Where the housing sector is in its own cycle, which timing stage is moving, and whether housing is diverging from the rest of the economy. Counts and records, never probabilities.
basis: as known Data through: September 2026 · updated Oct 20 9 voters · 3 timing stages · since 1980 comparator: EB ex-housing
Live census reads Mixed; the label changes after three months.
Colored by housing-favorable direction (declared per driver, not voter polarity), on the heatmap's own five bands — one print, one color (HC-DR-1 Amendment 1).
Spell rule: 3+ consecutive months in one state, symmetric for all four; horizon 24 months.
| Lens | Today | Since | Past runs |
|---|---|---|---|
| Divergence state | Both expanding | 1 month (no spell yet) | 29 |
| Stage label | Downturn spreading | Aug 2026 · 2 months | 11 |
| Both lenses lit | no — the divergence lens has no spell yet | 8 episodes with both | |
Union of the two records: 31 distinct episodes since Mar 1980 — 11 stage-lens runs and 29 divergence spells, the two populations counted separately; 8 where both lenses lit. Of the 8 closed episodes where both lenses lit, 3 were followed by an NBER peak within 24 months and 5 were not — counts, not probabilities.
Faded = shallow (gap depth below these spells' median, the 66th percentile of all gaps). Stacked = both lenses matched today (8).
| Episode | Stage lens | Divergence lens | Both | Stage then · triple | Held | Resolved into | Economy within 24 mo | Starts, 24 mo |
|---|---|---|---|---|---|---|---|---|
| Aug 2026 – open | Aug 2026 – Sep 2026 | — | — | Downturn spreading · →↓→ | 2 months | open | open | — |
| Feb 2024 – May 2024 | — | Feb 2024 – May 2024 | — | Expansion · ↑→→ | 4 months | Expansion | no NBER peak within 24 months | −19.0% |
| Nov 2022 – Dec 2023 | Nov 2022 – Sep 2023 | Jun 2023 – Dec 2023 | ✓ | Downturn spreading · ↓↓↑ | 14 months | Expansion | no NBER peak within 24 months | +27.7% |
| Jun 2021 – Jun 2022 | — | Jun 2021 – Jun 2022 | — | Late expansion · →↑↑ | 13 months | Late expansion | no NBER peak within 24 months | −19.9% |
| Aug 2020 – Apr 2021 | — | Aug 2020 – Apr 2021 | — | Expansion · ↑→↑ | 9 months | Late expansion | no NBER peak within 24 months | −11.0% |
| Dec 2018 – Mar 2020 | Dec 2018 – Sep 2019 | Apr 2019 – Mar 2020 | ✓ | Downturn spreading · ↓↓↑ | 16 months | Expansion | NBER peak Feb 2020 (+14 months) | +64.3% |
| Oct 2012 – Oct 2018 | — | Oct 2012 – Oct 2018 | — | Recovery · ↑↑↓ | 73 months | Expansion | no NBER peak within 24 months | +14.4% |
| Nov 2010 – Aug 2012 | Nov 2010 – Jun 2012 | May 2012 – Aug 2012 | ✓ | Downturn spreading · →↓→ | 22 months | Recovery | no NBER peak within 24 months | +25.9% |
| Aug 2006 – Jan 2007 | Aug 2006 – Jan 2007 | — | — | Downturn spreading · ↓↓→ | 6 months | Contraction | NBER peak Dec 2007 (+16 months) | −55.5% |
| Jun 2003 – Apr 2006 | — | Jun 2003 – Apr 2006 | — | Late expansion · ↑→↑ | 35 months | Late expansion | no NBER peak within 24 months | +13.6% |
| Feb 2003 – Apr 2003 | — | Feb 2003 – Apr 2003 | — | Recovery · ↑↑→ | 3 months | Late expansion | no NBER peak within 24 months | +37.7% |
| Jul 2002 – Sep 2002 | — | Jul 2002 – Sep 2002 | — | Recovery · ↑↑↑ | 3 months | Recovery | no NBER peak within 24 months | +26.6% |
