Can stocks fall 80%? Spitznagel's crash warning tested against market history

Yes: between 1870 and 2020, 9 of the 16 national stock markets in the Jordà-Schularick-Taylor database lost 75% or more after inflation at year ends, in 12 separate falls. On monthly data with dividends and after inflation, US stocks lost 76.8% from September 1929 to June 1932. Consumer prices more than doubled in 11 of the 12 national falls. Where bond data exist, a 60/40 mix of the country's own shares and bonds lost more than 70% too.

Mark Spitznagel, who founded the tail-hedging firm Universa Investments, told Reuters he expects "an 80% crash", but only "after a massive, euphoric, historic blow off rally", Moneywise reported via Yahoo Finance in December 2025. In September 2026 he told Business Insider he expected "one more really big, risk-on, insane, euphoric rally to come across risk assets", then "a bigger crash than we've seen in our lifetime".

The Nasdaq Composite closed 11.9% below its 6 October 2026 record on 5 November and 7.7% below it on 11 November, on Nasdaq's closing prices. This article does not test Spitznagel's forecast; no dataset can date the next crash. It answers the question a sell-off raises: how often falls of that size have happened, how long they lasted, and which holdings came through them.

Key takeaways

1. The 75% falls were national, often spanning war or sustained inflation. Consumer prices more than doubled in 11 of the 12 cases, and none began after 1973.
2. An 80% loss needs a 400% gain to recover. At a steady 5% a year after inflation that takes 33 years; one such loss in 30 years turns 5% a year into a 0.5% annual loss.
3. Home bonds were no shelter when the home currency failed. In the 10 collapses with bond data, a domestic 60/40 portfolio still lost between 71.3% and 96.2% after inflation.
4. The cross-country average had smaller losses, before currency and access constraints. A real-GDP-weighted average of the 16 markets never closed a year more than 46.8% below its peak, but it did not avoid falls of 40% or more.
5. Borrowing can exhaust capital before prices recover. In a simplified price-only margin example, a 1929 buyer with half the purchase borrowed was wiped out by December 1930; unleveraged total returns recovered in real terms by November 1936.

How often have national stock markets lost 75% or more?

Twelve times in 2,256 years of data across 16 countries, counting year-end values with dividends reinvested and after local inflation. The Jordà-Schularick-Taylor Macrohistory Database (release 6) has annual equity returns for 16 of its 18 countries, most from the 1870s to 2020. We chained them into real indices and recorded every fall from a year-end peak.

Twelve real falls of 75% or more in 9 of 16 markets; consumer prices more than doubled in 11 of them. Source: our calculations from the JST database, release 6 (equity total return and consumer prices); episode table (CSV).

The dates line up with the two world wars and the inflation of the 1970s. Germany's 1944-1948 fall includes an 88.4% nominal loss in 1948, the year of its currency reform. The US, UK, Switzerland, the Netherlands, Australia, Norway and Denmark never closed a year 75% below a previous real peak. Their deepest year-end falls ranged from 71.7% for the UK in 1972-1974 to less than 50% for Denmark.

Year-end values understate how deep a fall went within a year. The US loss of 1928-1932 is 51.9% at year ends, against 76.8% on Shiller's monthly data below. The count is also a floor: markets that closed for good are not in the database, which follows countries that remained market economies. Falls of 50% or more were common, with 37 episodes in 15 of the 16 markets.

Were the famous bubble crashes as deep after dividends and inflation?

The US and Japanese total-return measures show smaller losses, although Japan's figures also cover a broader market and use year ends. The US real recovery was much shorter; Japan's wait remained long. Three peacetime falls that reached roughly 80% on price indices were the US after 1929, Japan after 1989 and the Nasdaq after 2000. Each looks different once dividends, inflation and a broader portfolio are counted.

US 1929-1932: −84.8% on price, −76.8% in real total return; Nikkei −81.9% and 34.1 years; Nasdaq −77.9% and 15.1 years; MSCI World −51.4% and 6.7 years. Source: Robert Shiller's data, FRED NIKKEI225, Nasdaq Composite closes, MSCI World index 990100 and the JST database; bubble-bust table (CSV).

The 25 years often quoted for 1929 is a price-only figure in dollars. Dividends and a 21.4% fall in US consumer prices from September 1929 to June 1932 meant a reinvesting holder regained the 1929 real value by November 1936, before the 1937 fall. Shiller's prices are monthly averages of daily closes, so the daily peak-to-trough loss was larger.

Japan's wait was long however it is measured. The Nikkei regained its 1989 record only in February 2024. With dividends and after inflation the worst year-end fall was a smaller 63.4%, yet Japanese equities were still 2.8% below their 1989 level at the end of 2020, the last year of the data. The recovery-time guide explains why the fall and the rebound are two separate clocks.

Did bonds and a spread across markets protect against these falls?

The US deflation and the national inflation episodes produced very different bond outcomes. The cross-country average fell less in the selected windows, though it is not a common-currency portfolio. From September 1929 to June 1932, US 10-year Treasuries gained 38.2% after inflation. A 60/40 portfolio rebalanced monthly lost 50.4%, against 76.8% for stocks alone. In Japan after 1989, a 60/40 mix of Japanese shares and government bonds lost at most 27.1% at year ends and was back above its 1989 real value by 2005.

