PHM — Q3 2026 Financial Report Analysis

Overview

PulteGroup's third quarter of 2026 shows the demand side turning down again after a better spring. Net new orders (contracts signed, minus cancellations) fell 6% to 6,208 homes even though the company had 7% more communities open for sale, and CEO Ryan Marshall blamed "the rapid increase in interest rates," which he says "has pushed even more consumers to the sidelines." Closings fell 9% to 6,876 homes, home sale gross margin slipped to 24.8% from 26.2% (about 0.7 points of that is a $26 million land write-down), and net income fell 19% to $476 million. Diluted earnings per share (EPS) fell less, 15% to $2.53, because PulteGroup bought back 5% of its share count over the past year.

The numbers

PulteGroup doesn't report an "operating income" line, so pre-tax margin is used as the closest equivalent measure of overall profitability (the same convention as our Q2 report). The release gives the community count only as 1,077 and "an increase of 7%," so the prior-year figure is left blank.

Why revenue fell

A homebuilder books revenue when a home closes — when the buyer takes ownership — not when the contract is signed. Home sale revenue fell 9.2% to $3.86 billion, and the release splits that into a 9% drop in closings and a 1% lower average selling price ($561,000 vs $564,000). So this is mostly a volume story: fewer homes delivered, at roughly the same price.

Land sales also shrank to $7.8 million from $53.2 million. These are lumpy sales of land parcels, not homes, and carry thin profit, so they matter for revenue more than for earnings.

Closings fell in every region except Florida, which was roughly flat (1,904 vs 1,932). The West was again the weakest, down 18% (1,118 vs 1,359).

Orders: the spring improvement didn't last

Florida — now PulteGroup's largest region — was the only one to grow orders. The Southeast, West and Northeast all fell about 15%.

Two details point in different directions. Orders in units fell 6.5%, but orders in dollars fell only 3.9%, so the average signed contract was worth about $564,000 versus about $548,000 a year ago (our calculation from the reported unit and dollar totals). That fits the CEO's point that "more than 60% of our business" is now move-up and active-adult buyers (Del Webb), who buy pricier homes and are less dependent on mortgage rates than first-time buyers. The less comforting read is per community: with 7% more communities open and 6% fewer orders, each community sold roughly 12% fewer homes than a year ago.

For the first nine months, orders are still up 1.4% (21,778 vs 21,486), which shows how much Q3 gave back from a strong first half.

Margins and costs

Home sale gross margin — the share of each home's sale price left after paying for the land, materials and construction — was 24.8%, down from 26.2%. The release says the quarter's figure includes "$26 million, or approximately 70 basis points, of land impairments" (a basis point is one hundredth of a percentage point, so 70 basis points is 0.7 points). Without that write-down, margin would have been about 25.5% — a little better than Q2's 25.0%, but still clearly below last year. The earnings release does not break out how much of the remaining decline came from sales incentives (such as mortgage-rate buydowns) versus higher land costs; the 10-Q, not yet filed, normally does.

Overhead got heavier relative to sales. SG&A (selling, general and administrative costs) was essentially flat in dollars at $397 million vs $401 million, but because revenue shrank, it rose to 10.3% of home sale revenue from 9.4%. That is consistent with carrying 7% more communities on fewer sales, though the release does not break SG&A down.

The mortgage arm was steady: financial services pre-tax income of $45.6 million vs $44.4 million, with 84.6% of PulteGroup buyers who financed their home using its mortgage company (the "capture rate," which the release rounds to 85%).

What the headline numbers hide

• Cash conversion is weak. For the nine months, operating cash flow was $637 million against net income of $1,295 million — under half. The main reason is land and homes in progress: cash spent on inventories was $1.15 billion, almost double the $636 million a year earlier, and house and land inventory has grown to $14.1 billion from $12.9 billion at year-end. Building inventory while orders are falling is a bet that demand comes back.
• Buybacks are now partly debt-funded. PulteGroup spent $1.03 billion on buybacks in nine months — more than its operating cash flow — and issued $795 million of new notes while repaying $613 million. Cash fell to $1.4 billion from $2.0 billion, and the company moved from net cash of $378 million at year-end to net debt of $401 million (debt-to-capital 12.1%, still low for the industry).
• About 4 points of EPS came from the smaller share count. Diluted shares fell 5.2% to 187.8 million, which is why EPS fell 14.5% while net income fell 18.8%. The tax rate did not help: about 23.9% vs 23.7% a year ago (our calculation from the income statement).
• Other income gave a lift. "Other income, net" was $19.7 million vs $1.8 million a year ago; the release does not explain it. Without that $18 million swing, pre-tax income would have fallen about 21% instead of 19%.
• One-offs cut both ways. The $26 million land impairment depressed margin by 0.7 points; nine-month land-related charges were $65 million vs $70 million, so write-downs are not new, but the release singles out this quarter's.
• No cancellation rate and no guidance. The earnings release does not give a cancellation rate or any numerical outlook, and the full 10-Q isn't filed yet, so segment profit by region and the incentive detail are missing here.

Did last time's read hold up?

Our Q2 report said the number to watch was whether home sale gross margin kept improving from Q1's 24.4% to Q2's 25.0%, and warned that "if incentives have to rise again to keep orders coming," earnings were not near the bottom. The reported margin slipped back to 24.8%, though excluding the land impairment it edged up to about 25.5%, so the margin trend roughly held. The order side did not: Q2's 6% order growth turned into a 6% decline. Our Q2 expectation that a larger backlog would make second-half closings compare better did not show up either — closings fell 9% this quarter, about the same as Q2's 8%.

Takeaway: PulteGroup's margin held up better than its demand did. Excluding the land write-down, gross margin was about 25.5%, slightly above Q2, but orders fell 6% even with 7% more communities open, and the backlog shrank by about 670 homes since June. With buybacks now outrunning operating cash flow, EPS is getting increasing support from a falling share count rather than from the homebuilding business itself.

Outlook

The release gives no numerical guidance. Management's tone has turned more cautious than in July: the CEO describes demand "increasingly pressured by rising mortgage rates, widespread inflation, and global unrest," and says rate increases have "made selling an existing home even harder" — which matters because many buyers need to sell their current home first. The company says it will "balance the price and pace of home sales, align production accordingly," which suggests fewer homes started if orders stay weak.

Our read: the backlog of 10,298 homes is still 4% larger than a year ago, so Q4 closings should compare better than Q3's did, but the backlog value per home (about $626,000 vs about $631,000, our calculation) is drifting lower, and fewer orders this quarter means a smaller starting point for early 2027. The mix toward move-up and active-adult buyers is the main reason margins are holding near 25% while rates rise. The things to check in the coming 10-Q are the cancellation rate (13% in Q2) and how much of the margin decline is incentives. If PulteGroup keeps buying back stock at this pace while inventory spending stays high, net debt will keep rising — still from a low base, but the opposite of where it stood at the start of the year.