How U.S. Homebuilders Became Quality Compounders

The homebuilding business has undergone a transformational change since the Great Financial Crisis (GFC). From small traditional players with poor capital allocation and balance sheets, to cloning the business model of what has been one of the best performing stocks in U.S. history, NVR. Focusing on capital allocation and discipline has made the industry quite attractive for the next few years, especially considering the low returns the indices are poised to have given current prices.

Looking at the industry, there are several reasons why the outlook for homebuilding is more robust than the current macroeconomic might suggest including:

• Unemployment remains reasonable. Although unemployment has been increasing in the last year, it remains at reasonable historical levels. An increase in unemployment would put some pressure on the demand for new homes and there would be more price pressure coming from existing homes.
• Constrained supply of existing homes. There’s a “locked-in” effect with existing homeowners with 80% of them having mortgage rates under 6%, and half of them under 4% thanks to the low rates between 2020 and 2022, which gave the chance to refinance their mortgages. This situation makes the housing market tighter with fewer existing homes in the market as it doesn’t make economic sense to give up a 3% mortgage rate for a 6% one.
• Economies of scale from big players. Big players have bargaining power and can lower costs per home built, something local players can’t. Another product big homebuilders can offer is mortgages or incentives, which in an environment with affordability issues, can give them a competitive advantage.

One of the largest impediments to household formation and homeownership has been high prices, so going forward, it will be crucial for builders who wish to grow to provide affordable housing.

With data going back to 1968, approximately 1.4 million privately-owned housing units have been completed in the U.S. per year. However, the number of completed units has fallen below that number for the past few years.

Overall, studies forecast an annual need of 1.5-1.6 million new housing units to meet demand, give or take, around 1.2 million from family formation, 200k rebuild due to obsolescence and another 200k for second homes (vacation homes). NAREIT estimates the current shortage is around 4 million housing units given underbuilding over the past decade, especially of singlefamily units, while JP Morgan estimates the shortage at around 3 million housing units and other studies like the one from Freddie Mac, estimate the shortage at around 2-4 million. The reality is that nobody really knows the right number but it is pretty obvious there is a shortage. To close the housing deficit, homebuilders would need to build around 1.8-2 million units, which none of the big players believe can be done. Even suppliers like Sherwin Williams and similar companies have commented that it is very unlikely that 2 million units a year can be built, due to labor and capacity.

Right now, the main problem for the demand is mortgage rates, which sit at 6.2% at the end of 2025. Lower rates would help with monthly payments and bring demand to the market, and proof of that is the incentives homebuilders have been giving in 2025 like mortgage buydowns. Another example could be quarters when mortgage rates fall, and in which homebuilders have noticed that demand picks up, and the other way around as well: quarters when mortgage rates increase, demand slows down a bit. First-time buyers are less affected by the state of rates at the time of purchase given that the opportunity cost is lower compared to someone with a mortgage at 3%. Finally, according to TOL (Toll Brothers), the average age of homes in the U.S. has increased from 25 years in 1989 to currently 40 years, which makes a newly built home relatively more attractive, especially as the cost to maintain an older home has risen with higher labor and construction material costs.

The Case of NVR and U.S. Homebuilders

Historically, homebuilders were managed in a traditional way. They bought the land, developed it, built the house and hoped for a good economic cycle so they could sell at a good price. This business model proved to be a bad one, in which the cost of capital was higher than the returns the companies could achieve, so no capital compounding was being done, and the homebuilders needed a lot of debt to finance the vast amounts of land and to develop them, which led to a typical boom-and-bust cycle sector.

NVR [NYSE: NVR] went bankrupt in 1992 after the 1990s real estate crisis. The problem was (as it usually is) too much debt on the balance sheet to acquire land and M&A operations like the Ryan Homes acquisition in 1987. The company was in a situation in which, it overpaid for assets that were decreasing in value and were financed with debt. After the Chapter 11 episode, NVR changed the business model, and decided to become an asset-light homebuilder. This new way of doing business consisted of controlling developed land lots, but this time through option agreements instead of buying the land outright. The way it works is, a homebuilder will pay a premium usually around 10% of the lot cost, but it can be even lower than that. It’s important to note that NVR will not acquire land that needs developing; it will buy developed land lots to minimize the capital employed in the operations and minimize operational risk. The price of the land is fixed, and homebuilders will use LPAs (lot purchase agreements) to buy/control land from third-party land developers. LPAs work exactly like a call option on a stock; homebuilders have the right but not the obligation to purchase the land at the price set in the contract, and a deposit or premium has to be paid. In case the homebuilder does not want to exercise the option, it will forfeit the contract and the economic loss will be the premium paid.

