“Spin,” said Morty. “It’s all about spin.” He pointed to the web interface where he was listening to a certain equipment manufacturing company try to explain anomalies in their reported expenses. Like hyenas, the analysts were picking apart the footnotes.

Turning to me, Morty said, “These scoundrels are masters of Orwellian doublespeak,” and then he exited the call.

About a year later, I chanced upon research that looks into actual word usage during earnings conference calls. Three authors, Paul Brockman, Xu Li, and S. McKay Price, examined transcripts from nearly three thousand such calls. One of the authors is interviewed below.

I mentioned the article to Morty, but he was not as interested as I hoped. He had dumped the stock months earlier.

“Don’t say ‘he dumped their stock’,” Morty said with a sly smile. “Say ‘he rebalanced his portfolio.’  ”

Let’s have a look at the research…

Interview with S. McKay Price

Q: Your paper, “Differences in Conference Call Tones: Managers vs Analysts,” (Financial Analysts Journal Vol. 71 No. 4 p. 24), examines the words used during earnings conference calls—whether they are optimistic, pessimistic, or neutral. What spurred your interest in this topic?

There have been quite a few studies in recent years which utilize textual analysis techniques in an attempt to further our understanding of how “soft” information affects financial decision making and outcomes.  For example, some researchers have been looking at media content, quarterly/annual reports, or IPO prospectuses.  Our focus has been on the conference call setting.  Unlike the more static sources of text-based information (i.e. newspaper articles and SEC filings), conference calls are a unique disclosure forum where you have multiple parties interacting with one another.  Prior to this study, textual analysis had only been applied to either conference calls as a whole (i.e. calculating one tone measure for an entire call) or to the introduction and Q&A portions of the calls separately (i.e. two tone measures per call).  However, manager and analyst roles are quite different from one another and we wondered about the separate effects of what each had to say during the Q&A session.  So, this study was really the next logical step in the progression.

Q: As part of the research, you ran transcripts of conference calls through an algorithm that analyzed the “tone” of speech of managers and analysts. The algorithm counted words that were positive, negative, and neutral.  Some words like “challenge” and “risk” and “small” are very context dependent.  How did you assign such words to positive, negative, or neutral categories?

Context is important.  To that end, it has been argued that how “dictionaries” (i.e. simple word lists of, say, positive or negative words) are created can be vital to the applicability of textual studies.  Most researchers in the area try to use specialized word lists.  They take a body of documents such as 10K filings or earnings announcements and scour them for words that would generally be considered positive or negative in a financial disclosure setting and compile them into a dictionary.  But even then there is some noise that is difficult to deal with.  Take the word “volatility” for example; a negative word in most contexts, but something desirable if you’re an options trader.

Textual analysis applied to financial research is a blunt instrument.  It’s more like a hammer than a scalpel, although it is becoming more sophisticated and refined with time.  I have conducted analyses using multiple word lists, some of which are finance-specific and some which are more generally oriented.  I have observed that while all the dictionaries tend to lead to the same general inferences, the context-specific dictionaries are more powerful (in a statistical sense.)


Q: You found that the most optimistic tone occurs during the prepared introduction. However, “once the session is opened up for questions, the tone becomes significantly more pessimistic.” Were you surprised by this?

This result was in line with our prior expectations.  Managers are going to put their best foot forward when given the opportunity.  Once challenged by analysts, maintaining their desired level of positivity becomes more difficult.  It’s somewhat like an adolescent reporting on the events of an evening’s activities to her/his parents.  They give the high points.  However, once the parents probe, the less-favorable nuances are revealed.


Click to continue to Part 2.