F for fake
Aliens invaded America on October 30, 1938. Radio announcers feverishly described the progress of Martian war machines across the country. Giant robots stalked the streets, firing heat-rays and releasing clouds of noxious smoke.
Orson Welles’ radio dramatisation of The War of the Worlds was so realistic that it was famously mistaken for a genuine news broadcast. Although stories of a mass panic were exaggerated, many listeners were genuinely frightened. One attempted to sue the Columbia Broadcasting System radio network for $500,000, citing nervous shock brought on by fears of an alien attack. Another claimed for the more modest expense of a train ticket bought to escape the Martian invasion.
The War of the Worlds served as an early warning of the chaos ‘fake news’ can cause. Welles’ intentions were to entertain, but today’s fakers have more nefarious ends – and the financial consequences can be far more expensive than the cost of a train ticket. State-sponsored social media accounts are interfering with democratic elections, while armies of Internet trolls are waging corporate warfare. The lines between genuine news and opinion are becoming blurred, hurting the reputation of technology firms and eroding trust in legitimate media sources.
However, experts are fighting back. Academics are building digital platforms to teach citizens how to recognise fake news, while innovative start-up companies are developing software to stem the flow of false headlines. As the technological arms race between truth and fakery gathers pace, we consider the implications for financial markets.
Fake news and propaganda
There are several reasons why fake news has become a hot-button issue in recent years. The most obvious is the advent of smartphones and social media. Advances in communications technology have always tended to be accompanied by worries that people might be exposed to the wrong messages.
In the 1930s, Adolf Hitler claimed The War of the Worlds broadcast revealed the “decadence and corrupt condition of Western democracy”.1 But he was well aware of the power of radio to shock and manipulate listeners, and radio broadcasts became a key tool in Nazi propaganda efforts during the Second World War. This prompted the Allies to work on methods to ‘inoculate’ their citizens against enemy messages, techniques that are being revived today in the battle against fake news.
As with wartime propaganda crackling over the wireless, the ubiquity of social media means false stories can now reach millions of people before they are debunked. Automated algorithms, or ‘bots’, inundate Twitter and Facebook users with information, potentially influencing the way they think and behave. The governments of Venezuela, Turkey and Ecuador have used bots to amplify partisan messaging in domestic election campaigns.2
Such tactics can have an outsized effect in emerging markets, due to a lack of established independent news sources and widespread distrust of politicians. In Brazil, for example, 66 per cent of citizens use social media as a news source, compared with less than 50 per cent in most Western countries (see figure 1). Far right presidential candidate Jair Bolsonaro reportedly made use of bots to garner support during the run up to the general election in October 2018.3
The Russian state, meanwhile, is known to run ‘bot-farms’ that disseminate cyber-propaganda abroad. According to the US Senate report on the issue, it may have used these methods in an attempt to interfere with two momentous Western votes in 2016: the US presidential election and the UK referendum on EU membership.4
Whether or not Russia affected the results of these polls, the relationship between fake news and populism appears to cut both ways: fake news can influence political choices, but political upheaval also makes people more susceptible to fake news.
“During periods of political and economic instability, people are more liable to accept information that’s false,” explains Sander van der Linden, assistant professor in social psychology at the University of Cambridge, who has studied the effects of fake news on human behaviour. “When we are under cognitive stress, we are more likely to accept information without thinking deeply about it. That is why, whenever society is undergoing changes or is under some sort of threat, there is generally more traction for propaganda.”
A technological arms race between truth and fakery is gathering pace
Social media and market volatility
Widespread political and economic disruption over the last decade, along with the rise of social media and the promise of clickbait-fuelled advertising revenues, created fertile ground for fake news to flourish. Fake stories also began to impact financial markets during this time.
