USA · snapshot 2026-10-19 00:23 UTC · data to 2026-08-01
US housing construction
Is US home building recovering, and what is holding it back?
US housing starts was 1,275 thousands of units, seasonally adjusted annual rate in 2026-08-01, down 1.2 percent over one year, a move at the 6th percentile of its one-year changes.
Section 1 of 9
What changed?
AI narrative Modeled
US housing starts stood at 1275 thousands of units seasonally adjusted annual rate in 2026-08-01 [obs:410402], down 16 thousands of units or 1.2 percent from 1291 thousands one year earlier at the 6th percentile of one-year changes [obs:169872]. Building permits reached 1403 thousands of units [obs:410506], up 56 thousands or 4.2 percent from 1347 thousands one year earlier [obs:164603]. Residential construction spending was 894597 millions of dollars [obs:410934], down 4.8 percent from one year earlier [obs:182533].
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run 9cd0ceb8, prompt 1.0.0. It adds no facts.
US housing starts
Privately owned housing units started, annual rate.
1,275 thousands of units, seasonally adjusted annual rate
2026-08-01
| Span | From | Change | Percentile |
|---|---|---|---|
| 1 period | 1,309 (2026-07-01) | -34 (-2.6%) | 25th |
| 1 year | 1,291 (2025-08-01) | -16 (-1.2%) | 6th of 799 |
| 5 years | 1,596 (2021-08-01) | -321 (-20.1%) | 49th of 751 |
| 10 years | 1,171 (2016-08-01) | 104 (+8.9%) | 10th of 691 |
US building permits
Permits lead starts by one to three months.
1,403 thousands of units, seasonally adjusted annual rate
2026-08-01
| Span | From | Change | Percentile |
|---|---|---|---|
| 1 period | 1,433 (2026-07-01) | -30 (-2.1%) | 36th |
| 1 year | 1,347 (2025-08-01) | 56 (+4.2%) | 21st of 787 |
| 5 years | 1,740 (2021-08-01) | -337 (-19.4%) | 44th of 739 |
| 10 years | 1,201 (2016-08-01) | 202 (+16.8%) | 16th of 679 |
US residential construction spending
Residential construction spending.
894,597 millions of dollars, seasonally adjusted annual rate
2026-08-01
| Span | From | Change | Percentile |
|---|---|---|---|
| 1 period | 884,995 (2026-07-01) | 9,602 (+1.1%) | 69th |
| 1 year | 940,035 (2025-08-01) | -45,438 (-4.8%) | 36th of 283 |
| 5 years | 821,402 (2021-08-01) | 73,195 (+8.9%) | 9th of 235 |
| 10 years | 487,744 (2016-08-01) | 406,853 (+83.4%) | 60th of 175 |
Section 2 of 9
How unusual is it?
AI narrative Modeled
The latest starts value sits at the 31th percentile of the series history [obs:169229] [obs:169676]. The one-year change ranks at the 25th percentile of past changes with robust z of -0.3725 and one period since a larger move [obs:169872].
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run 9cd0ceb8, prompt 1.0.0. It adds no facts.
- Latest value percentile
- 31st of 811 values
- Latest change percentile
- 25th of 810 changes
- Robust z of latest change
- -0.37
- Largest move
- 1 period since a larger move
- Record level
- no
- Previous comparable move
- 2026-07-01: -130 thousands of units, seasonally adjusted annual rate (obs 410401, 410400)
- Prior maximum
- 2,494 (1972-01-01, obs 169229)
- Prior minimum
- 478 (2009-04-01, obs 169676)
Section 3 of 9
Why might it be happening?
AI narrative Modeled
Documented mechanism: a higher policy rate raises borrowing cost which reduces construction financing and residential building [mechanism:7006539d-a8a2-4038-86d0-112809badec4]. Statistical model drivers are not separated in the packet. Hypotheses are not listed. Unknowns: construction labour availability is not stored as a series; local permitting and zoning constraints do not appear in national totals.
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run 9cd0ceb8, prompt 1.0.0. It adds no facts.
Documented mechanisms
- A higher policy rate raises borrowing cost, which reduces construction financing and residential buildingMechanism supportedpolicy-interest-rate → construction-financing → residential-construction → housing
Model-supported drivers
No active forecast for this topic records driver contributions (the champion may be a baseline with no drivers).
Hypotheses
No candidate or hypothesized relationship touches this topic's entities.
Unknowns
- Construction labour availability is not stored as a series.
- Local permitting and zoning constraints do not appear in national totals.
Section 4 of 9
What has not changed?
AI narrative Modeled
US unemployment rate at 4.2 percent shows year-over-year change inside the 10th to 90th percentile band of -1.1 to 1.7 percent [obs:411088]. Average hourly earnings at 37.81 dollars per hour show year-over-year change inside the 10th to 90th percentile band of 0.44 to 1.47 dollars [obs:410816]. Nonfarm payroll employment at 159044 thousands shows year-over-year change inside the 10th to 90th percentile band of -1112.3 to 3450.9 thousands [obs:410852].
