Speaking of the short-term: disclosure horizon and managerial myopia

We mostly use the term “short-termism” but also occasionally refer to it as “myopia,” another commonly used word to describe excessive focus on the short term in the corporate world and capital markets.

StreetEvents also includes full transcripts from conference presentations that are excluded from the population of conference call transcripts that we use.

An electronic survey was sent to 170 business undergraduate and graduate students. The response rate was 47 %. Students were asked the following questions: “Rate the following words based on whether they refer to short or long time horizons for decision-making. Use your judgment.” We use a 1-to-5 Likert scale, with one referring to very short-term decisions and five to very long-term decisions. Students had the sixth option of responding “cannot say if the word refers to either the short- or long-term.” They were required to give an answer for all words in our dictionary and were given unlimited time to complete the survey, though the average response time was approximately 4 min.

The word “quarter” is the keyword that appears with the highest frequency in the conference call transcripts and exhibits the highest score among all short-term keywords (i.e., is classified as the least short-term oriented). In robustness tests, we construct our proxy for short-termism excluding this keyword, and our results hold (untabulated test).

Chuk et al (2013) document coverage biases in First Call. Specifically, they document that only 51 % of hand-collected earnings-forecast press releases are picked up by First Call. Furthermore, we obtain our guidance data by merging our sample with that of Brochet et al. (2011), who examine S&P 1500 firms. Hence our measure understates actual guidance issuance. While we cannot be sure how this coverage bias might influence our variable, it is conceivable that it helps capture short-termism (i.e., if firms that issue frequent forecasts are more likely to be picked up by First Call.) However, in untabulated tests, we find that our inferences remain unaffected if we limit our sample to S&P 1500 firms.

Ideally, we would like to use executive pay duration measures as developed by Gopalan et al. (2014) or Edmans et al. (2014). However, those measures can only be constructed from 2006 onward, thereby excluding a large portion of our sample.

When our dependent variable is performance-adjusted accruals, we use an OLS model. When our dependent variables are loss avoidance and small positive earnings surprises, we use probit models.

We deflate with the total number of short- and long-term oriented keywords rather the total number of words in conference calls so that our proxy is not driven by company size (i.e., conference calls of larger companies are longer).

Prior studies have explored the role of leadership and different managerial styles in influencing firms’ investment strategies (Bertrand and Schoar 2003). However, organizational inertia and path dependence are likely to limit managers’ effectiveness in determining or changing firms’ investment horizons (Liebowitz and Margolis 1995). To investigate the role of individual managers in inducing short-termism, we identify companies in our sample that experience a CEO turnover in 2002–2008, using data on corporate boards from the Corporate Library database. We choose CEOs as the unit of analysis because CEOs set the tone in an organization and are responsible for the overall performance of the company. We identify 12 instances of CEO turnover in our sample where the newly hired CEO also comes from a firm with complete earnings conference call disclosure data. We track the differences (distance) in the short-termism that these 12 pairs of companies exhibit before and after the CEO move. In untabulated results, we find that the correlation between the short-termism that a CEO’s past and current company exhibit significantly increases after the turnover (0.11 vs. 0.36 before and after CEO’s move). The average short-termism distance of past and current CEO’s employer is 0.28 before the turnover and 0.20 afterward. However, the difference of the means is not statistically significant (t-stat = 1.59), potentially due to the small number of observations.