Human behavior analysis under financial information science. Evidence from corporate social responsibility

Published date19 November 2018
DOIhttps://doi.org/10.1108/LHT-11-2016-0130
Pages685-704
Date19 November 2018
AuthorFeng Jui Hsu,Yu-Cheng Chen
Subject MatterLibrary & information science,Librarianship/library management,Library technology,Information behaviour & retrieval,Information user studies,Metadata,Information & knowledge management,Information & communications technology,Internet
Human behavior analysis under
financial information science
Evidence from corporate social responsibility
Feng Jui Hsu and Yu-Cheng Chen
Department of Insurance and Finance,
National Taichung University of Science and Technology, Taichung, Taiwan
Abstract
Purpose The purpose of this paper is to investigate the relationships among corporate social responsibility
(CSR), analyst forecast accuracy and firmsearnings management behavior using US-based firms.
Design/methodology/approach The authors use the Kinder, Lydenberg, Domini (KLD) database to
construct CSR performance scores and divide all firms into ten groups from high to low as a proxy for CSR
performance. The authors obtained an initial sample of 33,364 firm-year observations from 1991 to 2012.
Filtering for records which exist in the KLD, Compustat, and Center for Research in Security Prices databases
lefts a total of 16,807 firm-year observations and CSR evaluation reports for 5,896 firms.
Findings The authorsfind that high CSR-scorefirms have lower ratesof analyst forecast errorthan their low
CSR-score counterparts, suggesting that CSR performance is a us eful means of forecasting earnings.
Furthermore, firms with better CSR performance have significantly loweraccrual-based earningsmanagement
behavior. However, the level of the manipulation behavior of real earnings management (REM) activities
increasedsignificantly in betterCSR firms, suggestingthat high CSR-score firmssubstituted REM methodsfor
accrual-based methods.REM methods are consistentwith the stipulationsof the Sarbanes-OxleyAct and allow
high CSR-score firms to bettermanipulate earnings behavior. These results hold after the authors control for
various factors related to firm financial characteristics.
Originality/value Overall, the findings have important implications for investors and regulators to more
easily assess firmsearnings manipulation behavior and earnings stability under CSR performance and
financial information in financial markets.
Keywords Information science, Social sciences, Corporate social responsibility,
Accrual-based earnings management behaviour, Analyst forecast accuracy,
Real earnings management behaviour
Paper type Research paper
1. Introduction
Over the past two decades, corporate social responsibility (CSR) has emerged as an
increasingly important issue, and an increasing number of independent firms track and rate
CSR performance and provide these ratings to investors. The rapid increase in demand
for CSR disclosure has raised some questions for researchers. Does the disclosure of
CSR-related intelligence help improve the accuracy of earnings forecasts? What is the
relationship between CSR and firmsmanipulation behavior of earnings information?
Previous studies have found that key organizational variables (e.g. top management, CEO
ability, corporate culture, and size), as opposed to environment- and industry-level variables,
are more positively associated with analyst forecast accuracy (Lang and Lundholm, 1996;
McEwen and Hunton, 1999; Sambharya, 2011), and forecast accuracy is positively
associated with analyst experience and employer size (Sinha et al., 1997; Clement, 1999).
Other studies have found that strong enforcement of accounting standards (i.e. such as
International Financial Reporting Standards) is associated with higher forecast accuracy
(Hope, 2003; Tan et al., 2011), while analyst earnings forecast accuracy is higher and
the forecast dispersion is smaller for firms audited by a Big 5 auditor (Behn et al., 2008),
and the issuance of stand-alone CSR reports is associated with lower analyst forecast error
(Yu, 2007; Dhaliwal et al., 2012).
Library Hi Tech
Vol. 36 No. 4, 2018
pp. 685-704
© Emerald PublishingLimited
0737-8831
DOI 10.1108/LHT-11-2016-0130
Received 13 November 2016
Revised 23 March 2017
Accepted 24 March 2017
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0737-8831.htm
685
Human
behavior
analysis
Furthermore, earnings management behavior entails efforts by insiders to protect their
private control of benefits to hide firm performance data or losses from outsiders through
financial information in an attempt to beat analyst forecasts (Leuz et al., 2003; Cheng and
Warfield, 2005). Studies have found that audit committee and board independence are found to
be negatively correlated with earnings management behavior, suggesting that boards
structured to be more independent of the CEO are more effective in monitoring the corporate
financial accounting process (Becker et al., 1998; Kinney and Libby, 2002; Klein, 2002). Besides,
the earnings management behavior is associated with how to analyze and interpret financial
information. Other studies found that board and audit committee members with corporate or
financial backgrounds are associated with firms that have smaller discretionary current
accruals (Xie et al., 2003). Also, CSR ratings have been found to be negatively correlated with the
level of earnings management behavior (Choi et al., 2013), and firm emphasis on the importance
of social responsibility and corporate ethics could reduce the prevalence of earnings
manipulation (Shafer, 2015). In addition, firms active in CSR are likely to issue higher quality
earnings reports (Pyo and Lee, 2013). In addition, overall CSR intensities reduce the dispersion of
earnings forecasts, volatility of stock returns and cost of capital, while increasing firm value
(Harjoto and Jo, 2015). Among various potential issues related to a firms CSR performance, we
focus on earnings quality because it plays an important role in a firms accounting decisions and
affects investor judgment. Furthermore, human behavior is a key factor to affect analyst
forecast behavior under financial and non-financial information science. This study seeks to
determine if the disclosure of CSR-related non-financial information improves the accuracy of
the earnings forecasts and how accrual-based and real earnings management (REM) are
manipulated in CSR-related disclosure, which includes information useful to investors to assess
afirms financial performance through how the firm handles issues related to stakeholders.
This study sheds light on the explored link among CSR performance, corporate earnings
quality (Lee, 2015; Hong and Andersen, 2011), and earnings management behavior to rich
the literature. We use three different proxies for earnings quality: analyst forecast accuracy
per share, accrual-based management manipulation, and REM manipulation. Specifically,
we address two research questions:
RQ1. How does CSR-related performance affect forecast accuracy?
RQ2. Does the relationship between CSR-related performance and earnings management
behavior differ with CSR-related score?
We build upon several prior studies in addressing these questions. With respect to the first
research question, CSR performance has been found to influence a firms earnings forecast
accuracy (Dhaliwal et al., 2012; Becchetti et al., 2013; Schreck, 2013). Second, we examine the
link between CSR performance and earnings management behavior to determine accounting
manipulation in high-score and low-score CSR performance (Chihet al., 2008; Cohen et al., 2008;
Prior et al., 2008; Kim et al., 2012; Grougiou et al., 2014).
To investigate our research questions, we employ social mining data to assess human
behavior analysis. First, we use CSR performance scores obtained from Kinder, Lydenberg,
Domini (KLD) STAT as proxies for firm CSR performance. In addition, we hypothesize that
the positiveassociation between higherearnings quality and responsiblebehavior is tempered
by CSR disclosure.Consistent with prior research(Dhaliwal et al., 2012), we measurefirm-level
earningsforecast error based on the averageof the absolute errors of all forecastsmade for the
year for target firms. For the earnings management behavior issue, following prior studies
(DeFond andJiambalvo, 1994; Dechow et al., 1998;Roychowdhury, 2006; Cohen etal., 2008), we
capture earnings management manipulation from accrual-based and real earnings
manipulation. Second, earnings quality was computed using accounting materials obtained
from Compustat, earnings forecast data from the Institutional BrokersEstimate System
(I/B/E/S), and stocks prices from the Center for Research in Security Prices (CRSP).
686
LHT
36,4

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT