KB Home (KBH) Q3 2026 earnings review
Q3 margin beat, Q4 step-up guided away as orders fall 12%
Housing revenue fell 19.9% to $1.29 billion, inside the range guided in June. Diluted EPS was $1.05, or about $0.96 at last year's tax rate and without a one-time investment gain (derived). Adjusted gross margin of 16.8% beat the top of its range, but the Q4 midpoint of 16.3% sits below it. Net orders fell 12% and full-year housing revenue was cut $100 million at the midpoint.
⚖️ Verdict: 🔴 Bearish
The case is worse — a bearish verdict — because the Q4 margin step-up the standing story rested on has been guided away: with Built to Order at its target mix, Q4 gross margin is set below Q3, and a third full-year revenue cut this year confirms the annual guides do not hold. Demand moved the same way: absorption fell 18% and 8% community growth no longer covers it. The backlog turning positive on schedule is the material item the other way.
Unsettled is whether the backlog rebuild is demand or composition: it is above a year ago, yet Q3 conversion fell to 60% as Built to Order homes sit longer and three of four regions shrank. Q4 deliveries against the 3,000–3,500 guide, and the conversion they imply, give the reading.
🐂 Bull Case
Built to Order Reset Finished a Quarter Early, and the Margin Followed Sequentially
Built to Order homes were nearly three-quarters of Q3 deliveries, up from 60% in Q2 and already at the roughly 70% management had targeted for Q4. The margin moved with the mix: adjusted housing gross margin rose to 16.8% from 15.7%, above the 16.0–16.6% guided in June, and adjusted operating margin recovered to 5.4% from 3.0%. SG&A fell to 11.3% of housing revenue from 12.7%, the bottom of its range, with dollars down 9.6% on a year ago after personnel reductions and lower performance-based pay.
Two sequential margin gains confirm Q1 was the bottom. Whether Q4 holds this level is the bear case below.
Backlog Turned Positive on Schedule, and the Flow Improved
Ending backlog of 4,398 homes stands 1.5% above a year ago — the first year-over-year increase in four years and the outcome management committed to in June — with backlog value up 3.2% to $2.05 billion. The flow improved as well: net orders less deliveries drained 128 homes from backlog in Q3 against 443 in the same quarter last year (derived). Reaching the Q4 delivery midpoint needs about 74% of this backlog to close, against the 84% that converted a year ago (derived), so the delivery guide is covered even if orders stay weak.
Standing positives the print did not read on
Three drivers in the standing story have no reading in the release or presentation; each has a disclosure that would move it.
- Build times: last stated at 100 days from start to completion in Q2; the Q3 call gives the number.
- Land option walk-aways: optioned lots fell to 24,452 from 27,475 at year-end, an 11% drop, but the lots abandoned in Q3 and the related charges are in the 10-Q.
- Sandstone in North Las Vegas and the first Atlanta community: both were scheduled to open, with no update in the release.
🐻 Bear Case
Orders Down 12% With Absorption at 3.1: Demand Weakened, and Three Regions Carried the Fall
Net orders of 2,604 fell 11.7% on a year ago, the weakest of the last five quarters, and the cause is pace rather than store count: monthly orders per community dropped to 3.1 from 3.8, an 18% decline, while an 8% larger average community count of 279 could no longer offset it as it did in Q1. Cancellations returned to 18% of gross orders from 12% in each of the two prior quarters, undoing what had looked resolved. Management says conditions weakened after June, and with the Q4 ending community count guided to 270–275, down from 277, the offset shrinks from here.
The decline is regional. West Coast orders grew while the other three regions fell 20% combined and their backlog sits 7.6% below a year ago (derived).
- West Coast: orders +7.7% to 937; backlog homes +22.8%.
- Central: orders -26.2%; backlog homes -11.2%.
- Southeast: orders -23.5%; backlog homes -10.1%.
- Southwest: orders -2.4%; backlog homes +3.1% but value -9.4%.
The Q4 Margin Step-Up Is Gone: 16.3% Guided, Below Q3, Against a 17.2% Path in June
The standing story had a back-end-loaded margin: what the June full-year guide left for Q4 was an adjusted gross margin near 17.2% (derived). The new Q4 range is 16.0–16.6%, a midpoint below the margin just printed, and the full-year range fell to 16.0–16.2% from 16.1–16.5%. Against Q4 2025's 17.8%, the recovery quarter is now guided 150 basis points lower. Management's own mechanism — a four-point spread on Built to Order homes, plus leverage, plus Northern California mix — has had the mix arrive and the margin guided down, which leaves the causes the release names: continued pricing pressure and higher relative land costs.
