The 96 Homes That Cost More Than They Sold For

What happened

A builder closed land in January 2021 at $7.2 million and budgeted at January 2021 prices: $261,000 of hard cost a home across 96 homes. The first 24 were built under a guaranteed maximum price signed before the worst of the run-up and landed essentially on budget. The next three phases went out to bid in mid-2022, when the construction-materials index peaked 37.7% above the budget basis. By that July months' supply of new houses had reached 10.3 from 4.0, and by October the thirty-year mortgage rate hit 6.66%, dropping the sales pace from the underwritten eight closings a month to about three.

The case, start to finish

The budget was not wrong about the homes. It was wrong about the date.

A plan priced on one day in January

This is an anonymized composite built from documented market data: a 96-home for-sale townhome community in a Sun Belt suburb, budgeted in January 2021 and finished in April 2024. The project's own figures are illustrative and on round numbers. The market moves around them are real and cited.

The land closed at $7.2M, $75,000 a home, and the budget was set at that month's prices: $261,000 of hard cost a home, $28,000 of soft cost, $1.4M of interest carry, $36.4M all in against a $455,000 average sale price. The plan showed $7.3M of profit. The world it was written in had a construction-materials index at 256.4, a 30-year mortgage rate of 2.65%, and 4.0 months' supply of new houses nationally.

Hold those three numbers, because they are the whole exposure. A for-sale developer collects nothing for three years, is bound to a cost base priced on day one, and then sells into a market whose buying power is set by an interest rate nobody in the deal controls.

The decision that felt like discipline

Phase 1, 24 homes, was built under a guaranteed maximum price signed before the worst of the run-up. It landed at $266,000 a home, essentially on budget, and it was the last good news in the file.

Here is the moment. In early 2022, with Phase 1 delivered on plan, the contractor would hold prices on the remaining 72 homes only for a contingency premium the developer could see printed on a page. Materials were up sharply and everyone expected them to normalize. The premium was a certain cost you have to defend in a meeting; the exposure it removed was an uncertain one that appears in no line of the budget. So the guaranteed maximum price was allowed to lapse and phases 2 through 4 went out to competitive bid.

They came back at $318,000 a home. The materials index the entire $36.4M budget rested on had reached 353.0 in May 2022, 37.7% above the January 2021 basis and the highest print in the series through 2023. That one choice accounts for $4.2M of an $8.5M swing.

The second shock, which turned the first one into a third

A cost overrun with a fast sell-out is a thinner profit. What made this fatal is that demand moved at the same time. The 30-year mortgage rate went from 2.65% to 6.66% by October 2022 and peaked at 7.79% a year later. Months' supply of new houses went from 4.0 to 10.3. The sales pace fell from the underwritten 8 closings a month to about 3.

Now watch the cost base, because this is the mechanism the case is really about. Three of its four lines are driven by time. Interest carry was budgeted at $1.4M for a twelve-month sell-out and ran to $3.6M across twenty-seven months. Soft costs, which include marketing and the HOA subsidy, stretched from $2.7M to $3.1M for the same reason. In a for-sale project a pace miss is not a delay, it is a second cost overrun, because the debt service keeps accruing on inventory that is not moving.

Add $1.63M of price cuts and mortgage-rate buydowns bought to restore absorption, and the four pieces sum to $8.46M. A plan showing $7.3M of profit produced $42.0M of revenue against $43.2M of cost, and a $1.1M loss.

The number that was never really an assumption

Eight closings a month was not an estimate of anything. It was reverse-engineered: the pace that made a twelve-month sell-out work, and the twelve-month sell-out was what kept the carry line small enough to ignore. When absorption is a plug rather than a forecast, the carry quietly becomes a plug too, and nothing in the model announces it. Nobody re-ran the numbers at four closings a month until the market had already delivered three.

What genuinely worked was phasing. Building in four blocks of 24 rather than all at once meant that when the pace halved, the developer was carrying partly-built inventory instead of 72 finished, fully-funded homes with the facility drawn. That is the only reason this ends as a $1.1M loss rather than a conversation with a workout desk.

