CORPORATE SUSTAINABILITY IN AUSTRALIA: PERFORMANCE, DISCLOSURE AND GOVERNANCE Zhongtian Li Master of Business (Research) Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy School of Accountancy QUT Business School Queensland University of Technology 2020
243
Embed
CORPORATE SUSTAINABILITY IN AUSTRALIA ...Corporate Sustainability in Australia: Performance, Disclosure and Governance iii a firm has disclosed sustainability information, and it investigates
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
CORPORATE SUSTAINABILITY IN
AUSTRALIA: PERFORMANCE,
DISCLOSURE AND GOVERNANCE
Zhongtian Li
Master of Business (Research)
Submitted in fulfilment of the requirements for the degree of
Doctor of Philosophy
School of Accountancy
QUT Business School
Queensland University of Technology
2020
Corporate Sustainability in Australia: Performance, Disclosure and Governance i
Keywords
Australia, Corporate Sustainability, Corporate Social Responsibility, Diction 7,
Table 4-14 Time Stability .................................................................................... 158
Table 4-15 Inclusion of Firm Age as a Control for H3 and H4a ............................. 159
Table 5-1 Summary of Research Questions, Hypotheses and Findings .................. 170
Table 5-2 Summary of Research Conclusions and Implications ............................. 171
x Corporate Sustainability in Australia: Performance, Disclosure and Governance
List of Abbreviations
ASX Australian Securities Exchange
ESG Environmental, Social and Governance
GMM Generalized method-of-moments
PSM Propensity Score Matching
Corporate Sustainability in Australia: Performance, Disclosure and Governance xi
Statement of Original Authorship
The work contained in this thesis has not been previously submitted to meet
requirements for an award at this or any other higher education institution. To the best
of my knowledge and belief, the thesis contains no material previously published or
written by another person except where due reference is made.
Signature:
Date: 13 July 2020
QUT Verified Signature
xii Corporate Sustainability in Australia: Performance, Disclosure and Governance
Acknowledgements
First, I thank my parents and grandparents for their support and love. Since 2009,
it has been a long haul for me to complete my Bachelor, Master and Doctor programs
in Australia. Without their encouragement and support, completion of such a long
journey would have been an impossible mission. I am proud to be the second doctor
in my family, and my mother is the first doctor.
Secondly, I appreciate my principal supervisor, Professor Ellie Chapple, for her
guidance, encouragement and support throughout the entire research journey. I met
Professor Chapple in 2015 when I enrolled in the Master of Business (Research)
program. I am lucky to have Professor Chapple as my supervisor. Her encouragement,
wisdom and advice greatly help me to surmount expected and unexpected challenges
in my research journey. Professor Chapple has also been a role model for me in relation
to how to research, how to write up a paper, and how to communicate with others. In
here, I would like to express my deepest gratitude to Professor Chapple.
Thirdly, I thank my associate supervisor, Dr. Elisabeth Sinnewe, and my external
supervisor, Dr. Shamima Haque. The time and effort which Dr. Elisabeth Sinnewe
devoted to my thesis is unquantifiable. Comments made by Dr. Elisabeth Sinnewe
helped me greatly to improve this thesis. I met Dr. Shamima Haque in 2015 when I
enrolled in the Master of Business (Research) program. Although Dr. Shamima Haque
moved to the United Kingdom in 2017, we still make a great team researching topics
within corporate social responsibility, and I received many comments and help from
her.
Fourthly, I thank my co-researcher and wife, Dr. Jing Jia, for so many things.
My life dramatically changed (in good ways) because of her. I never thought my
Doctor program could grant a degree as well as make me meet my wife. Without her
love and support, my research journey would have been much bumpier and more
challenging.
Fifthly, I thank A/Professor John Nowland and Dr. Ammad Ahmed for their help
and advice in my research journey. It is noteworthy that part of this thesis has been
presented in the 2017 A-CSEAR conference, the 2018 FIRN PhD symposium and the
Corporate Sustainability in Australia: Performance, Disclosure and Governance xiii
2019 APIRA conference. Professor Jacquelyn Humphrey, Professor Charl de Villiers
and other participants also give valuable comments about this thesis.
Sixthly, I also thank my friends at the Queensland University of Technology who
were always there with me sharing support, friendship, knowledge, and locations of
nice restaurants in Brisbane.
Seventhly, I acknowledge the financial support from Australian Government
Research Training Program Scholarship and QUT Excellence Top-up Scholarship.
Last but not least, I thank Clare Moore and Marita Smith for their professional
copy editing and proofreading advice in relation to this thesis.
Chapter 1: Introduction 1
Chapter 1: Introduction
Corporate sustainability has been endorsed by some international organizations,
including the United Nations.1 Academic literature defines corporate sustainability in
different ways. For example, Dyllick and Hockerts (2002, p. 131) define this concept
as “meeting the needs of a firm’s direct and indirect stakeholders ……, without
compromising its ability to meet the needs of future stakeholders as well”; Eccles and
Serafeim (2013, p. 66) connect it with the society: a sustainable firm “creates value for
its shareholders” and leads to a sustainable society; from the broader perspective, Van
Marrewijk and Werre (2003) and Bansal (2005) and Montiel (2008) elaborate it as a
tridimensional construct that include economic prosperity, environmental integrity and
social equity. Many dominant Australian firms, including ANZ and AGL Energy, are
signatories to the UN Global Compact,2 a network that promotes sustainability; the
Australian Securities Exchange (ASX) also joined the Sustainable Stock Exchanges
Initiative that improves transparency of corporate sustainability. Therefore, corporate
sustainability in Australia is expected to continue to grow in importance.
Corporate sustainability performance is defined as “the extent to which a firm
embraces economic, environmental, social and governance factors into its operations,
and ultimately the impact they exert on the firm and society” (Artiach, Lee, Nelson, &
Walker, 2010, p. 32). As Wood (1991a, 1991b, 2010), Tregidga, Milne, and Kearins
(2014), Global Reporting Initiative (2014), Rao and Tilt (2016) and Guthrie (2016)
suggest, there is a relationship between sustainability disclosure and sustainability
performance: disclosure of sustainability information not only mitigates information
asymmetry between firms and their stakeholders but also holds corporate insiders
accountable. Transparency of corporate sustainability connects with sustainability
disclosure.3 It is considered as “communicating an understanding of how the flows of
material, resources and services between corporations, capital markets, society, the
1 The United Nations Global Compact and the Sustainable Stock Exchanges Initiative are backed up
by the United Nations to promote corporate sustainability. 2 More information about the UN Global Compact in Australia can be found on
http://www.unglobalcompact.org.au/ (access date is 1 July 2019). 3 This thesis uses sustainability disclosure and sustainability reporting interchangeably and does not
distinguish the two terms.
2 Chapter 1: Introduction
economy and the environment affect the mutual ability of those systems to continue
and flourish” (Guthrie, 2016, p.11).
Those charged with the governance of a firm are responsible for the operational
decisions affecting firm performance and are also responsible for communicating
aspects of performance to stakeholders. In the area of sustainability performance,
established frameworks of reporting and disclosure are not necessarily mandated, nor
entrenched. Accordingly, reporting and disclosure of sustainability performance is a
governance choice; increasingly observed through governance mechanisms referred to
as sustainability committee4 or chief sustainability officer (Miller & Serafeim, 2015).
By integrating sustainability with corporate governance, the sustainability committee
is expected to focus on (at least some if not all) sustainability issues on the committee’s
agenda. Aiming to advise the board and management about corporate sustainability
(Peters & Romi, 2014, 2015), the sustainability committee has a quite long (but not
necessarily pervasive) history in Australia. For example, as an early adopter, CSR
Limited formed a relevant committee in 1994, and Arrium Limited formed a similar
committee in 2001.
The remainder of Chapter 1 will expand upon this thesis as follows. Regulatory
background regarding corporate sustainability in Australia is detailed in Section 1.1,
and motivations of the thesis are elaborated upon in Section 1.2. Research purposes
and research questions are in Section 1.3. Contributions and practical implications are
explained in Section 1.4. Findings are reported in Section 1.5. Section 1.6 summarizes
this thesis in the ‘Libby Boxes’ format (Libby, Bloomfield, & Nelson, 2002) and a
Chapter 2 draws on sustainability literature from the US and Europe, whereas
this thesis investigates the Australian market and context. The regulatory framework
in Australia has similarities and differences with both the US and the European Union,
while this thesis elaborates nuances as pertains to Australia’s regulatory environment.
Australia is arguably more aligned to the regulatory environment of the US than of the
4 Alternative names can include “public policy committee, sustainability committee, corporate social
responsibility committee, environmental health and safety committee, etc” (Peters & Romi, 2015, p.
173).
Chapter 1: Introduction 3
European Union, as the latter is more stringent.5 Internationally mandating
sustainability disclosure, Directive 2014/95/EU has been introduced and enforced by
the European Union, and Directive 2014/95/EU deals with a range of sustainability
issues, including health and safety, pollution emissions, and charity. In contrast, the
US does not impose such comprehensive mandate on sustainability disclosure, and
issue-specific mandates are introduced, including California Transparency in the
Supply Chain Act. In relation to the Asia-Pacific region, several countries use
comprehensive mandate on sustainability disclosure to improve the transparency of
corporate sustainability. For example, Hong Kong introduced mandatory sustainability
disclosure in 2016, namely Appendix 27 of Main Board Listing Rules and Appendix
20 of GEM Listing Rules. Thus, compared with firms operating in the European Union
and the Asia-Pacific region whose regulatory environment is more stringent,
Australian firms remain largely subject to discretionary and voluntary choices in terms
of corporate sustainability disclosure. Although Australian firms do not have
comprehensive mandate on sustainability disclosure, they still face some issue-specific
mandates, which are discussed in the following paragraph. I discuss the Australian
regulatory framework in two parts, one component focuses on sustainability disclosure
and the other on corporate governance and sustainability.
Sustainability disclosure remains voluntary in Australia, and the regulations
regarding sustainability disclosure are evolving. Section 299(1)(f) of the Corporations
Act 2001 was introduced in 1998 to require firms to disclose activities subject to
environmental regulations. Frost (2007) highlights that this section still leaves much
discretionary leeway for managers. Introduced in 2011, Section 1013DA of the
Corporations Act 2001 focuses on product disclosure statements, requiring financial
product issuers to disclose whether labour standards as well as environmental, social
or ethical considerations were considered in the selection, retention or realisation of
the investment. Section 299A of the Corporations Act 2001 was added in 2013, to
require public firms to present their operating and financial review in their directors’
5 United Nations Environment Programme and KPMG (2006) and United Nations Environment
Programme, KPMG, Global Reporting Initiative, and Centre for Corporate Governance in Africa
(2010, 2013, 2016) suggest that firms in the European Union are subject to more stringent regulations
regarding corporate sustainability, and firms in the US are more flexible regarding how to practice and
disclose sustainability. Information about Directive 2014/95/EU can be found at https://eur-
lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32014L0095 (access date is 12 June 2020).
4 Chapter 1: Introduction
reports. The regulatory guide released by the Australian Securities and Investments
Commission articulates: operating and financial review shall contain environmental
and other sustainability risks that may be financially material (see Regulatory Guide
247). Regulatory guides 68 and 65, corresponding to Section 299(1)(f) and 1013DA,
were also issued. It is reasonable to argue that firms in Australia have much discretion
in whether, what and how to disclose sustainability performance.
Regulations with regard to sustainability disclosure outside of corporate reports
are relevant. For example, the National Pollutant Inventory was established in 1998 to
ask operators of all facilities that emit above minimum threshold levels to submit
annual reports which quantify their emissions of various land, air and water pollutants.6
Another example is the National Greenhouse and Energy Reporting Act 2007 – this
Act requires facilities that emit above certain levels of greenhouse gas to report their
emissions.7 In summary, firms need to report their environmental impact to regulators,
if they are above a designated threshold. The social impacts of corporate activities are
also reportable. The Workplace Gender Equality Act 2012 requires firms with 100 or
more staff to submit reports about gender diversity in their workforce.8
To a quite limit extent, financial reporting standards may incidentally require
sustainability disclosure. For example, AASB 6 and 137 are relevant to mining firms,
requiring recognition and disclosure of the restoration costs due to abandoned mine
sites. Thus, this thesis suggests while Australian firms are currently subject to select
regulations about sustainability disclosure, firms continue to retain some flexibility
regarding the scope and content of such disclosure.
The integration of sustainability and corporate governance remains on the
agenda of Australian legislators. An early example of attempts at such integration is
the Corporate Code of Conduct Bill 2000 that attempts to regulate how firms conduct
their business in a socially responsible manner overseas, and it was eventually rejected
in 2002. In 2006, whether and how to better integrate sustainability with corporate
6 More information about the National Pollutant Inventory can be found on http://www.npi.gov.au/
(access date is 1 July 2019). 7 More information about the National Greenhouse and Energy Reporting Act 2007 and Clean Energy
Regulator can be found on http://www.cleanenergyregulator.gov.au/NGER (access date is 1 July
2019). 8 More information about the Workplace Gender Equality Act 2012 and Workplace Gender Equality
Agency can be found on https://www.wgea.gov.au/about-the-agency (access date is 1 July 2019).
Chapter 1: Introduction 5
governance was raised by two committees, namely the Parliamentary Joint Committee
on Corporations and Financial Services9 and Corporations and Markets Advisory
Committee10, and no changes to corporate law were recommended at that time. In
2014, the Governance Institute of Australia11 again invited discussion about whether
clarifications to the relationship between the interests of shareholders and interests of
the corporation should be explicitly made in corporate law, although no actions or
changes followed. Deva (2011) suggests that absence of clarity about the duties of
directors and purpose of the corporation defy attempts to integrate sustainability with
corporate governance at corporate law level. A survey conducted by Anderson et al.
(2006) identifies directors often hold quite diverse and inconsistent views regarding
their duties. To summarise, this thesis suggests that whether and how to integrate
sustainability with corporate governance is largely determined by board of directors in
firms with discretion, and corporate law (nor other regulation) does not explicitly
regulate how firms regard sustainability.
In summary, the regulatory framework in Australia is introduced in two parts:
how sustainability disclosure is regulated, and how sustainability is considered at the
level of corporate law and other regulations. Compared with countries which impose
stringent regulations on sustainability disclosure (e.g. South Africa and the European
Union), Australia grants flexibility in scope and content of sustainability disclosure to
firms. Aligning with other countries (Sjåfjell et al., 2015), Australia does not position
sustainability in corporate governance at the level of corporate law. This thesis argues
that Australian firms have freedom to corporate sustainability. Motivations of this
thesis are elaborated in following section.
1.2 RESEARCH MOTIVATIONS
Policy discussion regarding sustainability disclosure in Australia attracts greater
global scrutiny. Some exchanges in the Asia-Pacific region, including the Hong Kong
Exchange, Singapore Exchange and Taiwan Stock Exchange, have recently mandated
9 This committee published a report, Corporate Responsibility: Managing Risk and Creating Value, in
2006. 10 This committee released a report, The Social Responsibility of Corporations, in 2006. 11 This organization published another report, Shareholder Primacy: Is There a Need for Change, in
2014.
6 Chapter 1: Introduction
sustainability disclosure.12 In Australia, two industry bodies have issued sustainability
disclosure guides.13 More countries, including Argentina and the European Union,
have mandated sustainability disclosure in small and medium-sized firms above a set
of legal threshold.14 Following this momentum of mandated disclosure15, legislators,
regulators, and market operators in Australia are considering whether and how to alter
the current regulatory framework.
The Senate Economics References Committee16 and Australian Securities and
Investments Commission17 emphasize sustainability, climate change and its negative
effects on firms, in their recent public reports. Highlighting concerns of legislators and
regulators, as commissioner of the Australian Securities and Investments Commission,
John Price discusses corporate sustainability as part of directors’ duties and encourages
firms to take into account sustainability in their decision making (McLeod & Hurley,
2018). The fourth edition of Corporate Governance Principles and Recommendations
of the ASX mentions ‘long term sustainable value’ and ‘standing in the community’
to encourage a corporate culture of ‘acting lawfully, ethically and responsibly’.18 Thus,
in examining sustainability disclosure and the role of governance mechanism (the
board committee) regarding sustainability performance, this thesis associates with the
current silo of literature about corporate sustainability in Australia, contributing to this
policy discussion.
As socially responsible/ ethical investment assets in Australia and New Zealand
increased by 247% from 2014 to 2016 to reach $516 billion (Foo, 2017), a reasonable
expectation is that corporate sustainability will attract increasing attention in the future.
Three reasons drive this boom of socially responsible investments in Australia. First,
12 See Listing Rules of corresponding exchanges. 13 Two versions are issued by the Financial Services Council and Australian Council of
Superannuation Investors, one in 2011 and the other in 2015. 14 See the Ley Nº 2594 de Balance de Responsabilidad Social y Ambiental (BRSA) and Directive
2014/95/EU. 15 According to United Nations Environment Programme and KPMG (2006) and United Nations
Environment Programme et al. (2010, 2013, 2016), there is indeed a trend of mandatory disclosure of
corporate sustainability. A recent example is the European Union that mandates disclosure of corporate sustainability in its members. 16 This committee published a report, Carbon Risk: A Burning Issue, in 2017. 17 As the market regulator in Australia, the Australian Securities and Investments Commission
released a report, Climate Risk Disclosure by Australia’s Listed Companies, in 2018. 18 The fourth edition of Corporate Governance Principles and Recommendations of the ASX can be
downloaded from https://www.asx.com.au/regulation/corporate-governance-council.htm (access date
is 10 August 2019).
Chapter 1: Introduction 7
new legislations discussed at Section 1.1, including Section 1013DA and Section 299A
of the Corporations Act 2001, promote investors’ and fund managers’ awareness about
socially responsible investment (Foo, 2017). Secondly, the third edition of Corporate
Governance Principles and Recommendations issued by the ASX further promotes this
awareness by suggesting that listed firms need to evaluate their own exposure to risks
of corporate sustainability and communicate how they manage these risks (Foo, 2017).
