Journal of Contemporary Accounting & Economics: Le Luo, Qingliang Tang

JCAE 57 No.
of Pages 15, Model 3G

5 September 2014
Journal of Contemporary Accounting & Economics xxx (2014) xxx–xxx
1
Contents lists available at ScienceDirect
Journal of Contemporary
Accounting & Economics
journal homepage: www.elsevier.com/locate/jcae
5
6
3 Does voluntary carbon disclosure reflect underlying carbon

4 performance?
7 Q1 Le Luo a, Qingliang Tang b,⇑
8 a
Newcastle Business School, University of Newcastle, Sydney, NSW, Australia
9 b
School of Business, University of Western Sydney, Sydney, NSW, Australia
10
a r t i c l e i n f o a b s t r a c t
1
2 2
1
13 Article history: Carbon information is becoming more and more important in the decision making of stake- 22
14 Available online xxxx holders, but there is growing concern regarding the reliability of corporate carbon disclo- 23
sure and a lack of empirical studies addressing this issue. The purpose of this paper is to 24
15 Keywords: examine whether voluntary carbon disclosure reflects firms’ true carbon performance. 25
16 Carbon disclosure Level of carbon disclosure was measured based on content analysis of Carbon Disclosure 26
17 Carbon performance Project (CDP) reports, and our carbon performance index focused on both carbon intensity 27
18 Signalling theory
of emissions and carbon mitigation. Based on a sample of 474 U.S., U.K., and Australian 28
19 Carbon Disclosure Project
20 firms, our findings show a significant positive association between carbon disclosure and 29
performance, suggesting that firms’ voluntary carbon disclosure in the CDP is indicative 30
of their underlying actual carbon performance. This result is consistent with signalling the- 31
ory. Our findings are useful for corporate stakeholders and governmental policymakers 32
who are concerned about the quality of voluntary greenhouse gas disclosure. 33
Ó 2014 Published by Elsevier Ltd. 34
35
36
37
38 1. Introduction
39 Globally, average temperatures have already risen by approximately 0.85 °C since 1880 (Intergovernmental Panel on Cli-
40 mate Change IPCC, 2013). We will experience significant increases in surface temperatures and atmospheric CO2 concentra-
41 tions, and sea levels are projected to rise between 26 and 81 centimetres by 2100 (IPCC, 2013). It is extremely likely (95%
42 certainty) that humans are the dominant cause of warming, and without aggressive cuts to emissions, the world is on track
43 for more than 2 °C or possibly 4 °C of warming by the end of this century (IPCC, 2013).
44 Corporations are expected to play an important role in stabilising climate change, and the control of greenhouse gas
45 (GHG)1 emissions is essential for sustainable corporate development2; thus, there is a growing demand for carbon-related infor-
46 Q2 mation (Organisation for Economic Co-operation and Development [OECD], 2010; Rankin et al., 2011). Although researchers
47 recognise the importance of voluntary carbon disclosure, concern has been expressed regarding the credibility of this informa-
48 tion. Because there is inherent uncertainty associated with emissions measurement and carbon reduction activities, some
⇑ Corresponding author. Address: School of Business, University of Western Sydney, Locked Bag 1797, Penrith South DC, NSW 2751, Australia.
E-mail addresses: Laura.Luo@newcastle.edu.au (L. Luo), Q.Tang@uws.edu.au (Q. Tang).
1
This paper uses the terms carbon, carbon equivalent, and GHG emissions interchangeably.
2
Sustainability can be defined in the sense of deep-rooted social justice and fair and responsible allocation and use of ecological resources (Tinker and Gray,
2003). Another definition is found in the (International Union for the Conservation of Nature, 1980): ‘‘the maintenance of essential ecological processes and life-
support systems, the preservation of genetic diversity, and the sustainable utilisation of species and ecosystems, with the overall aim of achieving ‘sustainable
development’ through the conservation of living resources.’’
http://dx.doi.org/10.1016/j.jcae.2014.08.003
1815-5669/Ó 2014 Published by Elsevier Ltd.