The pricing evidence sits on the West Coast, where the average selling price of homes delivered fell 6.2% to $641,800 and the backlog is valued at about $641,000 a home, flat on a year ago (derived). The Q4 guides imply a blended price near $477,000 (derived), not the roughly $500,000 the June guides pointed to. The year-on-year margin gap narrowed to 210 basis points from about 400 in Q2, but that is last year's comparison easing, not this year's margin improving.
Land Spend Up 40% on a Drawn Revolver: Net Debt Rose $486 Million This Year
Land and development investment of $722 million was 40% above a year ago and 46% above Q2, in a quarter when orders fell, and it was funded with debt: revolver borrowings reached $415 million from $275 million in May and none at year-end, and net debt rose $486 million since November to $1.95 billion (derived). Debt to capital of 35.7% is the highest in the presentation's five-year series. Operating cash flow after land spend ran about $75 million over the last twelve months against $335 million in fiscal 2025 (derived from the presentation).
Buybacks slowed to $50 million, the floor of the quarterly envelope, and $300 million of notes mature in June 2027; liquidity of $942 million covers that, but the direction has turned from returning cash to borrowing for land.
Standing risks the print did not read on
Five standing concerns got no reading in the release or presentation; the Q3 10-Q or the call carries each.
- Leadership: no finance executive is named in the release and no succession has been disclosed; the 10-Q signature page gives the reading.
- Headquarters relocation to Tempe: the release flags 2027 costs; the amount recognized in Q3 and any staff attrition are 10-Q items.
- Impairment watch list: inventory-related charges were $3.0 million, down from $5.6 million in Q2, but the number of communities under recoverability review is a 10-Q note.
- West Coast segment gross margin, the lowest of the four regions in Q2: not in the release, and it is the region carrying the mix story.
- DOJ ENERGY STAR subpoena and the 45L energy-credit expiry: no update; the contingencies note and the Q4 tax rate carry them.
👓 Other Themes
A 19.6% Tax Rate and an Investment Gain Added About $0.09 to EPS
The effective tax rate was 19.6% against 23.3% a year ago, which the release attributes to excess tax benefits from stock-based compensation, and pretax income included a $3.5 million gain on the sale of a stake in a private technology company. At last year's tax rate and without the gain, diluted EPS would have been about $0.96 rather than $1.05 (derived). The Q4 rate guide of approximately 26% says neither item repeats.
Conditions Weakened After June: Rates, Geopolitics, Tariffs
The release says the market weakened since the June report, with higher mortgage rates pressing affordability and geopolitical uncertainty and broader economic headwinds making buyers more cautious; the risk factors again list tariffs on building materials and the Middle East and Ukraine conflicts. None of it is quantified in the materials, so the demand card above carries the numbers; the company's stated response is to balance price and pace rather than to chase volume.
💲 Other KPIs
67.7 million in Q3 2025, 65.0 million in Q4, 63.7 million in Q1 2026, 61.8 million now, an 8.8% reduction on a year ago that turned a 41% net income decline into a 35% EPS decline. The pace is slowing: 0.9 million shares repurchased this quarter against 3.3 million in Q3 2025. Book value per share $62.56, up 4% on a year ago, on about 60.8 million shares outstanding; $725 million of authorization remains.
65,251 a year ago, 64,612 at Nov 30, 63,257 at Feb 28, 59,106 at May 31, now 61,581: down 5.6% on a year ago but up 4.2% in the quarter, the first sequential increase in the series (derived). Owned lots are flat at 37,129 since year-end; the optioned count is 24,452, so the quarter added about 2,000 optioned lots and 480 owned. Mix 60% owned, 40% under contract, from 62/38 in May.
$5.84 billion a year ago, $5.67 billion at Nov 30, $5.70 billion at Feb 28, $5.73 billion at May 31, $5.98 billion now: up 5.5% since year-end and 2.4% on a year ago, with $249 million of the increase in Q3 alone, while nine-month housing revenue is down 23%. Inventory is being built ahead of a delivery base that the guide has shrinking.
Fifth straight year-on-year decline, but the rate of decline has narrowed for two quarters: -26% in Q1, -18% in Q2, -15% now, tracking deliveries. The release attributes the quarter to lower title and insurance results; equity income from the mortgage joint venture was $4.1 million against $4.3 million.
🔮 Guidance
Midpoint 3,250 implies a 10.2% decline on Q4 2025's 3,619 and a 19% rise on Q3 (derived). It matches what the full-year range leaves for Q4, 3,003–3,503 (derived), so the two guides reconcile. The year-on-year decline narrows from -19.5% in Q3 because Q4 2025 was already down 9% on its own comparison.