Thirdly, this boom of socially responsible investments in Australia also mirrors the
global trend of socially responsible investment (Jones, et al., 2008; Foo, 2017). Pérez-
Gladish, Benson, and Faff (2012), surveying investors sensitive to social responsibility
in Australia, and Eccles, Kastrapeli, and Potter (2017), surveying investors in multiple
countries, identify an upward trend of socially responsible investment and indicate that
investors consider low quality of sustainability disclosure as a key obstacle. Therefore,
greater demand for socially responsible investment needs better disclosure to improve
market efficiency. In this context, Australian firms are expected to prepare more and
better sustainability disclosure. Interviewing stakeholders, KPMG and SustainAbility
(2008) and EY and Global Reporting Initiative (2013) find that more stakeholders
focus on how sustainability is integrated with corporate governance. Thus, examining
corporate sustainability disclosure, the role of the board committee as to sustainability
and sustainability performance, this thesis contributes to the interests of stakeholders:
(1) the reflection as to relationships between disclosure and performance sheds light
on the disclosure quality, a concern to stakeholders; (2) the sustainability committee
may be a solution to integrating sustainability into corporate governance, providing
another perspective to stakeholders in corporate sustainability.
Corporate sustainability generates practitioner and academic research; the extant
literature provides fresh research opportunities. The global surveys of KPMG (2013,
2015, 2017) and the local surveys conducted by Australian Council of Superannuation
Investors (2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017) reveal that more
Australian firms are engaging in sustainability disclosure over time. Using the Global
Reporting Initiative (GRI), Frost, et al. (2005) and Beck, Frost, and Jones (2018) assess
sustainability disclosure by matching disclosure content with the GRI criteria. Taken
together, prior literature19 is interested in research questions about content (e.g. which
19 Readers may refer to Chapter 2 for more information.
8 Chapter 1: Introduction
sustainability themes are reported on, and how much information is released in each
theme). While linguistic characteristics of disclosure have been frequently investigated
in accounting and finance literature, the characteristics in sustainability disclosure are
largely omitted. Identifying this literature gap, Beattie, McInnes, and Fearnley (2004),
Cho, Roberts, and Patten (2010), Beattie (2014), Jain and Jamali (2016) and Loughran
and McDonald (2016) suggest that the literature could benefit from linguistic analysis
of sustainability disclosure, as linguistic analysis can provide additional insights about
how firms communicate their sustainability performance. My thesis aims to contribute
to addressing this literature gap and is motivated to augment prior literature by
investigating sustainability disclosure characteristics using various linguistic
characteristics.
While many studies investigate how corporate governance affects sustainability,
few studies consider specifically how board committees relate to firms’ sustainability
performance (Jain & Jamali, 2016; Walls, Berrone, & Phan, 2012). A research gap of
this group of studies, including Rodrigue, Magnan, and Cho (2013) and Dixon-Fowler,
Ellstrand, and Johnson (2017), is that rich characteristics of sustainability committee,
including composition, authority, resources, and diligence, have been omitted. For
example, in Australian literature, Rankin, Windsor, and Wahyuni (2011) analyse how
the presence of sustainability committee relates to disclosure of carbon dioxide
emissions, revealing that simply the existence of committee does not affect how
Australian firms disclose relevant information. Thus, other meaningful characteristics
of the committee largely are omitted. Following the encouragement of Peters and Romi
(2014, 2015) as well as Dixon-Fowler et al. (2017), my thesis aims to contribute to
addressing this literature gap by investigating the sustainability committee in Australia
with two measures, one focuses on the committee’s presence and the other on
committee effectiveness based on various committee characteristics.
In summary, there are three motivations for this thesis. First, it is motivated by
current and anticipated changes to the regulatory framework in Australia, contributing
to relevant policy discussion. Secondly, it is motivated by the trend that corporate
sustainability continues to grow in importance, delivering the information that matters
to stakeholders. Thirdly, the thesis is motivated by research opportunities highlighted
in prior literature, utilizing linguistic analyses in exploring dimensions of disclosure
and better understanding the role of the sustainability committee in driving corporate
Chapter 1: Introduction 9
sustainability performance. The research purpose and questions are discussed in the
following section.
1.3 RESEARCH PURPOSE AND QUESTIONS
1.3.1 Research Purpose
This thesis aims to examine linguistic characteristics of sustainability disclosure,
firms’ experience of sustainability disclosure and the role of sustainability committees,
which are related to corporate sustainability performance. To be specific, seven textual
characteristics include how much sustainability information, quantitative information
and information about environmental impact are disclosed, to what extent disclosure
is readable, and to what extent disclosure is communicated in optimistic, certain and
clear tones; experience of disclosure is operationalized as number of years that a firm
has disclosed sustainability information; the ESG ratings from the Thomson Reuters
Asset4 (Asset4) are operationalized as sustainability performance; two elements of the
sustainability committee, namely presence and effectiveness, are considered.
This thesis examines these seven textual characteristics for two reasons. First, as
many studies (Cho et al., 2010; Lehavy, Li, & Merkley, 2011; Loughran & McDonald,
2014a; Parker, 2011) suggest, this PhD thesis researches the most frequently-examined
textual characteristics. Secondly, I use Diction 7. The availability of word dictionaries
and analytical software affect which textual characteristics that can be examined. The
ESG ratings from Asset4 are a legitimate data source for sustainability performance,
as demonstrated in many seminal studies (Cheng, Ioannou, & Serafeim, 2014; Ioannou
2013a), knowledge about corporate sustainability is perceived to improve quality of
assurance services; financial analysts that are interested in socially responsible
investing may find the thesis useful, as it examines the sustainability committee, a
nexus between board of directors/senior management and sustainability issues. Third,
from a broad aspect, this thesis aligns with Bowen (2014) who argues that symbolic
action has social cost; as the resources devoted to potentially symbolic actions, such
as sustainability disclosure and sustainability committee, ought be switched to
substantial action. Even though the resources devoted to sustainability disclosure and
the sustainability committee are bearable by each firm, such resources collectively
could be substantial to society. If they do not deliver substantial benefits to society, it
may be reasonable to reorientate resources to more productive places.
Chapter 1: Introduction 17
In summary, this thesis delivers practical implications in two ways. First, as its
findings suggest, future regulatory framework can pay more attention on readability
of sustainability disclosure. Secondly, as this thesis demonstrates the positive effects
of a sustainability committee on environmental performance, this committee can offer
an opportunity for firms that want to better integrate sustainability with corporate
governance.
This thesis also contributes to prior literature. It enriches knowledge in two key
silos of corporate sustainability, namely the sustainability committee and sustainability
disclosure, using innovative analytical techniques and better research designs. Further,
it considers the arguments of different theories. Findings of the thesis are summarized
in following section.
1.5 SUMMARY OF THE FINDINGS
The first research question is about whether and how an Australian firm’s
sustainability performance is associated with its sustainability disclosure. Analysing
2,076 firm-year observations between 2002 and 2016, this thesis finds a positive
correlation between selected textual characteristics and sustainability performance. In
other words, good performers release more sustainability disclosure, more quantitative
information, more environmental-impact information and present in optimistic, certain
and clear tones. Analysing 2,067 firm-year observations (2002 – 2016), this thesis
demonstrates that good performers communicate in a more readable way.
Extending prior literature, this thesis performs some robustness tests, including
an alternative performance measurement (by principal component analysis), disclosure
channels, sectoral environment, the Granger causality tests in relation to dependent
and independent variables, and time stability. Findings rendered by main analysis
remain qualitatively unchanged when compared with an alternative performance
measurement. As sustainability disclosure is extracted from a standalone report or
annual report, it is interesting to see whether the findings are affected by disclosure
channels. This thesis reveals different channels do influence relationships between
textual characteristics and performance. In terms of sectoral environments, this thesis
shows that different sectoral environments affect how sample firms communicate
sustainability. As this thesis includes a sample period of 15 years, it was interesting to
check if findings remained stable over time. Different sample periods are found to
18 Chapter 1: Introduction
affect the relationship between textual characteristics and performance. The Granger
causality tests confirm that textual characteristics in one period do not relate to
performance in the next period, and vice versa.
The second research question asks how firms’ experience of disclosure is related
to their sustainability performance. Analysing 2,076 firm-year observations from 2002
to 2016, this thesis reveals a positive relationship between experience of disclosure
and performance. The third research question asks how the sustainability committee is
related to sustainability performance. Analysing 2,166 firm-year observations based
on the same period, this thesis finds that there is a positive relationship between the
existence of a sustainability committee and sustainability performance; sustainability
committee effectiveness is positively related to environmental performance.
Extending prior literature, this thesis performs some robustness tests, including
endogeneity, alternative performance measurements (nine sub-categories of the ESG
ratings), sectoral environment and time stability. Regarding endogeneity, dynamic
generalised-method-of-moments (GMM) and propensity score matching (PSM) are
used, and findings rendered by this thesis remain qualitatively unchanged. Regarding
alternative performance measurements, the findings remain qualitatively unchanged.
In terms of sectoral environment, I find that different sectoral environment influences
effects of sustainability committee on performance. As this thesis includes 15 years in
sample period, it is interesting to check whether findings remain stable over time.
Different sample periods are found to not affect the relationship between committee
and performance. This thesis’s structure is presented in following section.
1.6 THESIS OUTLINE
Chapter 1 presents an overview of this thesis. Chapter 2 discusses various silos
of literature that are related to research questions. Chapters 3 and 4 execute the research
designs to answer research questions proposed in Chapter 1. Chapter 5 concludes this
thesis by summarizing Chapters 2, 3 and 4, whilst discussing limitations and avenues
for future research. As Libby, Bloomfield and Nelson (2002) suggest, ‘Libby Boxes’
are presented in Figure 1-1 to demonstrate research designs, and Figure 1-2 follows to
provide a roadmap.
Chapter 1: Introduction 19
Figure 1-1 Libby Boxes for Research Questions
20 Chapter 1: Introduction
Chapter 1: Introduction 21
22 Chapter 1: Introduction
Chapter 1: Introduction 23
Chapter 1: Introduction 25
Figure 1-2 Thesis Roadmap
Chapter 2: Literature Review 27
Chapter 2: Literature Review
As mentioned in Section 1.4.2, there are many studies that investigate corporate
sustainability from diverse perspectives in countries. Prior studies specifically relevant
to Chapters 3 and 4, respectively, are reviewed in each chapter, and other studies are
reviewed in Chapter 2 to facilitate understanding about how this thesis is positioned
within wider literature surrounding corporate sustainability. Chapter 2 discusses the
pertinent studies regarding sustainability disclosure and sustainability performance in
Section 2.1. Linguistic studies about financial information that inspired this thesis are
discussed in Section 2.2. Section 2.3 considers studies of corporate governance and
sustainability performance, and Section 2.4 reviews studies specific to governance and
sustainability disclosure. Section 2.5 accounts recent literature development from 2014
to 2018 in Australia. Chapter 3 reports original research into textual characteristics of
disclosure and performance.
2.1 SUSTAINABILITY DISCLOSURE AND SUSTAINABILITY
PERFORMANCE
If sustainability disclosure is defined as the “process of communicating the social
and environmental effects of organisations” (Gray, Owen & Adams, 1996, p.3), it and
sustainability performance ought to be coupled – the former should present a true and
fair view of the latter. But prior studies in diverse countries seem to contradict this
idea. Section 2.1 tabulates seminal studies into the relationship between sustainability
disclosure and sustainability performance.
28 Chapter 2: Literature Review
Table 2-1 Overview of Seminal Studies about the Relationship between Sustainability Disclosure and Sustainability Performance Author(s) –Year Country Measurement on Sustainability
Disclosure
Measurement on Sustainability Performance Result
Panel A – Environmental Disclosure and Environmental Performance
[1] Ingram and Frazier
(1980)
US Environmental disclosure – extent of
disclosure
Environmental performance – CEP indices None
[2] Freedman and Jaggi
(1982)
US Environmental disclosure – extent of
pollution disclosure
Environmental performance – CEP indices None
[3] Deegan and Rankin
(1996)
Australia Environmental disclosure – length of
disclosure
Environmental performance –success of prosecution Negative
[4] Fekrat, Inclan, and
Petroni (1996)
18 countries Environmental disclosure – extent of
disclosure
Environmental performance – CEP indices None
[5] Brown and Deegan
(1998)
Australia Environmental disclosure – length of
disclosure
Environmental performance –number of media articles
reporting on corporate environmental performance
Negative
[6] Neu, Warsame, and
Pedwell (1998)
Canada Environmental disclosure – length of
disclosure
Environmental performance – environmental fines levied
against firm and number of articles that had environmental
criticisms of corporate activities
Positive and
Negative
Chapter 2: Literature Review 29
Author(s) –Year Country Measurement on Sustainability
Disclosure
Measurement on Sustainability Performance Result
[7] Cormier and Magnan
(1999)
Canada Environmental disclosure – extent of
disclosure
Environmental performance – excess pollution, fines and
penalties, orders to conform or legal actions
None
[8] Hughes, Anderson,
and Golden (2001)
US Environmental disclosure – extent of
disclosure
Environmental performance – CEP indices None
[9] Patten (2002) US Environmental disclosure – extent of
Environmental disclosure –under CDP intensity of emissions and carbon mitigation under CDP Positive
[25] Herbohn et al.
(2014)
Australia Environmental disclosure – extent of
disclosure
Environmental performance – information extracted from
corporate disclosure
Positive
[26] Meng, Zeng, Shi,
Qi, and Zhang (2014)
China Environmental disclosure – extent of
disclosure
Environmental performance – firms listed by Ministry of
Environment as serious offenders
Non-linear
[27] Arena et al. (2015) US Environmental disclosure – tone of
language
Environmental performance – number of environmental
concerns listed by KLD Positive
32 Chapter 2: Literature Review
Author(s) –Year Country Measurement on Sustainability
Disclosure
Measurement on Sustainability Performance Result
[28] Cormier and
Magnan (2015)
Canada and US Environmental disclosure – extent of
disclosure
Environmental performance – toxic release inventory and
national pollution release inventory
None
[29] Braam, de Weerd,
Hauck, and Huijbregts
(2016)
Netherland Environmental disclosure – extent of
disclosure
Environmental performance – greenhouse gas emissions,
production of waste and total water consumption
Negative
Panel B – Sustainability Disclosure and Sustainability Performance
[30] Lanis and
Richardson (2012a)
Australia Sustainability disclosure – length of
disclosure
Tax aggressiveness – a firm has been accused of tax
aggressiveness by the ATO, resulting in the issue of an
amended tax assessment
Positive
[31] Lanis and
Richardson (2012b)
Australia Sustainability disclosure – extent of
disclosure
Tax aggressiveness – a firm’s effective tax rate Positive
[32] Wang et al. (2018) US Sustainability disclosure – readability of
disclosure
CSR performance – ESG and KLD ratings Positive
Chapter 2: Literature Review 33
Table 2-1 includes several prior studies that examined the relationship between
sustainability disclosure and sustainability performance. As indicated in this table, this
chapter suggests that they focus on thematic content in disclosure, and other disclosure
characteristics are comparatively omitted in the literature. Of the studies analysed in
this table, eight investigate this relationship in Australia. Five studies, [17], [24], [25],
[30] and [31], found a positive relationship, and three, [3], [4] and [5], concluded with
a negative result no correlation at all. Such non-consensus status is also reflected by
studies conducted in other countries. For example, of the 14 studies investigating the
US data, only five conclude a positive relationship. Literature-review or meta-analysis
studies were consulted to see how this relationship is reflected in the wider literature.
Ullmann (1985) included seven studies about this relationship and found that
contradictory results can be due to an absence of theory, inappropriate definitions or
lack of empirical data. Berthelot, Cormier, and Magnan (2003) and Alrazi, De Villiers,
and van Staden (2015) document that this relationship is negative. Following analysis
of 186 studies over decades, Fifka (2012, 2013) concluded that the available literature
produces mixed results regarding this relationship. Synthesizing 178 studies from 1999
to 2011, Hahn and Kühnen (2013, p. 16) confirm Fifka’s (2012, 2013) conclusion and
suggest further studies are encouraged to understand whether sustainability disclosure
“conveys a true and fair view of corporate sustainability performance”.
In summary, as Table 2-1 suggests, there is an extant mature body of literature
investigating the relationship between sustainability disclosure and sustainability
performance. But prior studies do not reach a clear consensus about this important
topic, leaving many opportunities for future studies. Chapter 3 further investigates this
body of studies. Linguistic studies into readability and tone of language in financial
information are reviewed in Section 2.2.
2.2 LINGUISTIC STUDIES ABOUT FINANCIAL INFORMATION
As suggested by Li (2010b) and Loughran and McDonald (2016), the linguistic
studies into financial information have been identified. In addition to thematic content
of sustainability disclosure, this thesis also considers two linguistic characteristics (i.e.
readability and tone of language) of disclosure. Thus, the linguistic studies into the two
characteristics in financial information are discussed, respectively.
34 Chapter 2: Literature Review
2.2.1 Readability
As a linguistic phenomenon, readability has a long history (DuBay, 2004). It has
long been investigated by researchers in the field of accounting. For example, Smith
and Smith (1971) measure the readability of annual reports made by US firms. As
defined in Dale and Chall (1949, p. 5), readability concerns “the sum total (including
all the interactions) of all those elements within a given piece of printed material that
affect the success a group of readers have with it. The success is the extent to which
they understand it, read it at an optimal speed, and find it interesting”. Within the field
of accounting, Lehavy, Li, and Merkley (2011, p. 1091) define it as “the costs incurred
by users to process and interpret a firm’s written communication”.
There are two stages in the development of readability studies in the accounting
literature (Loughran & McDonald, 2016). Following the review of early studies into
the readability of financial information, Jones and Shoemaker (1994), Courtis (1995),
Merkl-Davies and Brennan (2007), Brennan, Guillamon-Saorin, and Pierce (2009), Li
(2010b) and Merkl-Davies, Brennan, and Vourvachis (2011) identified that: studies at
the first stage are mixed20 and sampled a quite limited number of firms.
Li (2008) is a typical study at the second stage. After analysing the data of 55,719
firm-years between 1994 and 2004, Li (2008) identified that changes in readability of
managerial-discussion-and-analysis section (measured by number of words and the
Fog index) relate to financial performance and to earnings persistency. Subsequent to
Li (2008), other researchers explored how readability of financial information can be
related to other variables. For instance, You and Zhang (2009) and Lee (2012) looked
into whether and how readability of annual reports is related to information efficiency
of share prices, Miller (2010) and Franco, Hope, Vyas, and Zhou (2015) analysed how
readability of annual reports is related to trading volumes, and Bonsall and Miller
(2016) and Ertugrul, Lei, Qiu, and Wan (2017) link readability of annual reports with
costs of borrowing.
20 For example, Subramanian, Insley, and Blackwell (1993), Ober, Zhao, Davis, and Alexander
(1999), Rutherford (2003) and Henry (2006) render inconsistent outcomes about relationship between
firm performance and readability of annual reports.
