Abstract: This paper investigates an unexplored linguistic feature, vocabulary richness, in earnings calls. By adopting the measurement from Yule (1944), we document a low vocabulary richness (high Yule’s K) causes a decrease in initial market reaction and an increase in abnormal trading volume after earnings calls. Both the LDA topic analysis and Yule’s K for pre-defined dictionaries support the idea that vocabulary richness captures the executive word choice, not the amount of information disclosed during earnings calls. Additional analyses using the change-on-change regressions and the shock-based instrumental variable approach, suggest the effect of vocabulary richness on market outcomes is causal. In terms of the underlying economic mechanisms, we show that the initial expectation of vocabulary richness plays an important role. Firms with a high market value, a significant analyst following, high R&D costs, and low current earnings are particularly prominent in the relationship between market outcomes and linguistic richness. Furthermore, we demonstrate that our richness proxy has long-term effects because the likelihood of future stock crash risk rises as word diversity decreases. Analysts also react to the linguistic diversity as a high Yule’s K (low linguistic diversity) decreases both analysts' forecast speed and forecast likelihood when the earnings surprise is big. We conclude that vocabulary richness provides value to investors by helping them understand executives' messages more easily.
Download Link: Full Paper