When companies release their quarterly earnings, they typically also hold a conference call in which managers present the results and respond to questions from analysts. The authors of this study examine whether the tone of language that managers use in these conference calls might reveal additional useful information to investors beyond what is contained in the accompanying press release.

The authors posit that managers are naturally more inclined to emphasize news that is good (“brightness”) over news that is bad (“bleakness”); therefore, bleakness in managerial tone is likely to be more revealing than brightness. This study explores whether changes in negativity in conference calls (measured as current-quarter negativity minus prior-quarter negativity) can be used to predict a company’s future fundamentals. It also looks at the way in which forecasts by financial analysts and stock prices respond to managerial tone.

The authors assess characteristics of managerial speech by analyzing written transcripts of earnings conference calls. They look at US common stocks—traded on the NYSE, AMEX, or NASDAQ from 2003 to 2016—for which call transcripts and analyst forecast data are available. They obtain the corresponding earnings press releases, company fundamentals, and stock price data. In total, their sample comprises 100,000 conference calls.

Tone is assessed by cross-referencing the transcripts and press releases against word lists containing 2,329 negative, 354 positive, and 297 uncertain words. Negativity is measured in two ways: net negativity (negative words minus positive words) and the negativity ratio (negative words divided by total words). The authors divide each transcript into sections that are analyzed separately: presentations, which are prepared managerial remarks, and answers, which are more spontaneous responses to questions from analysts.

The authors run regressions to assess the relationship of managerial tone changes with earnings, uncertainty, analyst forecast revisions, and forecast errors. They also look at how tone affects stock returns, both immediately after the conference call and over the following two months.

The regressions control for variables that might influence the findings—notably, the tone of the earnings press release and analysts’ questions. They also control for several other relevant factors, such as quantitative earnings surprise and a number of company, industry, and quarter fixed effects.

In a further analysis, the authors look at whether monthly calendar-time portfolio strategies can be constructed to capitalize on any underreaction to managerial tone by investors.

The authors consistently find that managerial negativity in conference calls is strongly linked to declining earnings and higher earnings uncertainty. This is true both for negativity in presentations and in answers. The results hold after controlling for the negativity of the press release and analysts’ questions as well a number of other variables. Managerial positivity shows the opposite relationship: With better earnings and lower uncertainty, the link is much weaker.

The results show that analysts do react to bleak changes in tone in the appropriate direction by revising their forecasts downward. For example, if negativity in the presentation section of the conference call rises by one standard deviation, it reduces the next quarter’s consensus earnings forecast, on average, by 1.85%. However, in both presentations and answers when managers speak more negatively, forecast errors increase. This finding implies that analysts do not adjust their forecasts downward sufficiently.

The immediate stock market reaction, meanwhile, is strongly responsive to managerial negativity. For example, a one standard deviation increase in bleak tone changes in presentations leads to a short-run abnormal stock return of –0.55%, which is in addition to the effect of negativity changes in the tone of analysts’ questions and the accompanying press release. In the days and weeks after the conference call, companies whose managers used a more negative tone underperformed a benchmark of other companies with similar characteristics. However, a postcall drift is evident—indicating that after the initial reaction, the market slowly continues to incorporate information conveyed by changes in managerial tone.

The authors find that a calendar-time portfolio strategy capitalizing on the price drift following managerial negativity changes generated risk-adjusted returns of 0.3% per month. This strategy performs more strongly for stocks with lower expected price efficiency, such as companies with less institutional ownership or fewer analysts.

This study suggests that investors would be prudent to consider changes in managerial tone when making forecasts. Even after controlling for the tone of the earnings press release and the tone of analysts’ questions, changes in managerial tone are significant in predicting future returns. In particular, the results strongly support the authors’ prediction that bleak changes in managerial tone contain more value-relevant information than bright changes.

The study finds that although analysts and investors do respond to such information to some extent by adjusting their forecasts, their response is often too muted. There is an immediate (one-day) price reaction response to a negative change in tone for a conference call and then a further drift in company price. Investors may be able to devise profitable trading strategies to capitalize on this drift.

“Our major finding is that bleak tone changes portend bad developments well beyond other available information, but neither the analysts nor the market recognizes how bad.”

The authors also point out that in the near future, machine learning techniques will increasingly be applied to big data sets to conduct more sophisticated and granular analysis of human speech. Such techniques will enable investment professionals to uncover new clues from managerial tone at very little cost.