Please cite this article in press as: Luo, L., Tang, Q. Does voluntary carbon disclosure reflect underlying carbon performance? Journal of
Contemporary Accounting & Economics (2014), http://dx.doi.org/10.1016/j.jcae.2014.08.003
JCAE 57 No. of Pages 15, Model 3G
5 September 2014
2 L. Luo, Q. Tang / Journal of Contemporary Accounting & Economics xxx (2014) xxx–xxx
49 authors claim that firms may take advantage of disclosure to alter the perceptions of stakeholders rather than make true efforts
50 to reduce harm to the environment (Aerts and Cormier, 2009; Gray et al., 1995; Neu et al., 1998). As a result, such disclosures
51 could be merely cosmetic or represent an attempt to deceptively promote a firm’s image or reputation, a phenomenon known as
52 ‘‘green-washing’’ (Andrew and Cortese, 2010; Frost et al., 2005; Kolk et al., 2008). However, some studies present evidence that
53 firms’ disclosures tend to reflect their commitment to ecological improvement and are thus relevant in decision making (Al-
54 Tuwaijri et al., 2004; Clarkson et al., 2008; Deegan and Gordon, 1996; Deegan and Rankin, 1996; Mitchell et al., 2006; Tilt, 2001).
55 This debate motivated us to investigate whether disclosed carbon information reflects actual underlying carbon perfor-
56 mance. A higher degree of correspondence between the two would suggest that the information is potentially useful for
57 stakeholders. In the present study, we examined how carbon information voluntarily disclosed in the Carbon Disclosure Pro-
58 ject (CDP)3 by large firms in the United States, United Kingdom, and Australia relates to their objectively measured carbon per-
59 formance. Based on signalling theory, we predicted that good performers would provide more voluntary carbon disclosure as a
60 signal to highlight their true carbon performance, which cannot be imitated by their poor counterparts. Thus, an observed posi-
61 tive association would mean that such disclosed information is indeed indicative of a firm’s true carbon position and relevant for
62 assessing the risk of carbon exposure, commitment to climate change, and achievement in reducing GHG emissions of the
63 reporting firm.
64 Our sample consisted of 474 large companies that were listed as Australian Securities Exchange 200 (ASX 200), Standard
65 & Poor’s 500 (S&P 500), and Financial Times UK Top 350 (FTSE 350) companies. These firms showed their commitment to
66 carbon transparency by participating in the CDP, 2010 annual survey. The extent of carbon disclosure was measured using
67 an index, based on each firm’s responses to the well-designed and uniform CDP questionnaire, which is standardised to a
68 0–100 scale (CDP Rating Methodology, 2010). Our carbon performance measure focused on objective, observable, and quan-
69 tifiable carbon emissions and reduction outcomes achieved by each company. Our results consistently indicate that firms
70 with better performance disclose a greater amount of overall carbon information, supporting signalling theory.
71 The present work makes several contributions to the literature. First, the study addresses the ongoing debate on the rela-
72 tionship between environmental disclosure and environmental performance (Clarkson et al., 2008, 2011b; Cowan and
73 Gadenne, 2005; Gibson and O’Donovan, 2007; Mitchell et al., 2006) and provides new empirical evidence that has not been
74 documented previously. Second, although climate change is one dimension of the environment, carbon pollution differs from
75 other types of chemical pollution in that it causes global warming, with long-term and probably irreversible harmful effects
76 (Lash and Wellington, 2007). Carbon control requires a financial investment in firm-specific capabilities and carbon activities
77 guided by different laws and regulations (Luo et al., 2012, 2013; Tang and Luo, 2014). Because environmental protection is a
78 multidimensional construct (Chatterji et al., 2009; Strike et al., 2006; Walls et al., 2011), we examine this issue in a tightly
79 focused manner by investigating only the carbon dimension. Third, we have chosen the United States, the United Kingdom,
80 and Australia as our research setting because the three countries are heavy emitters. For example, the United States is the
81 second largest emitter after China, and the United Kingdom is one of the top 10 carbon emitters in the world. Australia is a
82 small country in terms of population but has among the world’s highest per capita GHG emissions, another indicator com-
83 plementary to absolute emissions of the severity of emission problems for a country. The serious consequences of climate
84 change on the Australian ecologic system (including biodiversity, tropical rainforests, etc.) are well recognised in the liter-
85 ature (Steffen et al., 2009; Williams et al., 2009). All three nations are combating global warming but have adopted some-
86 what different approaches and programs to prevent climate change from spawning social and economic problems. For
