Why Homebuilder Land Strategies Diverged After 2008

Picture a mid-size homebuilder's balance sheet in late 2008: thousands of finished and unfinished lots, bought at the top of the cycle with borrowed money, now worth a fraction of what the company paid and costing real cash to carry every month — property taxes, HOA dues on empty streets, interest on land loans nobody would refinance. Some builders wrote off land at a pace that erased years of prior profit. A few didn't survive long enough to finish writing it off.

That memory reshaped how the industry manages the single riskiest asset on a builder's balance sheet. It split homebuilders into genuinely different business models for handling land — and the differences matter to an investor as much as backlog growth or order counts, because two builders can report similar closings and similar margins while carrying wildly different amounts of land risk, on or off the balance sheet.

I've come to think land strategy is the more useful lens here, more useful than the headline growth numbers analysts tend to lead with. It's also the lens this piece uses to sort the sector: not by who's building the most homes this quarter, but by who owns what they're building on, and what happens to each company if housing demand stalls for two years instead of two quarters.

How 2008 Rewrote the Land Playbook

Before the crash, the dominant model was straightforward: buy raw land, entitle it, develop it into finished lots, build on it, sell it — financing and warehousing every step of that chain on the builder's own balance sheet. Land was inventory, but it was also collateral, and in a rising market that leverage looked like operating discipline. It wasn't. When national home prices fell starting in 2007, builders holding land at boom-era cost marks took writedowns that in some cases dwarfed a full year's operating profit, and several builders that were household names in their regional markets did not make it through — not because they built bad homes, but because they owned too much dirt at the wrong price.

NVR had already been running a different playbook since the 1990s. Rather than buying and developing land itself, NVR acquires finished lots through what its filings call Lot Purchase Agreements — paying a deposit, typically up to 10% of the aggregate purchase price, that it forfeits if it chooses to walk away from a contract, but never carrying the underlying land on its own balance sheet while it waits. NVR's 2008 was uncomfortable. It wasn't existential. The rest of the industry took note, and the slow migration toward that model is the real story of homebuilder land strategy over the past decade and a half.

The Land-Light, the Land-Heavy, and Everyone Between

Homebuilders today sit somewhere on a spectrum between two poles, and where a given builder sits tells you more about its downside risk than its geography or price point does.

• Land-light option specialists. NVR remains the clearest example of a builder that essentially never owns land. Lennar has spent over a decade moving toward the same model, and accelerated sharply in February 2025 when it completed the spin-off of Millrose Properties — a separately managed company that now buys, entitles, and develops land, then sells finished homesites back to Lennar under option contracts with set takedown schedules. By November 30, 2025, Lennar controlled roughly 98% of its homesites through options and joint ventures rather than direct ownership.
• National-scale hybrids. D.R. Horton and PulteGroup own meaningfully less land than they control, but they haven't gone fully land-light either. D.R. Horton's land and lot portfolio totaled 639,800 lots at the end of fiscal 2025 (the year ended September 30, 2025), with 76% controlled through purchase contracts and just 24% owned outright — a deliberate middle position that keeps some pipeline certainty without carrying the full land bank.
• Land-heavy regional builders. Green Brick Partners runs the inverse strategy on purpose. As of the end of fiscal 2025, Green Brick owned 75.8% of the roughly 48,828 lots it owned or controlled outright, concentrated in Dallas-Fort Worth and Atlanta — betting that direct land ownership and development in supply-constrained submarkets is the edge itself, not a risk to be optioned away.

None of these is a strictly better model, and the tradeoff runs in both directions. Land-light frees up capital, cuts balance-sheet exposure heading into a downturn, and lets a builder flex production without eating a land writedown when it doesn't work out — but it means giving up some of the upside if that land appreciates, and it depends on a functioning market of third-party land developers willing to hold entitlement risk in exchange for an option premium. Land-heavy works when a builder has genuine, local expertise in identifying and entitling land ahead of where demand is actually going, often in fast-growing metros where finished, buildable lots are themselves the scarce resource — but it puts the balance sheet back in roughly the position that hurt the industry the last time the cycle turned.

Reading Lot Position and Gross Margin Together

Two figures buried in a builder's 10-K tell you more about its land risk than almost anything in the earnings-call script: what share of the lot pipeline is owned versus merely controlled, and which direction gross margin is moving.