Take the ‘hack crash’ of April 23, 2013. Gaining access to the Twitter account of the Associated Press, hackers posted a tweet suggesting then-US president Barack Obama had been injured in a bomb attack on the White House. The story was quickly discredited, but not before the S&P 500 had fallen sharply, temporarily wiping $136.5 billion off the value of the index.5
A few months earlier, Bloomberg had started incorporating Twitter into its software platform, and the fake tweet appeared on its terminals. The AP hack is also thought to have triggered automated trading algorithms designed to scrape text from websites and adjust portfolios in reaction to major news events. US equities quickly recovered after the tweet was deleted, but the incident raised questions over the potential of social media to stoke market volatility.
“The vast majority of the financial world uses Bloomberg as its trading backbone and research platform, so tweets that appear on Bloomberg terminals can have a big influence on markets,” says Jason Bohnet, equities portfolio manager at Aviva Investors in Chicago. “This is not just affecting individual day traders; institutional investors see these messages – and this means fake news on Twitter can be market moving. It is a growing problem that has gone underappreciated in recent years.”
Shifts in market structure could further exacerbate the problem. In the post-crisis period, central banks implemented massive bond-buying programmes and held interest rates low, which tended to dampen volatility. As monetary support is removed, however, equity and fixed income markets are likely to be subject to more-frequent volatility spikes, if not ‘mini-crashes’, as investors begin to respond more dynamically to news flow – especially the kind of headlines that are subject to political spin.
Signal and noise
Not everything posted on social media platforms is fake news, in the commonly-accepted sense of stories made up for financial gain. But unlike mainstream media outlets, true stories on social media tend to sit alongside rumours and speculation, which can make it difficult to distinguish fact from fiction. Tweeters with millions of followers can sway debate, influence asset prices and even shift impressions of what qualifies as ‘truth’.
The US president is an inveterate tweeter. Donald Trump uses his Twitter account to deliver policy announcements to his 55 million followers, as well as his personal opinions on the activities of countries, individuals and companies (which may or may not tally with the official government line). Adding to the confusion, some of the stories Trump brands as “fake” are legitimate reports from mainstream news organisations.
Trump’s tweets – initially at least – had significant market repercussions. On December 12, 2016, the president criticised the cost of Lockheed Martin’s contract to make fighter jets for the US government, which wiped two per cent, or $4 billion, off the company’s value;6 on April 2, his tweeted attacks on technology giant Amazon knocked its stock price by 5.2 per cent.7
However, the market has started to react less dramatically to Trump’s tweets in recent months, suggesting investors are beginning to discount their influence. On January 28, 2017, the president’s comment that drug makers were “getting away with murder” caused the Nasdaq Biotechnology Index to fall almost three per cent due to fears of a regulatory crackdown; a year later, Trump’s tweeted pledge to bring drug prices down barely dented the index.8
“This shows how important it is to sort signal from noise,” says Giles Parkinson, global equities fund manager at Aviva Investors. “It used to be said that you couldn’t be in finance without following Trump on Twitter. But investors have realised that most of what Trump tweets is irrelevant.”
Parkinson says it is becoming ever more important for investors to filter out irrelevant or misleading information, and redouble their focus on the true drivers of long-term returns.
“As sources of market information become more diffuse – and potentially unreliable – you need to carefully curate your news sources so that you keep up with the developments that are really important. For example, while I no longer follow Trump on Twitter I do keep an eye on Scott Gottlieb, head of the Food and Drug Administration (FDA), whose tweets on regulation are likely to be more consequential for my portfolio holdings than anything Trump says.”
Investors also need to keep track of the social media posts of corporate executives. In August 2018, for example, Tesla CEO Elon Musk tweeted that he had “funding secured” to take the electric car firm private. The Securities and Exchange Commission deemed this statement misleading. On September 27, the SEC announced it was taking action against Musk over the tweet. Shares in the company fell sharply on the news, although they recovered on October 1 after a settlement was announced. The US regulator ordered Musk and Tesla to pay a fine of $20 million each; Musk was also instructed to step down as Tesla’s chairman for three years (he will continue as CEO), and clear any future tweets about the company’s prospects before posting them.
