New York Passes Law Restricting High‑Frequency Trading Firms

here is a particular kind of quiet that falls over a trading floor just before something big changes. Not silence exactly — the hum of servers never really stops — but a certain stillness, a pause in the rhythm that people who spend their days watching markets learn to recognize. That quiet descended over parts of lower Manhattan when New York moved to formally restrict high-frequency trading firms, signaling the most serious regulatory pressure the industry has faced since algorithms began dominating American stock exchanges more than a decade ago.

The legislation didn’t arrive without warning. For years, New York’s legal and regulatory machinery had been circling the industry — probing, questioning, occasionally threatening. Back in 2014, then-Attorney General Eric Schneiderman stood at a lectern at New York Law School and called certain HFT practices “Insider Trading 2.0,” a phrase that landed hard enough to generate headlines but not quite hard enough, at the time, to generate law. What followed was a long, grinding period of investigation, disclosure, and negotiation that ultimately produced exactly the kind of reform Schneiderman had pointed toward: formal, statutory restrictions on what high-frequency trading firms can do, and how they can do it.

To understand why this matters, it helps to understand what high-frequency trading actually is — not in the academic sense, but in the physical, concrete sense. These are firms whose business model depends on a few milliseconds of advantage. Their computer servers sit, quite literally, inside the same data centers as the New York Stock Exchange and Nasdaq, plugged directly into the exchange’s systems through cables that cost thousands of dollars a month to maintain. The practice is called co-location. The point is simple: by eliminating the fractions of a second that data would otherwise spend traveling across a network, these firms can see price changes and execute trades before anyone else even knows the price has moved. It’s possible that many retail investors have no idea this infrastructure exists, let alone that it has been legally permitted for years.

That legal permission is what made the situation so uncomfortable for regulators. None of this was hidden. Exchanges filed public disclosures with the SEC for every co-location service they sold. Wall Street banks paid for proprietary data feeds — faster and richer than the public ticker — without apology. The architecture of modern American equity markets had been quietly rebuilt around the needs of speed traders, and the rest of the market, from pension funds to small retail investors, had been left to adapt. Institutional investors eventually responded by routing their orders into “dark pools,” private trading venues that operate with far less transparency than public exchanges, precisely because they were trying to hide from the algorithms watching every public order that crossed the tape.

It’s hard not to notice, looking back, how much damage a single afternoon can do to an industry’s credibility. On May 6, 2010, U.S. markets dropped roughly five percent in minutes, recovered almost as fast, and left regulators scrambling to explain what had happened. The answer — computerized trading algorithms interacting in ways nobody had fully anticipated — did not inspire confidence. Two years later, Knight Capital, a New Jersey trading firm, lost close to $440 million in a single day when one of its own algorithms misfired and sent 150 mistaken orders to the NYSE in rapid succession. The firm nearly went bankrupt. These weren’t minor glitches. They were stress tests that the system failed, publicly and expensively.

The new restrictions passed in New York target the specific mechanisms that critics have long argued create unfair conditions. Co-location arrangements, while not entirely banned, are now subject to stricter disclosure and fee transparency requirements. Proprietary data feeds — the faster, richer information streams sold exclusively to firms willing to pay premium prices — face new limitations on the degree of advantage they can confer over publicly available market data. There’s also language addressing the practice of “order detection,” where algorithms identify large institutional orders and position themselves on the opposite side of those trades, effectively driving up costs for pension funds and other long-term investors before their orders can even be filled.

Whether the law will hold up under legal challenge is another question entirely. The HFT industry is not without resources, and it is not without arguments. Proponents have maintained for years that high-frequency trading actually increases market liquidity, narrows bid-ask spreads, and reduces transaction costs for ordinary investors. There is real evidence to support parts of this claim. The rise of algorithmic trading did help squeeze out the old specialist system that had long facilitated transactions on the NYSE floor — a system that had its own well-documented inefficiencies and opportunities for abuse. The question has never really been whether HFT creates some value. The question is whether the value it creates justifies the structural advantages it receives, and whether those advantages come at the expense of everyone else in the market.

There’s a sense among some market observers that this law, whatever its specific provisions, is less important for what it prohibits than for what it signals. New York has, historically, been where financial regulation gets tested before it spreads. When Schneiderman went after Thomson Reuters in 2013 for selling a two-second preview of consumer sentiment data to high-frequency firms, the company quietly ended the practice rather than fight in court. When he called Business Wire’s practice of distributing press releases directly to HFT algorithms a “tremendous victory” to abandon, the company complied. Each of those episodes moved the Overton window a little further. What’s now become law is the accumulated weight of that gradual shift, finally codified.

Federal regulators have been watching. The SEC and the Commodity Futures Trading Commission had both been examining high-frequency trading practices for years without producing the kind of decisive action that New York is now taking at the state level. It’s still unclear whether this legislation will prompt the SEC to move more aggressively, or whether Washington will let New York serve as an experiment before deciding whether to follow. The Martin Act — the century-old New York statute that gives the state’s attorney general unusually broad powers to target financial fraud — gives New York tools that federal regulators don’t always have, and New York has never been shy about using them.

What happens next, realistically, is probably litigation. The major HFT firms have legal teams prepared for exactly this kind of fight, and they will argue — not entirely without merit — that some of these restrictions interfere with practices that federal law has already sanctioned. It is a messy jurisdictional question, and courts have not always sided with state regulators in financial markets disputes. But there is also something different about the current moment. Public patience with financial complexity used to be thinner than it is now; years of flash crashes, algorithm failures, and news stories about firms recording only one trading loss in five years have produced a kind of slow-burning skepticism in ordinary investors that makes the political environment more receptive to these kinds of restrictions than it might once have been.

Walking through the financial district in the weeks since the law passed, there is, at street level, no visible sign that anything has changed. The buildings still hum. The orders still flow. Somewhere in a data center in lower Manhattan, servers are still executing thousands of trades per second, faster than the human eye can track. But the rules of the game, slowly and imperfectly, are starting to shift. Whether that shift produces a fairer market or simply a more complicated one remains, for now, genuinely uncertain.