Here is what we know about US equity markets. First, they are large — the S&P 500 alone has a capitalisation of $20tn. Second, they are — meant to be — extremely liquid. Billions of shares trade every day. The spread between the best bid and offer ought to be slender, and behind these prices will cluster a shoal of others that closely compete. The indices also support an array of futures contracts for those needing exposure to the broader market.

We also know that on May 6 2010 these markets, briefly, went berserk. Over the course of a few minutes, they wheeled and soared and swung, crashing 6 per cent before recovering. Household names like Procter & Gamble dipped by a quarter, the consultants Accenture at one point swapped hands for one cent. It was the purest example of market malfunction in recent times. Despite the brewing crisis in Greece, nothing fundamental happened to explain such a spasm — no rumour of war or assassination, not even a clumsy misstatement by a central bank.

Numerous, complementary theories explain how inexplicably low prices failed to attract the preponderance of buyers that would normally dampen such a move. A large sale of futures contracts hit an air-pocket of missing bids. Computer algorithms malfunctioned, selling rather than holding back as the price fell. “High frequency” traders withdrew their interest all at once. Everywhere illiquidity begat further illiquidity, spreading in a feedback loop between cash and futures markets.