| Mar 2002 – May 2002 | — | Mar 2002 – May 2002 | — | Recovery · ↑↑↑ | 3 months | Recovery | no NBER peak within 24 months | +26.3% |
| Sep 2000 – Apr 2001 | Sep 2000 – Apr 2001 | — | — | Downturn spreading · →↓↑ | 8 months | Recovery | NBER peak Mar 2001 (+6 months) | +21.0% |
| Dec 1999 – Apr 2000 | — | Dec 1999 – Apr 2000 | — | Expansion · →↑↑ | 5 months | Expansion | NBER peak Mar 2001 (+15 months) | −6.5% |
| Aug 1997 – Oct 1999 | Aug 1997 – Jul 1998 | Oct 1997 – Dec 1997; Feb 1998 – Oct 1999 | ✓ | Downturn spreading · →↓→ | 27 months | Expansion | no NBER peak within 24 months | +15.0% |
| May 1995 – Sep 1996 | May 1995 – Sep 1995 | Aug 1995 – Sep 1996 | ✓ | Downturn spreading · →↓→ | 17 months | Expansion | no NBER peak within 24 months | +7.9% |
| Apr 1994 – Jul 1994 | — | Apr 1994 – Jul 1994 | — | Recovery · ↓↑↑ | 4 months | Recovery | no NBER peak within 24 months | +0.4% |
| Jul 1993 – Feb 1994 | — | Jul 1993 – Feb 1994 | — | Bottoming · ↑↑↑ | 8 months | Recovery | no NBER peak within 24 months | +4.3% |
| Aug 1992 – Jan 1993 | — | Aug 1992 – Jan 1993 | — | Bottoming · →↑→ | 6 months | Bottoming | no NBER peak within 24 months | +10.8% |
| Feb 1992 – Apr 1992 | — | Feb 1992 – Apr 1992 | — | Bottoming · ↑→↓ | 3 months | Bottoming | no NBER peak within 24 months | +2.8% |
| Jun 1989 – Nov 1990 | Jun 1989 – Nov 1990 | Dec 1989 – Apr 1990 | ✓ | Downturn spreading · →↓↑ | 18 months | Contraction | NBER peak Jul 1990 (+13 months) | −10.2% |
| Dec 1988 – Mar 1989 | — | Dec 1988 – Mar 1989 | — | Late expansion · →↓↑ | 4 months | Late expansion | NBER peak Jul 1990 (+19 months) | −33.6% |
| Aug 1988 – Oct 1988 | — | Aug 1988 – Oct 1988 | — | Late expansion · →→↑ | 3 months | Late expansion | NBER peak Jul 1990 (+23 months) | −25.4% |
| Apr 1988 – Jun 1988 | — | Apr 1988 – Jun 1988 | — | Late expansion · ↑→↑ | 3 months | Late expansion | no NBER peak within 24 months | −14.0% |
| Jun 1987 – Oct 1987 | — | Jun 1987 – Oct 1987 | — | Late expansion · ↓→↑ | 5 months | Late expansion | no NBER peak within 24 months | −10.9% |
| Nov 1985 – Apr 1987 | — | Nov 1985 – Apr 1987 | — | Recovery · ↑→→ | 18 months | Late expansion | no NBER peak within 24 months | +6.9% |
| Feb 1985 – May 1985 | — | Feb 1985 – May 1985 | — | Late expansion · ↑↑→ | 4 months | Recovery | no NBER peak within 24 months | +19.0% |
| Nov 1983 – Mar 1984 | — | Nov 1983 – Mar 1984 | — | Recovery · →↑→ | 5 months | Recovery | no NBER peak within 24 months | −6.0% |
| Aug 1981 – Sep 1983 | Aug 1981 – Nov 1982 | Nov 1982 – Sep 1983 | ✓ | Downturn spreading · ↓↓→ | 26 months | Recovery | no NBER peak within 24 months | +89.0% |
| Mar 1980 – Feb 1981 | Mar 1980 – Feb 1981 | Sep 1980 – Nov 1980 | ✓ | Downturn spreading · ↓↓↑ | 12 months | Late expansion | NBER peak Jul 1981 (+16 months) | −3.8% |
How to read it: every row is an episode where at least one lens matched today's state. Each lens column states the dates that lens actually matched; a dash means it did not match during the episode. Both ✓ marks the stacked-lane cases. “Economy within 24 mo” lists the first NBER peak within the horizon of the episode's start, or none — individually, never as a rate.
| Spell | Held | Max gap | Depth pct | Economy within 24 mo | Starts, 24 mo |
|---|---|---|---|---|---|
| Feb 2024 – May 2024 | 4 months | 19.7 | 40th | no NBER peak within 24 months | −19.0% |
| Jun 2023 – Dec 2023 | 7 months | 26.3 | 52nd | no NBER peak within 24 months | −1.6% |
| Jun 2021 – Jun 2022 | 13 months | 48.0 | 78th | no NBER peak within 24 months | −19.9% |
| Aug 2020 – Apr 2021 | 9 months | 32.8 | 62nd | no NBER peak within 24 months | −11.0% |
| Apr 2019 – Mar 2020 | 12 months | 39.9 | 69th | NBER peak Feb 2020 (+10 months) | +21.3% |
| Oct 2012 – Oct 2018 | 73 months | 52.5 | 81st | no NBER peak within 24 months | +14.4% |
| May 2012 – Aug 2012 | 4 months | 38.9 | 69th | no NBER peak within 24 months | +25.9% |
| Jun 2003 – Apr 2006 | 35 months | 45.0 | 75th | no NBER peak within 24 months | +13.6% |
| Feb 2003 – Apr 2003 | 3 months | 28.1 | 56th | no NBER peak within 24 months | +37.7% |
| Jul 2002 – Sep 2002 | 3 months | 12.3 | 28th | no NBER peak within 24 months | +26.6% |
| Mar 2002 – May 2002 | 3 months | 20.5 | 41st | no NBER peak within 24 months | +26.3% |
| Dec 1999 – Apr 2000 | 5 months | 24.6 | 48th | NBER peak Mar 2001 (+15 months) | −6.5% |
| Feb 1998 – Oct 1999 | 21 months | 41.5 | 71st | no NBER peak within 24 months | +1.5% |
| Oct 1997 – Dec 1997 | 3 months | 36.3 | 66th | no NBER peak within 24 months | +16.2% |
| Aug 1995 – Sep 1996 | 14 months | 48.4 | 78th | no NBER peak within 24 months | −3.3% |
| Apr 1994 – Jul 1994 | 4 months | 57.8 | 85th | no NBER peak within 24 months | +0.4% |
| Jul 1993 – Feb 1994 | 8 months | 35.6 | 65th | no NBER peak within 24 months | +4.3% |
| Aug 1992 – Jan 1993 | 6 months | 22.2 | 45th | no NBER peak within 24 months | +10.8% |
| Feb 1992 – Apr 1992 | 3 months | 4.4 | 11th | no NBER peak within 24 months | +2.8% |
| Dec 1989 – Apr 1990 | 5 months | 21.4 | 43rd | NBER peak Jul 1990 (+7 months) | +4.2% |
| Dec 1988 – Mar 1989 | 4 months | 60.3 | 87th | NBER peak Jul 1990 (+19 months) | −33.6% |
| Aug 1988 – Oct 1988 | 3 months | 35.9 | 66th | NBER peak Jul 1990 (+23 months) | −25.4% |
| Apr 1988 – Jun 1988 | 3 months | 47.0 | 77th | no NBER peak within 24 months | −14.0% |
| Jun 1987 – Oct 1987 | 5 months | 37.2 | 68th | no NBER peak within 24 months | −10.9% |
| Nov 1985 – Apr 1987 | 18 months | 58.1 | 85th | no NBER peak within 24 months | +6.9% |
| Feb 1985 – May 1985 | 4 months | 33.8 | 62nd | no NBER peak within 24 months | +19.0% |
| Nov 1983 – Mar 1984 | 5 months | 75.6 | 95th | no NBER peak within 24 months | −6.0% |
| Nov 1982 – Sep 1983 | 11 months | 51.3 | 80th | no NBER peak within 24 months | +26.9% |
| Sep 1980 – Nov 1980 | 3 months | 30.3 | 59th | NBER peak Jul 1981 (+10 months) | −32.4% |
The expand view keeps the two lanes on top and adds, on the same clock beneath them, the two lines behind the divergence lane (housing net and EB ex-housing) with Momentum as a lighter third line. The divergence lane lights on every run of 3 months or more with both lines on today's sides of zero; the stage lane lights where the stage label matches today's. Hover any block for its row.
