The characteristic
What actually broke on Black Monday?
You assume your stop-loss saves you. In 1987, the stop-loss was the crash.
The thing built to protect portfolios is what destroyed them. On 19 October 1987 the Dow fell 22.6% in a single afternoon — still the largest one-day drop on record — and the accelerant was portfolio insurance, an automated strategy that sold index futures as prices fell. It executed exactly as specified: that was the failure mode.
Each desk’s hedge was rational alone. Run in parallel, it became a feedback loop: selling forced prices down, lower prices triggered more selling. The same reflexive correlation reappears in faster form during the March 2020 liquidity seizure and the model-driven AI disruption scenario.
The mitigation taken
What did the loss actually cost?
It cost the exit. The mitigation at the time — circuit breakers to break the loop — treated the symptom, not the mechanism. Liquidity vanished when holders needed it: spreads blew out, order books emptied, and a 22.6% repricing arrived faster than any human could act. For a concentrated holder the math is geometric: a 50% drawdown demands a 100% gain to flat. You do not wait out a crash you cannot trade through.
What does engineered resilience look like here?
A known failure surface, not a faster reflex — you read the drawdown that triggers your hedge and the carry you pay before the shock. Downside bounded, velocity intact. This is the convex tail-risk structure Spitznagel formalized: pay a small, known cost continuously so the Spitznagel-style payoff is largest when liquidity is gone.
The influence today
How it works
The same failure mode is live now. Today’s market is majority non-discretionary — index rebalancing, ETF flows, volatility-targeting, algorithmic selling that assumes continuous liquidity. The 1987 engine never left; it scaled. A concentrated single-ticker position sits downstream of every seller that fires when correlation breaks.
The Fail-Safe Circuit
Programmatic put-option protection that executes when correlation breaks down — priced as static carry, not as a reflexive sell into a falling tape.
model · dynamic GPD signal · stated failure rate
Static protection that sells into weakness is the 1987 pathology. Dynamic protection beats it net of cost: on 11 years of DAX futures a GPD-signal strategy lifted the Sharpe ratio to 0.4587 versus 0.3022 for static buy-and-hold (Packham et al., 2017) — the quantitative method behind the signal is documented in full. The diagnostic carries a published 4.2% false-comfort rate — runs that understate the realized tail. Rouault caged his clown in thick black contour lines; the 1987 trader was caged the same way, by an automated rule he did not control.
The objection
“Circuit breakers fixed this — why does it still matter?”
They did not fix it; they relocated it. Halts pause one venue, but the response seeded flash-crash dynamics across correlated assets. The engine is larger now — a halt buys minutes, not an exit.
3 fields · 48-hour document · no call, no sequence.
Frequently asked questions
What caused the 1987 Black Monday crash?
Portfolio insurance — an automated strategy that sold index futures as prices fell — turned a decline into a 22.6% one-day collapse on 19 October 1987. Each desk’s hedge was rational in isolation; run in parallel they formed a feedback loop where selling forced prices down and lower prices triggered more selling. It executed exactly as designed, which was the failure.
Could a 1987-style crash happen again today?
Yes, and the engine is larger. Today’s market is majority non-discretionary — index rebalancing, ETF flows, volatility-targeting and algorithmic selling that all assume continuous liquidity. A concentrated single-ticker position sits downstream of every automated seller that fires when correlation breaks, so the 1987 mechanism never left; it scaled.
What was the structural failure behind Black Monday?
The failure was reflexivity, not a one-off panic. Protection that sells into weakness amplifies the move it was meant to absorb, and liquidity vanished precisely when holders needed to exit: spreads blew out and order books emptied. Circuit breakers introduced afterward paused single venues but treated the symptom, not the mechanism.
How would a tail-risk hedge have behaved in 1987?
A convex put-option hedge pays the largest when liquidity disappears, the opposite of portfolio insurance that sold into the falling tape. On 11 years of DAX futures a dynamic GPD-signal strategy lifted the Sharpe ratio to 0.4587 versus 0.3022 for static buy-and-hold (Packham et al., 2017), with a published 4.2% false-comfort rate.
What should a concentrated investor watch for today?
Watch the drawdown that triggers automated selling and the carry you pay to stay hedged before the shock, not after. A 50% drawdown demands a 100% gain to return to flat, so the priority is a bounded, known downside rather than a faster reflex. Map the failure surface while the market is calm.