High-Frequency Trading (HFT): How It Works, Key Strategies, and Market Impact
What Makes HFT “High-Frequency”
Speed is the defining feature. While a retail investor’s order might take a few hundred milliseconds to reach the exchange, an HFT firm’s order arrives in microseconds — millionths of a second. This isn’t just about fast internet. HFT firms invest hundreds of millions of dollars in infrastructure designed to shave time at every step:
| Infrastructure | What It Does | Why It Matters |
|---|---|---|
| Co-location | Placing servers physically inside the exchange’s data center, feet from the matching engine. | Eliminates network distance. Even a few hundred meters of cable adds microseconds of latency. |
| Direct data feeds | Paying for the exchange’s raw, unprocessed market data feed instead of the public consolidated feed. | Direct feeds arrive faster and contain more granular information than the public SIP (Securities Information Processor). |
| Microwave / laser networks | Custom point-to-point communication links between exchanges (e.g., Chicago to New Jersey). | Microwave and millimeter-wave signals travel faster through air than light through fiber optic cable. |
| FPGA / custom hardware | Field-programmable gate arrays that execute trading logic directly in hardware, bypassing software overhead. | FPGAs can process market data and generate orders in under a microsecond — far faster than any general-purpose computer. |
Core HFT Strategies
Market Making
The most common HFT strategy. The firm acts as an electronic market maker, continuously posting bid and ask quotes and earning the bid-ask spread. Speed allows the firm to update quotes instantly as conditions change, managing inventory risk far more precisely than a traditional market maker. This strategy adds liquidity and generally tightens spreads.
Statistical Arbitrage
HFT firms exploit short-lived statistical relationships between correlated securities. If two stocks that normally move together temporarily diverge, the algorithm buys the lagging stock and shorts the leader, betting on mean reversion. These opportunities exist for fractions of a second — far too brief for human traders to capture.
Latency Arbitrage
This is the most controversial HFT strategy. Because prices on different exchanges update at slightly different times, an HFT firm with faster connections can see a price change on one exchange and trade on another exchange before that exchange’s price has adjusted. The firm essentially “picks off” stale quotes. Critics call this a tax on slower market participants.
Index Arbitrage
When an ETF or index fund drifts from the combined value of its underlying holdings, HFT firms buy the cheap side and sell the expensive side, locking in a risk-free profit and pushing the prices back into alignment. This is a form of arbitrage that helps keep ETF prices accurate.
Event-Driven / News Trading
Algorithms parse news feeds, SEC filings, economic data releases, and even social media in real time, trading within milliseconds of a market-moving headline. Natural language processing and machine learning extract sentiment from text faster than any human can read the first sentence.
HFT by the Numbers
| Metric | Approximate Figure |
|---|---|
| Share of U.S. equity volume | ~50% (down from ~60% peak around 2009–2010) |
| Average holding period | Milliseconds to minutes |
| Profit per trade | Fractions of a penny — profitability depends on massive volume |
| Co-location cost | $5,000 – $25,000+ per month per rack at major exchanges |
| Number of major HFT firms | Roughly 100–300 globally, with ~15–20 dominant players |
| Speed advantage | Measured in microseconds (millionths of a second) |
The Debate: Does HFT Help or Hurt Markets?
