Learn the stock market › Stock market crashes in history › Lesson 7 of 7
How do you spot a stock market bubble?
What is a bubble?
The word is used for prices that have risen far more than can be explained, and then fall. The economist Eugene Fama argues that such bubbles can't be shown, because high prices aren't followed by low average returns.
Ben Bernanke, then a Federal Reserve governor, said in 2002 that the Fed cannot reliably identify bubbles in asset prices.
The rises before the crashes
How much the S&P 500 had risen before each high, and how far it then fell.
| The high | 12 mo before | 3 yrs before | The fall after |
|---|---|---|---|
| Sep 1929 | +49% | – | −86% |
| Aug 1956 | +15% | +100% | −21% |
| Dec 1961 | +28% | +36% | −28% |
| Feb 1966 | +8% | +43% | −22% |
| Nov 1968 | +15% | +18% | −36% |
| Jan 1973 | +16% | +31% | −48% |
| Nov 1980 | +32% | +46% | −27% |
| Aug 1987 | +36% | +102% | −34% |
| Mar 2000 | +20% | +93% | −49% |
| Oct 2007 | +16% | +39% | −57% |
| Feb 2020 | +22% | +43% | −34% |
| Jan 2022 | +30% | +96% | −25% |
The pattern is weak. Large rises came before 1929 and 1987, but before nine of the twelve crashes the rise over a year was under 30%. Before the financial crisis it was 16%. A dash means our closes don't reach that far back.
Turned around: after big rises
The question can also be turned around. We took every month end since 1929 and compared those after a rise of at least 30% in twelve months with all months. The 30% is our own choice.
| Month ends | After a rise of 30% or more | All months |
|---|---|---|
| Count | 122 | 1,151 |
| At least 20% lower within two years | 27% | 21% |
| A year later, median | +5% | +10% |
| Higher a year later | 62% | 69% |
After a big rise a fall of at least 20% was somewhat more common, 27% against 21%, but the difference lies within what chance can give. In 48 of the 122 months the index also still stood at least 20% below its high, so it was on the way back up after a crash.
A year later the middle of them stood at +5%, against +10% for all months.
What researchers have found
Greenwood, Shleifer and You looked at every US industry since 1928 that had risen at least 100% in two years, 40 cases. They found Fama mostly right about averages: returns afterwards were not unusually low. But the risk of a crash rose. The chance of a 40% fall within two years went from 20% to 53% as an industry's rise beyond the market went from 50% to 100%. The shape of the rise, with volatility, turnover and new shares, helped too.
That is about industries, not the whole market. Health stocks rose over 100% between 1976 and 1978 and then went on rising for three more years.
What the crashes had in common
These can be seen afterwards. They don't mean a rise always ends in a crash.
- Borrowed money. In 1929 buyers typically put down 10 percent.
- Something new, untested in a crisis. Portfolio insurance in 1987, and new mortgage products before 2008 whose risk, the Federal Reserve says, may not have been well understood.
- Belief that prices will keep rising. The Federal Reserve describes how rising house prices led people to expect still higher prices.
- Talk of a new era. Irving Fisher spoke of a permanently high plateau after the 1929 high.
Test yourself
What do Greenwood, Shleifer and You show about industries that have risen a lot?
The index leaves out dividends. Closes through Nov 6, 2026.
Sources
- Greenwood, Shleifer and You, Bubbles for Fama, NBER working paper 23191, published in the Journal of Financial Economics in 2019, read November 5, 2026.
- Ben Bernanke, Asset-price "bubbles" and monetary policy, speech of October 15, 2002, read November 5, 2026.
- Federal Reserve History, Stock Market Crash of 1929, Stock Market Crash of 1987 and Subprime Mortgage Crisis, read November 5, 2026.