Cafeteria table with view of the docks.

Computer says...sell, sell, sell!

14 February 2017. Published by Alex Farrow, Associate

False news stories and the stock market - a toxic combination or a risk-taker's dream?

In 2016, "post-truth" was declared the word of the year, so it is unsurprising that facts are taking a back seat. We face an alarming proliferation of fake news stories. These fictions have been blamed for everything from Brexit to Trump and may yet be blamed for the demise of Brangelina.

As amusing or outrageous as these stories are, there may be one cause for concern: the stock market. It is estimated that nearly three quarters of trades on the stock market are conducted by computer algorithms. These codes are at least in part, and arguably extensively, reliant on news media outlets and the ebb and flow of real time news stories.

In recent months a dual-faceted concern has been identified: on the one hand the toxic mix of false news stories may be driving down share prices; and on the other, news stories in themselves are catalysing a chain reaction in world markets.

The pound will probably not win an Oscar this year for its performance, but on Friday 7 October 2016 its descent was unexpected. The pound suddenly started diving between 07:07 and 07:09 from $1.26 to $1.1819 before rallying to $1.24 at 07:39. The official cause is yet to be released but in the immediate aftermath the suspected culprit was a toxic mix of rogue algorithms and the release of an FT article reporting President Hollande's call for tough Brexit negotiations.

The mathematical formulae that provide the bedrock of trading algorithms are designed to follow and monitor such fluctuations. As soon as one decides to sell it sets off a whole series of other algorithms and commands to sell, or buy. This chasing sequence thus exacerbates market fluctuations.

On 22 November 2016 it was Vinci's turn to suffer at the hands of the market, but this time the news story itself was at fault. A hoax press release alleging that an accounting error had been discovered and that Vinci's Chief Financial Officer had been fired was picked up by major news outlets including Bloomberg, prompting an almost immediate 18% plunge in share price. Even when the story had been identified as false its share price was down 3.8% at the close of the day, proving that confidence is far slower to recover than algorithms are to chase stock down. All this, even when there was no truth in the story.

According to the most recent news reports (if you now dare believe them), the next trading algorithm is being designed to follow Donald Trump's tweets. His early morning twitter tirade over the price of Lockheed Martin's F-35 jets led to an initial fall of $4 billion in Lockheed's market value. It managed to recover relatively quickly but such spikes will tempt the more aggressive investor to chase the markets. Whilst very useful for the investors, such instability is not good for the companies involved, as it reduces their market value and damages their public perception.

No doubt as the algorithms are developed some of these 'small' events will be resolved, and the way in which they chase the buy and sell decisions of other algorithms will become more sophisticated. But how can a market which is so sensitive to news adapt to an environment where truth is an intangible and falsehoods are like wolves in sheep's clothing? And when a company's value plunges and doesn't recover who will the litigators look to?