87 instance, the United Kingdom is one of the major countries within the European Union that has unveiled various climate
88 change legislation, including an international Emissions Trading Scheme (ETS). In contrast, neither the United States nor Aus-
89 tralia4 had a national ETS during the study period. Although both Australia and the United Kingdom signed the Kyoto Protocol,
90 the United States did not. Despite this, we find consistent evidence, after controlling for the country effect, suggesting that the
91 pattern of the association is not caused by country-specific institutions. Thus, our multinational design reflects the global nature
92 of climate change and provides a useful complement to a merely national setting. Fourth, this study relies on stand-alone CDP
93 reports, which have been considered more comprehensive sources of data than alternative data sources, such as annual reports
94 or sustainability statements (Luo et al., 2012). In the absence of internationally accepted standards, the CDP has adopted a set of
95 rules that all participating firms must follow and thus significantly reduces the opportunity for managers to manipulate carbon
96 data. Finally, the paper makes a theoretical contribution by providing new evidence that enhances the validity of the application
97 of signalling theory in carbon studies. Overall, we offer additional insight regarding corporate behaviour that should be useful in
3
The CDP in its current incarnation is a system that produces assessable reports of firms’ carbon activities and emissions. The CDP encourages the
development of web-based forms of corporate accountability and has successfully used institutional investors to mobilise the world’s largest firms to disclose
carbon information (Kolk et al., 2008). An increasing number of studies have utilised CDP information in their research (Luo et al., 2012; Peters and Romi, 2009;
Reid and Toffel, 2009; Stanny, 2010; Stanny and Ely, 2008; Tang and Luo, 2011).
4
The implementation of a carbon price in Australia has undergone dramatic developments. In 2008, the Australian government, under Labour Party Prime
Minister Kevin Rudd, proposed the Carbon Pollution Reduction Scheme (CPRS). However, on 4 May 2009, the prime minister announced a 1-year delay in the
implementation of the CPRS. On 27 April 2010, the Australian government announced that the CPRS would not be reintroduced to parliament until after the
Kyoto emissions commitment period ended in 2012. On 24 February 2011, the subsequent Prime Minister, the Hon. Julia Gillard MP, announced a broad
architecture for a carbon-pricing mechanism; this carbon-pricing regime was finally approved by the Senate on 8 November 2011 and became effective on 1
July 2012. Nevertheless, after 1 year’s implementation in 2013, this carbon-pricing mechanism now faces huge uncertainty because the subsequent
administration, led by Prime Minister Tony Abbott, has sought to fulfil a pledge to repeal the carbon tax.
5 September 2014
L. Luo, Q. Tang / Journal of Contemporary Accounting & Economics xxx (2014) xxx–xxx 3
98 the debate surrounding the efficacy of voluntary/mandatory reporting requirements for carbon information and the develop-
99 ment of other public climate policies and initiatives.
100 The remainder of this paper is organised as follows. Section 2 presents a review of the literature and outlines the hypoth-
101 esis. Section 3 discusses the research design, sample selection, and data for this study. Section 4 presents the findings, and
102 Section 5 summarises our conclusions.
103 2. Literature review and hypothesis development
104 There is a considerable amount of research investigating the relationship between environmental disclosure and environ-
105 mental performance (Al-Tuwaijri et al., 2004; Cho and Patten, 2007; Clarkson et al., 2008, 2011b; Deegan and Rankin, 1996;
106 Freedman and Jaggi, 1982; Hughes et al., 2000; Ingram and Frazier, 1980; Patten, 2002; Wiseman, 1982). In recent years, the
107 focus has shifted to carbon-related issues, and many studies have examined the motivations toward voluntary carbon dis-
108 closure (Cotter and Najah, 2012; Freedman and Jaggi, 2005; Kolk et al., 2008; Liao et al., 2014; Pinkse and Kolk, 2009; Prado-
109 Lorenzo et al., 2009; Reid and Toffel, 2009; Stanny, 2010; Stanny and Ely, 2008). For example, Luo et al. (2012) found that
110 corporate carbon disclosure was correlated with social, economic, and legal/institutional pressures, and the attitude of
111 the general public and government appeared to be the decisive determinants. In addition, Luo et al. (2013) showed a system-
112 atic difference in firms’ propensity for carbon disclosure between developing and developed countries, and financial
113 resources were likely the major constraint. Another important stream of carbon studies involves investigations of the asso-
114 ciation between a firm’s carbon performance (emissions) and its valuation (Chapple et al., 2013; Griffin et al., 2012; He et al.,
115 2013; Krishnan, 2003; Luo and Tang, 2014).
