Black Monday, 1987Start here

The largest one-day percentage fall in modern equity history, with no news to explain it. A strategy sold as insurance turned out to be a synchronised sell order.

What happened

• Through 1987 — equities rise strongly. A technique called portfolio insurance grows rapidly in institutional use, promising equity participation with a floor.
• Mid-October 1987 — markets weaken over several sessions on a mix of rate concerns, trade-deficit figures and tax proposals. None of it is individually remarkable.
• 19 October 1987 — the US market falls more than 20% in one session. Markets around the world fall similarly, in some cases more, over a 24-hour period that spans multiple time zones.
• Order systems overwhelm. Quotes go stale, futures trade at enormous discounts to the cash index, and the arbitrage that normally links the two breaks down because the cash market cannot be traded.
• 20 October 1987 — central bank liquidity support and a partial recovery. Markets recover the loss over the following two years.
• No single triggering event was ever identified, and that is the most important fact about the episode.

The mechanism: insurance that was a sell order

• Portfolio insurance was dynamic replication of a put option. Rather than buying protection from someone, the strategy synthesised it: sell futures as the market falls, buy them back as it rises. Mathematically sound, and it was the delta-hedging logic in the option pricing model applied at portfolio scale.
• The assumption was continuous trading. Delta hedging requires the ability to adjust smoothly. If the price gaps, the hedge is adjusted only after the loss it was meant to prevent — see volatility and gamma.
• The strategies were correlated because the rule was the same. Many institutions ran the same logic on the same index with the same triggers. On the way down, they all sold at the same moments — not because anyone panicked, but because the model said to.
• That converts insurance into a feedback loop. Selling depresses the price, which triggers more selling. The strategy that was supposed to buy protection from the market instead demanded liquidity from it, in size, precisely when there was none.
• The buyer of last resort had not been priced. The strategy implicitly assumed someone would take the other side at reasonable prices. Nobody had agreed to.

What it teaches

• A hedging strategy that many people run identically is not a hedge; it is a crowd. The correlation of strategies matters as much as the correlation of assets — the point in how products fail.
• Synthetic protection is not the same as bought protection. A purchased put transfers risk to a counterparty who is obliged to pay. A replicated put depends on your ability to trade, which is exactly what disappears in the scenario you are protecting against.
• Liquidity is an assumption, and it is the one that fails first. Every model built on continuous prices has this dependency, and it is rarely stated as a risk — see what liquidity costs.
• Markets can move enormously with no news. The search for a cause after a large move assumes there is one. Structure and mechanics are sufficient explanation, and they are the harder thing to see.
• Recovery is not vindication. The market recovered within two years, which is true and irrelevant to anyone who was forced to sell — the sequence-risk point.

What changed afterwards

• Circuit breakers — coordinated trading halts at defined decline thresholds, designed to interrupt exactly this kind of feedback loop and to give participants time to assess.
• Cross-market coordination between cash and derivatives venues, after the futures–cash disconnection made the price of the index ambiguous for hours.
• Order-handling and capacity requirements, after systems demonstrably could not process the volume.
• Portfolio insurance largely disappeared as a marketed product, and its logic reappeared under other names — most visibly in the volatility-targeting and leveraged products that produced 2018.

The uncomfortable part

• Every element of the strategy was defensible individually. The maths was right, each institution's decision was rational, and the aggregate outcome was a crash.
• That combination — locally rational, globally destabilising — is the hardest risk to manage, because no participant can see it from their own position.
• It has recurred. Different instruments, same shape: 1998, 2008, 2018, 2022.