Black Monday 1987: The Largest One-Day Decline on Record

Key Takeaways

• The Dow Jones Industrial Average fell 508 points, 22.6 percent, on 19 October 1987, which the Federal Reserve records as the largest one-day decline in history and the sharpest United States market downturn since the Great Depression at the time.
• Computed from daily closes, the S&P 500 fell 20.47 percent that session, the worst single day in its own record examined for this library.
• It was global. The Federal Reserve calls it the first contemporary global financial crisis and records that New Zealand's stock market fell 60 percent.
• Portfolio insurance, a product using options and derivatives that mechanically sold into declines, accelerated the crash as initial losses triggered further rounds of selling.
• The rebound was immediate. The Federal Reserve records the Dow regaining 288 points, 57 percent of the Black Monday decline, in just two trading sessions.
• The closing low did not come on Black Monday. Computed from daily closes, the S&P 500's cycle low was 223.92 on 4 December 1987, marginally below its 224.84 close on 19 October.
• No recession followed. This is the only episode in this library with a decline of over 20 percent and no NBER-dated contraction attached.

What Happened on Black Monday?

The Federal Reserve's account is unusually vivid for an institutional history. It describes a chain reaction of market distress sending global stock exchanges plummeting in a matter of hours, with traders racing each other to the trading pits to sell. In the United States the Dow crashed at the opening bell and finished down 508 points, or 22.6 percent.

What distinguishes 1987 from every other episode here is the compression. This was not a decline that unfolded over months with an acute phase in the middle. Essentially the entire event happened inside one session, and the recovery began in the next one.

Chronology

The 1987 episode with S&P 500 closes. Index closes computed from daily closing values; Dow figures from Federal Reserve records.

The 4 December low is the detail most retellings skip, and it changes the shape of the story. The market did not bottom on the famous day. It rebounded sharply, fell again on 26 October, and only reached its closing low six weeks later. An investor who bought on 20 October in the belief that the capitulation had happened was underwater for another six weeks, then had to wait until July 1989 to break even.

What Did the Setup Look Like Before the Crash?

A strong advance already showing strain. Computed from daily closes, the S&P 500 peaked on 25 August 1987 at 336.77 and had already fallen 16 percent to 282.70 by 16 October, before Black Monday. The market was not at its high when the crash happened. It had been declining for nearly two months.

A new product that sold mechanically into declines. The Federal Reserve identifies portfolio insurance directly, describing it as a new product from United States investment firms that had become very popular, involving extensive use of options and derivatives, and states that it accelerated the crash's pace as initial losses led to further rounds of selling. The design intent was to limit downside by reducing equity exposure as prices fell. The design consequence, when many holders ran the same rule at once, was a large population of forced sellers activated by exactly the condition they were trying to protect against.

Increased international participation. The Federal Reserve notes that international investors had become increasingly active in United States markets, accounting for some of the rapid pre-crisis appreciation in stock prices. Cross-border participation makes a decline transmit faster and further, which is why this became the first contemporary global financial crisis rather than a domestic one.

Trade-clearing protocols that were not uniform across products. This was invisible to nearly everyone before it mattered. The Federal Reserve records that after Black Monday, regulators overhauled trade-clearing protocols to bring uniformity to all prominent market products, which is a plain statement that they had not been uniform beforehand.

How Far Did Prices Fall?

The 19 October 1987 session and the full peak-to-trough decline. Dow figure from Federal Reserve records; S&P 500 figures computed from daily closing values.

Two things are worth separating here. The famous number is the single-day move, and it is genuinely the extreme of the record. The full drawdown of 33.5 percent is comparable to the 2020 COVID crash at 33.9 percent and considerably shallower than 2008 at 56.8 percent. The distinctiveness of 1987 is entirely in the concentration, not in the depth.

The worst sessions

Computed from daily closes, the four worst S&P 500 sessions of 1987 were 19 October at 20.47 percent, 26 October at 8.28 percent, 16 October at 5.16 percent and 30 November at 4.18 percent. The 19 October figure is more than twice the next worst, which is the statistical signature of a mechanical selling cascade rather than a repricing of fundamentals. No plausible revision to expected corporate earnings occurred over one weekend that would justify a fifth of the market's value.

Which Warning Signs Were Visible in Advance, and Which Only in Hindsight?

1987 is the cleanest case in this library of a risk that was fully visible in structure and completely invisible in timing.

Signals classified by whether they were usable at the time.

The generalizable observation is about the shape of the risk rather than the specific product. When a large number of participants adopt the same rule that says sell when prices fall, the aggregate effect is a selling function that activates automatically at the worst possible moment. This does not require anyone to panic, and it does not require the rule to be badly designed for an individual holder. It is a composition problem: a strategy that is sensible in isolation becomes destabilizing when everyone runs it simultaneously.