Data through: September 2026 · updated Oct 20 · today +11.1 · 46th percentile of its own history · Downturn spreading · median member at the 58th percentile of its own history
The bars are the declared breadth bars (narrow inside ±20, moderate to ±50, broad beyond) — regions of a count, nothing about what follows. The last 20 years shown; expand the chart to reach the full record back to 1980. Where is housing? Occupancy since 1980: Expansion 186 · Downturn spreading 120 · Recovery 102 · Late expansion 62 · Bottoming 46 · Contraction 33 · Plateau 10 months.
Each panel is the stage's mean own-history percentile rank (0–100, expanding window). Default: today's window, so the panels locate now; the 2004–2010 toggle shows the same three panels at the same scale for that window — the 2006 shape shown, not asserted. How is the pipeline emptying?
| Peak (12-mo smoothed) | Decline at trough | Trough | NBER peak within 24 mo? | Lead (peak → onset) | Lead (rule fired → onset) | Regained | Starts, 24 mo |
|---|---|---|---|---|---|---|---|
| May 2024 → Sep 2026 to date | −10.1% to date | open | open | 28 months since peak | does not qualify (the >15% rule) | ||
| Jul 2021 | −25.5% | May 2023 | no | — | Feb 2023 · no NBER peak within 24 mo | open | −12.1% |
| Feb 2006 | −75.7% | Jun 2011 | yes · Dec 2007 | 22 months | Dec 2006 · NBER peak +12 months | open | −59.9% |
| Mar 1987 | −35.5% | Jun 1991 | no | — | May 1990 · NBER peak +2 months | Apr 1998 | −18.5% |
| Dec 1978 | −60.7% | Jul 1982 | yes · Jan 1980 | 13 months | Jan 1980 · NBER peak +0 months | Dec 1998 | −36.0% |
| Feb 1973 | −44.4% | May 1975 | yes · Nov 1973 | 9 months | Jan 1974 · recession under way (−2 months) | Aug 1977 | −50.3% |
| Nov 1968 | −16.5% | Jun 1970 | yes · Dec 1969 | 13 months | Apr 1970 · recession under way (−4 months) | Mar 1971 | +5.6% |
| Feb 1964 | −30.0% | Apr 1967 | no | — | Oct 1966 · no NBER peak within 24 mo | May 1971 | −21.7% |
7 qualifying declines since 1960 — counted from the smoothed peak, 4 followed by an NBER peak within 24 months, 3 not; counted from the month the rule fired, 3 followed by an NBER peak within 24 months, 2 already inside a recession when it fired, 2 with none. The record includes every no-recession case: it is not conditioned on recessions existing. In all, 16 smoothed peaks were examined; 4 near-misses fell below the rule: Dec 1960 (−6.0%), Apr 1984 (−7.2%), Sep 1994 (−8.4%), May 2024 (−10.1%). The peak lead is measured from the smoothed peak, which the 12-month rule can name only about a year after it occurs. The rule fired at the month-end by which the breach month's permits had been published — the month after the smoothed series first fell more than 15% below the peak. What followed is counted from there: a minus means the recession had already begun. Today’s row is a position on the same axes.
The limitation of record. The census describes where the housing sector is; it does not forecast where it goes. Measured before the roster was chosen, the correlation between the census's net breadth and the following twelve months' change in single-family starts is about 0.1 (0.07 to 0.26 for every roster examined). Every record here lists what followed as individual rows — counts, never probabilities.
Deliberately out: price forecasts · buy/sell timing · any probability · S&P outcomes (other pages’ job) · demographic demand · rental and regional views (filed for v2).
9 voters, 3 per timing stage · deferred: Existing-home sales (joins on burn-in, ~2030-04) · cell of record: 3-month window, p50 band · persistence 3 months · comparator EB ex-housing · descriptive only — counts, never probabilities