Besides reducing the capital employed in the operations, another focus for NVR was to reduce the risk of economic cycles. They did this by selling homes on a pre-sold basis rather than selling spec homes (homes built before were sold). This meant that cash flow would come in before NVR started building and because they didn’t have to buy land in advance, the balance sheet was debt-free. NVR had a negative working capital model, no debt, and free cash flow which was put to use in the form of share repurchases, a lot of them.

NVR has gone one step further relative to efficient operations, and that is offering 4-5 home models and fewer customization options, which gives them two important things: economies of scale and thus bargaining power with suppliers, and quicker operations. NVR could deliver a home in 3 months compared to some other homebuilders that take 6 months, which means higher asset turnover that more than compensates for the lower margins due to the LPAs model.

Finally, homebuilders provide mortgage-related services in which they originate mortgage loans almost exclusively for their homebuyers. They generate revenue primarily from origination fees, gains on sales of loans and title fees. Homebuilders usually sell almost all of the mortgage loans they close into the secondary market as loans are underwritten following the standards of future buyers like big banks and loans are held for no longer than 30 days on average on the balance sheets of homebuilders.

Homebuilders didn’t clone NVR until after the GFC. Even during this period NVR was profitable and cash flow positive, and with more than a billion dollars in net cash, which made many homebuilders change their business models. The focus in the industry became reducing the land and debt on the balance sheet. Companies started to use more LPAs, and many saw their returns on equity increase, and the industry overall is in better shape now than 15 years ago. However, there are some differences between them.

DR Horton [NYSE: DHI] and Lennar [NYSE: LEN], which are the largest homebuilders in the country, have been buying smaller builders and gaining market share, thanks to their economies of scale and their ability to reduce costs per home built compared to smaller local players. Both homebuilders focus on entry-level or first-time buyers with average selling prices of around $370,000-$390,000, which is below the U.S. average of around $500,000. Both companies are present in multiple states and can be a “safer” bet from a geographic standpoint. The use of LPAs has increased from 25%, pre-GFC, to 75% for controlled land, while debt has decreased from 3x net debt/EBITDA to 1x, or even net cash.

NVR is the efficient operator that focuses on markets where competition is lower and close to their factories, so that way they can reduce costs. The company is more concentrated on the East Coast, and especially around the D.C. metro area.

Pultegroup [NYSE: PHM] builds homes with a higher price point; the average selling price is around $550,000 and pretty diversified between customer types, first-time, move-ups and active-adult. It is also present in multiple states and what differentiates it the most from DR Horton or Lennar is the use of LPAs. PulteGroup has less land controlled through options (56%) and builds more spec homes, which in good times increases returns, although the goal is to increase the use of LPAs.

Toll Brothers [NYSE: TOL] has an average selling price of around a million dollars and thus exposure to a different type of customer compared to the other big homebuilders. The use of LPAs is at 55% of land lots but because the type of customer and home, the need for owned lots and more customization is higher.

All of these companies have been changing their business to a more asset-light/NVR type of model, and it’s something quite consistent throughout the industry, but there are some small players like Green Brick Partners [NYSE: GRBK], which buys the land, develops it, builds the houses and sells them, thereby capturing the development profit. They buy low cost land lots and finance the development with MUD (Municipal Utility District) and PID (Public Improvement District) which are special-purpose local government entities that finance water,sewer, drainage, roads or parks projects by issuing tax-exempt bonds and the repayment of those bonds is done with property taxes. These vehicles cover 70-80% of the development costs, which makes the company less capital-intensive. They operate in Texas and Atlanta, and although they could have a higher risk on an economic downside, the good inventory management and the industry tailwinds give them room to grow.

The Financials of U.S. Homebuilders

If we focus on financials, we can see that the homebuilders have been able to increase returns on capital employed as LPAs were put to use. Return on equity (ROE) for DR Horton went from low double-digit returns with the use of leverage, to high teens without leverage on the balance sheet. Margins are lower with the asset-light model because of the premium paid for the land option and don’t receive the profit margin from the land developer, but the higher asset turnover more than compensates and the lesser debt gives more financial stability to the company.

Another thing NVR led in was share buybacks. From the year 2000 to 2025, NVR has bought back 70% of the shares outstanding. A good business model, high free cash flow, low capital needs, good management and a cheap valuation are what every investor wants, but when the company buys back shares aggressively at the right price, returns explode. Multiple examples come to mind: AutoZone [NYSE: AZO], Murphy Oil [NYSE: MUR], AutoNation [NYSE: AN], Washington Post…

DR Horton, Lennar, PulteGroup, and most of them followed the teachings of NVR and started to buy back shares at very attractive prices, and still do. Since 2020, DR Horton reduced the share count 20%, and 9% alone in the last year; Lennar bought back 21%, PulteGroup 30%, Toll Brothers 33%, and so on.