| Dimension | The Case For HFT | The Case Against HFT |
|---|---|---|
| Spreads | HFT market making has dramatically narrowed spreads over the past two decades. Retail investors pay less to trade than ever before. | Tight spreads may be an illusion if HFT firms pull quotes during stress. Liquidity is “fair weather” — available when you don’t need it, gone when you do. |
| Price discovery | Prices adjust to new information in milliseconds, making markets more efficient. | Speed-based information extraction benefits HFT firms at the expense of fundamental investors. |
| Volatility | Routine HFT activity dampens intraday volatility by continuously providing liquidity. | Algorithmic feedback loops can amplify flash crashes and dislocations. The 2010 Flash Crash is the most cited example. |
| Fairness | Anyone can invest in speed technology. HFT levels the playing field by eliminating human inefficiency from market making. | Only the wealthiest firms can afford co-location and microwave networks. The speed arms race is effectively a private tax on slower participants. |
| Market stability | HFT firms add depth to order books under normal conditions. | HFT algorithms can withdraw simultaneously during stress, creating sudden liquidity vacuums. |
Notable HFT Events
| Event | Date | What Happened |
|---|---|---|
| Flash Crash | May 6, 2010 | The Dow Jones dropped nearly 1,000 points in minutes, then recovered almost as quickly. HFT withdrawal and algorithmic feedback loops amplified the sell-off. Led to the implementation of circuit breakers. |
| Knight Capital | August 1, 2012 | A software deployment error at Knight Capital caused the firm’s algorithms to execute millions of erroneous trades in 45 minutes, losing $440 million. Knight was acquired by Getco shortly after. |
| Treasury Flash Rally | October 15, 2014 | Treasury yields swung wildly in minutes with no apparent cause. Blamed partly on HFT activity in the Treasury market. |
| IEX Exchange launch | 2016 | IEX introduced a 350-microsecond “speed bump” to neutralize HFT latency advantages. Featured in Michael Lewis’s book Flash Boys. Now an SEC-approved exchange. |
Regulation
HFT regulation remains a work in progress. The SEC and FINRA have implemented several measures:
Regulation SCI (Systems Compliance and Integrity) requires exchanges and major trading platforms to maintain robust technology systems and report disruptions. Market Access Rule (Rule 15c3-5) requires broker-dealers to implement pre-trade risk controls. Circuit breakers (Limit Up-Limit Down and market-wide halts) pause trading when prices move too far too fast, preventing cascading algorithmic sell-offs.
Some jurisdictions have gone further. The European Union’s MiFID II imposes algorithmic trading registration requirements and mandates that market-making HFT firms maintain quotes for a minimum period. Italy and France have implemented small financial transaction taxes partly aimed at curbing HFT.
Key Takeaways
- HFT uses ultra-fast technology (co-location, FPGAs, microwave links) to trade in microseconds, accounting for roughly half of U.S. equity volume.
- Core strategies include electronic market making, statistical arbitrage, latency arbitrage, index arbitrage, and event-driven trading.
- HFT has demonstrably tightened bid-ask spreads and lowered trading costs for retail investors under normal conditions.
- Critics argue HFT creates “fair weather” liquidity that evaporates during stress, amplifies flash crashes, and constitutes a speed tax on slower participants.
- Regulation includes circuit breakers, pre-trade risk controls (Rule 15c3-5), and system integrity requirements (Reg SCI).
- The 2010 Flash Crash and Knight Capital blowup are the most prominent examples of HFT-related market disruptions.
Frequently Asked Questions
Is high-frequency trading legal?
Yes. HFT is fully legal. Specific practices like spoofing (placing orders with the intent to cancel them to manipulate prices) and front-running (trading ahead of known client orders) are illegal, but the use of speed and algorithms to trade is not. HFT firms are subject to the same market rules as all other participants.
Does HFT affect long-term investors?
Minimally. If you’re buying index funds or holding stocks for months or years, HFT’s impact on your returns is negligible. The tighter spreads HFT produces actually benefit you slightly. HFT primarily affects other short-term traders and institutional investors executing large orders.
How much money do HFT firms make?
Aggregate HFT revenue in U.S. equities has been estimated at $3–6 billion per year in recent years, down from higher levels in the early 2010s as competition intensified and spreads compressed. Individual firms like Citadel Securities and Jane Street are highly profitable, but many smaller HFT firms have been squeezed out.
What is spoofing?
Spoofing is an illegal practice where a trader places large orders they intend to cancel before execution, creating a false impression of supply or demand to manipulate prices. It was explicitly banned by the Dodd-Frank Act in 2010. Several HFT traders and firms have been prosecuted for spoofing.
What is a flash crash?
A flash crash is an extremely rapid, deep price decline followed by a quick recovery — typically lasting minutes. They’re often triggered or amplified by algorithmic trading, including HFT. The most famous is the May 6, 2010 event, when the Dow fell nearly 1,000 points in minutes. Circuit breakers and Limit Up-Limit Down rules were implemented in response.