Michael Lewis
Financial markets underwent a profound transformation as electronic systems replaced human floor traders. This automation created an environment where trading algorithms compete for microsecond advantages rather than fundamental asset valuation. Companies invested hundreds of millions of dollars to build dedicated fiber optic networks, laying straight line cables between financial hubs to shave mere milliseconds off communication times. This arms race incentivized extreme capital expenditures focused solely on executing trades faster than competing firms.
Government regulations designed to ensure investors received the best available prices inadvertently fragmented the stock market. Because orders had to be routed across multiple exchanges to secure the best price, structural delays occurred based on the physical distance between exchange servers. High-frequency traders exploited this fragmented infrastructure by identifying a large order at one exchange and racing ahead to buy the available shares at other exchanges. They then resold these same shares to the original buyer at a higher price.
Large institutional trades rarely execute on a single exchange. When an investor places a substantial buy order, the request hits the closest exchange first, signaling the intent to buy. Predatory algorithms detect this initial signal and utilize their superior network speed to intercept the remaining shares on more distant exchanges. By the time the investor's original order reaches those subsequent exchanges, the shares have already been purchased and marked up. This invisible tax drains capital from traditional investors and funnels it to firms possessing the fastest network connections.
Traders began noticing that visible market liquidity vanished the moment they attempted to execute large orders. Computer screens displayed available shares, but pressing the buy button consistently resulted in incomplete fills and immediate price increases. To combat this predatory behavior, developers created routing software designed to introduce intentional delays. By calculating the exact network transit time to each exchange, this software released order fragments at staggered intervals so they arrived at all exchanges simultaneously, neutralizing the speed advantage of predatory algorithms.
Frustration with systemic market disadvantages led to the creation of a new stock exchange explicitly designed to prevent high-speed arbitrage. This alternative trading venue instituted a mandatory delay on all incoming and outgoing messages using a physical coil of fiber optic cable. This physical hurdle slows down transactions by a fraction of a millisecond. The delay gives the exchange matching engine enough time to update its prices based on broader market data before high-frequency algorithms can exploit outdated quotes.
Empirical analysis of intentional access delays reveals concrete improvements in market quality. When an exchange implements a universal speed bump, securities with high trading volumes on that venue experience reduced quoted and effective spreads. By neutralizing latency advantages, the speed mechanism prevents algorithms from sniping stale quotes during volatile price shifts. This structural friction actually improves the efficiency of price discovery by forcing market participants to compete on price and fundamental analysis rather than pure technological speed.
Critics of the narrative against automated trading argue that high-frequency trading ultimately benefits retail investors by dramatically reducing transaction costs. Before automated trading dominated the landscape, human market makers maintained artificially wide spreads that extracted significant fees from every trade. The intense competition among algorithmic market makers collapsed these spreads to pennies, vastly cheapening the cost of basic stock purchases. From this perspective, the minor latency arbitrage extracted from institutional block trades is heavily outweighed by the systemic liquidity and tight pricing enjoyed by the broader market.
The extreme value placed on trading speed fostered a culture of paranoia within major financial institutions. Proprietary trading code became the ultimate competitive asset, leading banks to weaponize the legal system against departing employees. In highly publicized cases, programmers who transferred standard open-source software architecture to new employers faced aggressive federal prosecution. These legal battles highlight a profound industry contradiction where institutions exploit regulatory loopholes to extract public wealth while fiercely utilizing federal law to protect their internal technical advantages.