Analysis of call transcripts suggests potential impact of language and sentiment on share prices
Earnings season is a crucial time for companies, as what they share in calls and presentations can motivate strong decisions to sell or buy their stock. And if an analysis conducted by AGF Investments is any indication, one specific topic will be of particular importance to investors.
In a new blog post, Grant Wang, senior vice president and Co-CIO AGFiQ Quantitative Investing said that since February, his team has been using natural language processing techniques to analyze worldwide earnings call transcripts.
Specifically, they searched for words and sentences highlighting sentiment toward the coronavirus pandemic, with the ultimate aim of quantifying company exposure to the coronavirus.
Noting that healthcare providers have been treated separately because of their unique role in the current environment, Wang said: “[C]ompanies in other sectors that have commented on the virus in their earnings presentations have tended to suffer from severe losses in the aftermath.”
Those that do not mention the virus, meanwhile, have performed better.
Citing U.S. banks as an example, he said that once they began reporting second quarter earnings earlier this month, the team saw an immediate spike in their impact score, with the KBW Nasdaq Bank Index dipping by 3% from April 14th to 17th. Over the same period, the S&P 500 Index advanced by 4%.
“That makes banking one of the most negatively impacted U.S. sectors so far, while utilities is the least impacted sector,” Wang said.
