Does Limited Attention Matter In Security Analysis?

Does Limited Attention Matter In Security Analysis?

Does Limited Attention Matter In Security Analysis? Evidence From Analysts’ Reliance On Categories

Hae Mi Choi
Loyola University Chicago

Swasti Gupta-Mukherjee
Loyola University Chicago – Department of Finance

Qualivian Investment Partners July 2022 Performance Update

stocks performance 1651757664Qualivian Investment Partners performance update for the month ended July 31, 2022. Q2 2022 hedge fund letters, conferences and more Dear Friends of the Fund, Please find our July 2022 performance report below for your review. Qualivian reached its four year track record in December 2021.  We are actively weighing investment proposals. Starting in November Read More

July 1, 2016


This paper examines security analysts’ tendency to rely on category-level (e.g. industry) information as opposed to firm-specific information in issuing earnings forecasts. We find that analysts who rely more on category-level information have larger forecast errors, issue less frequent forecasts, have less impact on stock prices, and are more likely to experience job turnover. Our results are consistent with theoretical models and psychological evidence, which predict that market participants who oversimplify information processing by relying on coarse category-level information are subject to limited attention. The empirical results suggest that analysts are subject to limited attention, which has a significant association with their forecasting ability, stock price impact, and career outcomes.

Does Limited Attention Matter In Security Analysis? Evidence From Analysts’ Reliance On Categories – Introduction

The value added of credit ratings and credit rating agencies (CRAs) has been seriously questioned in recent years. CRAs’ main function is to reduce asymmetric information by assessing the credit risk of firms and securities. They communicate these assessments to capital markets by way of credit ratings, watch listings, outlooks, and rating confirmations. They explain their view in detailed reports that are released with the rating announcements.

In this paper, we investigate whether the tone of credit rating reports relates to future rating changes, earnings, and stock returns of firms in the United States during the period 2000-2012. Following previous studies on automated text recognition we measure tone through the frequency of negative and positive words as this provides an effective characterization of a text’s linguistic tone (Tetlock, 2007; Tetlock, Saar-Tsechansky, and Macskassy, 2008; Loughran and McDonald, 2011). Motivated by recent studies on the impact of different types of news (see Tetlock 2014 for a review), we hypothesize that the market reaction can be subject to behavioral biases. Our study relates to Agarwal, Chen and Zhang (2015), who show that reports on rating changes contain information to which the market reacts in a consistent manner. We investigate whether there is overreaction to salient reports, and underreaction to less salient reports. We take advantage of an institutional feature of CRAs’ communication that has not received attention in the literature so far: CRAs issue reports on rating changes (“change reports”) and reports on other subjects, such as rating confirmations, bond issues, and rating methodology (“no-change reports”).

We obtain the following principal results: First, the tone of change reports and no-change reports is equally predictive for future downgrades beyond the observable indicators of default risk. This finding suggests that the reports contain valuable information, and that the information content does not depend on the type of report. Second, the tone of change and no-change reports predicts firms’ quarterly earnings. Third, the tone affects the short-term reaction to negative rating changes, but not the one to positive rating changes and no-change reports. Most importantly, we show that the stock market overreacts to the tone of negative change reports, while it underreacts to the tone of positive change reports. In a pooled analysis of positive and negative rating changes, as conducted by Agarwal, Chen and Zhang (2015), the overreaction and underreaction effects cancel each other out. Fourth, we find that the intensity of stock market reaction to the tone of rating reports increases with issuer size and analyst coverage. We attribute the results on market reactions to differences in investor attention across report types, as well as differences in the efficiency of information processing across firms. We regard negative changes as the most salient announcements because they are more relevant for institutional investors, who are frequently subject to investment restrictions such as minimum rating requirements. We therefore interpret the results as overreaction to salient news, and underreaction to news that is less salient. The finding that the market reaction varies with size and analyst coverage corroborates the view that the information processing is not fully efficient, but dependent on investor attention.

Our paper contributes to several strands of the literature. First, research on credit ratings documents that rating announcements, especially negative rating changes, have a significant impact on capital markets, corporate finance, and firms’ future ratings. Rating announcements affect prices in stock, bond, credit default swap (CDS), and derivatives markets (e.g., Katz, 1974; Weinstein, 1977; Pinches and Singleton, 1978; Holthausen and Leftwich, 1986; Hand, Holthausen and Leftwich, 1992; Dichev and Piotroski, 2001; Hull, Predescu and White, 2004; Norden and Weber, 2004). Furthermore, the reasons for rating changes and the direction of the rating change matter for their impact (e.g., Goh and Ederington, 1993; Dichev and Piotroski, 2001). Under certain conditions, the announcement or anticipation effects are bigger or smaller (e.g., Norden, 2015). Rating announcements also have an ex ante and ex post impact on corporate finance and investment decisions (e.g., Kisgen, 2006, 2009) and explain future rating changes because of rating momentum (e.g., Altman and Kao, 1992; Carty and Fons, 1994; Lando and Skodeberg, 2002). The evidence from this literature is based on rating, outlook, and watchlist changes, while Agarwal, Chen and Zhang (2015) are the first to examine textual information from credit rating reports.

Second, there are a large number of studies with a textual analysis of news wires, analyst recommendations, newspaper stories, press releases, etc. (e.g., Tetlock, 2007; Fang and Peress, 2009; Engelberg and Parsons, 2011; Griffin, Hirschey and Kelly, 2011; Dougal et al., 2012; Norden, 2015). We directly build on the dictionary of financial negative and positive words suggested by Loughran and McDonald (2011).

Third, several studies document an overreaction to salient news and/or an underreaction to less salient news or passage of time with no news (e.g., Klibanoff, Lamont and Wizman, 1998; Palomino, Rennebog and Zhang, 2009; Tetlock, 2011; Giglio and Shue, 2014). Hirshleifer, Lim and Teoh (2009) explain why bad news is more salient than good news and Hirshleifer (2001) sheds light on the psychological effects of overreaction. Barber and Odean (2008) show that individual and institutional investors are subject to behavioral biases. Hillert, Jacobs and Müller (2014) document that the momentum effect depends significantly on media coverage, relates to the tone in media, and reverses in the long run. These results suggest an overreaction-based explanation for the momentum effect.

The rest of the paper is organized as follows. In Section 2, we describe the data, textual information, and the empirical methodology. In Section 3, we present the main results on the influence of the tone of reports on future rating changes, earnings and stock returns. In Section 4, we summarize the findings from further empirical checks and tests of robustness. We conclude in Section 5.

See full PDF below.

Updated on

No posts to display