Chapter 2: Literature Review 35
In addition to annual reports, other types of information are explored by prior
studies. For example, Brochet, Naranjo, and Yu (2012) examined the relationship
between the readability of conference calls and trading volumes, and Bradbury, Hsiao,
and Scott (2018) analysed the readability of summary annual reports issued by local
governments in New Zealand. Regarding methodology, behavioural experimentation
was used by Rennekamp (2012) and Lawrence (2013) to investigate how readability
of financial information can shape the decision-making process of investors. It is
reasonable to suggest that the readability of literature exponentially grew after the
research conducted by Li (2008).21 In analysing readability of sustainability disclosure,
this thesis arguably contributes to this body of studies.
2.2.2 Tong of Language
As a linguistic phenomenon, tone of language is defined as “the affect or feeling
of a communication” (Henry, 2006, p. 376). Kearney and Liu (2014, p. 172) consider
textual tone as “the degree of positivity or negativity in texts”. It has been investigated
in three types of disclosure, namely corporate disclosure, media articles and internet
posting (Kearney & Liu, 2014). As most of studies about tone of language investigate
it in financial information, this thesis places importance on how it is operationalized
by prior studies. Henry and Leone (2016) identify that there are two approaches in
current literature, namely word frequency and machine learning. It is noteworthy that
Kearney and Liu (2014) already give a comprehensive review of how tone of language
is measured.
Many studies operationalize tone of language in financial information based on
word-frequency measures/ dictionary-based measures/ bag-of-words measures. Such
measures are based on “a mapping algorithm in which a computer program reads text
and classifies the words, phrases or sentences into groups based on pre-defined
dictionary categories” (Kearney & Liu, 2014, p. 175). Two programs, namely General
Inquirer and Diction, are frequently used in analysis of relevant literature. For
example, Tetlock, Saar-Tsechansky, and Macskassy (2008), Loughran and McDonald
21 For example, a number of readability studies, including You and Zhang (2009), Biddle, Hilary, and
Verdi (2009), Moffitt and Burns (2009), Miller (2010), Lehavy et al. (2011), Brochet et al. (2012),
Lee (2012), Rennekamp (2012), Lawrence (2013), Loughran and McDonald (2014b), Jennings, Seo,
and Tanlu (2014), Bozanic and Thevenot (2015), Lang and Stice-Lawrence (2015), Franco et al.
(2015), Bonsall and Miller (2016), Guay, Samuels, and Taylor (2016) and Bushee, Gow, and Taylor
(2017), are performed.
36 Chapter 2: Literature Review
(2011) and Yukselturk and Tucker (2015) utilised the General Inquirer program, and
Henry (2006), Cho et al. (2010), Craig and Brennan (2012), Davis, Piger, and Sedor
(2012), Davis and Tama-Sweet (2012) and Arena et al. (2015) used the Diction
program. In addition to these two dominant computer programs, other programs are
used. For instance, the Oxford Concordance Program is used by Smith and Taffler
(2000), manual analysis is used by Lang and Lundholm (2000), Li (2006) and
Schleicher and Walker (2010), a customised program was developed by Abrahamson
and Park (1994), and Demers and Vega (2010) jointly apply both Diction and General
Inquirer.
In addition to word-frequency measures or dictionary-based measures or bag-of-
words measures, different machine-learning measures22 are used in current literature.
Antweiler and Frank (2004, 2006), Das and Chen (2007) and Li (2010a) use machine-
learning measures to analyse tone of language in a range of disclosure formats,
including online posting, news articles and forward-looking statements, respectively.
This expanding literature has two concerns when it comes to measuring the tone of
language. First, which measure should be used: does bag-of-words or machine-
learning function better? Henry and Leone (2016, p. 155) conclude that there are
“minimal differences in the power of the tests across these alternative tone measures”.
Secondly, which term weighting, equal/proportional, works better? Henry and Leone
(2016) find that equal or proportional weighting does not distort findings. By analysing
the tone of language in sustainability disclosure, the thesis extends this silo of linguistic
literature from financial information to sustainability disclosure. In the following
section, prior studies about corporate governance and sustainability performance are
reviewed.
2.3 CORPORATE GOVERNANCE AND SUSTAINABILITY
PERFORMANCE
Whether and how to integrate sustainability with corporate governance depends
on how sustainability is related to firm performance – if sustainability undermines firm
performance, there is no need to incorporate sustainability into corporate governance.
In the examined literature, many studies into sustainability performance and firm
22 Machine-learning measures are defined as “statistical techniques to infer the content of documents
and to classify them based on statistical inference” (Kearney & Liu, 2014, p. 175).
Chapter 2: Literature Review 37
performance were undertaken. Thus, this section consults meta-analysis and literature-
review studies to better understand how sustainability performance associates with
firm performance. As found by Pava and Krausz (1996), Margolis and Walsh (2001),
Orlitzky, Schmidt, and Rynes (2003), Allouche and Laroche (2005), De Bakker,
Groenewegen, and Den Hond (2005), Margolis, Elfenbein, and Walsh (2007), Orlitzky
(2008), Van Beurden and Gössling (2008), Stefan and Paul (2008), Molina-Azorín,
Claver-Cortés, López-Gamero, and Tarí (2009), Schreck (2009), Horváthová (2010),
Albertini (2013), Goyal, Rahman, and Kazmi (2013), Endrikat, Guenther, and Hoppe
(2014), Lu, Chau, Wang, and Pan (2014), Friede, Busch, and Bassen (2015), Malik
(2015) and Wang, Dou, and Jia (2015), there is a positive link between sustainability
performance and firm performance.
Moreover, many meta-analysis and literature-review studies, including Orlitzky
and Benjamin (2001), Godfrey (2005), Husted (2005), Godfrey, Merrill, and Hansen
(2009), Minor and Morgan (2011), Oikonomou, Brooks, and Pavelin (2012), Jo and
Na (2012), Albuquerque, Durnev, and Koskinen (2013), Mishra and Modi (2013),
Bouslah, Kryzanowski, and M’Zali (2013), Sun and Cui (2014), Koh, Qian, and Wang
(2014), Harjoto and Laksmana (2016), Al‐Hadi, Chatterjee, Yaftian, Taylor, and
Monzur Hasan (2017) and Lins, Servaes, and Tamayo (2017), find that sustainability
performance is negatively related to firm risks. In summary, prior studies indicate that
it is beneficial to incorporate sustainability into corporate governance, as sustainability
improves firm performance and mitigates firm risks.
In extant literature, venues to integrate sustainability with corporate governance
are discussed. Regarding the role of the board of directors, prior studies do not reach
consensus (Jain & Jamali, 2016). On one side, some studies, including Huang (2010)
who analysed 297 electronics firms in Taiwan, Sánchez, Sotorrío, and Díez (2011)
who sampled 125 firms in Spain, and Walls et al. (2012) who tested 2,002 firm-year
observations in the US, found that a board of directors which better serves the interests
of shareholders, also has positive associations with sustainability performance.
Contrastingly, other studies, including Cespa and Cestone (2007), Surroca and Tribó
(2008) and Chintrakarn, Jiraporn, Kim, and Kim (2016), suggest that sustainability
performance is related to managerial discretion and entrenchment. In other words, a
board of directors that does not diligently serve the interests of shareholders
encourages sustainability performance. Thus, whether a board of directors is an asset
38 Chapter 2: Literature Review
to the integration of sustainability with corporate governance is in debate. This thesis
improves this silo of studies by exploring whether and how a board committee can be
an avenue to better integrate sustainability with the decision-making at senior level
and contribute to better sustainability performance. In Section 2.4, prior studies about
corporate governance and sustainability disclosure are discussed.
2.4 CORPORATE GOVERNANCE AND SUSTAINABILITY DISCLOSURE
Corporate governance focuses on how to moderate the behaviour of boards and
senior executives (Mees & Smith, 2019), and corporate governance can be reduced to
issues of regulation and disclosure. Instructed by Jain and Jamali (2016), this section
tabulates seminal studies about the relationship between corporate governance and
sustainability disclosure. It is reasonable to suggest that this thesis just covers a fraction
of the relevant literature.
Chapter 2: Literature Review 39
Table 2-2 Overview of Seminal Studies about Relationship between Sustainability Disclosure and Corporate Governance Author(s) –Year Country Measurement on Sustainability
Disclosure
Measurements on Corporate
Governance
Result
[1] Prado-Lorenzo and
Garcia-Sanchez (2010)
28 countries Scores assigned by the Carbon
Disclosure Project about disclosure of
greenhouse gas emissions
Board independence, CEO duality
and gender diversity of board
Negative or no relationship between
corporate governance and disclosure
are found
[2] Frias‐Aceituno,
Rodriguez‐Ariza, and
Garcia‐Sanchez (2013)
15 countries Different types of disclosure:
financial disclosure only, financial
disclosure and corporate social
responsibility disclosure, and
integrated disclosure
Board size, board independence,
meeting frequency and gender
diversity of board
Board size and gender diversity are
found to be positively related to the
use of corporate social responsibility
disclosure and integrated disclosure
[3] Khan, Muttakin, and
Siddiqui (2013)
Bangladesh Disclosure index Board independence, CEO duality,
presence of audit committee
Board independence and the presence
of an audit committee are found to be
positively related to disclosure
[4] Ntim and Soobaroyen
(2013) South Africa Length of black economic
empowerment disclosure (measured
in number of words)
Board size, board independence,
CEO duality and board diversity
Board size, board independence and
board diversity are positively related
to disclosure
[5] Amran, Lee, and Devi
(2014)
12 countries Disclosure index Board size, board independence, gender diversity of board and
presence of corporate social
responsibility committee
Only the presence of a corporate social responsibility committee is
found to be positively related to
disclosure
40 Chapter 2: Literature Review
Author(s) –Year Country Measurement on Sustainability
Disclosure
Measurements on Corporate
Governance
Result
[6] Fernandez‐Feijoo,
Romero, and Ruiz‐Blanco
(2014)
22 countries Scores assigned by the KPMG
International Survey of Corporate
Social Responsibility Reporting
(2008)
Gender diversity of board (at least
three women on board)
Gender diversity of the board is found
to be positively related to length of
disclosure
[7] Jizi, Salama, Dixon, and
Stratling (2014)
US Disclosure index Board size, board independence,
meeting frequency and CEO
duality
Board size, board independence,
meeting frequency and CEO duality
are found to be positively related to
disclosure
[8] Jizi (2017) UK Disclosure index Board size, board independence,
gender diversity of board and CEO
duality
Board independence and gender
diversity of board are found to be
positively related to disclosure
Chapter 2: Literature Review 41
In order to depict a relatively comprehensive picture about relationship between
corporate governance and sustainability disclosure, this thesis includes four studies at
international level (i.e. [1], [2], [5] and [6]). It is noteworthy that international studies
do not reach a consensus about this relationship. For example, Amran, Lee, and Devi
(2014) report gender diversity does not exert influence on sustainability disclosure, yet
Fernandez‐Feijoo, Romero, and Ruiz‐Blanco (2014) find the effect of gender diversity.
As Table 2-2 shows, four studies at national level (i.e. [3], [4], [7] and [8]) suggest that
there is a positive relation between corporate governance and sustainability disclosure.
Taken together, inclusion of corporate governance characteristics in a research design
can make the design more robust, and future studies with regard to this relationship
are encouraged. As many studies are not included in this chapter, literature-review or
meta-analysis studies are consulted to check how this relationship is reflected in the
literature. Informed by Dienes, Sassen, and Fischer (2016), Jain and Jamali (2016) and
Rao and Tilt (2016) who comprehensively considered relevant studies, this chapter
concludes that there is no true consensus in terms of how corporate governance (e.g.
ownership structure and board of directors) relates to sustainability disclosure. Future
research is expected to therefore explore many opportunities in this field. In Section
2.5, Australian studies (2014 – 2018) published on nine regional accounting journals
are retrieved and reviewed.
2.5 CORPORATE SUSTAINABILITY: AUSTRALIAN EVIDENCE SINCE
2014
As my research setting is Australia, this final section of Chapter 2 (Section 2.5)
provides a survey of the relevant corporate sustainability/ social responsibility studies
that examine how Australian firms react to various incentives and pressures to conduct
sustainability practice. Inspired by Benson, Clarkson, Smith, and Tutticci (2015) who
reveal corporate sustainability/ social responsibility is one of seven major fields where
accounting research has impacted practices, Section 2.5 updates this silo of literature
by reviewing relevant studies since 2014 in the same nine nominated journals:23
Accounting, Auditing and Accountability Journal (AAAJ);
23 I recognise that relevant Australian studies may be published in other journals, such as Business
Strategy and Environment, Journal of Cleaner Production, Journal of Business Ethics and Accounting
and Business Research. But I have retained the same scope of accounting journals as recognised by
Benson, et al. (2015) as the journals’ influence in the accounting discipline in Australia.
42 Chapter 2: Literature Review
Australian Accounting Review (AAR);
Abacus (Abacus);
Accounting and Finance (AF);
Australian Journal of Management (AJM);
Accounting Research Journal (ARJ);
Journal of Contemporary Accounting and Economics (JCAE);
Managerial Auditing Journal (MAJ);
and Pacific Accounting Review (PAR).
Note, here and throughout this study, these journals are presented alphabetically
according to their widely used abbreviations, and no hierarchy is intended or implied
by this ordering.
I inputted keywords, including Australia*, sustain*, social*, envir* and different
combinations of various keywords, into EBSCOhost, Scopus and ProQuest databases
to retrieve the studies that meet all following criteria:
(1) They are published on the aforementioned nine accounting journals;
(2) They are published between 2014 and 2018;
(3) They examine corporate sustainability/social responsibility at firm level;
(4) They are not literature review studies or editorials.
After manually screening the searching outcomes, there are 54 studies that meet
all four criteria. Basic information about these 54 studies is presented below. As Table
2-3 shows, AAAJ, AAR and AF are the top three journals that publish accounting
studies with regard to corporate sustainability/social responsibility using or including
Australian data between 2014 and 2018; and the number of publications gradually
increased over this period of time.
Chapter 2: Literature Review 43
Table 2-3 Basic Information about 54 Studies from 2014 to 2018 in Nine Designated
Accounting Journals Panel A – Journal of Publications
Journal of Publications Number of Publications
AAAJ 14 AAR 11
ABACUS 1
AF 8
AJM 5
ARJ 5
JCAE 2
MAJ 3
PAR 5
Panel B – Year of Publications
Year of Publications Number of Publications
2014 7
2015 8
2016 8
2017 16 2018 15
As Table 2-4 presents, firms’ sustainability disclosure is the most examined topic
(23 studies), followed by sustainability performance (14 studies). In addition to these
two topics, economic consequence and corporate governance with regard to corporate
sustainability also attract much attention. The research themes of these 54 studies are
shown in Panel A of Table 2-5, and research methods used are presented in Panel B of
Table 2-5. As Panel A of Table 2-5 shows, the most researched theme is corporate
sustainability/ social responsibility (31.48%) and climate change/carbon (as a sub-
category of corporate sustainability) also attracts much attention (24.07%). As Panel
B of Table 2-5 presents, regression is the most frequently used method, and qualitative
methods also are frequently used.
44 Chapter 2: Literature Review
Table 2-4 Summary of 54 Studies from 2014 to 2018 on Nine Designated Accounting Journals Code Journal Name (Year) Research Aim(s) / Question(s) Key Findings
[1] PAR Canny (2014) The annual report disclosure of contributions by
Australian firms to the relief appeal in context of the
South-East Asian tsunami of 26 December 2004
• There is a strong relationship between public awareness of
the contributions and disclosing behaviour;
• Firm size and profit are related to some aspects of
disclosure;
• There is no relationship between size of the cash donation
and disclosing behaviour.
[2] AAAJ Egan (2014) How were a heterogeneous range of water efficiency
responses driven across a field of seven water consuming
organisations in Australia at a time of acute drought
conditions into the late 2000s?
• A loosely coordinated range of drivers motivated
pervasive water efficiency responses in few case
organisations;
• Would-be leader organizations sought to invoke a water efficiency field;
• While the field lacked effective champions for change, an
institutionalisation of novel water efficiency practices
continued across the field into 2010.
[3] PAR Hazelton (2014) The labelling of the water footprint of products in an
Australian context • Water footprint reporting could make a significant
contribution to public water literacy;
• Labelling of complex products is currently infeasible, but
existing and emerging solutions may make it possible in
the future.
[4] ABACUS Herbohn, Walker, and
Loo (2014)
The relationship between sustainability performance and
sustainability disclosure within the Australian extractive
industries
• There is a positive relationship between sustainability
performance and sustainability disclosure in the
Australian extractive industries.
[5] PAR Luo and Tang (2014a) The impact of the proposed carbon tax on the financial
market return of Australian firms, and the differential tax
effect on individual firms with different carbon profiles
• The proposed tax has negative impact on shareholder
wealth in Australia;
• The most significant effect is found in the materials,
industrial and financial sectors;
Chapter 2: Literature Review 45
• A firm’s direct carbon exposure is found to be
significantly related to abnormal returns, yet its indirect
exposure is not;
• The influence of proposed carbon tax is more notable
during the early stages of the development of the carbon
tax.
[6] JCAE Luo and Tang (2014b) Whether voluntary carbon disclosure reflects international
firms’ true carbon performance? • There is a positive relationship between carbon disclosure
and carbon performance, suggesting that firms’ voluntary
carbon disclosure in the Carbon Disclosure Project is
indicative of their underlying actual carbon performance.
[7] AAR Tang and Luo (2014) The implementation of carbon management systems by
large Australian firms • For firms that have higher quality carbon management
systems, they achieved better carbon mitigation;
• The most effective elements of carbon management
systems include adequate assessment of carbon risk and
opportunity, the presence of reduction targets, the strength
of carbon programs and enhanced external disclosures.
[8] AAR Fernandez‐Feijoo,
Romero, and Ruiz
(2015)
Factors that explain the decision of intranational firms to
assure their sustainability disclosure and of the choice of a
Big 4 auditor as assuror
• A European Union country affects the decision of a firm
to have sustainability disclosure externally assured and
hire a Big 4 as assurance provider;
• Industry affiliation, firm size, listed status and Global
Reporting Initiative application level are related to use of
external assurance;
• Industry affiliation, firm size, listed status, use of integrate
reporting and Global Reporting Initiative application level
are related to employment of a Big 4 as assurance
provider.
[9] AF Linnenluecke, Birt,
Lyon, and Sidhu
(2015)
Implications of changes in planetary boundary conditions
for increasing the risks of impairment that are contingent
on the impacts of breaches or violations of planetary boundaries with a consequent loss of a social or regulatory
licence to operate
• The Australian top 10 metals and mining firms by market
capitalisation in the 2013/2014 year would encounter
averaged $1.144 billion of impairment loss.
