Does An Analyst’s Access To Information Vary With The Favorableness Of Their Language When Speaking To Management?

Jonathan A. Milian
Florida International University (FIU)

Antoinette L. Smith
Florida International University (FIU)

Elio Alfonso
Florida International University

May 2016

Abstract:

We examine whether analysts that use more favorable language during earnings conference calls subsequently issue more accurate earnings forecasts. Using a large sample of earnings conference calls from the 2003 – 2013 period for S&P 500 firms, we find a significantly positive relation between an analyst’s tone during a firm’s call and the accuracy of their next quarterly earnings forecast for that firm. We find a similar relation for analysts that praise a firm’s management during the call. Our findings are consistent with the favorableness of an analyst’s language reflecting their access to a firm’s management. In additional analyses, we find that female analysts, analysts with less general experience, analysts at smaller brokerage firms, and analysts that cover more industries, on average, use significantly more favorable language during earnings conference calls. Overall, our findings suggest that analysts can use their words to potentially curry favor with management.

Does An Analyst’s Access To Information Vary With The Favorableness Of Their Language When Speaking To Management? – Introduction

It has been well-documented that management-provided information (publicly and privately) has a positive impact on analyst forecast accuracy (e.g., Lang and Lundholm 1996; Bowen, Davis, and Matsumoto 2002; Chen and Matsumoto 2006; Ke and Yu 2006). Recent research finds that accessing information through private communication with a firm’s management remains an important research strategy for analysts in the post-Regulation Fair Disclosure (“Reg FD”) period (e.g., Soltes 2014; Brown, Clement, Call, and Sharp 2015).1 We expect that analysts try to maximize the value of their private communication with a firm’s management by building relationships with them. One manner in which an analyst can build a relationship with a firm’s management (and that we as researchers can measure to infer the quality of this relationship) is through the use of favorable language when publicly communicating with a firm’s management. In this study, we examine whether the favorableness of an analyst’s language during a firm’s earnings conference call is associated with an analyst’s access to management’s private information (using the analyst’s subsequent earnings forecast accuracy as a proxy for their access to management’s private information).

Prior literature supports the notion that earnings forecast accuracy conveys information about analysts’ access to management and that analysts can curry favor with management during the pre-Reg FD period to increase their access (e.g., Chen and Matsumoto 2006; Ke and Yu 2006).2 For instance, Ke and Yu (2006) find that analysts bias their early forecasts to obtain better access to management’s private information, thereby increasing the forecast accuracy of their later forecasts. Similarly, Chen and Matsumoto (2006) find that analysts who issue more favorable stock recommendations have a greater increase in subsequent forecast accuracy than analysts who issue less favorable stock recommendations. We propose that the favorableness of an analyst’s language when speaking to a firm’s management represents an additional and previously unexplored method in which analysts can curry favor with management.

Even though Regulation FD (enacted in October 2000 by the Securities and Exchange Commission) prohibits managers from selectively disclosing material, private information, recent research suggests that some analysts continue to benefit from better access to management in the post-Reg FD period (e.g., Green, Jame, Markov, and Subasi 2014; Soltes 2014; Brown et al. 2015). Green et al. (2014) find that an analyst’s access to management at broker-hosted investor conferences leads to more informative, more accurate, and timelier research by those analysts. Soltes (2014) examines the records of a large firm’s interactions with analysts, and finds a total of 75 private interactions during the year examined. In a survey of analysts, Brown et al. (2015) find that analysts rate private communication with management as a more useful input to their earnings forecasts than their own primary research. Thus, given that private communication between analysts and management continues in the post-Reg FD period, we expect that the value of this private communication to an individual analyst will vary with the quality of the relationship between the analyst and the firm’s management.

We test for an association between an analyst’s relationship with management (using the favorableness of their language during the earnings conference call as a proxy) and that analyst’s access to management’s private information (using the analyst’s earnings forecast accuracy as a proxy). Using a sample of 15,620 earnings conference calls from the 2003 – 2013 period for S&P 500 firms, we find a significantly positive relation between an analyst’s tone (i.e., counts of the positive and negative words spoken by the analyst) during a firm’s earnings conference call and the accuracy of their next quarterly earnings forecast for that firm.3 In addition to tone, we measure the favorableness of an analyst’s language with an indicator variable for analysts that praise or compliment management during the call, using the Milian and Smith (2016) list of common praises or compliments that analysts use during earnings conference calls (e.g., “congratulations,” “good job,” “great quarter,” etc.). Similar to our analysis based on an analyst’s tone, we find that analysts that praise management during the call issue more accurate earnings forecasts after the call. Our findings are consistent regardless of whether we examine within-firm variation (i.e., using firm fixed effects to compare across the analysts that cover the same firm) or whether we examine within-analyst variation (i.e., using analyst fixed effects to compare across the firms that are covered by the same analyst).

Analyst's Language

Analyst's Language

Analyst’s Language

See full PDF below.