116 Few attempts have been made to examine the relationship between carbon disclosure and performance. Freedman and
117 Jaggi (2004) examined whether adequate disclosures were made by utility companies and provided weak support for a posi-
118 tive association between pollution disclosure and pollution emissions in a U.S. context. Then Freedman and Jaggi (2010)
119 extended the sample to include the European Union, Japan, and Canada. Carbon disclosure was measured based on content
120 analysis involving 148 companies. The results demonstrated no significant relationship between GHG performance and dis-
121 closure. Freedman and Jaggi (2011) reported similar insignificant results for 510 firms in various industries across Europe,
122 Japan, Canada, India, and the United States.
123 Kim and Lyon (2011) examined how U.S. firms reported their carbon reductions to the voluntary GHG Registry Program
124 developed by the Department of Energy. Specifically, they compared reported reductions in this voluntary program to actual
125 emissions, which were based on Federal Energy Regulatory Commission data. The sample consisted of 98 investor-owned
126 U.S. electric utilities during the period 1995–2003. The results showed that participants in the Department of Energy’s vol-
127 untary program actually increased emissions over time but reported reductions, whereas nonparticipants decreased their
128 emissions over time, indicating that participants engaged in selective reporting of successful emissions reduction projects.
129 Overall, prior studies have yielded conflicting results and failed to systematically theorise the carbon disclosure/performance
130 relationship.
131 The previous literature has traditionally appealed to one of two conceptual frameworks: signalling theory (also referred
132 to as voluntary disclosure theory) or legitimacy theory (also referred to as socio-political theory). Legitimacy theory posits
133 that an organisation is a member of society, which provides the organisation with scarce resources. Thus, the organisation is
134 expected to operate in a way that meets the expectations of the community; this may mean that the company must do more
135 than merely comply with laws and regulations. If companies fail to behave in accordance with social ethical standards, they
136 will face threats to their legitimacy. Hence, firms with poor environmental records tend to make increased disclosures, and
137 thus there is an inverse relationship between environmental performance and voluntary disclosure (Deegan, 2002; Gray
138 et al., 1995; Patten, 2002). Some authors have pointed out, however, that this disclosure is likely intended to create an
139 impression of sensitivity of the firm to important nonmarket influences to enhance the perception of compliance with social
140 expectations rather than to make actual changes to strategies, adopt energy-efficient processes, or produce environmentally
141 friendly products (Dowling and Pfeffer, 1975).
142 However, Belkaoui and Karpik (1989) argued that when a firm truly engages in socially responsible activities involving an
143 outlay of resources that is consistent with the long-term interests of its shareholders, managers, naturally, are eager to reveal
144 this information to interest groups and the public. One practical way to advertise this news is through some form of social
145 disclosure. Thus, signalling theory predicts that firms with better carbon performance will tend to voluntarily disclose more
146 to distinguish themselves from poor performers to avoid the problem of adverse selection.
147 Previous studies have provided evidence showing that there is a positive relationship between carbon performance and
148 financial performance/market value (Al-Tuwaijri et al., 2004; Clarkson et al., 2011a; Klassen and McLaughlin, 1996;
149 Krishnan, 2003; Matsumura et al., 2014; Schaltegger and Figge, 1997). Thus, firms that have achieved superior carbon per-
150 formance via proactive carbon strategies and substantial green investments would have strong incentives to inform inves-
151 tors of that good news through voluntarily revealing more carbon information. However, there is information asymmetry
152 between managers and investors or other stakeholders. Managers have private information about firms’ carbon profile,
153 including carbon strategy, carbon emissions, and carbon reduction activities, that is not directly accessible by outsiders.