46 Chapter 2: Literature Review
[10] AF Loh, Deegan, and
Inglis (2015)
The corporate social and environmental disclosure
practices of a sample of gambling firms operating within
Australia in time of three specific interrelated Australian
government initiatives
• Corporate social and environmental disclosure is a
response to social pressures created around the time of
these initiatives.
[11] AAR Martínez-Ferrero,
Gallego-Álvarez, and
García-Sánchez (2015)
The connection and possible bidirectional relationship
between corporate social responsibility and earnings
management
• The existence of an inverse bidirectional relationship
between corporate social responsibility and earnings
management;
• This bidirectional relationship is more important in
countries where there is significant institutional pressure with regard to corporate social responsibility and in
countries with greater investor protection.
[12] AAAJ O'Neill, McDonald,
and Deegan (2015)
Whether the different procedures for organising subsets of
a set of accounting data may lead to different conclusions
about (the same) reality?
• The authors demonstrate that different representations of
reality may result not only from accounting choices as to
“what” is measured, but also from accounting choices as
to “how subsets of measured data are organised”.
[13] MAJ Soh and Martinov-
Bennie (2015)
The nature and extent of internal audit functions’
involvement in environmental, social and governance
assurance and consulting in Australia
• Internal audit functions are very involved in providing
assurance on governance issues and reasonably involved
in social issues, but internal audit functions perform a very limited role in providing assurance on environmental
issues;
• Internal audit functions are limited involved in consulting
activities with regard to assurance;
• Environmental issues are widely expected to increase in
importance to internal audit functions;
• Internal audit functions’ skills and competencies with
regard to environmental issues are in greatest need of
further development;
• There is a divergence in usage of standards by internal
audit functions in performing relevant engagement.
Chapter 2: Literature Review 47
[14] AAR Vafaei, Ahmed, and
Mather (2015)
The relationship between gender diversity at board level
and firms’ financial performance
• Board diversity is positively associated with financial
performance.
[15] AAAJ Vesty, Telgenkamp,
and Roscoe (2015)
How is a carbon number (a dollar value derived from
physical units) elevated to become a pivotal actor in
organizational practice?
• The authors highlighted the importance of carbon number
in the newly emergent and evolving carbon market.
[16] MAJ Bepari and Mollik
(2016)
The degree to which assurance statements in sustainability
disclosure enhance and uphold organisational transparency
and accountability to stakeholders
• The assurance statements in sample lack stakeholders’
engagement assurance process, are limited in the scope,
and exhibit reluctance of the assuror to address to the
stakeholder groups;
• As the assurance statements in sample focus on internal
systems, process, data generation and data capture,
assurance practice is serving more as an internal control
tool than as a social accounting/auditing instrument.
[17] ARJ Kumarasiri and Jubb
(2016)
Use of management accounting techniques by Australian
large listed firms in constraining their carbon emissions
• Relevant regulations drive top management and boards to
use management accounting techniques to set targets,
measure performance and incentivise emission mitigation.
[18] AAAJ McPhail and Adams
(2016)
How respect for human rights is emerging and being
operationalized in the discourse of 30 Fortune 500 firms in the mining, pharmaceutical and chemical industries at two
key points in the recent evolution of the United Nations’
business and human rights agenda?
• Corporate constructions of human rights are broad: from
labour rights, through social and political rights, to the right to health and a clean environment;
• The corporate discourse is one of promoting, realizing and
upholding rights that construct the corporation as an
autonomous source of power beyond the state.
[19] AAAJ Moore and McPhail
(2016)
How and to what extent was the development of carbon
accounting frameworks at the policy, industry and
organizational levels enabled by external structures as
conditions of action?
• Soft power and trust are two drivers of the development of
carbon accounting frameworks.
[20] ARJ Ong, Trireksani, and Djajadikerta (2016)
The quality of sustainability disclosures in the current leading environmentally sensitive industry in Australia,
the resources industry
• Australian resources firms report more soft disclosure items than hard disclosure items;
48 Chapter 2: Literature Review
• Australian resources firms report the most sustainability
information in the economic aspect.
[21] ARJ Sands, Rae, and
Gadenne (2016)
The feasibility of integrating the social, environmental and
innovation processes within the four-perspective
sustainability balanced scorecard model
• It is feasible to integrate environmental, social and
innovation-orientated value-creating process into the
internal process of the four-perspective sustainability
balanced scorecard model.
[22] AAAJ Tello, Hazelton, and
Cummings (2016)
The perceptions of potential users about water accounting
reports prepared under Australian general-purpose water
accounting.
• Users perceive the introduction of Australian general-
purpose water accounting as useful and believe that the
benefits will outweigh the costs;
• Government agencies were likely to be the main users of Australian general-purpose water accounting;
• Users were also concerned about the degree of judgement
required to determine the identity and boundaries of a
“water report entity”;
• There was little consensus that Australian general-purpose
water accounting collectively discharged the
accountability of water managers.
[23] MAJ Yunus, Elijido-Ten,
and Abhayawansa
(2016)
Determinants of carbon management strategy adoption
among Australia’s top 200 listed firms • For firms that have carbon management systems, they are
more likely to have an environmental management
system, an environmental committee, larger board and greater board independence.
[24] AAAJ Adams (2017) What is the perceived relationship between the
management and governance of environmental, social and
governance risk, strategy development and value creation?
What role does corporate reporting (including the
processes to develop corporate reports) and cognitive
framing play in mitigating these relationships?
• There is an increased awareness of the impact of
environmental, social and governance issues together with
a broader view of value creation despite investor
disinterest;
• Contemporary reporting processes, and in particular those
set out in the King III Code and the International
Thirdly, non-economic consequences of sustainability disclosure are considered
by O'Neill, et al. (2015) and Adams (2017). O'Neill, et al. (2015) examined how work-
related injury data can be interpreted to depict quite different pictures about reality. By
interviewing business leaders in South Africa and Australia, Adams (2017) found that
sustainability disclosure contributes to an integration of sustainability with corporate
governance.
Fourthly, usefulness of integrated reporting (a type of sustainability disclosure)
to financial analysts has been explored by Abhayawansa, et al. (2018). Hazelton (2014)
investigated usefulness of another new type of sustainability disclosure, namely water
footprint labelling.
Fifthly, relationship between (a specific type of) sustainability disclosure and (a
specific area of) sustainability performance has been tested by Herbohn, et al. (2014)
(environmental disclosure of mining and energy firms), Luo and Tang (2014b) (carbon
disclosure of firms in three countries) and Luo (2017) (carbon disclosure of firms in
multiple countries). It is noteworthy that while Luo and Tang (2014b) and Luo (2017)
focus on relationship between carbon disclosure and carbon performance, they reach
Chapter 2: Literature Review 57
opposite conclusions. Last, Lodhia and Stone (2017) normatively highlighted potential
contribution of Internet to integrated reporting.
2.5.2 Studies about Sustainability Performance
There are 13 studies ([2], [7], [17], [29], [32], [38], [39], [41], [44], [46], [49],
[53] and [54]) that focus on sustainability performance. First, Hutchinson, et al. (2017),
examining 41, 655 executive-year observations from 2002 to 2013, found existence of
gender inequity in directors’ remuneration in Australia. Secondly, a number of studies
focus on how corporate governance (dis)encourages (a specific area of) sustainability
performance. Tang and Luo (2014) examined effect of carbon management system on
firms’ carbon performance. Powell and Tilt (2017) highlighted need for new business
model to solve tensions due to power transfer among managers in terms of keeping a
balance between economic sustainability and environmental sustainability. Biswas, et
al. (2018) examined effect of three corporate governance characteristics, namely board
gender diversity, board independence and existence of sustainability committees, on
firms’ sustainability performance. Sundin and Brown (2017) and Phan, et al. (2018)
discussed importance of internal management systems to environmental performance.
Heggen, et al. (2018) revealed organizational culture and integration of sustainability
with corporate governance would impact firms’ environmental performance. Thirdly,
Ahmed, et al. (2018) investigated a number of firm characteristics that are expected to
relate to gender diversity on board. Fourthly, how individuals within firms would (dis)
encourage (a specific area of) sustainability performance has been examined by three
studies. By surveying 101 energy firms in Australia, Bremer and Linnenluecke (2017)
reported that managers’ environmental attitudes, climate change knowledge and risk
perception affect how they perceive importance of climate change adaptation. Egan
and Tweedie (2018) examined how accountants were engaged in firms’ sustainability
initiatives. Salignac, et al. (2018) focused on how institutional entrepreneurs promote
gender diversity within construction industry in Australia. Last, Egan (2014) explored
how institutions at field level motivate firms’ water efficiency, and Kumarasiri and
Jubb (2016) examined how regulation inspires firms to use management accounting to
improve their carbon performance.
58 Chapter 2: Literature Review
2.5.3 Studies about Economic Consequence of Corporate Sustainability
There are 10 of 54 studies ([5], [9], [11], [14], [26], [37], [42], [50], [51] and
[52]) that examine economic consequence of corporate sustainability. First, how
carbon-related public policy and climate change affect firm value has been studied by
Luo and Tang (2014a) (proposed carbon tax in 2011), Nguyen (2018) (ratification of
Kyoto Protocol in 2007) and Linnenluecke, et al. (2015) (impairment loss caused by
climate change). Secondly, Martínez-Ferrero, et al. (2015) tested relation between
firms’ sustainability performance and earnings management at international level, and
Martínez-Ferrero, et al. (2017) analysed whether investors can detect linkage between
sustainability performance and managerial entrenchment. Nguyen et al. (2018) shed
light on relation between sustainability performance and information asymmetry.
Thirdly, relationship between (a specific area of) sustainability performance and
financial performance has been examined by Vafaei, et al. (2015) (how gender
diversity of board of directors is related to financial performance) and Krishnamurti,
et al. (2018) (how firms’ sustainability performance is related to corruption risk).
Fourthly, relationship between sustainability disclosure and financial performance has
been analysed by Al‐Hadi, et al. (2017) and Beck, et al. (2018).
2.5.4 Studies about Assurance Services
Three studies ([8], [13] and [16]) examine assurance services. First, Fernandez‐
Feijoo, et al. (2015) investigated determinants at national and firm levels with regard
to use of assurance services. Secondly, Soh and Martinov-Bennie (2015) explored how
internal auditors participate in assurance services. Thirdly, Bepari and Mollik (2016)
discussed accountability consequence of external assurance services by analysing the
assurance statements.
2.5.5 Studies about Corporate Governance with regard to Corporate
Sustainability
There are two studies ([21] and [23]) that focus on corporate governance devices
or structures with regard to corporate sustainability. First, Sands, et al. (2016) explored
how we can better use sustainability scorecard. Secondly, Yunus, et al. (2016) analysed
determinants at firm level in use of carbon management systems.
Chapter 2: Literature Review 59
2.5.6 Other Studies about Corporate Sustainability
In addition to the aforementioned five research topics (disclosure, performance,
economic consequence, assurance services and corporate governance), other topics are
also considered. Focusing on relevant public policy, Moore and McPhail (2016) shed
light on development of carbon accounting frameworks, and Tello, et al. (2016) elicit
users’ views about general-purpose waster accounting. In context of a large Australian
water utility, Vesty, Telgenkamp, and Roscoe (2015) discussed effect of quantifying
carbon emissions.
2.5.7 How this Thesis is Positioned in the Australian Literature
My first question (RQ1: Whether and how is an Australian firm’s sustainability
performance associated with its sustainability disclosure?) aligns with Herbohn, et al.
(2014), Luo and Tang (2014b) and Luo (2017) that focus on the relationship between
disclosure and performance. Chapter 3 (in which I examine RQ1) differentiates from
these studies by focusing on various textual characteristics of sustainability disclosure
together and using more recent and comprehensive data. Thus, Chapter 3 can provide
much richer and more generalizable conclusions about this frequently examined topic,
namely how disclosure relates to performance.
The second research question (RQ2: How is experience of disclosure related to
sustainability performance?) develops results of Adams (2017) by empirically testing
how firms’ experience of disclosure is related to their sustainability performance. My
approach differentiates from Adams (2017) who interviewed directors in Australia and
South Africa by analysing a set of panel data collected and synthesized. Collectively,
my work in addressing RQ2 and Adams (2017) explores the relationship between
disclosure and performance from a novel view: instead of examining how
sustainability disclosure is related to or affected by sustainability performance, I
explore or consider the other way around, namely how engagement in sustainability
disclosure affects or relates to sustainability performance.
The third research question (RQ3: How is the sustainability committee related
to sustainability performance?) relates to Tang and Luo (2014), Powell and Tilt (2017),
Biswas, et al. (2018), Sundin and Brown (2017), Phan, et al. (2018), and Heggen, et
al. (2018) by examining the effect of sustainability committees on firms’ sustainability
performance in detail. My work of addressing RQ3 differentiates from the
60 Chapter 2: Literature Review
aforementioned studies in two ways. First, I investigate sustainability committee as a
governance mechanism devoted to corporate sustainability, which is not
systematically studied yet has implications for practice. Second, I use a more robust
research design that is built on a longer sample period, richer committee
characteristics, inclusion of firms in different industries, and the use of different
performance measurements to render comprehensive evidence in this regard.
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance 61
Chapter 3: Textual Characteristics of
Sustainability Disclosure and their
Relationship to Sustainability Performance
3.1 INTRODUCTION
As Chapter 1 illustrates, corporate sustainability includes economic prosperity,
social equity and environmental integrity (Bansal, 2005). Prior literature emphasises
environmental and social impact due to firms’ operational activities (Dahlsrud, 2008).
Ostensibly, sustainability disclosure is an important part of transparency in corporate
sustainability, which is a key concern of many stakeholders (Gray, Owen, & Adams,
1996). There is currently a momentum of mandatory sustainability disclosure24 at a
global level. For instance, sustainability disclosure has been mandated according to
firm size in Argentina and the European Union.25 Exchanges in the Asia-Pacific region
are taking action on mandatory sustainability disclosure.26 The Australian Securities
Exchange is a partner of the Sustainable Stock Exchanges Initiatives27, an initiative
backed by the United Nation to improve transparency of corporate sustainability, and
it also acknowledges ‘long term sustainable value’ and ‘acting lawfully, ethically and
responsibly’ (ASX, 2019, p. 16). In addition to the Australian Securities Exchange,
legislators and regulators have begun to address firms’ sustainability and disclosure,
as discussed in Chapter 1. For instance, a firms’ contributions to climate change have
been discussed by the Senate Economics References Committee28 and the Australian
Securities and Investments Commission.29
24 According to United Nations Environment Programme and KPMG (2006) and United Nations
Environment Programme et al. (2010, 2013, 2016), there is indeed a trend of mandatory disclosure of
corporate sustainability. 25 See the Ley Nº 2594 de Balance de Responsabilidad Social y Ambiental (BRSA) and Directive
2014/95/EU. 26 The Hong Kong Exchange, Singapore Exchange and Taiwan Stock Exchange also mandate sustainability disclosure to their listed firms, and more information can be found at
http://www.sseinitiative.org/ (access date is 8 May 2019). 27 More information about the Sustainable Stock Exchanges Initiatives and who are partner exchanges
can be found at http://www.sseinitiative.org/ (access date is 8 May 2019). 28 This committee published a report, Carbon Risk: A Burning Issue, in 2017. 29 As the market regulator in Australia, the Australian Securities and Investments Commission
released a report, Climate Risk Disclosure by Australia’s Listed Companies, in 2018.
62 Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
From a broader perspective, as firms and stakeholders incur costs in preparing,
disseminating and using sustainability disclosure, transparency of firms’ sustainability
performance can be undermined by corporate symbolic action. Bowen (2014) explains
that symbolic action has social cost, as resources invested on symbolic action could
alternatively be devoted to substantial action. Therefore, research on sustainability
disclosure is also meaningful to society in general.
In practice, Australian firms are increasingly engaged in sustainability disclosure
(KPMG, 2015, 2017), and stakeholders actively utilise such disclosure. Pérez-Gladish
et al. (2012) and Eccles et al. (2017) found that institutional investors are increasingly
engaged in socially responsible investment, and they view low quality of sustainability
disclosure as a significant obstacle. KPMG and SustainAbility (2008), surveying non-
financial stakeholders, find that they use sustainability disclosure in a number of
scenarios. It is a reasonable expectation that sustainability disclosure will find more
users in the future, as socially-responsible-investment assets in Australia and New
Zealand rose by 247% from 2014 to 2016 to reach $516 billion (Foo, 2017).
Practitioner surveys and prior literature generally identify the sustainability
themes disclosed, how much information is reported, and where such information is
disclosed. The practitioner surveys on disclosure practices of Australian firms find that
more firms are engaged in sustainability disclosure, and more disclosure is being
released (Australian Council of Superannuation Investors, 2009, 2010, 2011, 2012,
2013, 2014, 2015, 2016, 2017, 2018; KPMG, 2013, 2015, 2017). Using the Global
Reporting Initiative, Frost et al. (2005) and Beck et al. (2018) evaluate sustainability
disclosure and find that Australian firms prefer standalone reports and their websites
as methods of communicating issues relating to sustainability and are able to present
them in a way that matches their performance.
This chapter extends on Cho et al. (2010), Arena et al. (2015), Beattie (2014),
Tregidga, Milne, and Lehman (2012) and Wang et al. (2018) by investigating textual
characteristics of sustainability disclosure (which cover linguistic characteristics) and
is informed by prior literature about textual characteristics in financial disclosure (see
e.g. Henry & Leone, 2016; Huang, Teoh, & Zhang, 2014; Li, 2010b; Loughran &
McDonald, 2016). As documented by prior literature about textual characteristics of
financial information (Ertugrul et al., 2017; Henry & Leone, 2016; Kearney & Liu,
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance 63
2014), it is argued linguistic characteristics of sustainability disclosure are meaningful,
as they affect how stakeholders understand sustainability disclosure.
As Wood (1991a, 1991b, 2010), Clarkson et al. (2008), Clarkson et al. (2011),
Tregidga et al. (2014), Rao and Tilt (2016), Global Reporting Initiative (2014) and
Guthrie (2016) indicate, whether and how textual characteristics of disclosure relate to
performance are in debate. Prior literature reviewed in Chapter 2 suggests different
theories or theoretical frameworks inform different views in terms of their relationship.
