US Single-Family Housing Starts 2026

U.S. construction starts for privately owned one-unit housing were at a seasonally adjusted annual rate of 895,000 in June 2026. The estimate was nearly level with May's 897,000 and below the 1.017 million rate reported for March. A housing start is stronger evidence of active construction than a permit, but it still is not a completed home or proof of new rental inventory.

Single-family housing starts in 2026

These observations come from FRED series HOUST1F, sourced from the Census Bureau and published through New Residential Construction. The measure is monthly, seasonally adjusted, and annualized. PortfolioRental verified it on July 23, 2026. We expanded FRED's “thousands of units” display into units but otherwise left the values unchanged.

The two latest rates differ by only 2,000 at an annual pace. Given survey variability and revisions, that small gap should not be presented as meaningful proof of direction. The March reading also shows why a single month should not anchor a long-term supply assumption.

When Census records a start

New Residential Construction defines a start by the beginning of construction activity, generally excavation for footings or foundations. The precise treatment varies with foundation and construction type, which is why analysts should use Census definitions rather than informal descriptions such as “announced,” “site work,” or “under development.”

HOUST1F covers privately owned housing units in one-unit structures. It does not separately identify detached homes, intended rentals, for-sale inventory, custom homes, or build-to-rent communities. Like the permit series, its unit is a SAAR. June's value does not mean 895,000 foundations were started during that month.

A start has a different informational role from a permit. Authorization shows administrative permission; a start shows that construction crossed a defined threshold; completion indicates that the structure can be occupied under the program's standard. Each stage reduces some uncertainty, yet delivery timing, tenure, and market positioning still need separate evidence.

A stage-based local pipeline

National starts can trigger a market check, but they cannot replace one. Create a project register that advances units only when there is dated support for each stage:

1. Authorized: valid building permit tied to a parcel or controlled project.
2. Started: foundation activity or a recognized start documented through inspections, local records, or verified site evidence.
3. Vertical construction: framing or equivalent progress is visible and dated.
4. Near completion: late inspections, certificates in process, or confirmed builder schedule.
5. Completed: completion or occupancy evidence meets the relevant definition.
6. Available: the home is actually marketed for rent or sale.
7. Occupied: a closing or lease commencement is documented.

This structure prevents the same units from being counted as new supply at every milestone. It also makes slippage visible. A project can remain “started” for months; delivery assumptions should age and be refreshed rather than automatically rolling forward.

For single-family rental analysis, add tenure confidence. Purpose-built rental communities should be supported by project documents or direct marketing, not inferred from architectural similarity. Scattered one-unit starts may ultimately be owner occupied. Maintain separate totals for confirmed rental, confirmed for-sale, and unknown tenure.

Compare this stage with single-family building permits and later-stage housing completions. All three are available through the research library, but none alone establishes local absorption. Teams maintaining construction records across markets can review financial reporting and analysis for a repeatable reporting workflow.

Applying starts to portfolio decisions

In acquisition underwriting, identify projects within a defensible competitive radius and estimate delivery in ranges. A 200-home community early in foundation work should not be treated the same as 200 homes already leasing. Test timing and product overlap, then connect only relevant deliveries to occupancy or rent scenarios.

In asset management, the pipeline can guide leasing calendars and competitive surveys. If a nearby project progresses toward availability, collect its advertised rents, concessions, unit plans, deposit terms, and actual opening date. If construction pauses, update the watchlist rather than leaving an obsolete delivery assumption in the budget.

For market selection, normalize carefully. A large starts count may simply reflect a large housing base. Ratios to households or existing units can add scale, but they require compatible geography and dates. Also distinguish growth from competition: construction may reflect demand while simultaneously adding alternatives for residents. The data do not resolve that tension without additional evidence.

Research method and uncertainty

We transcribed five HOUST1F observations and reviewed Census definitions, release materials, historical tables, and survey documentation. The latest values may be preliminary and subject to revision. The Survey of Construction is sample-based; Census publishes estimates, standard errors, and details on estimation. Users should consult those materials before treating short-run differences as statistically meaningful.

Weather, construction practice, reporting, and seasonal adjustment can influence monthly movement. National aggregation masks regions where activity diverges. The series also omits essential underwriting attributes: land basis, financing, builder capacity, sale price, rent, quality, expected completion, and tenure.

Reproducibility requires a dated source package. Save the New Residential Construction release, the HOUST1F export, and the observation vintage. Keep local project records separately because national revisions do not update parcel conclusions. Where two sources disagree on stage, document both, investigate the definition and date, and reduce confidence until reconciled.

Verification checklist

• Confirm the code is HOUST1F and not total housing starts.
• Retain the seasonally adjusted annual-rate unit.
• Do not sum monthly SAAR observations.
• Check preliminary months for later revisions.
• Review survey standard errors for short-run comparisons.
• Apply a written start definition to local projects.
• Prevent duplicate counting across permit, start, and completion stages.
• Record construction status and evidence date.
• Separate confirmed rental tenure from unknown tenure.
• Refresh delivery ranges before each material decision.

Frequently asked questions

What was the June 2026 single-family starts rate?

The published estimate was 895,000 one-unit homes at a seasonally adjusted annual rate.

Is a housing start the same as a permit?

No. A permit is an authorization. A start records the beginning of construction under the Census program's definition.

Does a start mean a home is ready to lease?

No. Substantial work can remain, and HOUST1F does not identify rental tenure or availability.

Why can a starts estimate be revised?

The monthly estimate is based on survey and administrative inputs that are updated as more information arrives. Census revises recent data accordingly.

How should a rental investor use the series?

Use it to frame national and regional construction context, then rely on dated local project stages, likely tenure, product match, and actual leasing evidence for decisions.

Sources and verification dates

1. Federal Reserve Bank of St. Louis: New Privately-Owned Housing Units Started: Single-Family Units, accessed July 23, 2026.
2. U.S. Census Bureau: New Residential Construction, accessed July 23, 2026.
3. U.S. Census Bureau: New Residential Construction Definitions, accessed July 23, 2026.
4. U.S. Census Bureau: Survey of Construction Methodology, accessed July 23, 2026.
5. U.S. Census Bureau: New Residential Construction Data, accessed July 23, 2026.
6. U.S. Census Bureau: Survey of Construction Release Schedule, accessed July 23, 2026.
7. Federal Reserve Bank of St. Louis: One-Unit Building Permits, accessed July 23, 2026.
8. Federal Reserve Bank of St. Louis: New One-Family Houses Completed, accessed July 23, 2026.
9. U.S. Census Bureau: Seasonal Adjustment, accessed July 23, 2026.
10. Federal Reserve Bank of St. Louis: ALFRED Vintage Data, accessed July 23, 2026.

For help maintaining a current local construction register, contact PortfolioRental.