154 For the poor performers, silence and maintaining an opaque carbon profile are perhaps the best strategies. However,
155 high-achieving performers tend to believe they deserve to be valued more, so they must send a message to avoid being trea-
5 September 2014
156 ted as a poor performer by stakeholders. In this regard, the CDP would be used as an effective communication tool to convey
157 the good news. This is because a CDP report is a stand-alone carbon statement that requires the company to provide com-
158 prehensive and specific carbon information. Because the CDP report is completed and presented in a standard format, it is
159 difficult for poor performers to mimic high achievers (Tang and Luo, 2014). It is noteworthy that investing in this signal (vol-
160 untary carbon disclosure) is not cost neutral. These costs include not only the costs of creating an emissions inventory and
161 emission footprint record and establishing a carbon management system (Leighton Holdings CDP Report, 2010; Lend Lease
162 CDP Report, 2010; Tang and Luo, 2014) but also the proprietary costs derived from disclosing information that may be used
163 by competitors and other parties in a way that is harmful for the reporting company (Verrecchia, 1983).5 Generally speaking,
164 good performers are assumed to incur fewer costs to produce the same level of carbon disclosure than poor performers, all other
165 factors being equal. This is because when good performers have created an effective carbon management system to control
166 emissions, the system already includes adequate carbon accounting and reporting elements (Tang and Luo, 2014). In contrast,
167 poor performers are likely to have an inadequate carbon reporting procedure, which prohibits them from disclosing sufficient
168 information to meet the requirements of CDP surveys. Thus, the cost of reporting carbon information is expected to be lower for
169 the good performers than it is for the bad ones. This differing cost structure ensures that it is both difficult and costly for poor
170 performers to mimic the signalling behaviour of good performers, so that the good performers are readily distinguished from
171 their counterparts, thereby offering higher firm valuation to investors.
172 Based on the above argument, we expect that there is a positive relationship between carbon disclosure and performance.
173 Our hypothesis is stated as follows:
174 H1. Good carbon performers disclose more overall carbon information than poor performers.
175 3. Research design
176 3.1. The measurement of carbon disclosure
177 A variety of methods have been implemented to measure general environmental disclosure. Some studies have quantified
178 the level of environmental disclosure using number of pages, sentences, and words (e.g. Deegan and Rankin, 1996), whereas
179 others have conducted analyses of content in publicly available reports according to preselected criteria (e.g. Cormier et al.,
180 2005; Cowan and Gadenne, 2005). Wiseman (1982) created an environmental disclosure index based on scores. They
181 assigned 3 (2) for quantitative (non-quantitative) disclosure and 1 (0) for a general mention (no disclosure). The Wiseman
182 index or its modification was used in subsequent studies (Bewley and Li, 2000; Hughes et al., 2001). Al-Tuwaijri et al. (2004)
183 analysed environmental information in four categories, namely, potential responsible parties’ designation, toxic waste, oil
184 and chemical spills, and environmental fines and penalties. Note that these disclosures are largely nondiscretionary. In addi-
185 tion, Clarkson et al. (2008, 2011b, 2013) developed another disclosure index based on the Global Reporting Initiative (GRI)
186 sustainability reporting guidelines. The index (including 95 equally weighted items) appears more comprehensive than the
187 methods used in previous studies.
188 As the focus of our study is carbon, we used the Carbon Disclosure Leaders Index (CDLI) to measure the extent or level of
189 disclosure based on content analysis. Most of the questions in the CDP questionnaire are binary questions. The participant
190 simply gives a ‘‘yes’’ (1 point awarded) or ‘‘no’’ (0 points awarded) answer to provide factual information, such as the exis-
191 tence of an environmental committee in the firm.
192 Other questions require qualitative or narrative answers, which are scored using content analysis with a set of standard-
193 ised scoring methodologies (CDP Scoring Methodology, 2010). Different types of predetermined scales (i.e., weights) are
194 applied to the responses by considering information quality in terms of ‘‘(a) the information details and relevance to the
195 company, (b) examples or case studies provided, and (c) quantitative or financial information’’ (Cotter and Najah, 2012,
196 p.177). The importance and materiality of specific climate change information to particular users are also considered. For
197 example, the disclosure of gross Scope 1 GHG emissions6 data is given 6 points, because this information is expected to be
198 highly important and relevant to investors and regulators. The CDLI for a firm is calculated using the total attainable score
199 divided by the total available score, and it increases with the level and comprehensiveness of carbon disclosure. A firm must
200 answer all applicable questions, and answers left blank reduce the awarded score. For example, if a firm states that it has a car-
201 bon reduction target, it must explain whether it has met the target; otherwise, it will lose 1 point. If a firm does not have a
202 reduction target, whether it achieves the target is irrelevant or inapplicable, so it will not answer this second question. The
5
Carbon emission information is proprietary because stakeholders could use it to impose costs on high emitters. For example, policymakers could use such
information as a pretext for investigations that would increase the costs of compliance. Moreover, disclosure could invite costly litigation by previously
uninformed victims of environmental incidents, affect the availability of debt and equity capital, benefit competitors’ green marketing strategies aimed at
environmentally conscious consumers, and provide ammunition for environmental protection groups to press for stricter legislation or boycotts of the
company’s products (Li et al., 1997).
