High-Frequency Trading: How Algorithms Trade in Microseconds

Speed as a Business Model

In 2010, Spread Networks spent $300 million to lay a fiber optic cable in the straightest possible line between Chicago and New Jersey, shaving three milliseconds off the communication time between the CME Group futures exchange and Nasdaq's data center. Three milliseconds — less time than a hummingbird wing beat — was worth that investment because high-frequency trading (HFT) firms earn profits measured in fractions of a cent per trade, executed millions of times per day. HFT accounts for roughly 50 percent of U.S. equity trading volume, yet the firms involved employ relatively few people. Citadel Securities, one of the largest, handles approximately 25 percent of all U.S. stock trading with about 1,400 employees.

How High-Frequency Trading Works

HFT firms use powerful computers located physically adjacent to exchange servers to analyze market data and execute trades in microseconds (millionths of a second). The strategies depend on identifying and exploiting tiny price discrepancies that exist for fractions of a second.

Core HFT Strategies

• Market making — The firm continuously places buy and sell orders, profiting from the bid-ask spread. Electronic market makers provide liquidity and earn a small margin on each transaction. Speed allows them to update quotes before adverse price movements cause losses.
• Statistical arbitrage — Algorithms detect and exploit temporary price discrepancies between correlated securities. If two stocks that normally move together briefly diverge, the algorithm simultaneously buys the cheaper one and sells the more expensive one.
• Latency arbitrage — Exploiting the microsecond delay between when a price changes on one exchange and when other exchanges reflect that change. The fastest firm captures risk-free profit.
• Event-driven trading — Algorithms parse news feeds, earnings releases, and economic data the instant they are published, executing trades before human traders can read the first sentence.

The Technology Stack

Speed determines survival in HFT. Every component of the technology infrastructure is optimized for minimal latency.

The arms race in speed has reached physical limits. Microwave and millimeter wave links between exchanges now approach the theoretical minimum latency dictated by the speed of light. Some firms have explored laser links and even hollow-core fiber optics to gain remaining nanosecond advantages.

The Flash Crash of 2010

On May 6, 2010, the Dow Jones Industrial Average plunged nearly 1,000 points in minutes before recovering most of the loss within 20 minutes. Individual stocks traded at absurd prices — Accenture dropped to one cent while Apple briefly exceeded $100,000 per share. The event exposed vulnerabilities in market structure that HFT had amplified.

The SEC and CFTC investigation found that HFT firms did not cause the crash but amplified the selling pressure through rapid position unwinding. The incident led to new circuit breaker rules and limit-up/limit-down mechanisms on U.S. exchanges.

Arguments For and Against HFT

Benefits

• Tighter spreads — Bid-ask spreads on major stocks have declined from an average of 6.25 cents in the late 1990s to fractions of a cent, reducing transaction costs for all investors.
• Greater liquidity — HFT market makers provide continuous liquidity, making it easier to buy or sell shares at any time.
• Price efficiency — Rapid arbitrage corrects mispricings faster, ensuring prices reflect available information more accurately.
• Lower volatility (normal conditions) — Studies from the SEC show that HFT generally dampens intraday price fluctuations under normal market conditions.

Criticisms

• Phantom liquidity — HFT orders are frequently canceled before execution. Critics argue this creates an illusion of liquidity that vanishes during stress.
• Unfair speed advantage — Institutional investors and pension funds cannot compete with microsecond execution, potentially disadvantaging long-term investors.
• Systemic risk — Interconnected algorithms can amplify market dislocations, as demonstrated by the Flash Crash.
• Arms race costs — Billions invested in speed infrastructure represent social costs with questionable public benefit.

Regulatory Responses Worldwide

Regulators have adopted varying approaches to HFT oversight. The European Union's MiFID II directive, implemented in 2018, requires HFT firms to register as investment firms, maintain records of all algorithmic strategies, and implement kill switches. Italy and France imposed financial transaction taxes partly targeting HFT. The U.S. has focused on market structure rules like Regulation NMS and circuit breakers rather than directly restricting HFT activity.

The IEX exchange, founded in 2012 and popularized by Michael Lewis's book Flash Boys, introduced a deliberate 350-microsecond delay ("speed bump") to neutralize latency advantages. The SEC approved IEX as a national securities exchange in 2016, legitimizing the concept that speed equality could coexist with efficient markets.

High-frequency trading has permanently reshaped financial markets. Whether that transformation represents progress or peril depends largely on which fraction of a second one examines — and who profits from it.

This article is for informational purposes only and does not constitute financial advice.