Since the days of the Flash Crash, anxiety over robots on Wall Street seems to have dwindled. Revenue from so-called high-frequency trades in the U.S. have dropped from $7.2 billion dollars at its peak in 2009 to $1.3 billion. A survey published this year by Convergex Group found that the number of financial professionals who felt the market isn’t fair dropped to 57 percent, from 70 percent a year before.
While regulatory scrutiny and declining margins have helped spur a slowdown of high frequency trades, the financial industry has continued to develop even better algorithms for trading. This means Wall Street should probably be pretty worried about a robot takeover, warned the World Economic Forum in a report released last week.
“As the adoption of smarter and faster machines accelerates the competition for speed in gathering, analyzing and acting on data, the role of humans in trade execution will diminish,” the report found. “Intelligent machines will replace largely human activities today.”
In other words: Wall Street, the machines are coming for you.
Since the 1970s, financial markets have increasingly relied on the use of algorithms in trading. High-frequency traders, for example, use computer algorithms to trade stocks milliseconds faster than competitors without much human intervention, turning investing into a game of speed.
For now, a lot of finance jobs are still done by humans — for example, humans are often still the ones charged with scouring the world's data, looking for trends and patterns on which to base trading strategies.
But the algorithms that control the finance industry in the future will probably be able to do that, too.
The World Economic Forum laid out a few possible scenarios for Wall Street’s robot takeover.
In one, trading algorithms would move away from simply responding to movements in market price the fastest — making decisions about what to buy or sell based on data about what's buying and selling. Instead, different algorithms might home in on different data sets to get an edge, searching sources outside of the market like the news for clues as to how the market might act. This is already starting to happen. Firms like Two Sigma Investments program machines to sort through information from newswires, earnings reports, weather bulletins and Twitter. Such differentiated algorithms, the report said, would create a market that’s still fairly competitive.
In another scenario, more intelligent trading algorithms would eventually all gain access to the same data, creating a single view of the market that could decrease both trading volume and risk. Think of it like a game where every player is using the same strategy. The opportunity for arbitrage is minimal.
But, without a human's judgement, algorithms can also be prone to making mistakes with major ramifications. When Tesla announced on April Fool's Day that it was creating a competitor to the Apple Watch, stock prices immediately jumped $2. Even more drastically, in 2013, a false rumor that circulated on Twitter about an explosion at the White House caused markets to dip. Critics argued that a human reading the headline might have paused and waited for more information before trading on it, whereas an algorithm reading a headline reacts immediately.
A third scenario outlined in the report may be the Wall Street stockbroker’s best chance for survival: in this one, the World Economic Forum imagines a future in which anti-algorithm backlash inspires regulators to stymie any further automation of Wall Street.
The World Economic Forum didn’t place any kind of bet on which outcome seems most likely. Already, it has found evidence of all three scenarios beginning to play out.
But trading today is already dominated by computers. Any way you slice it, for the humans of Wall Street, the future seems pretty bleak.