Causal language intensity in performance commentary and financial analyst behaviour

• Author: Huifeng Pan,Shuyu Zhang,Walter Aerts
• Date: 01 January 2019
• DOI: http://doi.org/10.1111/jbfa.12351
• Published date: 01 January 2019

DOI: 10.1111/jbfa.12351

Causal language intensity in performance

commentary and financial analyst behaviour

Shuyu Zhang1Walter Aerts2,3 Huifeng Pan4

1WenlanSchool of Business, Zhongnan Univer-

sityof Economics and Law, 182# Nanhu Avenue,

EastLake High-tech Development Zone, Wuhan

430073,P.R.China

2Accountancyand Finance, Antwerp University,

Antwerp,Belgium

3Schoolof Economics and Management

Accountancy,University of Tilburg, Tilburg,

theNetherlands

4Schoolof Banking and Finance, University of

InternationalBusiness and Economics, Beijing,

China

Correspondence

ShuyuZhang, Wenlan School of Business, Zhong-

nanUniversity of Economics and Law, 182#

NanhuAvenue, East Lake High-tech Development

Zone,Wuhan 430073, P.R.China.

Email:onething1984@126.com

Fundinginformation

EconomicPolicy Uncertainty index, Grant/Award

Number:15QN12

Abstract

We use automated techniques to measure causal reasoning on

earnings-related financial outcomes of a large sample of MD&A sec-

tions of US firms and examinethe intensity of causal l anguagei n that

context against extentof analyst following and against properties of

analysts’ earnings forecasts. We find a positive and significant asso-

ciation between a firm's causal reasoning intensity and analyst fol-

lowing and analyst earnings forecast accuracy respectively. Corre-

spondingly, analysts’ earnings forecast dispersion is negatively and

significantlyassociated with causal reasoning intensity. These results

suggest that causal reasoning intensity provides incremental infor-

mation about the relationship between financial performance out-

comes and its causes, thereby reducing financial analysts’ informa-

tion processing and interpreting costs and lowering overall analyst

information uncertainty.Additionally, we find that decreases in ana-

lyst following are followed bymore causal reasoning on performance

disclosure. We also find that firms with a considerable increase of

causal disclosure especially attract new analysts who already cover

many firms. Overall,our evidence of the relationship between causal

reasoning intensity and properties of analyst behaviour is consistent

with the proposition that causal reasoning is a generic narrative dis-

closure quality characteristic, able to provide incremental informa-

tion to analysts and guide analysts'behaviour.

KEYWORDS

accuracy, analyst following, causal reasoning, dispersion, narrative

disclosure

1INTRODUCTION

Explanation of performance outcomes is central to management commentary in annual reports. The Securities and

Exchange Commission (SEC) and the International Accounting Standard Board (IASB) insist on providing meaning-

ful causal explanation and related discussion of performance results in management commentary. The SEC has long

argued that investor understanding should be fostered by providing meaningful contextand interpretation of its per-

formance, market position and progress within their Management Discussion and Analysis report (SEC, 1989, 2002).

J Bus Fin Acc. 2019;46:3–31. wileyonlinelibrary.com/journal/jbfa c

2018 John Wiley & Sons Ltd 3

4ZHANG ET AL.

In its ‘Practice Statement’ on management commentary,the IASB (2010) voices similar arguments and contends that

firms should use management commentary to include their own perspective on how their business is evolving with an

analysis of how different factors are interacting to assist market participants in interpreting the financial statements

and in comprehending their content in relation to management's objectives and strategies and action to achievethose

objectives.Causal reasoning on performance is central in this respect (IASB, 2010). It should provide incremental infor-

mation byelaborating causes, reasons or motives, and other facilitating antecedents for results of operations which are

not straightforwardly apparent from the financial statements themselves. Clarifying and complementing the numeri-

cal information and accompanying notes in the financial statements, these disclosures are generally seen as a useful

extension of the financial reporting package (Baginski, Hassell, & Hillison, 2000; Clarkson, Kao, & Richardson, 1999;

Cole & Jones, 2005).

Both the SEC and the IASB present causal reasoning as an overall quality of management commentary that under-

pins its usefulness to market participants. Their depiction of useful causal reasoning embraces a repertoire of perfor-

mance explanation instances that may arise in management commentary.1Wewill investigate the assertion that the

intensityof causal language with regard to performance outcomes is relevant and useful to market participants by look-

ing into the association of causal reasoning intensity and different characteristics of the behaviourof financial analysts

as primary beneficiaries of firm-specific disclosures. Prior research documents that quality disclosure is able to attract

more analysts and to positively affect the quality of their work. In concert with regulators’ recommendations for more

causal explanation, we argue that causal reasoning on performance has the capacity to significantly increase the pre-

cision of public information and decrease the costs of financial analysts’ work, including their information processing,

interpretation and private search costs. Both higher information precision and lower costs would enhance the analysts’

ability to correctly impound all pertinent information in their recommendation and forecasting work. If public causal

reasoning on performance is useful for analysts in these respects, we expect its intensity to be associated with more

extensiveanalyst following, more consensus in their earnings forecasts, and more accurate earnings prediction.

We focus on the performance-related sections of the annual MD&A report of a large sample of US firms and use

automated content analysis techniques to capture the extentand intensity of causal reasoning in these narratives. Our

sample coversthe period 1998 to 2012. Wedocument a positive association between a firm's causal reasoning on per-

formance and the number of analysts following the firm. We also find a positive relationship between causal reasoning

intensity and the consensus of analysts’ earnings forecasts. The specific usefulness of more performance explanation

for the quality of the analysts’ work is further corroboratedby a positive association between causal reasoning and the

accuracy of earnings prediction.

Insupplemental tests, we elaborate on the association between causal reasoning intensity and analyst following. We

find that a decrease in analyst following is followed bymore intense causal reasoning on performance. This relationship

seems to be more significant for thinly-covered firms for which the marginal value of analyst following would be larger.

In additional tests, we also estimate whether causal reasoning disclosure is more attractiveto analysts who cover many

firms. Our results suggest that firms with a considerable increase of causal disclosure especially attract new analysts

who already cover manyfirms.

As these analysts can probably allocate less time to each individual firm, they may benefit more from causal reason-

ing intensity reducing information processing and interpretation time and related costs.

This study makes contributions along several dimensions. First, we extend the literature on causal reasoning

in financial disclosure (Koonce, Seybert, & Smith, 2011) by measuring causal reasoning intensity as an overall

characteristic of management commentary and investigating its relevance in large-scale archival research. Second,

we add to the literature on voluntary disclosure and properties of analyst behaviour. Sedor (2002) and Kadous,

1Priorresearch has studied aspects of management's explanatory behaviour from an impression management perspective (Merkl-Davies, Brennan, & McLeay,

2011).A basic assumption of our research is that impression management and information sharing are not necessarily mutually exclusive. Although managerial

impressionmanagement behaviour is often portrayed as deceptive self-presentation, impression management devices may also be used more innocently, as a

naturaltendency to resonate with an audience, or more proactively in order to signal management's private information and, by doing so, reduce information

asymmetries.It is our contention that impression management related devices in periodic accounting narratives need to subsume a substantial degreeoftruth

andveracity to stay credible. Otherwise its persuasiveness would be limited.