Black Monday 1987: A Crash With No News

On 19 October 1987 the Dow Jones Industrial Average fell 508 points — 22.6% in one session.

That record still stands. Not “the largest in recent decades” — the largest ever. For comparison, the worst single day of the 1929 crash was −12.8%.

The unsettling part is this: no piece of news explains it.

No war broke out. No country defaulted. No major institution failed. No economic release surprised anyone. The official inquiry that followed — the Brady Commission — spent months on it and could not identify a triggering event.

The setup

US equities had risen roughly 40% over the first eight months of 1987. The NASDAQ Composite peaked at 455 on 26 August. The market was in the euphoric stage of a long bull run and valuations were stretched — but stretched valuations are not a reason for a crash. They can persist for years.

The real hazard was a strategy that had spread rapidly through institutions in the preceding years: portfolio insurance.

Portfolio insurance: a policy that destroys itself

The pitch was attractive: keep your upside, but limit your losses on the way down.

The mechanism was dynamic hedging. A model continuously computed the appropriate exposure; when prices fell, it automatically sold index futures to reduce exposure, buying back as prices recovered. In effect, synthesising a put option through programmatic trading.

Taken alone, the logic is sound.

The problem was that by 1987, tens of billions of dollars were running the same strategy. And that strategy’s core instruction was: sell when prices fall.

Which closes a loop:

Price falls → programs sell → selling pushes price lower → more programs sell → …

That is not insurance. It is an accelerant. When enough participants buy the same policy, the policy becomes the risk.

What actually happened that day

Sell orders flooded in from the open. But what made the day legendary is that the market’s machinery stopped working:

• NYSE quote systems fell far behind. Prices on screen lagged actual executions by tens of minutes or more.
• Index futures traded at enormous discounts to cash. Arbitrage broke down and the two markets effectively decoupled.
• Large numbers of NASDAQ market makers simply stopped answering their phones.

That last point left a clear fingerprint in the data.

The Dow fell 22.6% on 19 October. The NASDAQ Composite fell only 11.4% that day (406.33 → 360.21).

This looks like NASDAQ held up better. The opposite is true — it could not fall, because it could not trade. Market makers are obliged to post two-sided quotes, but in the panic many simply let the phone ring. Quotes were notional. Sell orders could not be executed.

The suppressed selling arrived the next day. On 20 October, NASDAQ fell a further 9.0% (360.21 → 327.79).

Combined, 19.3% over two days — roughly the Dow’s single-day move. The same crash, executed a day late.

What transfers

One: liquidity is conditional, and the condition fails exactly when you need it

A market that trades freely in normal times is not thereby a market that trades freely in a crisis. Liquidity is not an intrinsic property; it comes from counterparties willing to take the other side — and in a panic, they leave.

This bears directly on a tool nearly everyone uses: the stop-loss order. A stop guarantees a trigger price, not a fill price. Through a gap or a liquidity vacuum, actual fills can be far below the level you set. On 19 October 1987, a great many stops filled at prices nobody anticipated.

Two: a strategy changes character once enough people use it

Portfolio insurance was a workable hedge at small scale. At tens of billions, it became crash fuel.

Any strategy that depends on “I can sell before the others” carries a hidden assumption: the others will not want to sell at the same moment. When that breaks, the strategy does not merely fail — it amplifies in reverse.

Three: crashes do not require reasons

If your framework assumes major declines always have an identifiable cause, 1987 is a direct counterexample. Market structure alone — leverage, programmatic trading, crowded positioning — can produce extreme moves with no external news whatsoever.

One more fact

Worth noting: US equities recovered the entire 1987 decline within about two years. For long-term investors who used no leverage and were not forced to sell, the crash is a sharp notch on a long-term chart.

What was destroyed were the participants who used leverage, were forcibly liquidated, or panic-sold at the lows. That distinction matters — the same market event produced completely different outcomes depending on position structure.

(But this is not a general rule. Japanese equities peaked in 1990 and took over three decades to reclaim that high. Whether and when a market recovers cannot be known in advance — treating 1987’s ending as typical is itself a survivorship bias.)

This is a historical review written for risk education. It is not investment advice and makes no inference about any current market conditions.