Market Analysis • November 18, 2026

October Housing Starts Rebounded 9.9%—But Permits Fell and the Gain Is Uncertain

•6 min read•Housing

The November 18, 2026 housing report delivered a headline-friendly rebound with a sizable caveat: housing starts rose 9.9% from September, but the estimate’s ±16.8 percentage-point margin means the underlying change could have been a decline. Meanwhile, permits—which offer a read on the construction pipeline—fell 7.4%. The release put both figures in plain view. This is a tension in the data, not evidence of misleading framing or a confirmed housing recovery.

Here’s what the report shows:

  • Total starts reached 1.309 million at a seasonally adjusted annual rate (SAAR), up from a revised 1.191 million.
  • Permits fell to 1.295 million SAAR, extending a decline that began before October.
  • Completions rose in the monthly estimate but remained below their year-earlier level; both changes are statistically uncertain.
  • Single-family permits were down 12.9% year over year, while permits for buildings with five or more units were down 0.4%.
  • Regional weakness was sharpest in the West on a year-over-year basis, while the South posted the steepest monthly permit decline.

Starts Rose; the Pipeline Pointed the Other Way

October’s headline figures pull in different directions. Starts rose by 118,000 from September’s revised estimate, while permits fell by 103,000. The starts estimate’s margin of error is wider than the reported increase, however, so the direction of the underlying monthly change is uncertain. The permit decline, by contrast, is not subject to sampling error, though the estimate can still contain nonsampling error.

MetricSeptember 2026, revisedOctober 2026, preliminaryOctober change
Permits1,398,0001,295,000-7.4%
Starts1,191,0001,309,000+9.9% ±16.8%
Completions1,194,0001,272,000+6.5% ±13.2%

These are annualized rates, not counts of homes built or authorized during October alone. Nor does the one-month gap between permits and starts reveal how many permitted homes will break ground. They are different flows, and their contrast is a reason for caution—not proof of a specific future drop in construction.

The permit sequence gives the weaker pipeline signal some context: total permits slipped from 1.433 million in July to 1.403 million in August, 1.398 million in revised September, and 1.295 million in preliminary October. October was not an isolated dip in an otherwise rising series.

More Homes Finished, but the Comparison Has Two Caveats

Completions rose 6.5% from September to 1.272 million SAAR, but the margin of ±13.2 percentage points includes zero. The October level was also 10.4% below October 2025, with a ±11.6-point margin. That year-over-year change, too, is statistically uncertain. The data therefore do not establish either a meaningful monthly increase or a definitive annual decline.

The mix matters. October completions were estimated at 900,000 single-family units and 362,000 units in buildings with five or more units, both at annualized rates. Completions mark homes reaching the completed stage; they do not directly measure affordability, occupancy, or the available stock of homes for sale or rent.

The distinction between categories is clearer in permits. Single-family authorizations fell 8.7% month over month and 12.9% year over year. Permits for buildings with five or more units declined 5.2% month over month but just 0.4% year over year. That leaves single-family construction authorization notably weaker on the annual comparison.

It does not explain the starts rebound. The supplied figures do not include comparable September starts for larger multifamily buildings, so they cannot show whether that category drove October’s increase.

Regional Weakness Was Not Evenly Distributed

The regional permit figures add another layer to the national decline. October permits in the West were 20.0% below October 2025, the steepest annual drop among the four regions. The South recorded the sharpest monthly fall, down 10.7% from September, and was down 8.3% year over year.

RegionOctober vs. SeptemberOctober vs. October 20252026 year-to-date vs. 2025
Northeast-6.9%-6.9%+8.7%
Midwest+0.5%+8.1%+2.0%
South-10.7%-8.3%-4.4%
West-4.6%-20.0%-2.5%

The Northeast’s 8.7% year-to-date increase and the Midwest’s modest gains show that the national picture does not describe every region equally. Still, the regional table is about permits, not a complete measure of housing demand or construction conditions. It is not enough to establish a housing bubble—or to explain the causes of the regional differences.

The unadjusted count of authorized units not yet started offers a related, but limited, clue. It stood at 276,000 in October, down from 284,000 in September but above 259,500 in October 2025. This stock includes permits issued earlier and excludes canceled, abandoned, expired, and revoked permits. It cannot by itself reveal how long projects take to begin or why some have not started.

What to Watch in the Next Release

The next report should help determine whether October’s starts increase persists or gives way to the permit trend. The key test is not simply whether starts rise again: the current 9.9% estimate is too uncertain to carry that conclusion alone. A further increase whose confidence interval excludes zero would make the rebound more persuasive; another uncertain estimate alongside continued permit declines would leave the recovery case unresolved.

The release also reports revised September estimates, but the supplied October 20 release excerpt does not provide the original September figures. The direction and size of those revisions therefore cannot be calculated here. That limits any judgment about how much the latest revision changed the supply picture.

A separate measurement change is on the calendar: the redesigned survey sample begins with January 2027 estimates, and estimates for non-permit-issuing areas will be discontinued. Those areas currently account for less than 1% of the affected estimates. The change is small in stated coverage but important when comparing future data with earlier estimates.

For now, the record is clear on the contrast and cautious on the rebound: October starts rose in the point estimate, permits fell, and the starts increase was statistically uncertain. The next release will show whether construction activity confirms the bounce—or whether the weakening permit pipeline remains the sturdier signal.

Produced with AI assistance from public data releases. For information only, not investment advice. How we work

Related Articles