Coronavius is the first major viral outbreak of the modern media era

There is nothing new about the virality and amplification of narratives. This element of human communication has been around nearly as long as biological viruses that spread from person to person. Across history, the spread of human disease and human narratives are driven by an important commonality: humans interacting with each other.

Much has changed over the last 16 years in terms of how information is generated and disseminated. We wanted to understand the impact of financial and economic narratives in the US media as it relates to the Wuhan coronavirus, which as of January 31, has grown to over 9,800 cases worldwide.

Photo by Macau Photo Agency on Unsplash

To conduct this analysis, Turbine Labs’ AI-intelligence platform analyzed over 35,000 articles published in January related to the business, trade, and economic impact of the coronavirus in major US-based publications. These articles comprised approximately 24 million words of text.

SARS, the last major global viral outbreak, took place in 2003 — a year before Facebook was founded and when newspaper circulation in the United States was more than double what it is today. Processing and understanding this volume of content during the SARS outbreak would have been nearly impossible, with the exception of the most advanced supercomputers. Today, through the use of artificial intelligence (AI) and Natural Language Processing (NLP), analyzing all media content related to the coronavirus topic can be completed in just hours.

At the same time, we’ve seen how media narratives are easily manipulated to the benefit of various external forces over the last five years. Today, the stakes are much higher. News about a viral epidemic can move billions of dollars in one direction or another globally and impact economic sentiment well after an event has occurred.

The Analysis

While media coverage of the virus began in late December and began to increase in January, it did not substantially correlate with market movements in the first three weeks of January.

However, on January 21, coverage spiked as the first case was reported in Washington State and as millions of Chinese prepared to travel for the Lunar New Year. Asian markets saw a see-saw of losses and gains as investors weighed in on the impact of the virus, and US markets followed closing lower. During the initial spikes of media, narrative was primarily focused on “fear,” “concern,” and “anxiety” themes.

Global stocks then fell again on January 23 as cases of the virus were confirmed in five countries and as the Chinese government moved to shut down three cities with a combined population of over 18–25 million, depending on the source. Again, narrative focused on “lockdowns,” “outbreak spreading,” and “containment.”

Image: Johns Hopkins University Center for Systems Science and Engineering

On January 23, the narrative expanded to include more substantive economic impact in terms of travel spending and fuel demand within China. However, some US publications played the economic and social impact off as hype. For example, the New York Post published a highly influential and viral article titled “Don’t buy the media hype over the new China virus.” The increase in content volume continued through January 23 and 24, including headline-grabbing, virality-inducing headlines, including this NBC News article “China expands lockdowns to cover 10 cities, builds hospital to treat coronavirus.”

Coverage expanded again on January 27 with publications focusing on flights leaving Wuhan bound for Canada, as well as additional cases being reported in the United States. The narrative turned more dismal and urgent with highly impactful coverage such as Reuters article titled “Stocks crumble as China virus toll mounts, safe havens in favor.”

A brief reprieve took place on January 28 as the markets rewarded Apple for positive earnings results, driving the broader tech sector. Coverage of the outbreak was pushed down within broader stories of other financial and economic news.

On January 29 and 30, global exchanges were again under pressure based on British Airways suspending all flights to China, followed by other international airlines. In addition, the World Health Organization again shifted the narrative, issuing an emergency decree, and calling it a “global health emergency.” Later on Thursday, January 30, The Economist accelerated the narrative stating that the coronavirus “is likely to become a pandemic.”

Much of the speed in the spread and amplification of narratives is driven by social media platforms and the algorithms they use to highlight and further spread content. In many cases, content that is perceived as dramatic or dangerous is naturally brought above the fray because humans are instinctively drawn to it, whether on or off social media. That presents additional risks for markets, as this virality creates its own news which further drives narratives, whether accurate or not. Social media platforms are becoming more responsive. Within the last 24 hours, Facebook and Google each announced they were taking steps to flag and remove misinformation related to coronavirus. But are they acting fast enough?


With a few exceptions, markets have generally counterbalanced the increasingly urgent narrative with positive economic and equities news. But the ability to ingest, process, and analyze the sheer volume of content related to this topic is impossible for any human. Indeed, it is still on the leading edge when it comes to machines.

Media narratives are certainly playing a role in how the markets and governments are responding to the coronavirus outbreak. More than the volume of stories published — often the standard measure of the interest in a topic — economists and policymakers need to understand the framing and impact of each narrative on its own merits. This is something technology is well-suited for today. Tracking these narratives is imperative to prevent over or underreaching when responding to an international outbreak.

Leigh Fatzinger is Founder and CEO of Turbine Labs

Founder/CEO of Turbine Labs. I write about information access, overload, and bias, as well as our AI-powered software. ( / @turbinelabs)