During periods of political and economic instability, people are more liable to accept false information
Cyber hoaxes and market manipulation
The idea made-up stories might influence financial markets isn’t new. In the early days of global trade, London-based investors found out how shipping companies were faring by sending scouts to the Cornish coast to catch an early sight of their vessels’ cargo. Some of these investors soon realised they could make more money by dispensing with the due diligence altogether – and started inventing reports that inflated the value of their equity holdings.
Social media has brought more sophisticated versions of this method. A recent academic study found messages posted by Twitter bots increased volatility and influenced pricing among companies listed on the FTSE 100, indicating efforts to manipulate the market. Although the effects of this activity are, for the most part, limited to intraday trading, it could potentially threaten market stability if it becomes more prevalent.9
“If you want to spread negative news about a company, what would be your strategy? You’d find a blogger or Twitter user to write a post and then use automated accounts to add wood to the fire,” says Oleksandr Talavera, professor of finance at Swansea University and one of the authors of the report. “It is currently difficult to build a trading strategy based on social media manipulation, but regulators should keep an eye on this area.”
Most cases of online market manipulation are discovered relatively quickly. Take French construction company Vinci, whose share price dropped by 19 per cent on November 22, 2016, after a fake press release that alleged accounting irregularities at the firm appeared online. The claim was disproved and Vinci’s stock recovered most of its value before the market closed.
Other examples of fake news have longer-lasting effects. In January 2012, a US pharmaceutical firm called ImmunoCellular Therapeutics indirectly paid the author of a story on the website Seeking Alpha that claimed a “revival could be in store” for the company thanks to its development of an experimental cancer treatment. ImmunoCellular Therapeutics’ share price gained 263 per cent over a period of six months after the article was published, before falling sharply thereafter following a disappointing clinical trial.10
The SEC revealed last year it had undertaken enforcement actions against 27 individuals and companies, including ImmunoCellular Therapeutics, for involvement in stock promotion schemes. In each case, investors were misled into thinking they were reading independent analysis on investment websites, while in fact companies were secretly paying writers to promote their services.11
A study conducted by experts at the Yale School of Management following the SEC’s action found deceptive ‘paid-for’ articles tended to boost share prices of smaller firms by an average of seven per cent over a period of months (stories about larger companies tended not to sway markets; probably because many more analysts keep a closer eye on those firms). Individual investors, who lack large risk management and due diligence departments, were likely to be more adversely affected than asset managers and institutional investors; the study found exposure to fake news may even cause individual investors to lose trust in legitimate sources of financial analysis.12
Bohnet believes the SEC will have to go further to ensure both professional and non-professional investors are protected against the dangers of fake news in the future. He expects the regulator will eventually update and expand its Regulation Fair Disclosure (or ‘Reg FD’) rules to provide more guidance on online messaging, for example.
“The hard-copy prospectuses we get from companies are bulky, strictly-templated items, and that’s because the rules require them to be. By contrast, social media is a free for all – but it is still used as a source of market news and information. I think there will have to be more of a uniform process that tells us what is acceptable going forward,” Bohnet adds.
Lockheed Martin’s share price fell $4bn following a Trump tweet
The problem of fake news may get worse before it gets better. Distinguishing truth from falsehood is likely to become even more difficult in an era of artificial intelligence and machine learning. Adobe recently debuted a piece of software, dubbed ‘Photoshop for audio’, that enables users to fabricate sentences by feeding in audio clips of a person’s voice. Researchers at the University of Washington went a step further in 2017, creating an AI-driven programme that can manipulate images and audio to create seamless fakes. They made a video in which President Obama delivers an entirely invented speech.13
There are ways for experts to debunk these inventions by examining the underlying digital code, but such methods take time – and a video of a politician or corporate leader saying something inflammatory could go viral on social media in a matter of minutes. Take the AP hack: the fallout might have been far more dramatic if the hoax tweet had been accompanied by photorealistic footage of Obama ‘confirming’ the fictional bombing at the White House.