For anyone building something that sells at the end

You do not need 96 homes for this to apply. Any project with a long build and a single collection window has the same shape, and the same three habits fall out of it.

Put a number on the cost certainty you are turning down, and write it beside the premium you refused to pay, so the comparison is between two figures rather than between a figure and a feeling. Release capacity against a trailing sales pace rather than a projected one. And stress the model at half the underwritten absorption before the land closes, because after that every input except the price you are willing to accept is already fixed.

Timeline

• Jan 2021 Land closes at $7.2M ($75,000 a home) and the budget is set at January-2021 prices: $261,000 of hard cost a home, $28,000 of soft cost, and $1.4M of interest carry, for $36.4M all in against a $455,000 average sale price. Backdrop: the BLS construction-materials index reads 256.4, the 30-year mortgage rate is 2.65%, and there are 4.0 months' supply of new houses nationally.
• Q3 2021 Phase 1, 24 homes, is built under a guaranteed maximum price signed before the worst of the run-up. It lands at $266,000 a home, essentially on budget. This is the last good news in the file.
• May–Jun 2022 The construction-materials index peaks at 353.0, which is 37.7% above the January 2021 budget basis. Phases 2 through 4 go back out to bid at exactly the wrong moment and come back at $318,000 a home.
• Jul 2022 Months' supply of new houses reaches 10.3, up from 4.0 when the budget was written. The market has stopped absorbing at the rate the pro forma assumed.
• Oct 2022 The 30-year mortgage rate hits 6.66%. The sales pace falls from the underwritten 8 closings a month to about 3. The builder starts paying rate buydowns and shaving list prices.
• Oct 2023 The 30-year rate peaks at 7.79%. The construction loan is still outstanding on unsold inventory, and every month of carry is now a real line item rather than a rounding error.
• Apr 2024 The last home closes, 27 months after the first, against a 12-month planned sell-out. Final tally: $42.0M of revenue against $43.2M of cost. A plan that showed $7.3M of profit produced a $1.1M loss.

You're in the owner's chair

Early 2022. Phase 1 is built and selling. Your general contractor will not hold prices for the remaining 72 homes without a contingency premium you can see on the page, and materials are up sharply. Your lender has approved the full facility. What do you do?

• Pay the premium for a guaranteed maximum price; release 24 homes at a time
• Build all 72 remaining homes now to lock in the current bid and capture the scale
• Keep the cost-plus contract and pre-sell hard at fixed prices to lock in your revenue

Where the $8.5M swing went

• Hard-cost overrun on phases 2–4 (re-bid at the index peak): 4.22 $ millions
• Extra interest from a 27-month sell-out instead of 12: 2.2 $ millions
• Price cuts and mortgage-rate buydowns to move inventory: 1.63 $ millions
• Extra soft costs over the longer sell-out: 0.41 $ millions

Illustrative project arithmetic, on round numbers. The four bars sum to $8.46M, the distance between a plan showing $7.3M of profit and a result showing a $1.1M loss.

Business model

Buy dirt, entitle it, build homes on it, sell them one at a time to retail buyers, and keep the spread. The developer is not a landlord and never collects rent, because the entire return arrives in a burst of closings at the end, which means the business is really a bet on two numbers three years out: what it will cost to build, and how fast people will buy.

Revenue model

96 closings at a planned $455,000 average, or $43.7M of revenue, all of it realised in the last third of the project's life. Revenue is not recurring, is not diversified, and cannot be trimmed: an unsold home produces nothing at all, and a home sold at a discount produces its discount permanently.

Cost structure

Land was $7.2M and fixed on day one. Hard costs were budgeted at $25.1M and came in at $29.3M. Soft costs (design, permits, impact fees, marketing, HOA subsidy through a longer sell-out) went from $2.7M to $3.1M. And interest carry, budgeted at $1.4M for a 12-month sell-out, reached $3.6M over 27 months. The structural point: three of those four lines are driven by TIME, and time is the one input the developer controls least.

