FX Traders Prove Twitter Might Have A Purpose After All

Someone let Jack Dorsey know there's still a reason to live.
(Getty Images)

(Getty Images)

Once upon a time, Twitter was supposed to “change the way we live.” Today, thankfully, we're beyond such lofty proclamations. Twitter is no Facebook. It's no Google. It's not even a LinkedIn. No longer must we delude ourselves that the road to enlightenment will be paved in tweets; we're more concerned with avoiding fake news and harassment. As Jack Dorsey recently told us, profitability itself is just a choice.

But Twitter might still have at least one purpose, a recent study tells us: signaling moves in foreign exchange markets. Analyzing 633 days in the harrowing wastes that make up FX Twitter, two researchers were able to construct a sentiment-based model that, unlike the service itself, might actually make money in the real world.

The result is all the more surprising you've ever spent any time perusing the forex side of Twitter, a largely humorless world of technical charts and vague speculation. But taken in the aggregate, there's something to be gleaned, at least if you have a giant data funnel and a proprietary model that separates the wheat from the chaff.

That's what economists Vahid Gholampour of Bucknell University and Eric van Wincoop of the University of Virginia constructed for the working paper, published by the National Bureau of Economic Research. They separated EUR/USD Twitter into informed vs uninformed traders (500 followers was the cutoff) and contentless vs “opinionated” tweets (there were around 44 opinionated tweets a day). After running some regressions they built a model that submitted hypothetical trades based on aggregated Tweet sentiment.

When the model followed only the most informed traders, it operated with a Sharpe ratio of 1.68. Compare that to the Sharpe ratio of the average carry trade, 0.91, and you have yourself some pretty decent risk-adjusted returns. From the study:

The large Sharpe ratios that we have reported suggest that there are significant gains from trading strategies based on Twitter Sentiment. The methodology developed here can easily be applied to other currencies or portfolios of currencies, as well as other financial markets such as the stock market.

Here's a graph:

Gholampour and Wincoop, 2017

Gholampour and Wincoop, 2017

At least Twitter is (hypothetically) making money for somebody out there.