For example, signalling theory proposes good performers use textual characteristics of
disclosure to signal their good performance (Connelly, Certo, Ireland, & Reutzel,
2011), yet institutional theory (decoupling) conjectures that bad performers use their
method of disclosure to cover up their unsatisfactory or bad performance (Delmas &
Burbano, 2011; Laufer, 2003; Ramus & Montiel, 2005). This chapter develops the
following research question:
RQ1: Whether and how is an Australian firm’s sustainability performance
associated with its sustainability disclosure?
Through analysing 2,076 firm-year observations (2002-2016), Chapter 3 models
seven textual characteristics (amount of disclosure, amount of quantitative information
in disclosure, amount of information in relation to environmental impact, tone of
optimism in disclosure, tone of certainty in disclosure, tone of clarity in disclosure,
and readability of disclosure) as variables of interest; where the dependent variable is
sustainability performance. I identified a positive relationship between sustainability
performance and six of the seven textual characteristics – good performers disclose
more information, more quantitative data, more environmental-impact information
and present their disclosure in a certain, clear and optimistic manner. These findings
lend support to the arguments of signalling theory.
In addition, this chapter identifies that firms with better performance present
their sustainability disclosure in a more readable way, the seventh characteristic. This
finding supports the incomplete revelation hypothesis proposed by Bloomfield (2002,
2008) wherein firms with better performance communicate their performance in ways
that stakeholders can easily comprehend. Chapter 3 argues that this finding aligns to
some extent with signalling theory – good performers prepare their disclosure in a way
that signals their superior performance.
64 Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
These findings are in line with Herbohn et al. (2014), who found that Australian
firms with better environmental performance release higher quality environmental
disclosure. The findings of this chapter are robust to different sensitivity checks that
have been performed in addition to the main analysis. From a theoretical view, Chapter
3 reveals the explanatory power of signalling theory and the incomplete revelation
hypothesis.
Ostensibly, this chapter enriches the current literature through focusing on the
relationship between disclosure and performance by investigating various linguistic
characteristics of sustainability disclosure (Beattie, 2014; Beattie & Davison, 2015)
and extending the research scope from content to include other textual characteristics.
For example, seminal studies in extant literature, Guthrie and Parker (1989) and
Deegan et al. (2002) analyse thematic content of disclosure via the lens of legitimacy.
Responding to Beattie et al. (2004) and Beattie (2014), this thesis extends the scope of
linguistic analysis from financial disclosure to sustainability disclosure, suggesting the
usefulness of linguistic analysis in different areas. Cho et al. (2010), Arena et al. (2015)
and Wang et al. (2018) examined the tone of optimism and readability, respectively,
in disclosure of the US firms. Chapter 3 extends the aforementioned studies in four
ways: (1) as a standalone report has more detailed sustainability information, I sampled
disclosure from two mediums (standalone report and annual report); (2) this chapter
covers both social and environmental disclosure, rather only environmental disclosure;
(3) this chapter samples firms from a wide range of industries over a time period of 15
years, while prior studies usually sample fewer industries and are generally conducted
over shorter time periods; (4) Chapter 3 takes into account seven textual characteristics
rather one characteristic at a time. Chapter 3 presents a more comprehensive picture
about sustainability disclosure.
The research findings are likely to be of interest to external assurance providers
of sustainability disclosure. In addition to content, the textual characteristics examined
in Chapter 3 are worthy of attention when performing assurance services on disclosure.
As sustainability disclosure has been a channel of communicating sustainability issues
30 Corporate legitimacy “is meeting and adhering to the expectations of a social system’s norms,
values, rules, and meanings” (Deephouse & Carter, 2005, pp. 331-332).
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance 69
(Boxenbaum & Jonsson, 2008) and impression management (Brennan et al., 2009). In
relation to RQ1, from an institutional theory view, worse performers use sustainability
disclosure as a shield from socio-political pressures due to worse performance. They
are expected to disclose sustainability information in ways that disguise their poor
performance and mislead stakeholders. To be specific, textual characteristics analysed
in this chapter consider how much sustainability information, quantitative information
and information about environmental impact are disclosed, to what extent disclosure
is communicated in optimistic, certain and clear tones, and to what extent disclosure
is readable.
In terms of the first six textual characteristics, as institutional theory instructs,
worse performers report information in a more comprehensive way (i.e. measured in
number of words), more quantitative information (operationalized by the Diction 7),
and more information about environmental impact (measured in number of words);
they would communicate in more ambiguous, optimistic and certain tone to mislead
stakeholders.
The second theory, signalling theory, highlights that information asymmetry can
be mitigated by “the party with more information signalling it to others” (Morris, 1987,
p. 48). In the first step, firms in a market are assumed to have more information about
their sustainability than stakeholders. If stakeholders have no information about firms’
sustainability but do have some general concerns (e.g. some of business activities can
be not sustainable), stakeholders measure firms’ sustainability at a weight average of
their concerns. In the second step, for firms that are in above average quality condition
with regard to sustainability, they incur an opportunity loss if stakeholders knew about
their superior sustainability performance, although firms of below average in terms of
sustainability make an opportunity gain. For firms of above average in sustainability,
they would use sustainability disclosures to signal their superior sustainability. In the
third step, as firms that are better in sustainability signal, stakeholders consider all the
remaining firms to be of poor. The best remaining firms then try to screen themselves
from the others via sustainability disclosure. As Connelly et al. (2011) instruct, this is
an iterative process that continues as long as the gain due to signalling is greater than
the signalling costs.
From the aspect of signalling theory, better performers use their sustainability
disclosure to distinguish themselves so that they may reap economic gains and pre-
70 Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
empt potential/future regulations (Delmas & Montes Sancho, 2010). To send credible
signals about their better performance, firms arguably present sustainability disclosure
in ways that are difficult to be mimicked by their peers. Regarding the first six textual
characteristics, as signalling theory instructs, better performers are likely to disclose
information in a comprehensive way (more sustainability disclosure, more quantitative
information and more information about environmental impact); they are perceived to
‘talk’ in a less ambiguous tone, actively utilise a more optimistic tone and exhibit more
certainty in their tone. Some studies, including Al-Tuwaijri et al. (2004) (the US),
Freedman and Jaggi (2004) (the US) and Clarkson et al. (2008) (the US) and Marquis,
Toffel, and Zhou (2016) (an international study), lend support to signalling theory.
Prior studies into Australian firms support institutional theory. As Section 3.2
indicates, studies in Australia, including Deegan and Rankin (1996), Mitchell et al.
(2006) and Clarkson et al. (2011), identify that worse performers put more effort into
their disclosure. Studies in other regions, including Cho et al. (2010) (the US), Marquis
and Qian (2014) (China), and Kim and Lyon (2015) (the US), support institutional
theory as well.
Thus, institutional theory informs that disclosure obfuscates worse performance,
and signalling theory posits a positive relationship between sustainability disclosure
and sustainability performance. As these two arguments are backed up by prior studies,
respectively, Chapter 3 does not make a directional hypothesis, and prior literature in
Australia does not give an adequate empirical foundation for a directional hypothesis.
The first non-directional hypothesis is presented below.
H1 There is no relationship between sustainability performance and six textual
characteristics of sustainability disclosure, a) amount of disclosure (i.e. measured in
number of words), b) amount of quantitative information in disclosure (i.e. measured
by the Diction 7), c) amount of information in relation to environmental impact (i.e.
measured in number of words), d) tone of optimism in disclosure (i.e. measured by the
Diction 7), e) tone of certainty in disclosure (i.e. measured by the Diction 7), and f)
tone of clarity in disclosure (i.e. measured by the Diction 7).
This chapter develops a second hypothesis about readability of disclosure and
finds its theoretical foundation in the incomplete revelation hypothesis developed by
Bloomfield (2002, 2008). This hypothesis indicates that information, which is costlier
to extract from corporate communication, would be less comprehended and therefore
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance 71
less absorbed by the audience (e.g. investors and other stakeholders). Information that
is costlier to extract exhibits lower readability (Loughran & McDonald, 2014a, 2014b,
2016). The literature about reliability of financial information, including Li (2008),
Bushee et al. (2017), and Jin, Luca, and Martin (2018), reveals that lower readability
of financial information is partially due to strategic decisions initiated by firms.
Following this silo of growing literature, Chapter 3 utilises incomplete revelation
hypothesis to develop the second hypothesis. Lower readability of sustainability
disclosure makes it more difficult for stakeholders to locate or uncover information
that firms attempt to hide (Bloomfield, 2002). It is thus reasonable to conjecture that
firms with better sustainability performance tend to present sustainability disclosure in
a more readable way (higher readability), ensuring that stakeholders comprehend their
performance in a less costly manner. While motivations behind why firms disclose
sustainability are reviewed by different and even conflicting theories31, a consensus
among them is that firms with superior performance are less likely to withhold their
performance. Better performers tend to present their performance in a more readable
way. Accordingly, the second hypothesis is shown below. The research design is
illustrated in following section.
H2 Readability of sustainability disclosure is positively related to sustainability
performance.
3.4 RESEARCH DESIGN
The sample period utilised in this thesis is between 2002 and 2016, as the
Thomson Reuters Asset4 covers Australian firms from 2002, and the data access ends
at 2016. As Section 3.1 indicates, data about sustainability performance was directly
extracted from the Thomson Reuters Asset4 (Asset4). I manually collected standalone
sustainability reports from the Sustainability Disclosure Database32 and various firms’
websites, and this collection took several months to undertake. If there were no
standalone reports, annual reports were examined to extract disclosure about corporate
sustainability. Control variables’ data were downloaded from Morningstar and
31 For example, socio-political theories consider legitimacy as a motivation behind voluntary
sustainability disclosure (Deegan, 2002), and voluntary disclosure theory is inclined to analyse
voluntary disclosure from the economic perspective (Clarkson et al., 2011). 32 Sustainability Disclosure Database is maintained by the GRI, and its URL is
http://database.globalreporting.org/.
72 Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
Worldscope. This sample consists of Australian firms that are constituents of the ASX
200.33 In total, I collected 2,639 items of sustainability disclosures for testing the
hypotheses. In terms of the first hypothesis, after synthesizing databases and removing
missing data, there is a final sample of 2,076 firm-year observations. Regarding the
second hypothesis, as the data from I/B/E/S are included as an additional control
variable, after synthesizing databases and removing missing data, the sample decreases
to 2,067 firm-year observations. Following Beck et al. (2018), the Global Industry
Classification Standard (GICS) is used to classify sample firms into ten industries.
Panel A in Table 3-1 presents the industry distribution of the sample, and Panel
B in Table 3-1 shows the year distribution of the sample. Panel C in Table 3-1 reports
how this sample was reduced from 2,639. Samples analysed in this chapter are quite
comparable with the samples examined by Nguyen, Agbola, and Choi (2018) who also
used the Thomson Reuters Asset4 to analyse corporate sustainability in Australia. It is
reasonable to suggest that the two samples analysed in this chapter are representative
in terms of industry distribution, compared with the complete ASX200.
33 ASX 200 represents about 82% of the market capitalization on the ASX. More details about ASX
200 can be found at https://au.spindices.com/indices/equity/sp-asx-200, and access date is 13 May
2019.
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance 73
Table 3-1 Sample Selection Panel A presents the sample distribution by industry, and Panel B presents the sample distribution by year. Panel C presents the synthesizing process of multiple databases.
Panel A Distribution of the First Sample Firms by Industry
Industrial Affiliation N Percent
Materials 501 24.13
Industrials 411 19.80
Financials 299 14.40
Energy 282 13.58
Consumer Discretionary 204 9.83
Consumer Staples 127 6.12
Health Care 122 5.88
Utilities 54 2.60
Information Technology 41 1.97
Telecommunication Services 35 1.69
Total 2076 100
Panel B Distribution of the First Sample Firms by Year
Year N Percent 2002 2 0.10
2003 2 0.10
2004 47 2.26
2005 55 2.65
2006 52 2.50
2007 57 2.75
2008 67 3.23
2009 152 7.32
2010 230 11.08
2011 249 11.99
2012 263 12.67 2013 255 12.28
2014 241 11.61
2015 218 10.50
2016 186 8.96
Total 2076 100
Panel C Sample Reduction Process
Observations
Observations available in the Thomson Reuters Asset4 (2002 – 2016) 2,639
LESS: missing data due to merging with Worldscope (191)
LESS: missing data due to merging with Morningstar (372)
Firm-year Observations for H1 2,076
LESS: missing data due to merge with I/B/E/S (9)
Firm-year Observations for H2 2,067
3.4.1 Corporate Sustainability Performance (CSP)
As suggested by Malik (2015) and Dragomir (2018), there are different ways to
operationalize sustainability performance. For instance, Galbreath and Shum (2012)
measure sustainability performance firm by firm by hiring experts’ opinion, and Lee,
Faff, and Langfield-Smith (2009), Humphrey, Lee, and Shen (2012) and Beck et al.
(2018) adopt ESG ratings from (rating) agencies to operationalize such performance.
As instructed by various seminal studies, including Ioannou and Serafeim (2010),
74 Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
Cheng et al. (2014), Eccles, Ioannou, and Serafeim (2014), Lys et al. (2015), Michelon,
Pilonato, and Ricceri (2015) and Liang and Renneboog (2017), this chapter measures
sustainability performance according to the ESG ratings assigned by the Thomson
Reuters Asset434 (i.e. Asset4). Asset4 has a multilevel structure in its assessment: at
the first level, there are more than 750 data points for firms in its universe; at the middle
level, these data points are synthesized into more than 250 performance indicators; and
at the very top level, performance indicators are synthesized into categories that are
composed of four pillars, (1) environmental performance, (2) social performance, (3)
corporate governance performance and (4) economic performance. In the Australian
literature, Asset4 ESG ratings have been used by Biswas et al. (2018), Nguyen et al.
(2018) and Krishnamurti, Shams, and Velayutham (2018) to operationalize Australian
firms’ sustainability performance.
While I acknowledge that the Asset4 considers firms’ disclosure in measuring
how firms perform regarding sustainability performance, it includes other sources,
including news, stock exchange filings and NGOs’ information. Following the studies
of Ioannou and Serafeim (2010), Cheng et al. (2014), Eccles et al. (2014) and Nguyen
et al. (2018) that also use the ESG ratings/ scores of Asset4 to measure sustainability
performance, I measure firms’ sustainability performance by averaging environmental
score/performance and social score/performance (i.e. the mean of the first two pillars).
It is noteworthy that the third pillar, corporate governance performance/ score, is also
included as control. Regarding sub-categories, environmental performance/ score
consists of three sub-categories: (1) emission reduction, (2) product innovation, and
(3) resource reduction; social performance/ score has six sub-categories: (1) product
responsibility, (2) diversity and opportunity, (3) health and safety, (4) community, (5)
training and development, and (6) employment quality; and, corporate governance
performance/ score has five sub-categories: (1) board functions, (2) board structure,
(3) compensation policy, (4) vision and strategy, and (5) shareholder rights. Although
I do not use economic performance/ score in Chapter 3, I still introduce sub-categories
of economic performance/ score for a comprehensiveness purpose: (1) performance,
34 More information about the Thomson Reuters Asset4 can be found on
https://libguides.mit.edu/sustainablebusiness/asset4 and https://www.sri-
connect.com/index.php?option=com_comprofiler&Itemid=4&task=userProfile&user=1007283, and
access date is 24 January 2019.
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance 75
(2) shareholder loyalty, and (3) client loyalty. The earliest available data is from 2002,
and the latest data ceases in 2016. The sub-categories of first two pillars are presented
in Figure 3-1. Figure 3-2 presents the second two pillars of Asset4, namely corporate
governance performance/ score and economic performance/ score. More details about
Asset4 are presented in Appendix A.
76Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
Figure 3-1 Asset4’s Scores Structure: Environmental Performance/Score and Social Performance/Score
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance 77
88Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
Table 3-4 Correlation Matrix of Variables A correlation matrix of variables is presented in this table. The variables tabulated below are described in Panel A and C of Table 3-2, as they are continuous variables.
Variables in Panel B of Table 3-2 are not included, as they are dummy variables. Variable CSP LENGTH NUMBER ENV IM OPTIMISM CERTAINTY CLARITY FRE FKGL SOMG CLI
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance 89
3.5.2 Equation (1) – Unexpected Part of Sustainability Performance (DA_CSP)
Findings derived from Equation (1) are tabulated below. As indicated by Table
3-5, firm size (Serafeim, 2013), corporate governance (Walls et al., 2012), years of
sustainability disclosure, and the presence of a social responsibility committee (Rankin
et al., 2011) are positively related to sustainability performance. Leverage is negatively
related to firms’ sustainability performance. Sample firms with higher leverage hold
less organizational slack, restricting sustainability investment (Arora & Dharwadkar,
2011). Krishnamurti and Velayutham (2017) find the existence of a risk management
committee is related to the disclosure of greenhouse gas emissions, yet this board
committee is found to have insignificant active impact on such performance. In terms
of economic significance, a one standard-deviation increase in firm size relates to a
4.44% increase in the performance; and economic significance of the other significant
variables also is apparent. The findings in relation to two hypotheses are presented in
following section.
90 Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
Table 3-5 Equation (1) – Unexpected Part of Sustainability Performance (DA_CSP) This table presents findings regarding Equation (1) that aims to estimate the unexpected part of sustainability performance (DA_CSP). Variables in this table are described in Table 3-2. In terms of
model fit, the Equation (1) arguably is fit, as the adjusted R-square is 0.62 that is comparable with
adjusted R-squares reported by prior studies.
CSP
Coeff.
(t-stat)
YEAR_REPORT 4.32
(8.65)***
RMC 0.09
(0.10)
CSR_C 6.16
(2.66)***
SIZE 1.48
(2.49)** LEV -3.49
(-1.67)*
ROA 0.001
(0.09)
MTB 0.01
(0.52)
NEW -0.85
(-1.01)
CAPEXP -0.01
(-0.68)
CG 0.26 (9.49)***
SHARECON 0.02
(0.60)
CONS -22.78
(-3.27)***
Year FE YES
Industry FE YES
Observations 2076
Adjusted R-square 0.62
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
3.5.3 Equation (2) – H1
Findings of H1 are tabulated below. As indicated by Table 3-6 (column 1 – 6),
the unexpected part of sustainability performance (DA_CSP) positively relates to the
first six textual characteristics. In terms of economic significance38, a one standard-
deviation increase in DA_CSP is related to a 32% increase in sustainability disclosure
(LENGTH) and is related to a 3.72% increase in quantitative information (NUMBER).
A one standard-deviation rise in the unexpected part of environmental performance
38 In terms of how to calculate economic significance, this chapter follows Miller (2005) and Gul,
Hutchinson, and Lai (2013).