6
Scope 1 emissions are direct emissions that occur onsite or are from sources that a company owns and controls. Scope 2 emissions are indirect emissions
that result from the generation of electricity, heat, or steam that a company purchases. Scope 3 emissions include all indirect emissions other than those
represented by Scope 2 emissions.
5 September 2014
203 available score does not include any inapplicable questions; therefore, there is no penalty for not answering inapplicable
204 questions.7
205 The CDLI provides insights into the characteristics, quality, and common trends of the leading companies in the world
206 regarding GHG footprints, carbon governance, and emissions-reducing activities. Overall, the CDLI has the following advan-
207 tages: first, because the CDLI methodology was designed and developed by the CDP with guidance from PricewaterhouseCo-
208 opers in its capacity as global advisor and expert, it appears superior to a self-designed methodology (Cotter and Najah,
209 2012). Second, the CDP questionnaire covers many different aspects of climate change issues, including carbon control mech-
210 anisms, carbon strategies, carbon accounting and auditing, carbon initiatives, carbon risks and opportunities, and carbon
211 communication and engagement. Thus, CDLI scores reflect both the depth and breadth of corporate carbon information.
212 Cotter and Najah (2012) pointed out that in their annual and sustainability reports, ‘‘companies tend to disclose a relatively
213 small portion of the information that they provide to the CDP’’ (p. 182). Third, the content analysis method utilised by the
214 CDLI methodology facilitates analysing and interpreting the relevance, importance, and substance of disclosures rather than
215 simply counting the amount and extent of disclosed information (Al-Tuwaijri et al., 2004; Ingram and Frazier, 1980). Fourth,
216 CDLI scores have gained widespread support. For instance, Google added CDP scores to its ‘‘Key Stats and Ratios’’ section of
217 Google Finance in April 2010. An increasing number of academic studies have applied CDLI scores in their research (e.g.
218 Griffin et al., 2012; Luo and Tang, 2014; Prado-Lorenzo and Garcia-Sanchez, 2010; Tang and Luo, 2011, 2014). Despite the
219 fact that the CDP leverages institutional investors to encourage companies to disclose carbon information, and the use of
220 existing methodology ensures external validity and consistency with prior studies (Cotter and Najah, 2012), it is recognised
221 that CDP data are provided voluntarily. Thus, it is possible that managers may use the CDP disclosure for legitimation pur-
222 poses rather than to truly and accurately reveal their carbon activities and outcomes. However, the format of the CDP sig-
223 nificantly reduces the chances of green-washing. When the CDP, on behalf of its signatory institutional investors, invites a
224 company to participate, the company does not have a statutory obligation to accept the offer. Therefore, participation in the
225 CDP is entirely optional. Nevertheless, if the company accepts, it must complete a standard questionnaire, which means that
226 the company cannot choose what to disclose and what not to disclose. In other words, CDP reporting is voluntary but not
227 entirely discretionary in the means by which managers make disclosures (Luo et al., 2012). If managers deliberately or neg-
228 ligently omit or provide misleading information, the responsible executives may be held liable if the disclosures are subse-
229 quently found to be incorrect.
230 In addition to total CDLI (CD1), we used sector-adjusted CDLI (CD2) to measure carbon disclosure. CD2 is calculated as a
231 firm’s CD1 minus its sector mean. Sector mean was calculated per the two-digit Global Industry Classification Standard
232 (GICS). The formula is
233
N
1X j
CD2i ¼ CD1i CD1i

235 Nj i¼0
236 where Nj is the total number of firm observations in sector j and j = energy, materials, industrials, consumer discretionary,
237 consumer staples, financials, health care, information technology, telecommunications, or utilities. Both CD1 and CD2
238 increase with the level of carbon disclosure, but CD2 uses the sector mean as a benchmark.
