year’s earnings are hostage to rates and the housing cycle. The durability and position of the business are predictable; the path of its profits is not. That duality must be carried forward: any valuation must use cycle-normalized earnings, not the current print. With that boundary condition attached, the company is judged inside our circle.
Gate 2: Moat
The moat is cost advantage plus efficient scale. As the largest builder, Horton buys materials and subcontractor labor cheaper, turns inventory faster, and spreads costs across a national footprint. The lot option and Forestar structure raises capital efficiency and reduces inventory-loss exposure when the cycle turns down.
Now the uncomfortable part: there is no evidence of pricing power — rather the opposite. Homes are commodities, and in Q4 FY2025, 73% of buyers received a mortgage rate buydown (to around 3.99%). That is a price cut wearing a nicer name. This is not a See’s Candies moat; it is a GEICO moat. The essence of the advantage is not the ability to charge more but the ability to endure longer. When an incentive war breaks out in the entry-level segment, the builder with the lowest costs and deepest pockets outlasts everyone else. That builder is Horton.
The moat is also widening, slowly. Large public builders have been gaining share structurally — their financing, land access, and capacity to fund incentives simply overwhelm smaller operators. The fact that Horton can afford rate buydowns through a rate-lock malaise is itself functioning as a barrier to entry.
On the AI question: neutral. Construction is not a domain where AI collapses marginal cost. Design, land selection, and process management may gain modest efficiency, but the moat’s substance — scale, land, and construction networks — lives in the physical world and faces little erosion risk.
Gate 3: Management
The capital allocation record is excellent. In FY2025 Horton repurchased $4.3 billion of stock (30.7 million shares), cutting shares outstanding by 9% year over year (to 294.5 million as of September 30, 2025). It paid $495 million in dividends and raised the quarterly payout to $0.45, a 13% increase. FY2026 guidance calls for roughly $2.5 billion more in buybacks plus around $500 million in dividends. There is no history of empire-building M&A. Critically, the repurchases have continued while the stock has been pressed lower during a margin downcycle — buying below intrinsic value when earnings are compressed is precisely the behavior one wants to see.
The retained-earnings test passes. Net income grew from $2.4 billion in FY2020 to $3.6 billion in FY2025, and with the share count sharply reduced over that period, per-share value growth is even larger.
On candor: management discloses the incentive escalation and margin compression in plain numbers, stating plainly that incentives will “remain elevated through fiscal 2026.” The balance sheet is run conservatively — debt-to-total-capital of 19.8% and $6.6 billion in liquidity — without optimistic packaging. Pass.
Gate 4: Price
Here is where the thesis stops. Reported owner earnings are about $3.55 billion, close to FY2025 net income of roughly $3.6 billion — a fair approximation, since the lot-light model keeps maintenance capex light. That yields an 8.3% owner-earnings yield against a 4.3% ten-year Treasury. At first glance, cheap.
But cycle adjustment is the whole game. Earnings have already fallen from $5.9 billion in FY2022 (ROE 34.5%, cycle peak) to $3.6 billion in FY2025 (21.5% margin), and the decline is ongoing — Q2 FY2026 home sales gross margin was 20.1%, with net income down 20% year over year. Current earnings are neither peak nor confirmed trough. And this is an industry that posted three consecutive years of large losses in 2007–2009, including a $2.6 billion loss in FY2008.
Normalized owner earnings, applying a conservative 8–9% net margin to roughly $34 billion of revenue, come to $2.7–3.0 billion. At 12 times that — low growth, and no 15x allowed for a cyclical — intrinsic value is $32–36 billion, or roughly $110–123 per share at the current share count.
The market disagrees. At around $157 in early July 2026 (market cap roughly $46 billion; $43 billion at the metrics date), the stock trades at a premium to even the top of the intrinsic value range — no margin of safety. A 14x PE on current earnings looks cheap, but that is the classic cyclical illusion: low multiples on falling earnings. Applying a 30% discount requirement gives a target buy price of $77–86 per share (a $22–25 billion market cap). Even the 52-week low of $129 does not reach it. Gate 4 fails.
Verdict
Watch. Gates 1 through 3 pass: a comprehensible business (with a cyclical caveat), a narrow but widening cost-and-scale moat, and exemplary capital allocation. The failure is Gate 4 — the 14x PE and 8.3% owner-earnings yield are unadjusted cyclical figures, and against cycle-normalized earnings of $2.7–3.0 billion, the market is paying a premium. Earnings have already fallen 40% from the FY2022 peak, yet margins are still declining and the bottom is unconfirmed. The buy zone begins near $85 or below.
What would change this verdict: sustained home sales gross margins above 22% without rate declines would suggest structurally higher normalized margins and force a re-estimate of intrinsic value upward — or, on the downside, a fall in price toward the $77–86 range with buybacks continuing, at which point the margin of safety would finally exist.
This analysis is AI-generated, educational, and not investment advice. Figures may contain errors or be delayed. Disclaimer