Big technology companies are keenly aware of the reputational risk – and the threat of heavy regulation – they face due to the presence of fake news on their platforms. In the US, executives from Facebook and Twitter have been hauled in front of Congress to explain how Russian bots were able to game their systems and influence voters. In India, WhatsApp is coming under pressure after the messaging service was used to spread fake stories that led to a spate of killings by lynch mobs.14
“Facebook, in particular, has been quite cavalier about the content it allowed on its platform and that is now coming back to haunt it,” says Parkinson. “The market is putting a question mark over the Facebook platform right now – and fake news may be part of that, with the slow drumbeat of regulation building as a risk in the background – even if investors are still ascribing a lot of value to the company’s other assets, such as Instagram.”
The prospect of new regulation is a live one. In the US, Internet companies are not currently held legally responsible for content posted on their platforms – under Section 230 of the Communications Decency Act of 1996, they are treated as intermediaries rather than publishers (the European Union’s E-Commerce Directive of 2000 offers similar protections). But these laws were designed in the early days of the Internet to shield small start-up companies from expensive legal costs, and there is a growing consensus they are outdated.
In congressional hearings earlier this year, politicians questioned Facebook CEO Mark Zuckerberg as to why his company continued to require legal protections. Both Republican and Democrat lawmakers have called for Internet companies to be held liable for the sale of illegal opioids on their platforms, which would weaken the Section 230 provision.15
Big technology companies are keenly aware of the reputational risk they face due to fake news
In response to growing scrutiny from politicians and even their own users, the operators of social media platforms and search engines are trying to eradicate fake news. Facebook and Google have moved to oust known fake news sites from their advertising networks, depriving them of their key revenue streams. They have also engaged in a series of acquisitions, or ‘acqui-hires’, of start-ups developing software to root out fakes, and hired new moderating teams to police their platforms.
In the short term, such investments could eat into big tech firms’ profit margins, says Bohnet: tens of thousands of extra content moderators do not come cheap. In July, Facebook’s weaker-than-expected quarterly results caused the company’s market capitalisation to fall by more than $100 billion. Its disclosure of higher costs associated with security, including the policing of content – operating expenses are set to increase by between 45 and 60 per cent this year – was held partly to blame.16
Aware of the damage fake news can cause to their brands, various companies are taking their own steps to defend themselves from false stories. This is creating opportunities for small, specialist firms – for example, UK-based Crisp and Texas-based New Knowledge are working with brands to sift through social media and identify bots spreading negative messages.
Bohnet believes that while there may be the odd ‘unicorn’ success among these young firms, larger companies with established AI development teams will have a head-start when it comes to creating programmes that can autonomously identify false stories. While fake news is currently causing significant problems for the big tech companies, their AI capabilities could provide them with solutions – and lucrative opportunities – over the longer term.
Some tech firms are already creating anti-fake news products as a sideline. One example is cybersecurity specialist Cisco: its unit Talos is developing technology that can detect the ‘stance’ of a news article, which could help identify subtle attempts to persuade readers with false information.
This may point to the future of fake news. As technology becomes more sophisticated, it is likely to involve an ever-more insidious erosion of facts, rather than grand deceptions in the style of War of the Worlds. Later in his career, Orson Welles made a film called F for Fake, a “documentary” about art forgers that gradually leads viewers to question the narrator’s reliability. Like that movie, fake news will continue to blur the lines between truth and illusion – and investors need to be vigilant to avoid the pitfalls.
The prospect of new regulation on tech firms is a live one
Continue reading AIQ
- ‘The War of the Worlds panic was a myth,’ The Telegraph, May 2016
- See ‘Spheres of Influence,’ AIQ 2.