239 3.2. The measurement of carbon performance
240 Prior studies have typically used a wide range of data, such as waste management; toxic emissions to air, land, and water;
241 and the existence of environmental management systems to measure environmental performance (e.g. Johnston et al., 2008;
242 Klassen and McLaughlin, 1996; Sutantoputra et al., 2012).8 More specifically, most early studies relied on rankings from the
243 Council on Economic Priorities (CEP). However, CEP data only cover a small number of U.S. firms in highly polluting industries
244 such as steel, oil, paper, and electric utilities (e.g. Freedman and Wasley, 1990; Shane and Spicer, 1983; Wiseman, 1982) and
245 apply inconsistent methodology across industries (Patten, 2002). Thus, more recent U.S. studies have employed chemical emis-
246 sion data as a proxy (Al-Tuwaijri et al., 2004; Clarkson et al., 2008; Patten, 2002). Similarly, Canadian and Australian studies
247 have sourced data from National Pollution Release Inventory (NPRI) and National Pollutant Inventory (NPI) databases to quan-
248 titatively measure environmental performance (Bewley and Li, 2000; Clarkson et al., 2011b).
249 Because public concern about climate change has focused on the release of CO2 emissions into the atmosphere, the level
250 of carbon emissions would seem to be an appropriate and objective measure of carbon performance. We constructed four
251 proxies: CP1, CP2, CP3, and CP4. Whereas CP1 and CP2 represent carbon emission performance, CP3 and CP4 measure carbon
252 mitigation performance. CP1 refers to total CO2 emission intensity (emissions relative to economic output) and is measured
253 as the ratio of total Scope 1 and Scope 2 GHG emissions to total sales (Clarkson et al., 2008; Patten, 2002; Sutantoputra et al.,
254 2012). Relative to absolute emissions, this intensity measure considers the variation in the output of products and services
7
CDP 2010 scoring methodologies provide different routes for each question. Companies need to identify the most appropriate route for them and answer all
questions within that route.
8
Various organisations or databases have been sourced to provide information on firms’ social or environmental performance, such as the Council on
Economic Priorities (CEP), the Toxics Release Inventory (TRI), the KLD Research and Analytics (KLD), the National Pollution Release Inventory (NPRI), and the
National Pollutant Inventory (NPI).
5 September 2014
255 and thus is more comparable across firms and between different reporting periods (Hoffmann and Busch (2008). CP2 is sec-
256 tor-adjusted CO2 emission intensity, calculated using firms’ total emission intensity (CP1) minus their sector mean. The for-
257 mula is
258
N
1X j
CP2i ¼ CP1i CP1i ;

260 N i¼0
261 where Nj is the total number of firm observations in sector j and j = energy, materials, industrials, consumer discretionary,
262 consumer staples, financials, health care, information technology, telecommunications, or utilities. A higher value for CP1
263 or CP2 suggests that a firm uses its resources, particularly energy, inefficiently and is therefore a poor performer.
264 CP3 is an index that measures whether a firm lowered its emissions relative to historical levels or other benchmarks (Tang
265 and Luo, 2014). The value of the index is determined on the basis of four subitems. First, 3 points are awarded if a firm’s
266 carbon intensity of emissions (CP1) was lower than in the previous year. This item objectively and directly captures the most
267 desirable mitigation outcome and thus is given the highest weight. Second, 2 points are awarded if a firm’s emissions inten-
268 sity (CP1) was lower than the median of its two-digit GICS sector, and another 2 points are given if the firm realised a carbon
269 reduction from at least one of its carbon reduction projects or activities.9 Finally, a firm earns 1 point if it met at least one of its
270 carbon reduction targets.10 Items 2, 3, and 4 each carry less weight than the first because they may not necessarily guarantee a
271 decrease in emissions. For example, a firm’s emissions may still increase from the previous year even if its emissions are lower
272 than those of its sector counterparts or if it has achieved one of its targets. In addition, the lowest weight is appropriate for item
273 4 because this item is less objective than the others, as the target is chosen by the firm itself. For instance, a firm may set a low
274 target that is easy to achieve. In sum, the value of the variable CP3 can range from 0 to 8 based on the points awarded.
275 However, these assigned weights are arguably arbitrary. Thus, we also employed an additional proxy (CP4) that is based
276 on an equally weighted index. A higher value for CP3 or CP4 indicates better performance in terms of carbon reduction.
