On March 11th, the Dow Jones Industrial Average plunged 1,485 points, ending the longest bull-market run in history, and sending the market into nosedive the likes of which has not been witnessed since the Great Recession. While it could take years to fully understand all of the factors that led to this recent crash, a consensus has emerged that fear of an economic downturn brought on by the coronavirus has played a large role. In the wake of the collapse, new research shows that dire predictions for a COVID-19 pandemic – and its potential impact on the economy – may have been fed by the market decline itself thanks to the non-stop flow of news about the novel coronavirus.
New research from Columbia Business School’s Harry Mamaysky, Associate Professor of Professional Practice, concludes that in times of volatility like the COVID-19 pandemic, financial markets can be highly influenced by anxiety reported in the media, even on days when fundamental conditions do not significantly deteriorate.
“We’re in the middle of what I would call a non-virtuous market cycle heightened by non-stop media reports,” said Professor Harry Mamaysky. “Sure stock prices should naturally be lower now than they were a month ago when we knew less about the virus, but the extreme swings we’re seeing are tied to the ceaseless news stories covering the crisis, which in turn is causing a hyper reaction in the market that we would not see otherwise. In short, market conditions are worse than they likely should be on account of the media’s justifiable concern with the virus and its resulting impacts. I say ‘justifiable’ because the media produces news its customers want to read, but the extreme market volatility seems to be a consequence of this dynamic.”
Mamaysky used natural language processing (NLP) – a process that can analyze vast amounts of human-generated text in real-time – to analyze Reuters news articles mentioning “coronavirus” or “COVID-19,” linking the tone of these news articles to changes across key market indicators: the S&P 500, VIX, high-yield corporate bond indexes, and US Treasury yields. Mamaysky then developed a statistical model that looked at how positive or negative framing of different aspects of the COVID-19 crisis mentioned in news stories tracked market fluctuations. His findings identify a significant correlation between topical sentiment and market price, suggesting that the coronavirus outbreak created an environment where the market is particularly responsive to news, more so than at other times. He identifies important links between today’s market reactions and tomorrow’s news flow about COVID-19, and vice versa. Finally, Mamaysky shows that despite their hypersensitivity to news flow, markets are nevertheless remarkably prescient about the future incidence of coronavirus cases.
Mamaysky’s findings could represent a dramatic shift in the consensus regarding the relationship between financial markets and news coverage. The study suggests that markets and media coverage can create a vicious feedback loop, that in effect feeds its own extreme volatility. News attitudes can undoubtedly have hypersensitive reactions to bad days on Wall Street, but Wall Street similarly reacts hypersensitively to an onslaught of negative news.