Hindsight check. Retrospectives frequently present portfolio insurance as an obvious flaw that a careful observer would have spotted. The product was widely sold, publicly discussed and understood in its mechanics by the people using it. What nobody could compute from outside was the aggregate notional running the same trigger, which is the number that actually mattered. Knowing that a mechanism exists and knowing its systemic scale are different pieces of information, and only the first was available. Cognitive biases in trading covers why the second gets retroactively assumed into the first.

Why Was the Fall So Concentrated?

Mechanical selling responded to price rather than to value. Portfolio insurance strategies reduced equity exposure as prices fell. When many portfolios executed the same reduction at the same time, the selling itself pushed prices lower, which triggered further required selling. The Federal Reserve describes exactly this: initial losses led to further rounds of selling. Nothing in the loop references whether the price is reasonable.

Clearing infrastructure was strained. Because trade-clearing protocols were not uniform across products, the plumbing that connects a futures position to an equity position did not behave predictably under extreme volume. The subsequent overhaul is the evidence.

There was no fundamental repricing to anchor to. This matters for interpretation. In 2008 the decline reflected a genuine and large change in expected outcomes. In 1987 there was no comparable news, which is why the rebound could be so fast: nothing had actually changed about the businesses, so once forced selling exhausted itself the prior valuations were still defensible.

That last point explains the otherwise strange combination of the worst day on record and no recession. A price decline caused by the mechanics of trading is a different object from a price decline caused by deteriorating cash flows, even when the number on the screen is identical. Distinguishing between the two in real time is genuinely hard, and 1987 shows why the distinction matters so much.

How Did the Federal Reserve Respond?

The response was one sentence, issued the following morning. The Federal Reserve records that on 20 October 1987, Chairman Alan Greenspan said the central bank had affirmed its readiness to serve as a source of liquidity to support the economic and financial system, consistent with its responsibilities as the nation's central bank. Behind the scenes, the Federal Reserve encouraged banks to continue to lend on their usual terms.

That is a strikingly small intervention relative to what came in later crises, and it is worth understanding why it was sufficient. The problem on 19 October was not solvency and was not a repricing of the real economy. It was that a mechanical selling cascade had exhausted the buying capacity of the market, and that the firms who might otherwise supply that capacity did not know whether their own funding would be there. A credible statement that liquidity would be available addressed exactly the constraint that was binding.

The structural response took longer and was more consequential. The Federal Reserve records that regulators overhauled trade-clearing protocols to bring uniformity to all prominent market products, and developed new rules known as circuit breakers, allowing exchanges to halt trading temporarily in instances of exceptionally large price declines. Those halts remain in force. They do not prevent declines, as 2020 demonstrated, but they interrupt the specific feedback loop in which mechanical selling and falling prices reinforce each other within a single session.

How Long Did the Recovery Take?

Two answers, and both are true.

Recovery measured two ways. Dow figures from Federal Reserve records; S&P 500 figures computed from daily closing values, price only.

The two-session rebound is the number that gets quoted, and it recovers a majority of one day's loss, not the drawdown. The full round trip took just under two years, which places 1987 as the second fastest recovery in this library after 2020.

The reason it was fast is the reason it is unrepresentative. Nothing fundamental had broken, so nothing fundamental had to be repaired. Corporate earnings did not need to recover, the banking system did not need recapitalizing, and no recession intervened. In episodes where those things were true, recovery took years to decades.

What Was Specifically Different About 1987?

No recession followed. The NBER dates the expansion continuing until July 1990, nearly three years later. Every other decline in this library either coincided with a recession or, in the case of 2022, coincided with a genuine change in the monetary environment. 1987 is the only pure market event.

The cause was market structure, not the economy. Portfolio insurance and non-uniform clearing were technical features of how trading worked. They have both been directly addressed.

The policy response was minimal and sufficient. One sentence and informal encouragement to banks. There was no rate cut on the day, no asset purchases and no legislation.

The concentration was extreme. One session accounted for more than the entire remainder of the drawdown put together in single-day terms. That has not recurred, and the circuit breakers introduced afterwards are specifically designed to make it less likely.

Circuit breakers did not exist yet. They exist now precisely because of this day. Comparing 1987 to a modern market means comparing a market without automatic trading halts to one with them.

Common Myths About Black Monday

"The market bottomed on Black Monday." It did not. Computed from daily closes, the S&P 500 closed at 224.84 on 19 October and reached its cycle low of 223.92 on 4 December, six weeks later, after another 8.28 percent session on 26 October.

"It recovered in two days." The Dow recovered 57 percent of the Black Monday decline in two sessions. That is not the same as recovering the drawdown, which took until July 1989 for the S&P 500 on a price basis.