- ‘How social media exposed the fractures in Brazilian democracy,’ Financial Times, September 2018
- Putin’s asymmetric assault on democracy in Russia and Europe: implications for US national security, US government publishing office, 2018
- ‘Lockheed Martin shares drop after Trump says F-35 program too expensive,’ CNBC, December 2016
- Social media bots and stock markets, Swansea University School of Management Working Paper
- ‘Fake news infiltrates financial markets,’ Financial Times, May 2017
- ‘SEC: Payments for Bullish Articles on Stocks Must Be Disclosed to Investors,’ SEC press release, April 2017
- ‘Does fake news sway financial markets?,’ Yale Insights, June 2018
- ‘Fake Obama created using AI tool to make phoney speeches,’ BBC News, July 2017
- ‘How WhatsApp helped turn an Indian village into a lynch mob,’ BBC News, July 2018
- ’How social media platforms dispense justice,’ The Economist, September 2018
- ‘Facebook's fight to kill fake news may hurt its profit margin,’ CNBC, November 2017
This document is for professional clients and advisers only. Not to be viewed by or used with retail clients.
Except where stated as otherwise, the source of all information is Aviva Investors Global Services Limited (AIGSL). As at 31 October 2018. Unless stated otherwise any views and opinions are those of Aviva Investors. They should not be viewed as indicating any guarantee of return from an investment managed by Aviva Investors nor as advice of any nature. Information contained herein has been obtained from sources believed to be reliable, but has not been independently verified by Aviva Investors and is not guaranteed to be accurate. Past performance is not a guide to the future. The value of an investment and any income from it may go down as well as up and the investor may not get back the original amount invested. Nothing in this material, including any references to specific securities, assets classes and financial markets is intended to or should be construed as advice or recommendations of any nature. This material is not a recommendation to sell or purchase any investment.
In the UK & Europe this material has been prepared and issued by AIGSL, registered in England No.1151805. Registered Office: St. Helen’s, 1 Undershaft, London, EC3P 3DQ. Authorised and regulated in the UK by the Financial Conduct Authority. In Singapore, this material is being circulated by way of an arrangement with Aviva Investors Asia Pte. Limited (AIAPL) for distribution to institutional investors only. Please note that AIAPL does not provide any independent research or analysis in the substance or preparation of this material. Recipients of this material are to contact AIAPL in respect of any matters arising from, or in connection with, this material. AIAPL, a company incorporated under the laws of Singapore with registration number 200813519W, holds a valid Capital Markets Services Licence to carry out fund management activities issued under the Securities and Futures Act (Singapore Statute Cap. 289) and Asian Exempt Financial Adviser for the purposes of the Financial Advisers Act (Singapore Statute Cap.110). Registered Office: 1Raffles Quay, #27- 13 South Tower, Singapore 048583. In Australia, this material is being circulated by way of an arrangement with for distribution to wholesale investors only. Please note that Aviva Investors Pacific Pty Ltd (AIPPL) does not provide any independent research or analysis in the substance or preparation of this material. Recipients of this material are to contact AIPPL in respect of any matters arising from, or in connection with, this material. AIPPL, a company incorporated under the laws of Australia with Australian Business No. 87 153 200 278 and Australian Company No. 153 200 278, holds an Australian Financial Services License (AFSL 411458) issued by the Australian Securities and Investments Commission. Business Address: Level 30, Collins Place, 35 Collins Street, Melbourne, Vic 3000, Australia.
The name “Aviva Investors” as used in this material refers to the global organization of affiliated asset management businesses operating under the Aviva Investors name. Each Aviva investors’ affiliate is a subsidiary of Aviva plc, a publicly- traded multi-national financial services company headquartered in the United Kingdom. Aviva Investors Canada, Inc. (“AIC”) is located in Toronto and is registered with the Ontario Securities Commission (“OSC”) as a Portfolio Manager, an Exempt Market Dealer, and a Commodity Trading Manager. Aviva Investors Americas LLC is a federally registered investment advisor with the U.S. Securities and Exchange Commission. Aviva Investors Americas is also a commodity trading advisor (“CTA”) and commodity pool operator (“CPO”) registered with the Commodity Futures Trading Commission (“CFTC”), and is a member of the National Futures Association (“NFA”). AIA’s Form ADV Part 2A, which provides background information about the firm and its business practices, is available upon written request to: Compliance Department, 225 West Wacker Drive, Suite 2250, Chicago, IL 60606.