These great economic characteristics, gave the chance for homebuilders to compound value at great returns, which can be seen in their book values. DR Horton 18% CAGR, Lennar 12.5%, PulteGroup 17%, NVR 16% and so on. Great investments are those that can compound at high returns for many years instead of paying dividends, and I do think this group of homebuilders can keep reinvesting their capital for some time at high rates. The structural housing deficit, in my opinion, provides a chance that they will keep achieving similar returns on capital employed.

When looking for great compounders, first thing companies must have, is high ROE, otherwise you won’t accomplish much compounding done with a 4% return on reinvested capital company. Terry Smith uses a quote from Charlie Munger that explains it best:

Over the long term, it’s hard for a stock to earn a much better return than the business which underlies it earns. If the business earns 6% on capital over 40 years and you hold it for that 40 years, you’re not going to make much different than a 6% return—even if you originally buy it at a huge discount. Conversely, if a business earns 18% on capital over 20 or 30 years, even if you pay an expensive-looking price, you’ll end up with one hell of a result.

I do think the market thinks incorrectly about homebuilders. Most investors I’ve spoken to about these companies don’t usually even take a look at them because of the thought of cyclicality and thus “poor quality”, and at most trade them based on what they think interest rates will do.

That’s why I think you can see wide movements in the stock price from time to time, which is a great thing for investors who understand these companies and look at the long term fundamentals. Right now I don’t view the industry as extremely cheap, but relative returns compared to an index like the S&P 500 should be very good.

The way I value these homebuilders, and any other company, is by using a residual income framework, which you can find more information about it from Professor Stephen Penman in his books or from Javier López Bernardo, another member of MOI Global who wrote a great article. The mental model is pretty simple, you can think of it as looking at a bond. What’s the coupon, what’s the price of the bond and what yield do I get from buying at the current price.

With a company is the same thing, what’s the capital the company employs, what return does it generate from that capital and how much am I paying for that capital.

If we take DR Horton, it has a book value of $82 trading at 1.8x book value. If we consider 18% ROE and a cost of capital of 10%, IRR should be around high-single digit to low double digit returns in the long term, assuming modest economic growth. Lennar should yield pretty similar results. PulteGroup should have higher returns than DR Horton even though they trade at the same multiple, but that’s because return on equity from PulteGroup is higher.

Finally, perhaps the most attractive one could be Green Brick, which trades at 1.6x book value but with returns on equity above 20%.

Another, and better, explanation of how I think about valuation is this one from the great investor, Chuck Akre:

To illustrate: If a business has an underlying 21% return on capital, but an investor paid a market price of 3x capital to acquire the stock, then a 100% dividend of all earnings would only yield the investor a 7% pre-tax return (the arithmetic: 21%/3 = 7%). This is very much an average return, even though the underlying business has excellent returns on capital. While the underlying business model may be outstanding, and while the business may be run by exceptional people, the lack of reinvestment opportunity and a decision to pay out 100% of earnings will cause shareholder returns to be just average.

On the other hand, if it were possible that all earnings could be retained and reinvested at the same 21% rate, the shareholder would receive no dividend, but the capital and earnings of the business would both grow at the rate of 21% per annum. Over the long term, if the stock is purchased at a reasonable multiple, its market price would grow more or less in parallel with the rate of growth in capital and earnings. The shareholder, in this example, would build capital gains at a rate in the neighborhood of the underlying 21% growth rate, for so long as the runway of excellent reinvestment opportunities persisted. To top it off, the shareholder would enjoy tax deferral on the gains. If you know of any businesses that can retain 100% of earnings and reinvest them at a 21% rate, please give us a call!

These returns have to be of course taken into consideration given the risks we are exposed to:

• Given the cyclicality of the business, an economic downturn would of course impact the industry’s profitability.
• High interest rates worsen the issue of affordability.
• The tightening of credit standards and limited availability of suitable mortgage financing could prevent customers from buying homes.
• Inventory risks. The market value of land lots can decline and housing inventories can fluctuate as market conditions change. Due to the LPAs (options on land), the impact is lower than if the company owned the land.
• Costs of construction materials.
• Execution risk. The move to an NVR-type of business model takes time and it’s always hard to clone a winner. There’s still some land on the balance sheet that cannot be sold quickly, because it would drive down the price, especially if the other players do it as well. So the risk of land value impairment is higher than with NVR for example.

 

About the author

Gianluca Ruano, CFA, is a fund manager at Welzia Asset Management, an investment fund management firm based in Madrid. He began his career as an equity analyst in 2023 after passing the CFA exams. He also teaches on the Master’s Degree in Financial Markets at the Universidad Autonoma de Madrid, where he was once a student.

Download the PDF version