As economies around the world shut-down in response to the coronavirus pandemic, global stock markets crashed and unemployment reached levels seldom seen since the Great Depression of the 1920s.

New research from Monash University attempts to understand just what’s going through the mind of investors during such turbulent times.

The experiment took place on the “StockTwits” platform, which is designed to help investors and traders share ideas.  The researchers mined posts made on the platform during the pandemic to explore trends in investor sentiment.

“There are several unexplored and exciting directions for further investigation of social media platforms such as the StockTwits dataset,” the researcher says.  “For example, John Maynard Keynes claims that when there is too much activity in the market, e.g., during a crisis, investors/agents trade mostly by looking at each other’s strategies rather than information.”

Investor sentiment

In total, nearly 3.7 million messages were assessed on the platform, and the analysis revealed that the disagreement between bulls and bears reached a nadir on the 23rd March, which was the day the market bottomed out and began a sharp recovery.

“If I want to invest based on the relationship between the disagreement and price process on March 23rd, I would take long positions in stocks that just had the highest disagreement,” the researcher says.  “The same strategies can be taken for the disagreement processes among sectors.  This time by taking a short/long position in ETFs (Exchange-Traded Funds) related to that sector if the disagreement is low/high.”

This pattern of sentiment and disagreement appeared to be homogeneous for all investors, regardless of their experience, philosophy, or horizon.  It wasn’t consistent across sectors, however, with investors most bearish about the financial sector, and most bullish about the healthcare sector.

Perhaps unsurprisingly, the analysis found that professional investors not only used the site more often, but were also more pessimistic than novice or intermediate investors.  The study marks another example of how social media data can be used to derive real-time sentiment analysis on a range of topics during the pandemic, with other works exploring things such as physical and mental health of specific communities.