Can Twitter predict the stock market? Derwent Capital thinks so. The London investment firm this week launched the first hedge fund that analyzes tweets for clues about which way the market is headed.
Now before I call my broker I have to blast out the following tweet: "It's the end of the world! Sell! Sell! Sell!"
Actually, the idea may not be as screwy as it sounds. The basic idea is to use proprietary software to comb through some million messages per day on Twitter to gauge public sentiment. The $40 million fund bases its investment decisions on the general mood captured in those tweets. A tide of "happy" tweets might augur a bullish day later in the week, while a predominance of "sad" messages might indicate a bearish swing.
Derwent was influenced in part by academic research showing that running tweets through an algorithm can help detect how investors are feeling. The resulting data in turn predicted the direction of the Dow Jones Industrial Average with surprising accuracy. As Derwent founder Paul Hawtin told The Atlantic:
For years, investors have widely accepted that financial markets are driven by fear and greed, but we've never before had the technology or data to be able to quantify human emotion.
Do we have that technology now? I'm not so sure. The problem isn't so much with software -- companies already mine all sorts of social media content to, say, track their ad campaigns or online branding -- although it's worth noting that algorithms have a disturbing tendency to go haywire (see "crisis, financial").
The issue is hardware, namely the flesh-and-blood kind. Simply, it's hard to tell what folks are really thinking, let alone infer their mood from 140 characters or less. The other thing is that investors lie. If there's money to be made by bombing the Twittersphere with bogus happy talk, they'll do it.
By definition, financial markets are also hard to predict. While it may be possible to deduce what makes them tick up or down on any given day -- a downbeat jobs report, strife in the Middle East, Arnold Schwarzenegger's love child -- figuring it out over the long haul is much harder.
Relative to the flood of quantifiable and less tangible inputs affecting the market at any given moment, even Twitter's entire daily content amounts to a tiny data set. Meanwhile, the advent of high-speed trading, where computers make second-to-second investment calls, raises questions about exactly what role human emotion plays in moving the market. There's a reason modern-day trading floors are so quiet.
Trading software guru John Bates touches on another reason to question Derwent's approach:
[A] sentiment algo is unlikely to be able to deal with unforeseen events, and these are increasingly common. Earthquakes, tsunamis, flash crashes, credit crises, volcanoes -- no one did a very good job of predicting any of these nor did they accurately foresee the global fallout and repercussions.
Who would have predicted that the largest one-week drop in crude oil prices in history would occur just after Osama Bin Laden's death? The markets saw his death as bearish since a source of geopolitical tension had been removed from the equation. The fact that Libya, an oil-producing nation, was still wracked by civil war -- a much more bullish indicator -- was of lesser emotional importance at that time.
Of course, the real test for Derwent will be the returns it generates for its investors. The firm is shooting for 15-20 percent. If that sounds good to you, don't forget to tweet it.
Image from Wikimedia Commons