Trading in taverns: the first futures market 

Here is the practical problem that shaped everything. A tulip bulb is lifted from the ground in summer, roughly June, and replanted in autumn. Between planting and lifting there is nothing to hand over. So if you wanted to trade a bulb in December, you were trading a promise about an object still in the soil, described by name and weight in a written note.

That is a forward contract, and by the mid 1630s the Dutch were writing them in volume. The venues were not exchanges. They were inns, where groups of traders met as informal associations known as colleges, with their own customs and their own procedures for setting prices. Sales ran through methods like the boards, where buyer and seller each wrote a number and intermediaries closed the gap, and every completed deal carried small charges, wine money and a payment to the college. The atmosphere, on the evidence of the pamphlets, was closer to a card room than a bourse.

Two features made this fragile. The first is that almost nobody paid up front. A contract typically promised payment on delivery in the summer, sometimes with a token deposit, sometimes with nothing but a signature. A trader could therefore commit to thousands of guilders while holding very little cash, which is leverage in its purest form: exposure without an outlay.

The second is that there was no institution standing behind any of it. The Amsterdam exchange, which handled the shares of the East India Company with far more formality, was not involved. There was no clearing house, no margin, no mechanism for closing out a position, and no reliable legal enforcement of the contracts. In fact the Dutch authorities had repeatedly frowned on speculative forward trading, which left tulip notes in an awkward zone of doubtful validity.

The composition of the crowd changed too. Alongside the wealthy connoisseurs who had always collected rare bulbs, the winter of 1636 and 1637 drew in weavers, bakers, artisans and small merchants, trading cheaper bulbs by weight rather than by named specimen. Goldgar's research shows this group was still socially narrower than legend suggests, more of a spreading network among people who knew each other than a national frenzy, but the market had clearly widened beyond its original circle.

So picture the structure as of late January 1637: rising prices, contracts that require no cash today, no institution to enforce settlement, and a fresh cohort of buyers whose confidence rests on the observation that prices have been rising. Every one of those elements works beautifully while new buyers keep arriving.

In the first days of February, in Haarlem, they stopped arriving.