Wall Street Had Five Years To Prove They Can Control Their Computers. They Failed.


Five years ago Wednesday, the American financial system got a quick and terrifying glimpse at the market panics of the future when the computer algorithms that handle most trading these days went haywire and destroyed a trillion dollars in paper wealth in a matter of seconds.

Fifteen minutes later, the rise of the machines subsided as quickly and mysteriously as it’d begun. The Dow closed that day down just 3 percent, despite a chaotic “flash crash” in the afternoon during which $862 billion of shareholder value vanished and one company’s stocks were selling simultaneously for one cent and $100,000.

It was the kind of near-calamity that demands a swift, thorough response. But on the five-year anniversary of the micro-panic, the institutions charged with making sure it can’t happen again have made almost zero meaningful progress. Regulators opted for a complex, technology-based response to the flash crash instead of a tougher, simpler alternative. Now, they are entangled in a jungle of technical challenges and unanswered questions, with no comprehensive response to the crash on the horizon.

The fix that the Securities and Exchange Commission (SEC) settled on is called the Consolidated Audit Trail (CAT), an elaborate and comprehensive data system that would allow regulators to track what the computers and high-frequency traders are doing as they’re doing it. The private companies that oversee financial markets in New York City are in charge of building the CAT, with the SEC looking over their shoulders. But after years of work, the CAT is still just an idea. No firm has been selected to create the thing, multiple deadlines have been pushed back, nobody has a clear plan to pay for the system, and the SEC has yet to sign off on a final plan.


It’s unsurprising that the CAT isn’t coming together quickly, given how absurdly complex the idea is. The proposed system would produce 58 billion separate records each day and actively track 100 million separate trading actors across dozens of separate public and private stock markets. And that’s without even incorporating futures trading into the system.

The CAT is an effort to keep up with the technical progress that ultra-fast computerized trading has made so that markets and regulators can’t get duped so easily by misfiring computer programs or criminal “spoofing” by dishonest firms. It’s a gamble that regulators and stock exchanges can win a footrace against high-frequency trading (HFT). But what if instead of trying to keep up with them so violations won’t go unnoticed, the government simply discouraged them from using fiberoptics and computers to cheat the markets?

Enter the Financial Transactions Tax (FTT), a minuscule tax on individual market transactions that would be too small to affect mom-and-pop stock pickers but big enough to cancel out the microscopic price changes that computerized trading preys upon. One version of the proposal would put a half-cent tax on every dollar traded in stocks and tax riskier derivatives trading by five one-thousandths of a penny per buck. Another draft would tax all financial transactions at the same three-tenths-of-a-cent rate. Either version would raise about $350 billion over 10 years that could be spent on any number of public goods including increased oversight of the financial system.

Multiple European countries have adopted an FTT already. The long list of people who have publicly endorsed the FTT as simple and effective includes former U.N. leader Kofi Annan, Nobel Laureate Paul Krugman, former Federal Reserve leader Paul Volcker, International Monetary Fund head Christine Lagarde, and dozens of other renowned voices on economics and human development. One former JP Morgan executive called it “a remarkably efficient tool to achieve needed reform.”

Regulators have struggled to explain the 2010 flash crash, first blaming it on legitimate trades made by a Kansas-based firm and then charging one individual trader with intentionally creating the crash from his parents’ basement. It’s not clear that either explanation is correct, though, and that very inability to convincingly diagnose what happened that day underscores how loose human control over digitized trading strategies has become. The microtax on transactions would restore much of that control by curbing the incentive that traders have to make gigantic bets on tiny price changes. An FTT would address either official version of what caused the crash, whether it was an innocent chain reaction stemming from Kansas or a nefarious get-rich-quick scheme.


In the five years that companies and the SEC have labored over the CAT, runaway computers and high-speed cheats have created other, smaller convulsions in the markets. In 2012, rogue code in the trading program used by a New Jersey firm nearly wiped out the entire company in less than an hour before staff tracked down the glitch and fixed it. A twitter hoax involving an attack on the White House in 2013 sent the computers into a brief frenzy that wiped out $130 billion in market wealth for shareholders before any human being could intercede. An energy trading company was fined $3.1 million in 2013 for using HFT tools to manipulate market prices and boost their own profit, and the trader responsible later faced criminal charges.

Stock markets and an efficient financial system are supposed to benefit every other part of the economy by delivering investment capital where it is needed, but HFT tricks contribute no such societal benefit. The practice “harms the average investor” according to one University of Michigan study and produces “no investment” in the real economy according to analysis by the think tank Demos. Even the guy who invented HFT thinks it’s now an out-of-control practice that “has absolutely no social value.”