Strategic challenge

Two independent shocks arrived in the same eighteen months and compounded. Costs rose because the materials index moved 37.7% between the budget basis and the re-bid. Absorption fell because the 30-year mortgage rate went from 2.65% to 6.66% and the national months' supply of new houses went from 4.0 to 10.3. Neither alone would have killed the deal. A cost overrun with a fast sell-out is a thinner profit; a slow sell-out at budgeted costs is a delayed profit. Both at once turns the interest line into a second cost overrun, because the debt is still there while the inventory is not moving.

Key decision

In early 2022, after Phase 1 came in on budget, the developer let the guaranteed maximum price lapse and put phases 2 through 4 out to competitive bid, expecting materials to normalise. The re-bid landed at the top of the index. That single choice, declining to buy price certainty because certainty carried a visible premium and the alternative carried none, accounts for $4.2M of the $8.5M swing.

What worked

Phasing. The community was built in four blocks of 24 rather than all at once, so when the sales pace halved the developer was carrying partly-built inventory rather than 72 finished, fully-funded homes. That is the only reason the loss was $1.1M instead of a lender workout. The Phase 1 guaranteed maximum price also worked exactly as intended; it just was not renewed.

What failed

The absorption assumption, and the way it was written. Eight closings a month was not derived from anything; it was the number that made the 12-month sell-out work, and the 12-month sell-out was the number that made the interest carry small. When absorption is a plug rather than an estimate, the carry line silently becomes a plug too. Nobody re-ran the model at 4 closings a month until the market had already delivered 3.

Risk factors

A cost base priced at one date and spent over three years; revenue realised entirely at the end; construction debt that accrues whether or not anything sells; retail buyers whose purchasing power is set by a mortgage rate the developer cannot hedge; and price concessions that are permanent while carrying costs are merely persistent, so cutting price to restore absorption trades a fixed loss for a variable one, which is usually right and always expensive.

Lesson summary

In development, absorption is the interest-rate assumption wearing different clothes. Every month the sell-out slips is another month of carry on the whole undrawn plan, so a pace miss compounds into a cost miss. Price the cost certainty you are declining to buy, phase construction against a trailing sales pace rather than a projected one, and stress the model at half the underwritten absorption before you close on the land, because after that every input except the price you accept is already fixed.

Key data

• 256.4 → 353.0 (+37.7%, BLS) Construction-materials index, Jan 2021 → May 2022
• 2.65% → 6.66% (Freddie Mac PMMS) 30-year mortgage rate, 7 Jan 2021 → 6 Oct 2022
• 4.0 → 10.3 (Census/HUD) Months' supply of new houses, Jan 2021 → Jul 2022
• $261,000 → $305,000 Hard cost per home, budget → actual average
• 12 months → 27 months Sell-out, planned → actual
• $1.4M → $3.6M Interest carry, budget → actual
• +$7.3M profit → −$1.1M loss Bottom line, planned → actual

Sources & basis

The business in this story is a stand-in, not a company you can look up. This case is an illustrative composite: the operator, the people and most of the dollar figures represent a pattern rather than reporting one firm's history. What the list below cites is the other half, the documented industry data and public reporting the composite was assembled from, including any real company whose published figures the case draws on by name. The mechanism and the arithmetic are real even where the business is not.

1. U.S. Bureau of Labor Statistics — PPI Commodity data, Special indexes: Construction materials (series WPUSI012011, 1982=100, not seasonally adjusted): 256.4 in January 2021, 353.0 in May 2022 View source ↗
2. Freddie Mac Primary Mortgage Market Survey, 30-year fixed rate, full weekly history as published by FRED (St. Louis Fed): 2.65% on 7 Jan 2021, 6.66% on 6 Oct 2022, 7.79% on 26 Oct 2023 View source ↗
3. U.S. Census Bureau and HUD, New Residential Sales — monthly supply of new houses, full history as published by FRED: 4.0 months in January 2021, 10.3 months in July 2022 View source ↗
4. Composite project budget and per-home figures — illustrative, on round numbers; see the Real Estate and Capital & Financing categories