Lennar delivered 20,519 homes in its second quarter and gave away roughly $55,000 per home to do it.1
That is the number worth sitting with. Not because it is large — though it is — but because of what a builder has to believe about its own market before it agrees to hand back twelve cents of every dollar at closing.
The incentive rate was 12.9% of the gross sales price.1 A year earlier, it was 13.3%. In the fourth quarter of 2025 it was 14.5%, and in the first quarter of 2026 it was 14.1%.2 Read that series quickly, and it looks like relief. Lennar reads it that way too, telling investors the gap between current incentives and a normalized 4-6% range is narrowing for the first time in three years.2
The series is misleading, and the reason it is misleading is the entire story.
The discount didn't shrink. It moved.
Lennar's average sales price fell from $389,000 to $371,000 year over year, a 5% decline the company attributes to continued weakness in the market.1 Back out the incentive rate, and the gross sticker went from roughly $449,000 to roughly $426,000. The discount below the line got smaller. The price above the line got smaller too.
The buyer paid less. Lennar simply stopped routing as much of the reduction through rate buydowns and closing credits and started routing it through the price tag itself. Gross margin fell from 17.8% to 15.6%, driven by lower revenue per square foot.1 That is not mix. That is price.
Important
Incentives were always a way of cutting price without admitting it. A builder who moves the cut back into the sticker is not signaling recovering demand. It is signaling that the concession has become permanent enough to capitalize.
And new orders still fell 4% year over year.1 Lennar cut the effective price, cut the nominal price, and sold fewer homes than a year ago while operating more communities. Full-year delivery guidance came down to 82,000-83,000.2
Builders reveal. Owners don't.
The cleanest evidence that this is a constrained-buyer market rather than a distressed-seller market sits in two inventory figures that almost never get printed next to each other.
-
New single-family homes: 488,000 for sale, a 9.6-month supply, with July sales running 6.3% below a year earlier.3
-
Existing homes: 1.54 million for sale, a 4.6-month supply, with July sales running 0.7% above a year earlier.4
The same country, the same buyers, the same mortgage rate. One segment is carrying more than double the months of supply.
The asymmetry is structural. A builder owns land, carries construction financing, runs a cycle-time-optimized pipeline, and answers to a quarterly delivery number. Lennar's cycle time is 121 days.2 Nothing about that business permits waiting for a better market. A homeowner with a 3.4% mortgage and no obligation to transact can wait indefinitely, and roughly 1.5 million of them collectively decided to keep waiting.9
So builders discover the clearing price, and existing owners refuse to. The gap between 9.6 months and 4.6 months is the width of that refusal.
2008 forced people to sell. 2026 prevents people from buying.
One was a foreclosure crisis. The other is a frozen market.
In 2008, the leverage sat with households. Resetting mortgages, negative equity, and job losses manufactured sellers who had no choice. Foreclosures added inventory, inventory pushed prices down, lower prices created more negative equity, and the loop fed itself.
The 2026 version runs in the opposite direction and never closes the loop. Price times rate times insurance times taxes exceeds what incomes support. Buyers withdraw. Transactions collapse. Owners, seeing no bid they like and facing no pressure to accept one, decline to list. Prices stay sticky because nothing forces them to move.
Existing-home sales are running at a 4.06 million annual pace with the median price up 2.0% year over year — the thirty-seventh consecutive month of annual increases.4 Prices rising into a transaction recession is not a contradiction. It is the definitional signature of a frozen market.
A homeowner says: I am not selling my $450,000 house for $390,000.
Lennar says: the house is $450,000, and here is $55,000 in buydowns and credits.
Economically, those are close to the same sentence. Statistically, they are nothing alike, and only one of them shows up in the index.
The correction is happening off the index
FHFA's purchase-only index rose 2.1% year over year in the second quarter, its measurement built from repeat sales of the same single-family properties.5 Forty-seven states posted gains. The East North Central division ran +4.5% while the Pacific division sat at roughly zero.5
That index cannot see what Lennar is doing. Repeat-sales methodology requires a property to trade twice. A new home sold once, with $55,000 of the price handed back at the closing table, enters the national price statistics either weakly or not at all.
Which means a meaningful share of the American housing correction is already complete and simply not visible in the series people quote. The builder segment has cut effective prices for three straight years. The resale segment has not cut them at all. The index is dominated by the segment that refuses to move.
Anyone waiting for a nominal crash to confirm that a correction happened is waiting for a confirmation the measurement apparatus is not built to deliver.
Two clocks
A frozen market thaws through one of two mechanisms. Incomes rise into prices, or financing costs fall into payments.
-
The income clock: nominal average hourly earnings are growing about 3.2% against national home price growth of 2.1%.56 Roughly 1.1 points of convergence per year. If the gap between prices and incomes is 15%, that pace closes it in about fourteen years.
-
The rate clock: a $360,000 mortgage costs $2,325 a month at 6.71% and $1,933 at 5.00%. Seventeen percent off the payment with no change in the price of the house.
Both clocks are currently running backwards.
The thirty-year fixed averaged 6.38% in late March and 6.71% on September 3, up from 6.50% a year earlier.7 Rates have not fallen in 2026. They have climbed to the year's high.
And the income clock is worse than the arithmetic suggests. CPI ran 3.4% in the twelve months through July against 3.2% wage growth, leaving real average hourly earnings down 0.1% year over year — the fourth consecutive month in which inflation outpaced pay.68
Affordability can improve on paper and deteriorate in the room where the decision gets made. NAR's affordability index rose to 103.3 from 98.3 a year ago while sales fell for a second consecutive month.4
What the timeline actually requires
The 2029-2031 thaw is a reasonable center of gravity, but only under assumptions that are not currently holding.
The frozen path is not stable in the way "frozen" implies. Every year of 4.06 million transactions is a year of household formation deferred, brokerage capacity destroyed, and builders shrinking their delivery guidance. Lennar just cut its own by thousands of homes. Supply response weakens while the underlying shortage persists.
The thing that converts the freeze into 2008 is not a 7% mortgage. It is sustained unemployment — the only mechanism that reliably turns unwilling sellers into forced ones. Absent that, prices do not need to fall for a correction to occur. They need to stand still while wages walk past them, and wages are currently not walking.
Which leaves an uncomfortable third possibility: neither clock ticks forward, the freeze extends into the 2030s, and the adjustment happens entirely through builders' margins and the quiet, unindexed erosion of the price a new house actually fetches.
| Path | Home prices | Nominal wages | 30-yr rate | Convergence |
|---|---|---|---|---|
| Current run rate | +2.1% | +3.2% | 6.7% rising | 12-15 years |
| Slow normalization | 0% | +3.5-4% | 5.5-6% | 4-6 years |
| Meaningful correction | -2%/yr | +4% | 5-5.5% | 2-4 years |
| Labor shock | -5-10% initially | +2% | 4.5-5% | 1-3 years |
Three dates settle the near term. August CPI on September 11 shows whether real earnings turn positive or extend to a fifth negative month. Lennar reports its third quarter in mid-September, against guidance of approximately 16% gross margin and further incentive moderation — watch whether the average sales price rises into the guided $375,000-$380,000 range or whether the base price keeps sliding.
And the FHFA monthly series will show whether the Pacific division's zero turns negative, because the divisions go first and the national number goes last.