Midpoint $1.55 billion implies -8.0% on Q4 2025's $1.68 billion and +19.9% on Q3 (derived). The full-year midpoint leaves $1.53 billion for Q4, $20 million less than the quarterly midpoint, so the two guides only reconcile in the lower half of the quarterly range (derived). Against the June full-year guide, which left about $1.65 billion for Q4, the quarter is $97 million lower at the midpoint (derived).
Midpoint 16.3% is 50 bps below Q3 and 150 bps below Q4 2025's 17.8%. The full-year guide leaves 15.9–16.5% for Q4 (derived), so the top of the quarterly range is not reachable inside the annual one. Q3 beat its own range by 20 bps and the average outcome over the last three quarterly margin guides is +0.2 pts (derived); if that holds, Q4 lands near 16.5%, still below Q3.
Midpoint 10.6% against 10.0% in Q4 2025, or 9.1% excluding the $16 million of accelerated equity compensation that quarter per the presentation, and 11.3% in Q3; the full-year range implies 10.6% for Q4 (derived), consistent. The improvement is volume: at the midpoints Q4 SG&A dollars are about $164 million, 12% above Q3's $146.6 million (derived), so the guide allows spending to rise into year-end.
Up from 19.6% in Q3 and 26.6% in Q2; the excess tax benefits that lowered Q3 are not assumed to repeat, and the loss of 45L energy credits on homes delivered after June 30 sits in the base. The full-year rate of approximately 23% against 20.6% for the nine months (derived) is consistent with a Q4 near 26%.
Down from 277 at Q3 and the 280 peak in May; midpoint 272.5 is 0.6% above the 271 a year earlier (derived), so year-on-year community growth, which ran 8% on average in Q3, is close to zero by year-end. The offset that community openings gave to falling absorption ends here.
Confirmed the June range; midpoint 10,750 implies -16.7% on 12,902 (derived). With 7,497 delivered in nine months, the range leaves 3,003–3,503 for Q4, the same as the explicit Q4 guide. The midpoint has moved 11,750 in January, then 10,750 in March, June and now: one cut, then held three times.
Cut the midpoint by $100 million to $5.0 billion, the third reduction this year: $5.60 billion in January, $5.15 billion in March, $5.10 billion in June, $5.00 billion now, moves of -$450M, -$50M and -$100M (derived). The new range sits inside the old one, which is what 'within the ranges we last provided' refers to; the top came down $200 million. The midpoint implies -19.5% on fiscal 2025's $6.21 billion and leaves Q4 at $1.53 billion, -9.2% (derived). Fiscal 2025 followed the same shape — three cuts, then a final guide beaten by 1% — and if that pattern holds the year lands near $5.05 billion (derived).
Cut 20 bps at the midpoint to 16.1% and 30 bps at the top; against fiscal 2025's 19.1% that is -300 bps. The nine-month adjusted margin is 16.1%, so the year is guided flat from here, and the range implies 15.9–16.5% for Q4 (derived).
Confirmed the 11.6% midpoint and narrowed the range; against 10.4% in fiscal 2025, +120 bps. Nine months at 12.1% leaves 10.6% for Q4 (derived), matching the quarterly guide.
Held at the midpoint of the June range, which was itself the third downward move this year (24–26% in January, 23–25% in March). The nine-month rate is 20.6% (derived), which squares with a Q4 near 26%.
❓ Key Questions
What is the West Coast doing to the Q4 margin?
How much of the 16.0–16.6% Q4 gross margin guide reflects West Coast pricing, and what share of Q4 and fiscal 2027 deliveries and gross profit comes from Northern California communities — the driver named in June — with their margin relative to the company average?
Who is the CFO?
Who is the chief financial officer of record, and has the executive who presented on the March call left the company? The release names no finance executive.
How much of the backlog is sold but not started?
Of the 4,398 homes in backlog, how many are sold but not yet started versus a year ago, and what conversion rate does the 3,000–3,500 Q4 delivery guide assume? Q3 conversion fell to 60% of opening backlog from 71%.
What does $722 million of Q3 land buy for 2027?
What community count does the Q3 land investment support for fiscal 2027 and 2028, given lots controlled are 5.6% below a year ago and the Q4 ending count is guided down to 270–275, and how much of the spend went to Sandstone, Atlanta and Northern California?
Why did cancellations return to 18%?
The cancellation rate went from 12% in each of the last two quarters back to 18%: is that buyers failing to qualify at higher rates, or Built to Order contracts aging out, and what incentive load — rate buydowns and closing-cost credits — sits inside the 3.1 monthly absorption?