As if busy CEOs and CFOs did not have enough to worry about, researchers have begun leveraging powerful computers and sophisticated algorithms to link executives’ word choice and behavior during earnings calls to subsequent stock performance. Two recent Market Watch reports discuss the application of so-called natural language processing, or NLP, to earnings calls.
An October 14 report summarizes a recent study by S&P Global Market Intelligence Quantamental Research that applied NLP to earnings call transcripts and found that executives use more words and complexity when explaining bad news. According to David Pope, managing director of Quantamental Research, “When management is trying to hide bad news, they use language that is not easily understood.” Beyond merely analyzing the language used by executives, the study also tracked executives’ pauses, repetition of words and other “metadata” to incorporate “tone” into the analysis. In addition, the study found a clear relationship between “analyst selectivity” during an earnings call and future stock price. According to Frank Zhao, the study’s lead author, “Firms whose executives exercised more analyst selectivity [i.e., called on a smaller proportion of Wall Street analysts] underperformed by about 2 percent relative to those firms whose executives called on a greater proportion of their analysts.”
Similarly, an October 13 report showcases Prattle, a company that “automates investment research by quantifying language.” The underlying principle of Prattle’s NLP algorithm is that linguistic patterns “manifest in the language of a corporate executive like a poker player’s tell.”
The takeaway? The importance of executives’ language and behavior during earnings calls cannot be overstated.