Markets · Issue 02
The crash from CAPE 27 was deeper than the one from 44
X is Shiller CAPE on 1 January. Y is the cycle drawdown that followed. 2007 launched from 27.2 and fell 57%. 2000 launched from 43.8 and fell 49%. 2022 from 36.9, −25%. A high multiple is not a date.
- Jan 2007 CAPE, then GFC depth
- 27 → 57%
- Jan 2000 CAPE, then −49%
- 43.8
- Jan 2022 CAPE, then −25%
- 36.9
- CAPE, 2 Sep 2026 — no crash row
- 41.93
The cheaper start was the worse crash
January 2007’s CAPE was 27.21. January 2000’s was 43.77. The crash that followed 2007 was −56.8% and 49 months back to the old high. The crash that followed 2000 was −49.1% and 56 months. Read that twice. The starting multiple did not rank the depth. 2007 was a banking-system event that launched from a multiple the 2010s would have called reasonable. 2022 left from 36.94 and fell 25% — the dearest start of the four, the shallowest hole. CAPE 42 in September 2026 is a statement about prospective real returns. It is not a timestamp.
Size is the wait. 2020 is the exception, again.
The disc is months from the closing low back to the prior peak, the Y-axis of Deeper takes longer. 2000 and 2007 are large, slow recoveries. 2022 is a medium 15. 2020 is a speck: −34% and whole in five months, from a January CAPE of 31. A high multiple plus a policy flood is a V. A middling multiple plus a broken banking system is a hole. We will not put 2 September 2026’s 41.93 on this wall with a made-up Y. There is no 20% episode to attach.
The long view is two pages, not one slogan
Hold a lower 10-year real equity assumption because 42 is not 17. Hold a crash playbook because 2007 launched from 27. Do not collapse them into ‘it must crash this year.’ What would change this page: a new cycle we can finally give a Y, or a January 1 CAPE back under 25. Until then four bubbles are the whole argument.
Investing lens
Horizon 10–15 years · Educational, not advice
CAPE 42 lowers prospective real returns. It does not date a −57%. Size for the GFC because it launched from 27, and lower the return assumption because 42 is not 17.
Where the map points
- A lower assumed 10-year real equity return than the 2010s
- The crash bubble’s 48-month clock, independent of CAPE
- Avoid a 2026 calendar that treats 42 as a date
- Quality cash-flow over multiple expansion
What can break it
- A further multiple expansion that makes 42 look early
- A crash from 42 that is 2000-shaped (years) not 2020-shaped (weeks)
- Using January 1 CAPE as if it were the peak print
CHART does not recommend securities, funds or trades. Figures can be revised by their publishers. Do your own research and consider regulated advice before allocating capital.
Sources
Every headline number traces to a named publisher. Contextual sources inform the essay, not the key stat.
- 01PrimaryMultpl (Shiller)2026-09-02Shiller PE Ratio
CAPE 41.93 on 2 Sep 2026. Record 44.19 in Dec 1999. Long-run mean 17.40.
- 02PrimaryMultpl2026Shiller PE Ratio by Year
January 1 prints used on the CAPE line: 2000 43.77; 2007 27.21; 2009 15.17; 2026 39.65.
- 03PrimaryRobert Shiller2026U.S. Stock Markets 1871–Present
Underlying CAPE series Multpl reprints.
- 04PrimaryHartford Funds / Ned Davis Research2025-0310 Things You Should Know About Bear Markets
S&P 500 close-to-close declines of 20% or more, 1929–2024. As of 31 Mar 2025. Average −35% across 20% episodes. Cycle depths (1929 −86%, 2008 −57%) are a different ruler, used on our crash bubble.
- 05PrimaryReuters2013-03-28S&P 500 closes above 2007 peak
28 Mar 2013 close 1,569.19, first close above 9 Oct 2007’s 1,565.15.
Keep reading
CAPE is 42. It has only been higher once.
Shiller via Multpl: January 2000 43.8, the record 44.2 in December 1999. 2 September 2026: 41.93. January 2009: 15.2. A high CAPE is not a crash date. 2007’s crash arrived from 27.
41.93
Shiller CAPE, 2 Sep 2026 (Multpl)
The deeper the crash, the longer the wait. Except twice.
S&P 500 price, peak to trough versus months from the low back to the old high. 1929 is a different animal. 2020 is the other exception: −34% and whole in five months. 2022 was −25% and fifteen. There is no 2026 crash on this wall.
5 mo
2020: trough to prior peak (price)
Minus 86, then 57, then 49. The modern crash is not 1929.
S&P 500 cycle drawdowns, price. 1929–32 −86.2%. 2007–09 −56.8%. 2000–02 −49.1%. 2020 −33.9%. 2022 −25.4%. Hartford’s 20% census averages −35%. That average is a slice, not the Depression.
−56.8%
GFC, 9 Oct 2007 – 9 Mar 2009
The year stocks and bonds occupied the same quadrant
Each bubble is a calendar year. X is Damodaran’s S&P total return. Y is his 10-year Treasury. Size is gold. 2008 is the upper-left hedge. 2022 sits on the 45-degree line, both down 18%.
−18.0 / −17.8
S&P and 10-year Treasuries, 2022
The 45-degree line is a 60/40 that did no work
X is Damodaran’s S&P calendar return. Y is 60% that S&P and 40% his 10-year. The dashed line is Y = X: the mix is the stock. 2008 sits far above it. 2022 sits on it.
0.08 pt
2022 60/40 vs 100% S&P