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run 9cd0ceb8, prompt 1.0.0. It adds no facts.
| Series | Latest | Basis | Normal band (p10 to p90) | Historical range | Reading |
|---|---|---|---|---|---|
| US unemployment rate Observed Unemployment rate; household income supports demand. | 4.2 percent 2026-09-01, obs 411088 | year-over-year change -0.2 (-4.5%) 931 past changes | -1.1 to 1.7 | 2.5 (1953-05-01) to 14.8 (2020-04-01) since 1948-01-01; latest at the 22nd percentile of levels | within its normal band |
| US average hourly earnings, all private employees Observed Average hourly earnings. | 37.81 US dollars per hour 2026-09-01, obs 410816 | year-over-year change 1.11 (+3.0%) 234 past changes | 0.44 to 1.47 | 20.04 (2006-03-01) to 37.81 (2026-09-01) since 2006-03-01; latest at the 100th percentile of levels | within its normal band |
| US nonfarm payroll employment Observed Nonfarm payrolls. | 159,044 thousands of persons 2026-09-01, obs 410852 | year-over-year change 496 (+0.3%) 1040 past changes | -1,112.3 to 3,450.9 | 29,923 (1939-01-01) to 159,044 (2026-09-01) since 1939-01-01; latest at the 100th percentile of levels | within its normal band |
Section 5 of 9
What absorbs the shock?
AI narrative Modeled
Financing buffers are evidenced by the federal funds effective rate at 3.75 percent [obs:410817] and 10-year Treasury yield at 5.1 percent [obs:415578]. Inventory of homes for sale is not measured. Substitution for softwood lumber lists 0 substitutes. Imports of lumber and fixtures are not ingested. Demand elasticity is not measured.
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run 9cd0ceb8, prompt 1.0.0. It adds no facts.
US federal funds effective rate
3.75 percent · 2026-09-01 · yoy -11.1% · obs 410817
The policy rate; mortgage rates follow with a lag.
US 10-year Treasury yield
5.1 percent · 2026-10-15 · yoy +25.9% · obs 415578
The 10-year yield anchors mortgage pricing.
Not measured
No stored series evidences this buffer.
Homes for sale and months of supply are not ingested.
Softwood lumber
The graph records no substitute for this entity.
Substitutes recorded on the graph for framing materials.
Not measured
No stored series evidences this buffer.
Lumber and fixture imports are not ingested.
Not measured
No stored series evidences this buffer.
Rate sensitivity of buyers is not measured here.
Section 6 of 9
What response is underway?
AI narrative Modeled
Residential construction spending shows year-over-year change of -4.8 percent [obs:410934]. Cement production shows year-over-year change of -1.2 percent [obs:263553]. Nonfarm payroll employment shows year-over-year change of 0.3 percent [obs:410852]. No modeled capacity pipeline values are given.
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run 9cd0ceb8, prompt 1.0.0. It adds no facts.
| Group | Series | Latest | Year earlier | Change |
|---|---|---|---|---|
| Capital spending | US residential construction spending Observed | 894,597 millions of dollars, seasonally adjusted annual rate (2026-08-01, obs 410934) | 940,035 (2025-08-01) | -45,438 (-4.8%) |
| Production | Cement production Observed | 84,000 thousand metric tons (2025, obs 263553) | 85,000 (2024) | -1,000 (-1.2%) |
| Workforce | US nonfarm payroll employment Observed | 159,044 thousands of persons (2026-09-01, obs 410852) | 158,548 (2025-09-01) | 496 (+0.3%) |
The US pipeline and response measures are on the capacity page; bottleneck assessments are on bottlenecks.
Section 7 of 9
Scenarios
AI narrative Modeled
At the 12-month horizon the model last_value estimates base case 1275 thousands of units with 80 percent interval 1158 to 1387 [forecast:6e11beba-0f1b-40e2-a7c0-cae478cf3e04]. At the 6-month horizon the same model estimates base case 1275 thousands with 80 percent interval 1143 to 1390 [forecast:ca67a30e-9325-4d40-80b4-dce1d9c11a68]. The model assigns 10 percent probability to values below the lower bound and 10 percent above the upper bound in each case.
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run 9cd0ceb8, prompt 1.0.0. It adds no facts.
US housing starts
Base case
1,275 thousands of units, seasonally adjusted annual rate
The point estimate: the model's central path.
Upside
1,386.82 thousands of units, seasonally adjusted annual rate
Upper bound of the interval: what would need to go better than the central path.
Downside
1,158.22 thousands of units, seasonally adjusted annual rate
Lower bound of the interval: what would need to go worse than the central path.
The model assigns 80 percent to the interval; the site assigns no other probability.
