Electronic Trading Transition

From the 1970s onward, electronic systems methodically displaced human intermediaries from the trading floor. Electronic trading replaced open-outcry pits and telephone brokers with matching engines, decimalized prices, fragmented market makers, and eventually algorithmic trading strategies that executed at microsecond speeds.

The floor trader era

Before electronics, a stock exchange was a physical marketplace. At the New York Stock Exchange, specialists stood at their posts on the trading floor, matching buy and sell orders by voice and hand signal. The specialist earned the spread between bid and ask, but also had a duty to maintain an orderly market—stepping in to buy if demand was weak, sell if supply was scarce. Order flow moved through telephone lines to the floor, where the broker relayed it to the specialist.

This system was transparent to the naked eye but opaque in its internals. Specialists could see the order book for their stock; they possessed information about imbalances that retail investors never saw. They had the right of first refusal on incoming orders—a market maker-like privilege. Trading volumes were modest by modern standards, and spreads were wide—often a full dollar per share, sometimes more.

NASDAQ and the electronic opening

The National Association of Securities Dealers launched NASDAQ in 1971 as an entirely electronic system. There was no floor. Instead, competing market makers displayed bid and ask prices on an electronic bulletin board; traders placed orders by telephone or electronic message. NASDAQ was initially slower and less prestigious than the NYSE, but it offered something revolutionary: competition among market makers. A stock like Apple could have dozens of market makers bidding for the spread, rather than a single specialist.

NASDAQ proved that electronic trading worked. Spreads were tighter than at the NYSE, and trading volumes were higher. But it was still fundamentally a telephone-and-screen market—market makers and brokers relied on human judgment to update prices and manage orders.

Decimalization and fragmentation

The next watershed was the SEC’s 1997 move toward fractional-penny pricing (eventually full decimalization in 2001). Before decimalization, stocks quoted in eighths or sixteenths; the minimum spread was economically visible—a penny or two mattered less. Decimals shrank spreads to the penny or less. Instantly, market-maker economics shifted. Firms that had made fat profits on spreads now had to compete for basis-point differentials. Many exited; others turned to algorithmic trading and pure volume to survive.

Simultaneously, the SEC opened the door to alternative trading systems (ATSs)—private electronic venues that could match orders without being full exchanges. The Regulation ATS rule (1998) required the SEC to treat ATSs like exchanges in terms of disclosure and transparency, but allowed them more flexibility. Suddenly, market structure fragmented. Orders could route to the NYSE, NASDAQ, regional exchanges, or dozens of private ATSs. Brokers needed order-routing logic to find the best execution across all venues.

Algorithmic strategies and speed

The floodgates for algorithmic trading opened in the 1990s–2000s. Large mutual funds and hedge funds needed to execute enormous orders without moving the market. Humans splitting orders into batches and manually routing them were too slow and left room for market impact. Algorithms could split orders intelligently, track market microstructure, and execute across venues in milliseconds.

Early algorithms were simple: participate in a fixed percentage of market trading volume, or execute a target quantity by a deadline while minimizing expected market impact. But they evolved. Algorithmic trading became a profit center itself. Firms built proprietary algorithms that detected patterns in order flow or volatility and exploited them. The boundary between execution (passive, cost-saving) and speculation (active, profit-seeking) blurred.

By 2010, algorithmic trading accounted for more than 50% of trading volume in equities. High-frequency trading (HFT)—strategies that held positions for milliseconds and competed on latency—had emerged as a significant market maker of last resort. Traditional specialists and market makers had nearly vanished.

The vanishing spread

For retail investors, the change was positive in one dimension: bid-ask spreads narrowed dramatically. A stock that once traded with a 10-cent spread might now have a one-cent spread, or less in heavy trading volumes. Execution costs fell. Passive indexing became cheaper, and the barrier to self-directed trading dissolved.

But structure became opaque. A retail investor no longer saw the specialist at their post; they saw a consolidated ticker from dozens of venues. Market makers were now algorithms offering microsecond-scale quotes. The visible spread might be tight, but hidden costs lurked: transaction fees, maker-taker rebates that distorted order routing, and the adverse selection risk of hitting a quote that was milliseconds stale.

Flash crashes and fragility

The transition was not frictionless. On 6 May 2010, the “Flash Crash” wiped $1 trillion from equities in minutes. Algorithmic trading strategies triggered rapid selling; other algorithms amplified it; and the fragmented market structure made it impossible for any participant to see the full order book or impose discipline. Stock prices disconnected from fundamentals for brief, terrifying moments.

Investigations revealed that the system had incentives to trade at any price, even non-economic ones, if the reward was fractional. Market makers algorithms withdrew liquidity the instant volatility spiked. The speed of electronic networks meant corrections could unfold faster than humans could process, let alone respond.

Modern markets

By 2020, electronic trading was the only system. The NYSE, once the apex of floor trading, had floor “traders” who were mostly ceremonial. NASDAQ, now a global exchange operator, ran at nanosecond speeds. Spreads on large-cap stocks were pennies or less; on bonds and commodities, they had widened in some markets where algorithmic trading provided less support.

The speed and efficiency came at a cost: fragmented markets, hidden complexity, latency-based arbitrage, and vulnerability to cascading failures in a system where humans had been almost entirely removed from the moment-to-moment execution decision.

See also