The 1987 stock market crash: the fastest fall in history, and the fastest recovery

On Monday 19 October 1987 the S&P 500 fell by more than 20 % in a single session, the largest one-day fall in its history before or since; the Dow lost 22.6 %. Nothing in the economy had changed over the weekend. There was no recession, no banking crisis, no war. Within two years the index was back at its high, and 1987 as a whole ended with a small positive total return. The 1987 crash is the clearest example of the other kind of crash: fast, violent, and over before most investors had decided what to do.

The crash in numbers

Sources: daily figures as recorded by the exchanges and widely documented; monthly and calendar-year figures from Shiller data and the series behind our withdrawal calculator; InvestWealth calculations.

The last two rows are the ones to remember. An investor who looked at the account on 1 January 1987 and again on 1 January 1988 saw a gain of about 5 %. Everything in between, the 40 % rise to August, the 34 % fall to December, happened inside one calendar year. The crash was real, the damage to anyone who held was temporary, and the damage to anyone who sold in late October was permanent.

What caused the 1987 crash

Again several causes, but with a clearer mechanical story than 1929:

• A fast run-up. The S&P 500 had risen more than 40 % from January to August 1987 on top of a five-year bull market; valuations were high and interest rates were rising through the summer.
• Portfolio insurance. Large institutions used a strategy that automatically sold stock index futures as prices fell, to limit losses. On a day when everyone's rule said "sell", the selling fed on itself. The crash was the first one caused in large part by rules executing together, which is a lesson in itself about rules that depend on liquidity being there when everyone needs it.
• Market plumbing. Order systems were overwhelmed, prices were stale for hours, and index arbitrage between futures and cash markets broke down. Circuit breakers that halt trading after large falls were introduced afterwards because of this day.
• The trigger was unremarkable: a weak trade figure, talk of a tax change for takeovers and a bad week before. Fast crashes rarely have a cause worth the name; they have a market positioned for one.

Year by year: stocks, bonds, inflation

Calendar-year total returns, US data, 60/40 rebalanced yearly, before costs.

What protected a portfolio

Less than in 1929, and for an instructive reason. On the crash day itself, Treasuries rallied hard as money fled to safety; from the August peak to November, 10-year bonds gained a little while stocks lost a quarter. But over the full year bonds lost 4 %, because rates had risen through 1987 before the crash. A 60/40 portfolio ended the year up 1.5 %: the bond half cushioned October and cost return in the months before it.

The real protection in 1987 was not doing anything. The fall was too fast for any monthly process to react to; by the time a month-end review came round, most of the fall had happened. Anyone who sold at the end of October locked in a loss of about a quarter and watched the market recover without them. This is the mirror image of 1929, where the fall was slow and the right action was to reduce risk early and stay out for years.

Fast crashes versus slow bear markets

This distinction is the most useful thing the two episodes teach together, and it is the organising idea of our research on trend following:

• Slow bear markets (1929–32, 33 months; 2000–02, 30 months; 2007–09, 17 months) give a rules-based process time to step aside and are where it earns its keep.
• Fast crashes (1987, two months; 2020, five weeks) do not. A monthly process takes most of the fall and then has to decide whether to re-enter; the only defence against a fast crash is the position size you held going in and the mix around it.

Of the twelve S&P 500 falls of more than 15 % since 1955, the three fastest were 1987, 1998 and 2020, and all three had recovered within two years. The slow ones took four to seven years. Speed and depth are related: the falls that destroy a decade are the slow ones.

What 1987 teaches today

• Do not confuse a crash with a bear market. A 20 % day is terrifying and, historically, has been followed by recovery within two years every time. A 20 % year inside a recession has not.
• Size for the fast fall, process for the slow one. No rule saves you from a one-day 20 % fall; only position size does. Rules earn their keep in the 1929 and 2008 kind. The crash protection guide sets out both.
• Rules that everyone runs at once can cause the thing they protect against. Portfolio insurance in 1987 is the historical warning; a process that depends on selling into a falling market at speed should be treated with suspicion.
• Check the two numbers that describe any fall: depth and time to recover. The max drawdown guide explains why both matter, and the Portfolio Crash Test shows what your own mix did in 1987 and the eleven other big falls since 1955.

Frequently asked questions

Hamann, H. (2026). The 1987 stock market crash: the fastest fall in history, and the fastest recovery. InvestWealth.com. https://www.investwealth.com/learn/guides/stock-market-crash-of-1987

<a href="https://www.investwealth.com/learn/guides/stock-market-crash-of-1987">The 1987 stock market crash: the fastest fall in history, and the fastest recovery</a> (InvestWealth.com)