Falsifiers and assumptions
- If the first published value of us-housing-starts (USA) for 2027-08 is below 1,158 thousands of units, seasonally adjusted annual rate or above 1,387 thousands of units, seasonally adjusted annual rate, the 80 percent interval is falsified.
- The model assigns a 10 percent probability to a value below 1,158 thousands of units, seasonally adjusted annual rate and a 10 percent probability to a value above 1,387 thousands of units, seasonally adjusted annual rate; interval misses should occur about one time in five over many forecasts, and a run of misses well above that rate falsifies the calibration.
- Assumption: No structural break in us-housing-starts between the data cutoff and the target period.
- Assumption: The source keeps publishing the series on the same definition; a rebasing or redefinition invalidates the comparison.
- Assumption: The most recent value is the best available guess: recent changes carry no persistent information.
US housing starts
Base case
1,275 thousands of units, seasonally adjusted annual rate
The point estimate: the model's central path.
Upside
1,389.72 thousands of units, seasonally adjusted annual rate
Upper bound of the interval: what would need to go better than the central path.
Downside
1,142.78 thousands of units, seasonally adjusted annual rate
Lower bound of the interval: what would need to go worse than the central path.
The model assigns 80 percent to the interval; the site assigns no other probability.
Falsifiers and assumptions
- If the first published value of us-housing-starts (USA) for 2027-02 is below 1,143 thousands of units, seasonally adjusted annual rate or above 1,390 thousands of units, seasonally adjusted annual rate, the 80 percent interval is falsified.
- The model assigns a 10 percent probability to a value below 1,143 thousands of units, seasonally adjusted annual rate and a 10 percent probability to a value above 1,390 thousands of units, seasonally adjusted annual rate; interval misses should occur about one time in five over many forecasts, and a run of misses well above that rate falsifies the calibration.
- Assumption: No structural break in us-housing-starts between the data cutoff and the target period.
- Assumption: The source keeps publishing the series on the same definition; a rebasing or redefinition invalidates the comparison.
- Assumption: The most recent value is the best available guess: recent changes carry no persistent information.
Section 8 of 9
What to watch
AI narrative Modeled
US building permits at 1403 thousands show year-over-year percent 4.1574 above threshold 0 with status easing [obs:410506]. US housing starts at 1275 thousands show year-over-year percent -1.2393 below threshold 0 with status worsening [obs:410402]. Federal funds effective rate at 3.75 percent is below threshold 4 with status easing [obs:410817]. 10-year Treasury yield at 5.1 percent is above threshold 4 with status worsening [obs:415578].
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run 9cd0ceb8, prompt 1.0.0. It adds no facts.
On the easing side of the threshold · rising confirms easing
Permits turning up year over year lead starts.
On the worsening side of the threshold · rising confirms easing
Starts rising year over year confirm the recovery in the metric itself.
On the easing side of the threshold · falling confirms easing
A policy rate below 4 percent has historically coincided with stronger building.
On the worsening side of the threshold · falling confirms easing
The 10-year yield below 4 percent brings mortgage rates toward the range that clears more buyers.
Section 9 of 9
Forecast track record
AI narrative Modeled
For the 6-month horizon the champion last_value model shows backtest MAE 91.119 over 42 cutoffs and 80 percent interval coverage 83 percent with skill versus last value 0 percent and no live evaluations yet. For the 12-month horizon the same model shows backtest MAE 87.000 over 36 cutoffs and 80 percent interval coverage 97 percent with skill versus last value 0 percent and no live evaluations yet.
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run 9cd0ceb8, prompt 1.0.0. It adds no facts.
| Forecast | Champion | Backtest MAE | Skill vs last value | 80% coverage | Live evaluations |
|---|---|---|---|---|---|
| us-housing-starts:USA:6 | last_value | 91.12 over 42 cutoffs | 0% | +83% | none yet |
| us-housing-starts:USA:12 | last_value | 87 over 36 cutoffs | 0% | +97% | none yet |
The full scorecard is at /forecasts/scorecard.
Provenance
- Snapshot
- 53 · 2026-10-19 00:23 UTC · navigator_view_v1+navigator_topics_v1
- Gaps stated
- 0
- AI narrative
- grok-4.3, run 9cd0ceb8-9341-427d-a28b-0a90a6655109, prompt 1.0.0, 17 ids cited
- Earlier snapshots
- 45 (2026-10-12 00:12 UTC), 37 (2026-10-05 00:12 UTC), 29 (2026-09-28 00:13 UTC), 21 (2026-09-21 00:13 UTC), 13 (2026-09-14 00:15 UTC), 5 (2026-09-07 22:35 UTC)
Every value on this page is a stored observation (obs ids shown), a forecast record, a mechanism or a relationship row, or a percentile computed from those the same way the relevance engine computes its features. Series the config references but the site does not store are listed as gaps under their section, never filled in.