In the national collapses, nominal government bonds also suffered. Their coupons and principal are claims in the home currency, so rising consumer prices erode what those payments can buy. Where bond returns are available for the 75% equity falls, combining the two domestic assets still left large real losses.

Home 60/40 portfolios lost between 71.3% and 96.2% after inflation in the 10 collapses with bond data; the 16-market average's deepest loss in the same windows was between 0% and 46.8%. Source: our calculations from the JST database, release 6 (equity and government bond total returns, consumer prices, real GDP); episode table (CSV) and 16-market average by year (CSV).

The 16-market average weights each market's real return by its real GDP in the previous year and includes the collapsing market itself. From 1871 to 2020 its deepest year-end falls were 46.8% from 1912 to 1920, 42.6% from 1928 to 1931 and 40.1% in 2008. For a dollar investor, the MSCI World price index fell 51.4% in 2000-2002 while the Nasdaq Composite fell 77.9%.

From 1901 to 2020 the average compounded at 6.07% a year after local inflation. Among the 14 markets with uninterrupted data, the US at 6.57%, Australia at 6.43% and Denmark at 6.28% did better; the other 11 did worse (growth by market, CSV). Germany and Japan have gaps, so their available-year figures are not continuous-period growth rates.

The average shows that local real equity outcomes differed, rather than what an investor could have earned across borders. It ignores exchange rates and trading costs, and wartime capital controls could prevent foreign investment. Our guide to a portfolio beyond the US examines current country weights and the concentration they create.

What does an 80% loss do to long-run growth?

An 80% loss leaves one fifth of the starting balance, so subsequent percentage gains act on much less capital. The gain needed after a loss of L is 1 / (1 − L) − 1, and the years needed at a steady growth rate follow from the logarithm. The table uses a hypothetical 5% a year after inflation as a scenario, not a forecast.

An 80% loss needs a 400% gain; at 5% a year that takes 33 years, and one such loss in 30 years of 5% growth leaves −0.48% a year. Source: our arithmetic; loss table (CSV).

The geometric growth rate is the exponential of the average log return, minus one, so a deep loss weighs heavily in the result. Here the shock is additional to 30 years of assumed 5% growth. A 50% fall costs 2.4 percentage points of annual growth over 30 years in this scenario; an 80% fall costs 5.5. Reducing a deep loss can improve the outcome if the cost of protection does not consume the gain over the full period.

Why do leverage and forced selling make a deep fall permanent?

Because a borrower can be forced to sell before the recovery, and the units sold never recover. We ran the US 1929-1932 fall through Shiller's monthly data with borrowed money, paying the annual US short-term bill rate from the JST database.

Real loss to June 1932: 60/40 −50.4%, stocks −76.8%, 1.5x −92.1%, 2x −97.7%; a separate price-only 50% margin example exhausted capital by December 1930. Source: our calculations from Robert Shiller's data (S&P Composite real total return, 10-year Treasury real total return, CPI) and the JST database (US bill rate); monthly paths (CSV).

In the separate static margin example, the price index fell below half its September 1929 level by December 1930, 18 months before the trough. That calculation omits dividends, interest and earlier margin calls. The monthly rebalanced 2x total-return simulation instead lost 97.7% in real terms and took 22.8 years to recover; its borrowing rate omits lender spreads.

A household that sells to meet spending or repay a loan at the bottom is in the same position on a smaller scale. Our withdrawal model shows how a cash reserve changes that, and the leveraged-ETF study finds that a simulated 3x Nasdaq 100 fund lost 99.98% from the March 2000 record.

What can a household protect, and what not?

A household can examine the conditions that might force it to sell during a fall: debt, near-term spending and concentrated exposure. These checks cannot guarantee recovery, but they distinguish a temporary fall in quoted value from a loss of capital that leaves fewer assets to participate in any rebound.

1. Find your largest single market and sector. The examples include national markets and the technology-heavy Nasdaq; check what share of your equities depends on one of them.
2. Ask whether your defensive holdings are claims on the same currency. Home bonds fell with home shares when prices multiplied; check their exposure to an inflation shock as well as to a recession.
3. List every debt a fall could force you to repay. Check repayment and collateral requirements alongside any leveraged funds; the static margin scenario exhausted capital before the recovery.
4. Check which foreseeable spending depends on the equity allocation. A fall only becomes a forced sale when bills must be paid from falling assets.
5. Judge any protection by whole-portfolio growth after its cost. Compare its effect on compounded wealth with its premiums, fees and liquidity benefits. A lower return may still serve a household that needs to fund spending during a fall.

The Spitznagel study covers Universa's own hedge and its critics, the crash-protection analysis compares premiums with bonds, gold and trend following, and the put-option guide checks a published put benchmark. The asset-class crash study shows what each asset class did in 2000, 2008, 2020 and 2022, and the November 2026 sell-off analysis covers the current fall.