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance 91
relates to an 18% increase in amount of environmental-impact information (ENV
_IM). A one standard-deviation increase in DA_CSP is related to a 65% increase in
tone of certainty (CERTAINTY), a 16% increase in tone of optimism (OPTIMISM)
and a 16% increase in tone of clarity (CLARITY). In terms of control variables, firm
size and status of corporate governance are positively related to the above textual
characteristics of disclosure. Thus, H1 is rejected.
92Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
Table 3-6 H1 and H2 This table presents results regarding analysis of the relationship between the unexpected part of sustainability performance (DA_CSP) and the seven textual characteristics of
sustainability disclosure. The unexpected part of environmental performance is estimated by replacing sustainability performance with environmental performance in
Equation (1). Variables tabulated below are described in Table 3-2. In terms of model fit, Equation (2) is based on instruction from prior studies, and the proportion of
variance is comparable with proportion of variance explained by prior studies; Equation (3) is based on instruction from prior studies, and he proportion of variance is
comparable with proportion of variance explained by prior studies.
(1) (2) (3) (4) (5) (6) (7)
LENGTH NUMBER ENV_IM OPTIMISM CERTAINTY CLARITY READ
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
DA_CSP 0.02
(10.93)***
0.23
(3.95)***
0.01
(3.72)***
0.01
(2.53)**
0.04
(4.13)***
0.01
(4.08)***
0.017
(3.12)***
RMC -0.01
(-0.02)
-2.27
(-1.49)
0.03
(0.57)
0.30
(2.15)**
0.09
(0.35)
0.09
(1.81)*
-
CSR_C 0.42
(4.62)***
6.68
(1.89) *
0.05
(0.71)
0.38
(1.60)
-0.52
(-1.43)
0.14
(1.91)*
-
SIZE 0.30
(11.06)***
0.23
(3.95) ***
0.04
(1.88)*
0.09
(1.01)
0.03
(5.19)***
-0.04
(-1.58)
0.120
(1.89)*
LEV 0.16
(0.94)
-4.34
(-0.98)
0.04
(0.36)
-0.14
(-0.32)
0.04
(1.39)
0.08
(0.58)
0.012
(0.05)
ROA 0.00
(-0.90)
0.02
(0.81)
0.00
(0.23)
0.00
(0.94)
0.00
(1.08)
0.00
(-0.11)
0.129
(0.56)
MTB 0.00
(1.86)*
-0.01
(-0.43)
0.00
(0.31)
0.02
(2.19)**
0.00
(-0.63)
0.00
(0.60)
-0.000
(-0.75)
CG 0.00 (-1.98)*
0.10 (2.70)***
0.00 (-0.88)
0.002 (0.65)
-0.10 (-0.09)
0.00 (-0.67)
-
SHARECON 0.01
(7.43)***
-0.05
(-1.79)
0.00
(0.90)
0.001
(0.49)
0.01
(1.79)*
0.00
(-1.04)
-
CROLIST - - - - - - 0.067
(0.89)
BUSSEG - - - - - - -0.019
(-0.54)
COVERAGE - - - - - - 0.004
(0.29)
CONS 0.00 6.23 0.00 0.01 0.00 0.00 -2.937
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance 93
and READ) relate to performance in different periods, and vice versa. As Arena et al.
(2015) instruct, I consider one-year and two-year lags in these tests: performance may
affect textual characteristics of disclosure in the following year and two years, and vice
versa. Findings are presented below in Tables 3-7.
96Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
Table 3-7 Signalling Effects of Textual Characteristics in Sustainability Disclosure This table presents the results of signalling effects of the seven textual characteristics on sustainability performance. Following Choi (2001), unit root tests for panel data are
conducted and reported in Panel A. Panel B presents results of panel Granger causality tests. One period of lag and two periods of lag are considered. In the Granger causality
test of ENV_IM, unexpected part of sustainability performance is replaced by unexpected part of environmental performance. Variables tabulated below are described in
Table 3-2.
Panel A: Fisher-type Augmented Dickey-Fuller Tests for Panel Data
Variable p-value
CERTAINTY 0.00
CLARITY 0.00
DA_CSP 0.00
ENV_IM 0.00
LENGTH 0.00
NUMBER 0.00
OPTIMISM 0.00 READ 0.00
Panel B: Granger Causality Tests (the Chi probability is reported)
Unexpected part of sustainability
performance does not Granger
cause a textual characteristic
(One Lag)
A textual characteristic does not
Granger cause unexpected part
of sustainability performance
(One Lag)
Unexpected part of sustainability
performance does not Granger
cause a textual characteristic
(Two Lag)
A textual characteristic does not
Granger cause unexpected part of
sustainability performance
(Two Lag)
CERTAINTY 0.357 0.812 0.558 0.950
CLARITY 0.186 0.814 0.401 0.597
ENV_IM 0.374 0.107 0.987 0.254
LENGTH 0.765 0.516 0.675 0.519
NUMBER 0.287 0.685 0.815 0.187
OPTIMISM 0.738 0.907 0.985 0.512
READ 0.721 0.129 0.219 0.176
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance 97
Panel A of Table 3-7 presents the Augmented Dickey-Fuller test (Choi, 2001):
the variables concerned are stationary. Panel B of Table 3-7 presents the findings of
the panel Granger causality tests. With respect to the first six textual characteristics, as
the critical value for one period of lag is 𝜒2 (1, N = 2076) = 3.841 and for two periods
of lag is 𝜒2 (1, N = 2076) = 5.991, it is obvious that there is no relationship between
sustainability performance and the lags in different time periods. Similar findings
regarding the seventh textual characteristic are also revealed: as the critical value for
one period of lag is 𝜒2 (1, N = 2067) = 3.841 and for two periods of lag is 𝜒2 (1, N =
2067) = 5.991, there is no relationship between sustainability performance and the
seventh characteristic over different time periods. Thus, the findings of two hypotheses
are substantiated.
3.5.6 Robustness Test Two – Alternative Measurement on Sustainability
Performance
In the main analysis of that two hypotheses, firms’ sustainability performance is
measured as an average of social score and environmental score. I use the principal
component analysis to extract a latent component from nine sub-categories which are
composed of a social score and an environmental score (see Section 3.4.1and Figure
3-1). It is used as the alternative measurement on sustainability performance in Chapter
3. The findings are presented below. As Table 3-8 indicates, the findings about the two
hypotheses remain qualitatively unchanged.
98Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
Table 3-8 Alternative Measurement on Sustainability Performance This table presents the results of the relationship between unexpected part of sustainability performance and seven textual characteristics of sustainability disclosure. Panel A
shows alternative measurement of sustainability performance based on the principal component analysis. Panel B shows the results of the relationship between alternative
measurement on the unexpected part of sustainability performance and seven textual characteristics of sustainability disclosure. Environmental performance is calculated by
extracting a latent component from the three sub-categories comprising it. The unexpected part of environmental performance is estimated by replacing sustainability
performance with environmental performance in Equation (1).
Panel A: Principal Component Analysis on Nine Sub-Categories of Sustainability Performance
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
100 Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
3.5.7 Robustness Test Three – Sectoral Analysis
Following Cho and Patten (2007) and Eccles et al. (2014), this test categorises
the GICS industries into two pairs of categories, (1) natural versus non-natural40; (2)
business-to-consumer (B2C) and others41. In terms of the first pair, I code the value of
one for industries whose products are deemed as sensitive to sustainable development
(the natural industries), and the value of zero otherwise (the non-natural industries)
(De Villiers, Naiker, & Van Staden, 2011). It is argued that firms in the natural
industries are subject to more pressures relevant to sustainability (Cho, Laine, Roberts,
& Rodrigue, 2015) so that the relationship between how firms communicate about
sustainability and how firms perform sustainability can consequentially be affected by
such pressures. With respect to the second pair, the value of one is coded for B2C
industries and zero otherwise. Following Eccles et al. (2014), it is argued that firms in
the B2C industries are more visible as they directly interact with consumers. Thus,
firms in the B2C industries are more likely to care about their image and reputation
than their peers in other industries. The relationship between textual characteristics
and performance can be affected by such concerns. The findings are presented in Table
3-9. Panel A of Table 3-9 shows natural versus non-natural, and Panel B of Table 3-9
reports B2C versus others.
40 Following de Villiers, Naiker and Van Staden (2011, p. 1650), natural industries include forestry,
metal mining, coal mining and oil and gas exploration, paper and pulp mills, chemicals,
pharmaceutical and plastics manufacturing, iron and steel, manufacturing and electricity, gas and
wastewater. 41 Following Eccles et al. (2014, p. 2850), B2C industries include consumer goods and finance.
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance 101
Table 3-9 Sectoral Analysis This table presents the results of sectoral moderation effects on the relationship between unexpected part of sustainability performance and seven textual characteristics of
sustainability disclosure. They are the number of words (LENGTH), the amount of quantitative information (NUMBER), the amount of environmental impact information
(ENV_IM), optimism (OPTIMISM), certainty (CERTAINTY), clarity (CLARITY) and readability (READ). Panel A presents how industry affiliation based on natural versus
non-natural affects the relationship between performance and textual characteristics. Natural is coded as a dummy variable – firms in natural industries are coded one, and
zero otherwise. Panel B presents how industry affiliation based on B2C versus others affects such relationship. B2C is coded as a dummy variable – firms in B2C industries
are coded one, and zero otherwise. Variables tabulated below are described in Table 3-2. Panel A Natural versus Non-natural
LENGTH NUMBER ENV_IM OPTIMISM CERTAINTY CLARITY READ
Coeff. (t-stat)
Coeff. (t-stat)
Coeff. (t-stat)
Coeff. (t-stat)
Coeff. (t-stat)
Coeff. (t-stat)
Coeff. (t-stat)
DA_CSP 0.14 (2.75)***
0.27 (3.79)***
0.01 (2.69)***
-0.01 (-1.22)
0.14 (2.75)***
0.01 (4.20)***
0.028 (2.93)***
Natural -0.43 (-1.16)
4.01 (1.34)
-0.08 (-0.64)
0.10 (0.30)
-0.43 (-1.16)
-0.21 (-1.28)
0.146 (0.13)
Natural × DA_CSP 0.09 (1.12)
-0.09 (-0.76)
0.00 (-0.14)
0.03 (2.98)***
0.09 (1.12)
0.004 (1.29)
-0.013 (-1.45)
Controls included YES YES YES YES YES YES YES Year FE YES YES YES YES YES YES YES
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
102 Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
Regarding natural versus non-natural affiliation, this chapter does not find that
the sectoral environment of the natural industries influences the relationship between
performance and disclosure characteristics except for optimistic tone of language. As
Panel A of Table 3-9 indicates, firms in the natural industries are likely to use more
optimistic tone of language than their peers in non-natural industries. As Clarkson et
al. (2011), Cho et al. (2010) and Patten (2002) instruct, industrial pressures can be a
key factor affecting the relationship between textual characteristics and performance.
This finding is in line with the expectation that firms in natural industries are subject
to more external pressures to demonstrate sustainability and face more scrutiny of their
sustainability practices, leading them to communicate in a more optimistic way.
Regarding B2C versus other affiliation, this chapter identifies that the sectoral
environment of B2C industries affects optimism, clarity and readability. As indicated
by Panel B of Table 3-9, firms in B2C industries are less likely to use an optimistic
tone of language and are more likely to present information in a clearer manner than
their peers in other sectors. As expected, firms in B2C sectors care more about their
public image and are likely to be more cautious (less use of optimistic tone) and clearer
in their disclosure. Moreover, firms in B2C industries do pay more attention on how
to present sustainability disclosure so that their disclosure is more readable. Thus, the
sectoral environment seems to influence the relationship between disclosure and
performance. As prior studies that analyse linguistic characteristics in disclosure,
including Cho et al. (2010) and Arena et al. (2015), incorporate annual reports only,
Chapter 3 explores whether a separation of annual report and standalone report affects
the findings in following section.
3.5.8 Robustness Test Four – Disclosure Channels
Chapter 3 further examines whether findings about the two hypotheses are
sensitive to various disclosure channels. Arguably, compared to putting sustainability
disclosure in annual reports, releasing standalone reports is a more proactive way of
communicating sustainability practices42. The use of standalone reports demonstrates
that firms that prepare them are more interested in sustainability communication. Thus,
42 A standalone report represents “a very clear engagement by corporations with the increasingly
critical issues of environmental stewardship, social responsibility, and planetary sustainability at a
time when society’s well-being and the planet itself are under unique levels of threat” (Gray &
Herremans, 2011, pp. 2-3).
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability
Performance 103
disclosure medium may affect the relationship between sustainability performance and
textual characteristics. Sample firms are divided into two groups that are determined
by whether they release a standalone report in a year. As presented in Table 3-10,
reporting practices between firms that use standalone reports and firms that do not use
them are very similar, except for three characteristics (i.e. LENGTH, OPTIMISM and
CLARITY). Firms that use standalone reports tend to disclose more comprehensive
information, which is presented in more optimistic terms and communicated in clearer
tones. The findings are consistent with prior literature, including Gray and Herremans
(2011) and Higgins et al. (2015), which highlights that the use of standalone report
demonstrates a proactive attitude towards sustainability communication and/or desire
to improve corporate reputation.
104Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
Table 3-10 Disclosure Channels: Standalone Report versus Annual Report This table presents the results of whether use of standalone reports affects relationship between unexpected part of sustainability performance and the seven textual
characteristics of sustainability disclosure, namely the number of words (LENGTH), the amount of quantitative information (NUMBER), the amount of environmental impact information (ENV_IM), optimism (OPTIMISM), certainty (CERTAINTY), clarity (CLARITY) and readability (READ). In the analysis of ENV_IM, the unexpected part of
sustainability performance is replaced by the unexpected part of environmental performance. Variables tabulated below are described in Table 3-2. A dummy variable,
Dummy for Use of Standalone Report, is coded the value of one for the use of a standalone report in a firm-year observation, and the value of zero otherwise.
LENGTH NUMBER ENV_IM OPTIMISM CERTAINTY CLARITY READ
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
DA_CSP 0.02
(8.36)***
0.01
(2.17)**
0.01
(1.40)
0.01
(0.69)
0.14
(2.41)**
0.01
(2.08)**
0.023
(2.12)**
Dummy for Use of Standalone Report 0.53
(6.27)***
0.34
(5.06)***
0.20
(2.85)**
0.51
(2.37)**
-0.15
(-0.33)
0.22
(3.14)***
0.723
(1.50)
Dummy for Use of Standalone Report × DA_CSP 0.01
(2.18)**
0.03
(0.92)
0.01
(1.21)
0.02
(1.85)*
0.05
(0.47)
0.01
(1.16)**
-0.011
(-1.21)
Controls included YES YES YES YES YES YES YES
Year FE YES YES YES YES YES YES YES Industry FE YES YES YES YES YES YES YES
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability
Performance 105
3.5.9 Robustness Test Five – Time Stability
While Australia maintains a voluntary disclosure status in relation to corporate
sustainability during this sample period, it is reasonable to argue disclosure practice
may change over 15 years. As discussed in the introduction, the increased awareness
of corporate sustainability and availability of increased sustainability guidance were
observed during this 15-year period. Therefore, it is possible that findings of the two
hypotheses may change over time. This test selects the first five years (2002 – 2006)
and the last five years (2012 – 2016) to examine time stability. Observations of the
first five years are coded as zero, and ones of the last five years are coded as one. As
Table 3-11 reports, two characteristics of sustainability disclosure change over this
sample period: firms tend to disclose more information about environmental impact
and present their information in clearer terms. I also re-run the main analysis on the
data from 2010 to 2016, which consists of about 80% of both final samples, and the
results remain qualitatively unchanged.
106Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
Table 3-11 Time Stability This table presents the results about whether time affects the relationship between the unexpected part of sustainability performance and the seven textual characteristics in
sustainability disclosure, namely the number of words (LENGTH), the amount of quantitative information (NUMBER), the amount of environmental impact information
(ENV_IM), optimism (OPTIMISM), certainty (CERTAINTY), clarity (CLARITY) and readability (READ). In the analysis of ENV_IM, the unexpected part of sustainability
performance is replaced by the unexpected part of environmental performance. Variables tabulated below are described in Table 3-2. The first five years (2002 – 2006) and
the last five years of the sample period (2012 – 2016) are selected to test time stability. A dummy variable (Dummy of Time Periods) is introduced: observations of the first
five years are coded as zero, and observations of the last five years are coded as one.
LENGTH NUMBER ENV_IM OPTIMISM CERTAINTY CLARITY READ
As Table 3-12 shows, the coefficient of the one-year lagged unexpected part of
sustainability performance is highly significant and equal to 3.328. The explanatory
power of the recursive equation is substantially higher than the Equation (1) in Table
3-12 with a R-squared of 89% versus 62% in Table 3-5. I re-run Equation (2) and (3)
using the unexpected part of performance determined using Equation (4). As Table 3-
12 presents, the findings regarding H1 and H2 remain qualitatively unchanged.
43 The lagged values of the unexpected part of sustainability performance (DA_CSP) are incorporated
into the Equation (4).
108Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
Table 3-12 Alternative Model Specification This table presents the results from alternative model specification. This alternative specification includes the lagged values of the unexpected part of sustainability
performance in Table 3-5 as additional explanatory variables in Equation (1). Seven textual characteristics in sustainability disclosure, namely number of words (LENGTH),
amount of quantitative information (NUMBER), amount of environmental impact information (ENV_IM), optimism (OPTIMISM), certainty (CERTAINTY), clarity
(CLARITY) and readability (READ) are included. In the analysis of ENV_IM, the unexpected part of sustainability performance is replaced by the unexpected part of
environmental performance. Variables tabulated below are described in Table 3-2.
Panel A Equation (4)
CSP
Coeff.