"Circuit breakers would have prevented it." Circuit breakers pause trading. They do not change valuations or remove sellers. They were introduced in response to 1987 to interrupt the intraday feedback loop, and the market still fell 11.98 percent in a session in March 2020 with them fully in place.

"It was caused by computers." Closer to true than most myths here, but imprecise. It was caused by many participants running the same sell-when-prices-fall rule simultaneously. The automation made execution faster; the composition problem would exist with the same rule executed by hand.

"A 22.6 percent day means economic catastrophe." The single strongest counterexample in the record. Nearly three more years of expansion followed. The size of a price move is not a reliable measure of the size of the underlying economic problem.

What a Reader Can Actually Carry Forward

1987 is the episode that most directly challenges the instinct to read a large price move as information about the world. It is worth studying precisely because the inference that felt obvious on the day was wrong.

What generalizes

• Crowded identical rules create their own risk. Any strategy that sells because prices fell becomes destabilizing when enough capital follows it, and the aggregate size is usually not observable from outside. This applies to stop-loss clusters, risk-parity deleveraging, margin calls and volatility targeting, not just to 1987's specific product.
• Price declines and economic declines are different objects. The worst single day in the record was followed by nearly three more years of expansion. Before treating a large move as news about the economy, ask whether anything about future cash flows actually changed.
• The famous day is rarely the low. This is true in 1987, in 1929 and in 2008. Capitulation feels like a bottom and frequently is not.
• Market structure is a real risk category. Clearing arrangements, halt rules and the mechanics of how orders interact can dominate outcomes over short horizons, entirely independently of valuation.
• A credible liquidity commitment can be enough when the problem is liquidity. One sentence stabilized a market that had fallen more than a fifth. That works when the constraint is funding availability and does not work when the constraint is solvency, which is the distinction 2008 made expensive.

What does not generalize

• The absence of a recession. 1987 is the outlier here, not the rule. Most 20 percent-plus declines in this library did coincide with an economic contraction.
• The two-year recovery. It was fast because nothing fundamental had broken. That condition is unusual.
• The minimal policy response being sufficient. It matched the specific problem. A solvency crisis does not respond to a statement of readiness.
• The specific product. Portfolio insurance in its 1987 form is not the current risk. The composition problem it illustrates has simply moved to other instruments.

The one question worth asking now

The portable exercise from 1987 is to look at your own holdings and ask which of them would be sold by someone else's automatic rule during a fast decline. Concentrated positions held largely by strategies with mechanical exit triggers behave differently in a stressed session than positions held by patient capital, and that difference is a property of the holder base rather than of the business. It is not always knowable, but the question is worth asking, and it is a far more useful thing to take from Black Monday than any expectation about how quickly a market rebounds. Risk management covers the position-level version of the same problem.

References

Every figure on this page was verified against the following sources, each retrieved on 23 August 2026:

• Federal Reserve History: Stock Market Crash of 1987: the 508 point, 22.6 percent decline on 19 October 1987, its characterization as the largest one-day stock market decline in history and the sharpest United States downturn since the Great Depression at the time, the description of the episode as the first contemporary global financial crisis, the 60 percent fall in New Zealand's market, the role of portfolio insurance and its use of options and derivatives in accelerating the crash, the increased activity of international investors, the 20 October 1987 Greenspan statement affirming readiness to serve as a source of liquidity, the encouragement to banks to continue lending on usual terms, the recovery of 288 points or 57 percent of the Black Monday decline across two trading sessions, the statement that United States markets surpassed pre-crash highs in less than two years, and the post-crash overhaul of trade-clearing protocols and introduction of circuit breakers.
• National Bureau of Economic Research: US Business Cycle Expansions and Contractions: the July 1990 business cycle peak, establishing that the expansion continued for nearly three years after Black Monday.

Figures deliberately not stated. This page gives no figure for the notional value of portfolio insurance outstanding in 1987, no trading volume figures for 19 October, and no total-return recovery date, because no source verified in this session supplied them. Where the shape of an event is known but the number is not, the shape is described and the number is left out rather than estimated.

Method note: index peak, trough, decline and recovery figures labeled as computed were derived by Swoopr Investment from daily closing values of the named index, retrieved from the Yahoo Finance historical chart API on 23 August 2026. Drawdowns are measured close to close, not intraday, so the intraday low of any episode is lower than the trough shown. Recovery means the first daily close at or above the prior peak close, price only, with no dividends reinvested. Figures labeled total return are computed instead from dividend-adjusted closing prices and are stated as such wherever they appear.

This is educational content about a historical episode. It is not investment advice, it is not a forecast, and nothing here should be read as a claim about how any future market decline will behave.

Frequently Asked Questions