Gross margin on home sales looks like a plain profitability number. For a homebuilder it's also a land-cost signal, and the two readings can point in different directions in the same quarter. D.R. Horton's home-sales gross margin fell to 21.5% in fiscal 2025 from 23.5% in fiscal 2024 — a decline the company attributed mainly to heavier sales incentives, particularly mortgage-rate buydowns, needed to keep homes moving as affordability worsened. That's a demand-side story this cycle more than a land-cost story. But the underlying mechanism is the one worth remembering: a builder sitting on land bought at a prior cycle's prices can defend margin through a soft patch in ways a builder still paying current market prices for every fresh lot option cannot.

Lot position tells the other half of the story, and it's worth reading against gross margin as a quality signal rather than in isolation. A builder with a high owned share, like Green Brick, has locked in land cost and doesn't need to renegotiate option terms every quarter — but it also can't shrink its land exposure quickly if demand craters. A builder with a high controlled share, like D.R. Horton or NVR, can walk away from an option contract and forfeit only the deposit, which effectively caps the downside on any single parcel. That asymmetry is the entire reason the land-light model spread across the industry after 2008: it converts a fixed liability into a series of small, capped bets, each one individually survivable.

Where the Land-Light Model Could Break

The land-light model isn't a free lunch, and the honest version of this piece has to say where it could fail. It depends on third-party land developers and financiers willing to hold entitlement and development risk in exchange for a builder's option premium — a market that itself tightens in a credit downturn, when the entities financing that land inventory can't roll their own debt as easily.

Millrose Properties is a useful test case precisely because it's new. It intends to qualify and be taxed as a REIT, which means it needs steady income from the option fees and land sales it collects from Lennar — and eventually, the company says, from other builders too — a structure that hasn't yet been tested through a full housing downturn. I'm less certain than I'd like to be about how cleanly that model holds up if home sales slow sharply and Lennar's contracted takedown pace slows with it. The whole appeal of a spin-off like this is that Lennar's own balance sheet no longer absorbs that land risk. Somebody's balance sheet still does, and it's worth knowing whose before assuming the risk has simply vanished.

There's a second, quieter risk, and it cuts the other way from the 2008 lesson: land-light builders are trading balance-sheet risk for supply risk. Where entitled, finished lots are scarce — already the case in several fast-growing metros — a builder that never built its own land-development pipeline has less control over where its next few years of lots come from than a builder like Green Brick that entitles and develops its own dirt on purpose. A cyclical stock doesn't stop being cyclical just because the balance-sheet risk moved to someone else's books; it just shows up as a different constraint at a different point in the cycle.

It's a structural echo of a lesson from a very different capital-intensive sector: telecom's capital-intensity cycle made a version of this same point a few months back — that where the fixed cost sits changes who bears the pain, but it rarely makes the pain disappear. Homebuilding's version of that fixed cost is land, and the housing-demand backdrop discussed in how housing markets affect stocks still ultimately decides who has to absorb it.

Practically, that means comparing land position before comparing headline growth. Two builders posting similar order growth this quarter can be underwriting very different bets — one leaning on a persistent secular tailwind in housing demand, one leaning on land bought cheaply years ago, one on execution discipline inside option contracts it's free to walk away from. Ask what share of a builder's pipeline is owned outright versus merely controlled. Ask what its deposits look like relative to total land spend. Ask whether gross margin is holding up because of pricing power or because of a favorable cost basis on land bought a cycle ago. The answer changes what you're actually underwriting when you buy the stock, even if the reported growth number looks identical across two names.

Key Takeaways

• Homebuilder land strategy split into distinct models after the 2008 crash exposed how dangerous direct land ownership is with leverage: land-light option specialists (NVR, and increasingly Lennar via its Millrose Properties spin-off), national-scale hybrids (D.R. Horton, PulteGroup), and land-heavy regional builders (Green Brick Partners) that hold land on purpose.
• Lennar controlled roughly 98% of its homesites through options and joint ventures by November 30, 2025; D.R. Horton controlled 76% of its 639,800-lot pipeline through contracts rather than ownership at the end of fiscal 2025; Green Brick owned 75.8% of its roughly 48,828 lots outright over the same period — three genuinely different risk postures on the same underlying asset.
• Gross margin and lot-ownership share should be read together, not separately: D.R. Horton's home-sales gross margin fell to 21.5% in fiscal 2025 from 23.5% in fiscal 2024 on heavier incentives, a reminder that even a disciplined land strategy doesn't insulate margin from a softening demand backdrop.
• The land-light model shifts balance-sheet risk onto land developers and financiers rather than eliminating it — Millrose Properties' still-unproven REIT structure is the clearest live test of whether that shift holds up through a real downturn.