(t-stat)
One-year Lagged DA_CSP 3.328
(4.24)***
Other Variables of Equation (1) Included YES
Year FE YES Industry FE YES
Observations 2076
Adjusted R-square 0.89
Panel B Equation (2) and (3) including the Unexpected Part of Sustainability Performance Estimated by Using Equation (4)
LENGTH NUMBER ENV_IM OPTIMISM CERTAINTY CLARITY READ
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability
Performance 109
3.5.11 Robustness Test Seven – Inclusion of Firms’ Experience of Sustainability
Disclosure as a Control
This chapter provides additional assurance that the findings of H1 and H2 are
not caused by excluding the experience of sustainability disclosure (YEAR_REPORT)
as a control. I re-estimate Equation (1) and (2) to include the experience of
sustainability disclosure (YEAR_REPORT) as a control. As Table 3-13 shows, the
findings of H1 and H2 remain qualitatively unchanged.44
44 I appreciate this suggestion from the thesis examiner.
110Chapter 3: Textual Characteristics of Sustainability Disclosure and their Relationship to Sustainability Performance
Table 3-13 Inclusion of Firms’ Experience of Sustainability Disclosure as a Control This table presents the results from including firms’ experience of sustainability disclosure (YEAR_REPORT) as a control. Seven textual characteristics in sustainability
disclosure, namely number of words (LENGTH), amount of quantitative information (NUMBER), amount of environmental impact information (ENV_IM), optimism
(OPTIMISM), certainty (CERTAINTY), clarity (CLARITY) and readability (READ) are included. In the analysis of ENV_IM, the unexpected part of sustainability
performance is replaced by the unexpected part of environmental performance. Variables tabulated below are described in Table 3-2.
LENGTH NUMBER ENV_IM OPTIMISM CERTAINTY CLARITY READ
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
DA_CSP 0.02
(11.23)***
0.24
(5.38)***
0.01
(3.93)***
0.01
(2.30)**
0.02
(4.41)***
0.01
(4.11)***
0.02
(3.15)***
YEAR_REPORT 0.12
(4.84)***
0.83
(1.77)
0.07
(3.50)***
0.08
(0.35)
0.01
(1.77)*
0.04
(1.58)
-0.07
(-1.27)
Constant and other Controls Included YES YES YES YES YES YES YES
Year FE YES YES YES YES YES YES YES Industry FE YES YES YES YES YES YES YES
Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability
Performance 137
4.5 RESULTS
4.5.1 Descriptive Statistics
Descriptive statistics are presented in Table 4-3. The first sample (with 2,166
firm-year observations) is presented in Panel A of Table 4-3. Clearly, there is much
variance in sustainability performance with minimum score at 0 and maximum score
at 97.89. Firms in the first sample have a maximum of three sustainability committees.
On average, sample firms have four years’ experience of disclosure. Compared with
the sample analysed in Rankin et al. (2011), firms in the first sample are smaller in
firm size, have higher leverage ratio and report lower return on assets. As presented in
Panel B of Table 4-3, about half of the firms in the first sample have a risk management
committee, and one fifth have a sustainability committee. The percentage of firms with
a sustainability committee in the first sample is comparable to the percentage of firms
with an environmental committee (18.7%) in Rankin et al. (2011).
The second sample of 430 firm-year observations regarding the sustainability
committees is presented in Panel C of Table 4-3. Regarding committee composition,
84% of committee members are reported as independent, and 11% are sustainability
experts. Regarding committee authority, average board tenure is 5.94 years, and an
average committee tenure is 3.59 years. Regarding resources, 17% of them are women,
49% sit on an audit committee, and 34% sit on a risk management committee. 56 On
average, sustainability committee members hold 2.38 higher education qualifications
and 2.92 positions in other organizations or firms. Compared with directors analysed
in Gray and Nowland (2013), sustainability committee members in this chapter hold
comparable number of higher education qualifications and are more active (i.e. they
sit on more other organizations). A committee consists of 4.12 directors. Regarding
diligence, a committee holds 3.6 meetings per year, average attendance rate is 96%,
and members in a committee hold 0.3% of ordinary shares. It is reasonable to argue
that sustainability committees in the second sample are comparable with committees
analysed in other studies, including Rodrigue et al. (2013), Peters and Romi (2014,
2015) and Dixon-Fowler et al. (2017). It is noteworthy prior studies do not include as
56 It is noteworthy that audit committees are compulsory to listed firms in Australia, and risk
management committees are strongly encouraged for listed firms in Australia.
138 Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance
many characteristics of a sustainability committee as this chapter does. Correlation of
variables is presented in Table 4-4. It is interesting to notice that firms’ sustainability
performance is positively related to the presence of a sustainability committee and
effectiveness of a sustainability committee.
Table 4-3 Descriptive Statistics Panel A presents descriptive statistics about important variables. Panel B presents the number of risk management committees and the number of sustainability committees. It is noteworthy that
audit committee is compulsory to public firms in Australia. Panel C presents the characteristics of
the sustainability committee that are used to calculate scores about committee effectiveness.
Market Capitalization (in Billions) 2166 4.30 0.03 4.22 4.38
CG 2166 61.82 26.65 0.00 97.51
NEW 2166 0.63 0.23 0.00 1.00
CAPEXP 2166 0.08 0.11 0.00 1.00
SHARECON 2166 62.83 24.04 0.00 97.39
Panel B Frequency TOTAL PERCENTAGE
SC 430 19.85
RMC 1115 51.48
Panel C Characteristics of
Sustainability Committee
N MEAN STD.DEV MIN MAX
SC_INDEP 430 0.84 0.20 0 1
SC_EXP 430 0.11 0.20 0 1
BOARD_TEN 430 5.94 2.49 0 20.67
SC_TEN 430 3.59 1.76 0 8.75
SC_AC 430 0.49 0.26 0 1 SC_RMC 430 0.34 0.32 0 1
SC_GENDER 430 0.17 0.18 0 1
SC_RELATION 430 2.92 1.38 0 6.75
SC_EDUCATION 430 2.38 0.78 0 6
SC_MEET 430 3.60 1.32 0 10
SC_ATTEND 430 0.96 0.10 0 1
SC_SHARE 430 0.003 0.02 0 0.15
Size of Sustainability Committee 430 4.12 1.45 1 10
Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance 139
Table 4-4 Correlation Matrix of Variables A correlation matrix of variables is presented in this table. The variables tabulated below are described in Panel A of Table 4-3. Variables in Panel B of Table 4-3 are not
included, as they are dummy variables. Variables in this table are defined in Table 4-2. Variable CSP SC_NO SC_EFFE SIZE LEV ROA MTB NEW CAPEXP CG SHARECON YEAR_REPORT
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
140 Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance
4.5.2 Experience of Sustainability Disclosure, the Presence of a Sustainability
Committee and Sustainability Performance – Equation (1) – H3 and H4a
Findings regarding H3 and H4a are shown in Table 4-5. As the number of years’
experience (YEAR_REPORT) is positively and significantly related to sustainability
performance (CSP), H3 is not rejected. Following Section 3.5.5 that uses Fisher-type
Augmented Dickey-Fuller Test to perform unit root test, I also used this test to check
stationarity of CSP data. At significance level of 1%, I rejected the null hypothesis that
CSP data is non-stationary. Thus, findings about H3 is not influenced by co-integration
issue, and sample firms with a longer duration of disclosure deliver better performance.
While H4a is about the presence of a sustainability committee, where a firm has
more than one sustainability committee, I substitute the presence of a sustainability
committee with the number of sustainability committees to provide more evidence
about this hypothesis. How the presence of a sustainability committee and the number
of sustainability committees are related to sustainability performance, respectively, are
shown in the first two columns of Table 4-5. The presence of a committee is
significantly related to sustainability performance, β = 9.928 (p<0.01), and the number
of committees is also significantly related to performance, β = 9.836 (p<0.01).
How the presence of this committee is related to environmental performance and
to social performance, respectively, is shown in the second two columns of Table 4-5.
The presence of a sustainability committee is significantly related to environmental
performance, β = 10.461 (p<0.01), and social performance, β = 9.097 (p<0.01). Thus,
H4a is rejected. Apparently, the effect of presence of a sustainability committee and
number of sustainability committees on firms’ sustainability performance is strong.
Regarding controls, firm size (Serafeim, 2013) and firms’ corporate governance
performance (Walls et al., 2012) are positively related to sustainability performance.
Leverage is negatively related to firms’ sustainability performance. Firms with higher
leverage hold less slack, restricting their investment in corporate sustainability (Arora
& Dharwadkar, 2011). Findings therefore support institutional theory and stewardship
theory. In terms of H3, Chapter 4 finds firms with more experience of sustainability
disclosure deliver better performance. Regarding H4a, this chapter reveals that the
presence of a sustainability committee can improve sustainability performance. The
aforementioned findings are consistent with prior literature in the US, including Walls
et al. (2012) and Dixon-Fowler et al. (2017).
Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability
Performance 141
Table 4-5 Experience of Sustainability Disclosure, the Presence of a Sustainability
Committee and Sustainability Performance – H3 and H4a This table presents the results of analysis of the relationship between experience of sustainability disclosure and sustainability performance and relationship between the presence of a sustainability
committee and such performance. The first two columns indicate the presence of a sustainability
committee and the number of such committees, respectively. The second two columns show
environmental performance and social performance, respectively. Variables in this table are described
in Table 4-2. In terms of model fit, the Equation (1) arguably is fit – it is based on prior studies, and
the proportion of variance is comparable with proportion of variance explained by prior studies.
(1) (2) (3) (4)
DV: CSP DV: CSP DV: ENV DV: SOC
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
SC 9.928
(3.68)***
- 10.461
(3.55)***
9.097
(3.02)***
SC_NO -
9.836
(3.89)***
- -
YEAR_REPORT 3.773
(7.03)***
3.795
(7.09)***
4.317
(6.82)***
3.296
(5.56)***
SIZE 2.493
(3.71)***
2.457
(3.66)***
2.441
(3.08)***
2.313
(2.98)***
LEV -4.837
(-1.72)*
-4.864
(-1.73)*
-3.211
(-0.95)
-6.313
(-2.09)**
ROA -0.017
(-1.13)
-0.017
(-1.11)
0.031
(1.77)*
0.005
(0.29)
MTB 0.036
(1.65)
0.036
(1.67)*
0.022
(0.91)
0.041
(1.61)
NEW -0.779 (-0.50)
-0.797 (-0.51)
-1.106 (-0.64)
-0.365 (-0.23)
RMC 1.565
(1.60)
1.570
(1.60)
1.849
(1.55)
1.720
(1.62)
CAPEXP -0.012
(-0.63)
-0.013
(-0.66)
0.004
(0.21)
-0.018
(-0.83)
CG 0.229
(7.74)***
0.226
(7.60)***
0.189
(5.49)***
0.273
(7.93)***
SHARECON 0.013
(0.41)
0.012
(0.38)
0.008
(0.23)
0.005
(0.12)
CONS -38.285
(-3.99)***
-38.104
(-3.99)***
-32.428
(-3.10)***
-32.178
(-2.81)***
Firm FE YES YES YES YES Year FE YES YES YES YES
Industry FE YES YES YES YES
Observations 2166 2166 2166 2166
Adjusted R-
square
0.63 0.62 0.57 0.55
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
142 Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance
4.5.3 Relationship between Sustainability Committee Effectiveness and
Sustainability Performance – Equation (2) – H4b
Findings regarding H4b are shown in Table 4-6. How sustainability committee
effectiveness as a whole is related to sustainability performance, environmental
performance and social performance is presented in the first three columns. As
presented in Table 4-6, sustainability committee effectiveness is positively related to
environmental performance, β = 1.499 (p<0.05), and effect of sustainability committee
effectiveness on firms’ environmental performance is also economically significant.
Thus, H4b is partially rejected, that is sustainability committee effectiveness is not
associated with sustainability performance as a whole. Additionally, how the four
components of sustainability committee effectiveness, namely composition, authority,
resources and diligence, are related to performance is presented in the second three
columns. As presented in Table 4-6, resources are positively related to environmental
performance, β = 3.177 (p<0.01), and authority is positively related to environmental
performance, β =1.829 (p<0.10). These two components of committee effectiveness
are economically significant. The control variables also behave as expected.
Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance 143
Table 4-6 Sustainability Committee Effectiveness and Sustainability Performance – H4b This table presents the results of the analysis of the relationship between sustainability committee effectiveness and sustainability performance. The first three columns show
relationships between committee effectiveness and sustainability performance, environmental performance and social performance, respectively. The second three columns
show relationships between the four components of effectiveness and three types of performance. Variables in this table are described in Table 4-2. In terms of model fit, the
Equation (2) arguably is fit – it is based on prior studies, and the proportion of variance is comparable with proportion of variance explained by prior studies. (1) (2) (3) (4) (5) (6)
DV: CSP DV: ENV DV: SOC DV: CSP DV: ENV DV: SOC
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
SC_EFFE 0.599
(1.18)
1.499
(2.04)**
-0.052
(-0.96)
- - -
Composition - - - -13.941
(-1.39)
-16.577
(-1.36)
-11.305
(-1.23)
Authority - - - 0.392 (0.36)
1.829 (1.93)*
-1.045 (-0.82)
Resources - - - 1.703
(1.52)
3.177
(2.68)***
0.229
(0.18)
Diligence - - - 1.437
(0.73)
-0.472
(-0.19)
3.346
(1.42)
YEAR_REPORT 3.762
(2.83)***
5.848
(3.45)***
0.505
(3.10)***
2.723
(1.47)
4.844
(2.10)***
0.602
(0.38)
SIZE 2.293
(0.21)
-0.575
(-0.40)
-0.100
(-0.82)
0.649
(0.47)
0.220
(0.17)
1.518
(0.72)
LEV -11.564
(-2.37)**
-16.083
(-2.96)***
0.027
(0.06)
-12.369
(-2.32)**
-18.213
(-3.11)***
6.525
(0.98) ROA 0.034
(0.63)
0.051
(0.75)
0.006
(1.26)
0.036
(0.67)
0.062
(0.94)
0.010
(0.17)
MTB 0.038
(0.38)
-0.041
(-0.57)
0.002
(0.45)
0.045
(0.48)
0.031
(0.41)
0.121
(1.01)
NEW -8.580
(-0.98)
-19.804
(-1.81)*
0.553
(0.94)
-11.478
(-1.26)
-22.405
(-1.92)*
-0.552
(-0.05)
RMC 1.281
(0.70)
3.148
(1.30)
-0.076
(-0.58)
1.063
(0.64)
2.883
(1.37)
-0.756
(-0.48)
CAPEXP 0.060
(0.50)
0.017
(0.11)
-0.001
(-0.05)
0.064
(0.56)
0.051
(0.34)
0.077
(0.65)
144Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance
CG 0.327
(5.42)***
0.278
(3.61)***
0.010
(2.19)**
0.350
(5.35)***
0.310
(3.86)***
0.390
(6.01)***
SHARECON 0.236
(2.48)**
0.248
(1.94)*
0.010
(1.82)*
0.228
(2.64)**
0.231
(1.87)*
0.225
(2.85)***
CONS -19.705
(-0.82)***
-38.236
(-1.47)
-2.234
(-0.85)
-6.601
(-0.21)***
-19.020
(-0.55)***
5.819
(0.16)
Firm FE YES YES YES YES YES YES
Year FE YES YES YES YES YES YES
Industry FE YES YES YES YES YES YES
Observations 430 430 430 430 430 430
Adjusted R-square 0.16 0.43 0.10 0.36 0.35 0.24
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability
Performance 145
Findings of H4b support committee effectiveness as discussed by DeZoort et al.
(2002) and the resources dependency theory. On one side, the sustainability committee
effectiveness and its two components (i.e., resources and authority) are significantly
related to environmental performance, reinforcing the argument that who sits on the
sustainability committee, and how they are grouped as a whole matter. On the other
side, sustainability committee effectiveness is not related to social performance. Social
performance includes customer and product responsibility, community, human rights,
diversity and opportunity, health and safety, employment quality, as well as training
and development (refer to Appendix 1). Thus, a speculation about this insignificant
relationship between the sustainability committee effectiveness and social
performance is that social issues are managed by the firms’ other departments,
mechanisms, or structures, rather than their sustainability committees. For example, in
relation to workers’ training and development and , labour relations, managers and
boards of directors may choose to supervise relevant these issues through other
mechanisms (e.g., risk management committee). Future studies are encouraged to
analyse if not by sustainability committee, which corporate governance mechanisms
are used to supervise social issues, and why social issues seem to be marginalized on
the agenda of the sustainability committee.
Arguably, this chapter extends on the studies of Gray et al. (2016) and Nowland
and Simon (2018) who examine director characteristics in Australia from board level
to committee level. Extending Rankin et al. (2011), Rodrigue et al. (2013), and Dixon-
Fowler et al. (2017) who investigate the presence of sustainability committee, Chapter
4 analyses a relatively novel direction, namely sustainability committee effectiveness.
Different from Peters and Romi (2015) who identify that composition of sustainability
committee (i.e., experts on committee) is related to use of assurance on sustainability
disclosure, I reveal that committee authority and resources are related to environmental
performance. In the following sections, Chapter 4 conducts robustness tests as
instructed by prior studies.
4.5.4 Robustness Test One – Endogeneity in H3 and H4a – PSM
Following Antonakis et al. (2010), Mishra (2014), Michelon et al. (2015) and
Nguyen et al. (2018), this chapter adopts PSM method to address omitted selection in
testing H3 and H4a. Two hypotheses are re-examined by matched-sample design. With
regard to H3, sample firms are grouped according to their experience of sustainability
146 Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance
disclosure (YEAR_REPORT). We code the value of one, if a firm-year observation’s
YEAR_REPORT is greater than the median of the full sample, and zero otherwise. A
propensity score is calculated from the first-stage model to predict whether a firm’s
YEAR_REPORT is above the median. The prediction model has all control variables
of Equation (1). Each firm-year observation whose YEAR_REPORT above median is
matched with another observation whose YEAR_REPORT is not. It is noteworthy that
this PSM analysis is based on 2,166 firm-year observations. Findings from this PSM
analysis are tabulated in Table 4-7. Panel A of Table 4-7 presents that the matching
procedure is valid, as all matched variables are statistically indistinguishable. Panel B
of Table 4-7 presents after using the PSM analysis to address for potential endogeneity,
YEAR_REPORT still is found to be positively related to CSP, β = 3.864 (p<0.01).
Regarding H4a, similarly, a propensity score is calculated from the first-stage
model to predict a firm with a sustainability committee. The prediction model includes
all control variables of Equation (1). Then each firm-year observation that corresponds
to a sustainability committee is matched with another observation that does not. It is
noteworthy that this PSM analysis is based on 2,166 firm-year observations. Findings
from this PSM analysis are tabulated in Table 4-8. As presented in Panel A of Table
4-8, matched variables at the first stage are statistically indistinguishable, suggesting
that the matching procedure is valid. Panel B of Table 4-8 shows that the presence of
a sustainability committee is positively related to sustainability performance, β = 7.424
(p<0.01); the firms’ experience of sustainability disclosure positively relates to
sustainability performance, β = 3.536 (p<0.01). Taken together, findings about H3 and
H4a remain qualitatively unchanged.
Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability
Performance 147
Table 4-7 PSM – H3 Panel A presents the results of propensity-matched variables when dependent variable is a dummy
variable that reflects whether a firm’s experience of sustainability disclosure is above median.
YEAR_REPORT_DUMMY is coded with the value of one, if a firm’s experience of sustainability
disclosure is above industry-year median, and the value of zero otherwise. Panel B presents the
regression results. Variables in this table are defined in Table 2.
Panel A: Propensity-matched Variables
Treated (SC) Controlled (no SC) t-stat P-value
SIZE 14.19 14.21 -0.27 0.79
LEV 0.47 0.47 -0.03 0.98
ROA 0.07 0.07 -0.46 0.65
MTB 2.52 3.06 0.75 0.45
NEW 0.55 0.55 -0.56 0.58 RMC 0.64 0.64 -0.06 0.96
CAPEXP 0.08 0.09 -1.32 0.14
CG 70.65 70.10 -0.61 0.54
SHARECON 68.77 68.73 0.04 0.97
Panel B: PSM Regression
DV: CSP
Coeff.
(t-stat)
YEAR_REPORT_DUMMY 3.864***
(8.42)
SC 5.959***
(4.48)
SIZE 2.381*** (4.45)
LEV 3.967
(1.41)
ROA 0.017
(0.91)
MTB 0.084**
(2.00)
NEW -1.464
(-0.49)
RMC 0.247
(0.28)
CAPEXP -0.100** (-2.34)
CG 0.249***
(10.12)
SHARECON 0.023
(0.71)
Constant -2.619***
(-3.06)
Observations 912
Adjusted R-square 0.62
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
148 Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance
Table 4-8 PSM – H4a This chapter uses propensity score matching (PSM) to re-test H3 and H4a, namely the relationship between experience of sustainability disclosure and sustainability performance, and relationship
between the presence of a sustainability committee and sustainability performance. Panel A presents
the results of propensity-matched variables when dependent variable is sustainability performance.
Panel B presents the regression results as to H3 and H4a. Variables in this table are defined in Table
4-2. Standard errors are corrected for clustering by firms.
Panel A: Propensity-matched
Variables
Treated (SC) Controlled (no
SC)
t-stat P-value
SIZE 14.76 14.61 1.18 0.240
YEAR_REPORT 5.192 6.018 3.16 0.020
LEV 0.45 0.47 -1.89 0.059
ROA 0.07 0.07 -0.46 0.647
MTB 2.52 3.06 0.75 0.453 NEW 0.61 0.60 1.04 0.296
RMC 0.62 0.61 0.37 0.714
CAPEXP 0.08 0.09 -1.32 0.138
CG 76.76 75.54 0.79 0.431
SHARECON 68.76 68.45 0.22 0.830
Panel B: PSM Regression
DV: CSP
Coeff.
(t-stat)
SC 7.424
(4.56)***
YEAR_REPORT 3.536
(7.08)***
SIZE 3.641
(4.78)***
LEV -12.748
(-3.40)***
ROA -0.021
(-0.51)
MTB -0.010
(-0.22)
NEW -14.264
(-3.38)***
RMC 2.783 (2.17)**
CAPEXP -0.040
(-0.47)
CG 0.324
(7.94)***
SHARECON 0.039
(0.84)
CONS -5.068
(-4.44)***
Observations 671
Adjusted R-square 0.64
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability
Performance 149
4.5.5 Robustness Test Two – Endogeneity in H3 and H4a – Dynamic GMM
Following Schultz et al. (2010), Wintoki et al. (2012) and Nadeem, Zaman, and
Saleem (2017), Chapter 4 uses dynamic GMM to control the effects of sustainability
performance in one period on performance in following period. Two hypotheses are
re-tested using dynamic GMM. Findings from dynamic GMM are tabulated below. As
shown in Table 4-9, the presence of a sustainability committee is positively related to
sustainability performance, β = 5.626 (p<0.01); experience of disclosure is positively
related to sustainability performance, β = 4.185 (p<0.01). Thus, the findings as to H3
and H4a remain qualitatively unchanged.
Table 4-9 Dynamic GMM – H3 and H4a This chapter uses dynamic generalised method-of-moments (GMM) to re-test H3 and H4a,
namely the relationship between experience of sustainability disclosure and sustainability
performance, and the relationship between the presence of a sustainability committee and sustainability performance. Variables in this table are defined in Table 4-2.
DV: CSP
Coeff.
(t-stat)
L1.CSP 0.117
(5.03)***
L2.CSP 0.001
(0.07)
SC 5.626
(3.54)***
YEAR_REPORT 4.185
(10.72)***
SIZE 2.209
(4.13)*** LEV 2.665
(1.28)
ROA -0.011
(-0.78)
MTB 0.014
(0.59)
NEW -8.013
(-2.98)***
RMC 2.034
(2.55)**
CAPEXP -0.013 (-0.55)
CG 0.201
(8.58)***
SHARECON -0.037
(-1.35)
CONS -14.641
(-1.48)
Observations 1266
No. of instruments 122
Arellano-Bond AR (1) -5.128***
Arellano-Bond AR (2) -0.644
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
150 Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance
4.5.6 Robustness Test Three – Endogeneity in H4b – Dynamic GMM
Following Schultz et al. (2010), Wintoki et al. (2012) and Nadeem, Zaman, and
Saleem (2017), Chapter 4 uses dynamic GMM to control the effects of sustainability
performance in one period on performance in the following period. H4b is re-tested
using dynamic GMM. Findings from dynamic GMM are tabulated below. As shown
in Table 4-10, the sustainability committee effectiveness is positively related to
sustainability performance, β = 7.803 (p<0.10). Thus, the finding as to H4b remains
qualitatively unchanged.
Table 4-10 Dynamic GMM – H4b This chapter uses dynamic generalised method-of-moments (GMM) to re-test H4b, namely the relationship between sustainability committee effectiveness and sustainability performance.
Variables in this table are defined in Table 4-2.
DV: CSP
Coeff.
(t-stat)
L1.CSP 0.074
(1.32)
SC_EFFE 7.803
(1.79)*
Observations 396
No. of instruments 122
Arellano-Bond AR (1) -5.04***
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
4.5.7 Robustness Test Four – Endogeneity in H4b – Two-Stage Least Squares
Regression (2SLS)
As I do not find matched samples to carry out PSM, I use a two-stage least square
(2SLS) approach to solve reverse causality and omitted variable concerns (Wooldridge
2010). As Section 4.5.1 indicates, firms in some industries are more likely to establish
sustainability committees than their peers in other industries. Thus, I expect that
sustainability committee effectiveness converges to an industry average. Restated, I
argue that sustainability committee effectiveness at firm level is influenced by year-
industry average of sustainability committee effectiveness (SC_EFFE_INDUS),
which is not directly related to sustainability performance at that firm level. Findings
rendered by the 2SLS approach are presented in Table 4-11. Panel A of Table 4-11
shows that the results of the first stage 2SLS – the endogenous variable sustainability
committee effectiveness at firm level (SC_EFFE) is significantly related to the
instrumental variable, SC_EFFE_INDUS. This supports that firm-level SC_EFFE is
closely related to its industry norms, as captured by SC_EFFE_INDUS. Panel B of
Table 4-11 shows the results of the second stage 2SLS. I find the results are consistent
Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability
Performance 151
with those reported in Table 4-6. To test the validity of my instrument, I conduct the
Wu–Hausman (Hausman, 1978; Wu 1974) test. The results are shown at the bottom of
EFFE) is exogenous – is not rejected, indicating the validity of my instrument.
Table 4-11 2SLS – H4b This chapter uses two-stage least squares regression (2SLS) to re-test H4b, namely the
relationship between sustainability committee effectiveness and sustainability performance.
Variables in this table are defined in Table 4-2.
Panel A Stage One of 2SLS
SC_EFFE Coeff.
(t-stat)
SC_EFFE_INDUS 0.76
(4.64)***
Other Controls Included YES
Year FE YES
Industry FE YES
Panel B Stage Two of 2SLS
DV: CSP
Coeff.
(t-stat)
SC_EFFE 8.55
(3.60)***
SIZE 3.83 (3.62)***
LEV 16.03
(2.20)**
ROA -0.04
(-0.56)
MTB -0.09
(-1.39)
NEW -18.15
(-3.33)***
RMC 4.45
(2.21)**
CAPEXP -0.02 (-0.01)
CG 0.32
(5.20)***
SHARECON 0.12
(1.57)
Constant -12.20
(-7.59)***
Year FE Yes
Industry FE Yes
Observations 430
Adjusted R-square 0.55 Wu–Hausman test for validity of instrument 0.177
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
152 Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance
4.5.8 Robustness Test Five – Alternative Measurement on Sustainability
Performance
Three types of performance, namely sustainability performance, environmental
performance and social performance, are used in the main analysis. Following Nguyen
et al. (2018), to further test how the presence of a sustainability committee is related
to sub-categories of sustainability performance, I test nine sub-categories separately.
The findings are presented in Table 4-12. The presence of a sustainability committee
is positively related to six sub-categories, resource reduction, emission reduction,
health and safety, product responsibility, training and development and community
(refer to Panel A of Table 4-12). Sustainability committee effectiveness is positively
related to two sub-categories, namely resource reduction and emission reduction (see
Panel B of Table 4-12). The findings presented in Table 4-10 reinforce the findings
concerning H4a and H4b.
Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance 153
Table 4-12 Alternative Measurement on Sustainability Performance This table presents the results of the relationship between a sustainability committee and nine sub-categories of sustainability performance. Panel A presents the results of
relationships between the presence of a sustainability committee and the nine sub-categories of sustainability performance. Panel B presents the results of relationships
between effectiveness of sustainability committee and nine sub-categories of sustainability performance. Resource reduction is abbreviated as ENRR, emission reduction
is abbreviated as ENER, product innovation is abbreviated as ENPI, product responsibility is abbreviated as SOPR, health and safety is abbreviated as SOHS, training
and development is abbreviated as SOTD, diversity and opportunity is abbreviated as SODO, employment quality is abbreviated as SOEQ, and community is abbreviated as SOCO. Other variables in this table are defined in Table 4-2.
Panel A The Presence of a Sustainability Committee (1) (2) (3) (4) (5) (6) (7) (8) (9)
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
154 Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance
4.5.9 Robustness Test Six – Sectoral Analysis
Following Cho and Patten (2007) and Eccles et al. (2014), Chapter 4 classifies
the GICS industries into two pairs of categories, (1) natural versus non-natural57; (2)
business-to-consumer (B2C) and others58. In terms of the first pair, this chapter codes
the value of one for industries whose products are deemed as sensitive to sustainable
development (the natural industries), and the value of zero otherwise (the non-natural
industries) (De Villiers et al., 2011). I argue that firms in natural sectors are subject to
more pressures about corporate sustainability (Cho et al., 2015); thus, the relationship
between the presence of a sustainability committee and sustainability performance
may be affected. In terms of the second pair, Chapter 4 codes the value of one for B2C
sectors and the value of zero otherwise. Following Eccles et al. (2014), I argue that
firms in B2C sectors are more visible as they directly interact with consumers. Thus,
firms in B2C sectors care more about image and reputation than their peers in other
sectors. The relationship between the presence of a sustainability committee and
performance can be influenced by such concerns. The findings are presented in Table
4-13. Panel A of Table 4-13 presents natural versus non-natural, and Panel B of Table
4-13 reports B2C versus others.
57 Following de Villiers, Naiker and Van Staden (2011, p. 1650), natural industries include forestry,
metal mining, coal mining and oil and gas exploration, paper and pulp mills, chemicals,
pharmaceutical and plastics manufacturing, iron and steel, manufacturing and electricity, gas and
wastewater. 58 Following Eccles et al. (2014, p. 2850), B2C industries include consumer goods and finance.
Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance 155
Table 4-13 Sectoral Analysis This table presents the results of sector moderation effects on the relation between the presence of a sustainability committee and sustainability performance. Panel A
shows how industry affiliation based on natural versus non-natural affects the relationship between sustainability performance and the sustainability committee. Natural is
coded as a dummy variable – firms in natural industries are coded one, and zero otherwise. Panel B shows how industry affiliation based on B2C versus others affects this
relationship. B2C is coded as a dummy variable – firms in B2C industries are coded one, and zero otherwise. Variables in this table are defined in Table 4-2.
Panel A Natural versus Non-natural
(1) (2) (3) (4) (5) (6)
CSP ENV SOC CSP ENV SOC
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
SC 0.263
(0.06)
2.523
(0.52)
-1.671
(-0.34)
-
- -
SC_EFFE - - - 3.184
(0.70)
3.618
(2.48)**
2.750
(0.21)
Natural -5.302
(-1.87)*
-6.526
(-1.93)*
-5.119
(-1.35)
46.405
(0.84)
37.217
(1.19)
5.593
(0.32)
Natural × SC 15.922
(3.29)***
13.158
(2.43)**
17.840
(3.30)***
- - -
Natural × SC_EFFE
- - - 3.837
(0.75)
3.162
(1.90)*
4.512
(0.96) CONS -35.898
(-3.82)***
-29.395
(-2.78)***
-29.881
(-2.71)***
-52.295
(-1.08)
-62.110
(-1.14)
-42.480
(-0.88)
Controls included YES YES YES YES YES YES
Firm FE YES YES YES YES YES YES
Observations 2166 2166 2166 430 430 430
Adjusted R-square 0.64 0.57 0.57 0.48 0.41 0.43
Panel B B2C versus Others
(1) (2) (3) (4) (5) (6)
CSP ENV SOC CSP ENV SOC
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
Coeff.
(t-stat)
SC 11.504
(4.23)***
12.587
(4.42)***
9.900
(3.02)***
-
- -
SC_EFFE -
- - 0.513
(1.08)
1.552
(2.10)**
-0.526
(-0.79)
156Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability
Performance 157
Regarding natural versus non-natural affiliation, this chapter identifies that the
sectoral environment of natural sectors strengthens the positive relationship between
the presence of a sustainability committee and sustainability performance, as presented
by the first three columns in the Panel A of Table 4-13. The sectoral environment
strengthens the positive relationship between sustainability committee effectiveness
and environmental performance, as presented by the fifth column in the Panel A of
Table 4-13. Regarding B2C versus other affiliation, this chapter finds that the sectoral
environment of B2C sectors affects neither the committee presence nor the committee
effectiveness, as shown by Panel B of Table 4-13. In the following section, the way in
which the sample period influences the findings regarding H4a and H4b is analysed.
4.5.10 Robustness Test Seven – Time Stability
Although Australia retained a voluntary attitude towards corporate sustainability
during the sample period, it is reasonable to argue that relevant practices have changed
within that period. As discussed in Chapter 1, awareness of corporate sustainability
increased, and available guidance about sustainability increased during the 15-year
sample period. Therefore, it is possible that findings regarding H4a and H4b may
change over time. This test selects the first five years (2002 – 2006) and the last five
years (2012 – 2016) to examine time stability. Observations from the first five years
are coded as zero, and observations of the last five years are coded as one. Findings
are presented in Table 4-14. As shown in Table 4-14, I did not find that the sample
period affects the findings regarding H4a and H4b. This chapter also re-runs the main
analyses with a sub-sample from 2010 to 2016 that consists of about 75% of data of
the presence of a sustainability committee and 70% of data of sustainability committee
effectiveness, and the findings remain largely unchanged. This chapter is concluded in
following section.
158Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability Performance
Table 4-14 Time Stability This table presents the results regarding whether time affects the relationship between sustainability performance and the presence of a sustainability committee. This chapter
selects the first five years (2002 – 2006) and the last five years (2012 – 2016) to test time stability. Time is coded as a dummy variable – observations of the first five years are
coded as zero, and observations of the last five years are coded as one. Other variables in this table are defined in Table 4.2.
(1) (2) (3) (4) (5) (6)
CSP ENV SOC CSP ENV SOC
Coeff. (t-stat)
Coeff. (t-stat)
Coeff. (t-stat)
Coeff. (t-stat)
Coeff. (t-stat)
Coeff. (t-stat)
SC 8.476
(0.91)
8.180
(0.74)
8.772
(1.01)
-
- -
SC_EFFE
- - - 2.369
(1.06)
4.565
(1.74)*
0.174
(0.06)
Time -17.464
(-5.09)***
-16.054
(-3.93)***
-18.874
(-4.66)***
34.985
(1.48)
61.641
(2.33)**
8.330
(0.26)
Time × SC 2.450
(0.25)
2.758
(0.49)
-0.859
(-0.10)
- - -
Time × SC_EFFE
- - - -1.318
(-0.64)
-3.382
(1.35)
0.746
(0.26)
CONS -20.707
(-1.40)
-17.639
(-1.03)
11.362
(0.53)
25.995
(0.61)
42.560
(0.84)
9.431
(0.19) Controls YES YES YES YES YES YES
Year FE NO NO NO NO NO NO
Firm FE YES YES YES YES YES YES
Observations 2166 2166 2166 430 430 430
Adjusted R2 0.53 0.46 0.48 0.06 0.02 0.26
***Significant at 0.01 level, ** Significant at 0.05 level, * Significant at 0.1 level
Chapter 4: The Sustainability Committee, Experience of Sustainability Disclosure and Sustainability
Performance 159
4.5.11 Robustness Test Eight – Inclusion of Firm Age as a Control for H3 and
H4a
This chapter provides additional assurance that the findings of H3 and H4a are
not caused by excluding firm age (lnAGE) as a control. I re-estimate Equation (1) to
include firm age (lnAGE)as a control. As Table 4-15 shows, the findings of H3 and
H4a remain qualitatively unchanged.59
Table 4-15 Inclusion of Firm Age as a Control for H3 and H4a This table presents the results from including firm age (lnAGE) as a control. Other variables in this
table are defined in Table 4.2.
(1) (2) (3) (4)
DV: CSP DV: CSP DV: ENV DV: SOC
Coeff. (t-stat)
Coeff. (t-stat)
Coeff. (t-stat)
Coeff. (t-stat)
SC 6.04
(4.14)***
- 5.14
(3.03)***
6.02
(3.44)***
SC_NO -
6.15
(4.44)***
- -
YEAR_REPORT
4.12
(4.95)***
4.14
(5.02)***
4.77
(5.11)***
3.41
(3.57) ***
lnAGE 3.50
(3.69) ***
3.46
(3.65)***
3.57
(3.12)***
2.80
(2.38) **
Constant and other Controls included YES YES YES YES
Firm FE YES YES YES YES
Year FE YES YES YES YES Industry FE YES YES YES YES