Key Takeaways:

  • In heightened, high-stress periods, the news and markets together can create a negative feedback loop. Negative economic news — or news with negative implications about the economy — causes the markets to react, which inspires more negative news. 
  • These feedback loops are temporary. In 2020, the negative feedback loop continued from January to March, before it “broke” when the Federal Reserve clearly messaged its market interventions to counteract the crisis. 
  • Investors can use this knowledge from the COVID-19 pandemic to avoid making reactive decisions during these feedback loops in the future. 

Adapted from “News and Markets in the Time of COVID-19,” by Harry Mamaysky from Columbia Business School. 

Why the research was done: The media’s financial reporting in the first year of the COVID-19 pandemic offers a unique opportunity to examine the interactions between news and markets. In the early months of the pandemic, financial markets experienced almost unprecedented volatility. Like many, during the first months of 2020, CBS Professor Harry Mamaysky was largely working from home, only exposed to the outside world through the media. The onset of COVID-19 was characterized by voluminous, negative news. “And in reading The Wall Street Journal and the Financial Times and Bloomberg News, I began to notice that markets would react to yet another story about a COVID outbreak or some kind of tragedy in a hospital system somewhere in the world,” says Mamaysky, professor of professional practice in CBS’s Finance Division. 

With that news coverage, Mamaysky noticed a negative feedback loop between the news and the markets. “As markets sold off, the news coverage about the pandemic also became very negative. And it wasn't negative because anything had factually changed about COVID — it was negative because the media was interpreting the market reaction to be a negative signal about the scope of the pandemic.” As a result, Mamaysky suspected, the fluctuations in the market inspired negative news, and that negative news inspired new fluctuations in the market, creating a downward spiral.

In this research, Mamaysky set out to see if he could analytically demonstrate that this feedback loop was occurring and perhaps help market forecasters better understand these interactions.

How it was done: To analyze the news coverage from 2020, Mamaysky used the newswire service Reuters — which he chose because, according to AllSides media ratings, it’s politically unbiased — and isolated every story that referenced “coronavirus” or “COVID-19,” which accounted for 189,548 articles in 2020.

Mamaysky then employed new computational linguistic tools to analyze the coverage. “This study is pre-ChatGPT,” Mamaysky says. “We used other language processing techniques that existed at the time, and which were just starting to gain prominence in the social sciences. So, going into this crisis episode, we had the capability as a research community, for the first time, to analyze important aspects of news flow.”

He characterized the news articles based on their “narrativity” — assessing whether the stories were heavily editorialized or simply presented objective information. To do this, he categorized the articles based on their similarity to highly editorialized articles describing the 1987 stock market crash and textual distance from more fact-based Federal Reserve communications. To map the market’s reaction to the flow of news, he analyzed the news coverage against the S&P 500 index, the VIX index, the FTSE US High-Yield Market index (tracking the performance of high-yield corporate bonds), and US two- and 10-year Treasury yields.  

What the researcher found: The analysis demonstrated that the feedback loop Mamaysky had observed was, in fact, occurring during the early months of the COVID-19 pandemic. He found that markets reacted more to high-narrativity news than to low-narrativity news, and that high-narrativity news topics were more frequently associated with hypersensitivity in the market. “Unless investors are completely irrational, narrative content can impact risk premia (for example, by making investors more risk averse due to a very colorful Great Depression analogy),” Mamaysky wrote. 

His analysis showed that the feedback loop continued from January to March of 2020 before the loop was broken. That “break” correlated to the Federal Reserve’s interventions to stabilize the market, marking its clear actions to stabilize the economy. “Once the regime break happened, markets went back to a much more ‘normal’ state,” Mamaysky says. 

Why it matters: This research offers a novel understanding of the relationships between news and markets during the COVID-19 pandemic. In understanding this feedback loop, this paper encourages an analysis of past crises to examine if they also were impacted by news-markets feedback, such as the crash of 1987, the dot-com crash of 2000-2001, the global financial crisis of 2008-2009, and the European sovereign debt crisis.

Looking forward, this study may also help the news media and investors alike understand these feedback loops as they occur. “The onus is also on the news media to report in responsible ways,” Mamaysky says. “They have to find that optimal balancing point between strictly factual reporting and appealing to their readership with narrative coverage.” But whether the media sticks to the facts or produces doom-and-gloom stories, investors can use this knowledge to avoid reactionary decisions in their decision-making in the midst of a crisis.