The Effect of Epidemiological Investor Sentiment on Financial Market Movements

Ruben Silverstone and Dr. Boragan Aruoba, Department of Economics, University of Maryland College Park, 3114 Preinkert Dr, College Park MD 20742

This paper will investigate the effect of public sentiment related to epidemiological crises on financial market movements. The outbreak of COVID-19 has acted as a natural experiment that provided ample evidence of the havoc that a pandemic can wreak on financial markets. The Ebola outbreak that occurred between December 2013 and January 2016 provides the ideal case study to isolate the particular vehicle of investor sentiment. Sentiment will be quantified with established text-processing methods, using news on viral events uncorrelated with other potential causes of market movements as data points, which will then be used in regression models.

This paper explores the intersection of Ebola, text-processing, and financial markets; existing research has only explored relationships between two of these three subject areas. Past studies into Ebola and financial markets provide a theoretical backing for the intuitive notion that the Ebola outbreak affected financial markets. More broadly, behavioral finance research helps to justify deviations from full trader rationality. Past studies into financial markets and text-processing sets the precedent for use of a variety of text-processing methods for analyzing financial markets. Past studies into text-processing and Ebola narrows down the broader list of financial processing methods to those appropriate for a virus-specific context.

Findings and methodologies from this investigation could contribute to the body of knowledge informing investor approaches to novel news on both current and imminent pandemics. In all likelihood, negative epidemiological sentiment will cause financial market indices to fall due to increased levels of panic and pessimistic moods, while positive epidemiological sentiment will cause the opposite effect. While the direction of these effects is fairly intuitive, quantifying the magnitude will be far more valuable. The text-processing method developed, including a novel sentiment dictionary created with the help of expert epidemiologists, can be applied in future research.

Additional Abstract Information

Presenter: Ruben Silverstone

Institution: University of Maryland College Park

Type: Poster

Subject: Economics

Status: Approved

Time and Location

Session: Poster 5
Date/Time: Tue 12:30pm-1:30pm
Session Number: 4123