High-frequency trading (HFT) refers to the use of sophisticated algorithms and ultra-low latency technology to execute large numbers of trades in very short timeframes — microseconds to milliseconds. HFT firms are not traditional investors: they typically end each day with flat (or near-flat) positions, profiting from the spread they earn as market makers or from tiny price discrepancies between venues.
The technological infrastructure
HFT is fundamentally a technology business. The primary constraint is latency — the time it takes for a price signal to be received, processed, and an order sent and executed. HFT firms co-locate their servers physically in the same data centres as exchanges, reducing communication latency to microseconds. They use custom-built Field Programmable Gate Arrays (FPGAs) instead of standard CPUs for order processing. They lease dedicated fibre or microwave links between exchange data centres — a straight-line microwave link between Chicago and New York is faster than the curved optical fibre routes.
This infrastructure investment is enormous — hundreds of millions of dollars for a major HFT operation. The competitive advantage is measured in microseconds.
HFT strategies
*Electronic market making*: The most common and arguably most socially valuable HFT strategy. The HFT firm continuously quotes bid and offer prices in thousands of instruments across multiple venues, profiting from the bid-offer spread. When a retail investor buys shares, they are often buying from an HFT market maker who earns fractions of a penny on the transaction. At millions of transactions per day, those fractions accumulate.
*Statistical arbitrage*: Exploiting small price discrepancies between related instruments — e.g. an ETF and its constituent stocks, or the same stock trading on two different venues. When prices diverge, the algorithm buys the cheaper and sells the more expensive, profiting when the gap closes.
*Latency arbitrage*: Using speed advantage to trade ahead of slower participants who are responding to the same market information. Controversial — some argue it is valuable price discovery; others argue it is a tax on slower investors.
The market structure debate
HFT has improved many market quality metrics: bid-offer spreads in liquid equities and futures have tightened dramatically since the early 2000s; market depth has increased; execution for small retail orders has improved. But the 2010 Flash Crash demonstrated that HFT liquidity provision is conditional — HFT firms withdrew from the market precisely when liquidity was most needed, amplifying the crash.
The debate about HFT continues: is it socially valuable market-making, or is it a tax on real investors, systematically extracting value from slower participants? The regulatory response has focused on market structure rules rather than restricting HFT directly: circuit breakers, tick size rules, and market maker obligations.