The first electronic market

This note is about the day a market first became a screen, and about a distinction that gets flattened in the retelling. On February 8, 1971, the National Association of Securities Dealers switched on the National Association of Securities Dealers Automated Quotations — NASDAQ — created by the body now known as FINRA. It is routinely called the world's first electronic market, and that is true in a specific and worth-preserving sense: what went electronic in 1971 was quotation, not execution. Terminals displayed competing market-maker quotes so a broker could see, in one place, who was bidding and offering where. The trades themselves were still arranged by phone. Automated execution phased in over the following decade. The precise claim is "electronic since 1971 for quotation," and the imprecise version hides the mechanism that actually changed.

That mechanism is worth stating carefully, because seeing the quotes was the whole revolution. Before the screen, a broker discovered a dealer's price by calling around, and the market was only as visible as one's phone list and memory. Putting the competing quotes on a terminal collapsed that search: the best bid and offer, and the dealers standing behind them, became legible at a glance. Price discovery did not move onto a central auction the way it had at the older venue; it moved onto a lit board of competing dealers, each posting where they would trade. The market became something you could read rather than something you had to assemble by telephone, and once a thing can be read, it can be compared, and once it can be compared, competition sharpens the number.

The structural choice underneath NASDAQ is the dealer model, and it is the clean opposite of the single-agent auction. Rather than one obligated agent per name, NASDAQ has many market makers, each posting two-sided quotes in the same security and competing with one another. Liquidity is not the affirmative obligation of a single designated party; it is the emergent product of inter-dealer competition. When it works, competition among dealers tightens spreads without anyone being commanded to make a market. The trade-off is the mirror image of the auction's: you gain redundancy and competitive pressure on price, and you give up the guarantee that one specific party is obligated to stand in with capital at the ugly moments. Two philosophies of who is responsible for a fair price, and NASDAQ picked the crowd of dealers.

The venue is all-electronic and has no physical floor, and its matching follows price-time priority — best price first, and among equal prices, the order that arrived first. That last rule is quietly consequential. Price-time priority is what turns latency into an edge worth paying for, because when two orders sit at the same price the clock decides who fills, and the clock rewards whoever reached the queue first. A large part of the modern structure of electronic trading — the racing, the co-location, the obsessive engineering of the path from decision to book — is a rational response to a matching rule that was, in its origins, simply a fair way to break ties. The design of the queue produces the behavior around the queue.

It is not an accident that NASDAQ became the natural home for technology and growth listings. A screen-based, all-electronic, competing-dealer market was culturally and mechanically suited to the companies that were themselves screens and networks, and the association compounded over decades. But the deeper point is architectural rather than sectoral: NASDAQ demonstrated that a market does not require a floor or a single obligated agent to discover a price. It requires a way for competing intentions to be posted, seen, and cleared under public rules — and once quotation went electronic, the floor became one option among several rather than the definition of a market.

What we keep from 1971 is the recognition that the screen changed the object we are trading. When quotes went onto terminals, the market stopped being a set of phone conversations and became a stream of posted, competing decisions — which is very close to how we described a market in the first note on this page: not the asset, but the compressed decisions other people are making about it. The dealer model and the auction model are two answers to the same question about who owes the market a price, and both are still running side by side. Understanding which one you are trading into, and how its rules turn into behavior, is not background knowledge. It is part of the process, and the process is the thing we actually build.