277 Carbon performance appears to be a complex and multidimensional concept. Hence, it is necessary to consider more than
278 one aspect of carbon performance. The adoption of a multifaceted approach will show the relative nature of the measure-
279 ment of carbon performance. Overall, if a firm has attempted to control its emissions, the results will be reflected in CP1,
280 CP2, CP3, or CP4, or all of these proxies. Following previous studies, carbon emission data were retrieved from the CDP data-
281 base, which provides relatively consistent, complete, and comparable information based on a set of well-recognised methods
282 (Chapple et al., 2013; Griffin et al., 2012; Krishnan, 2003; Luo and Tang, 2014; Luo et al., 2013; Tang and Luo, 2014).
283 3.3. Empirical models
284 The following four ordinary least squares (OLS) regression models are used in the paper.
285
CD1i ¼ a0 þ a1 CP1i þ a2 SIZEi þ a3 ROAi þ a4 LEV i þ a5 UK i þ a6 USAi þ Sector Effect ð1Þ
CD2i ¼ a0 þ a1 CP2i þ a2 SIZEi þ a3 ROAi þ a4 LEV i þ a5 UK i þ a6 USAi þ a7 INTENSIVE ð2Þ
CD2i ¼ a0 þ a1 CP3i þ a2 SIZEi þ a3 ROAi þ a4 LEV i þ a5 UK i þ a6 USAi þ a7 INTENSIVE ð3Þ
287 CD2i ¼ a0 þ a1 CP4i þ a2 SIZEi þ a3 ROAi þ a4 LEV i þ a5 UK i þ a6 USAi þ a7 INTENSIVE ð4Þ
288 All of the left- and right-hand variables for disclosure and performance were described in the previous section, and i denotes
289 the individual company. Model (1) tests whether low-carbon-intensity emissions firms (i.e., high performers) tend to dis-
290 close more, so a1 is expected to be negative. Model (2) tests whether firms with low-carbon-intensity emissions relative
291 to the sector average tend to disclose more relative to the sector average. In other words, model (1) tests the association
292 between disclosure and performance in the entire sample, whereas model (2) tests the association at the sector level.
293 Because each sector may have different carbon regulations and institutions, the association patterns might differ across sec-
294 tors. Models (3) and (4) are similar to model (2), but CP3 and CP4 replace CP1 and CP2 as proxies for carbon performance. We
295 used these four models because it allowed us to determine whether the association between disclosure and performance is
296 sensitive to alternative measures of disclosure and performance and to potential influences that are sector specific. There-
297 fore, the four models are not duplicates and should not be used interchangeably.
298 The selection of control variables was guided by prior studies. We first controlled for firm size (SIZE, calculated as the nat-
299 ural logarithm of total assets) because large firms are under greater public scrutiny and are expected to be more transparent
300 (Freedman and Jaggi, 2005; Luo et al., 2012).11 Next, profitability was controlled (ROA, return on assets), as many prior studies
301 have reported a positive relationship between profitability and voluntary disclosure (Bewley and Li, 2000; Cormier and Magnan,
302 1999; Li and McConomy, 1999; Luo et al., 2012). We also controlled for leverage (LEV), as debt holders with large claims are
303 likely to demand carbon information to help them evaluate the risk associated with their debt contracts (Clarkson et al.,
304 2008; Dhaliwal et al., 2011). LEV was calculated as total debt divided by total assets. All of the financial data were obtained from
9
The data were obtained from firms’ responses to question 9.7 in the CDP 2010 questionnaire, which describes the actions undertaken to reduce GHG
emissions, including the nature of the actions, annual energy savings, annual emissions reduction, and reduction achieved or anticipated.
10
The data were obtained from responses to question 9.2 in the CDP 2010 questionnaire, which requires companies to disclose the type of target, the target
amount, and whether the target was met.
11
We used lagging data for SIZE, ROA, and LEV because the CDP 2010 questionnaire requests carbon-related data for the year ending 2009.
5 September 2014
Table 1
Sample distribution.
GICS code Sector Frequency Percentage (%)

Panel A: Distribution by industry
10 Energy 33 6.96
15 Materials 43 9.07
20 Industrials 62 13.08
25 Consumer Discretionary 68 14.35
30 Consumer Staples 51 10.76
35 Health Care 31 6.54
40 Financials 91 19.2
45 Information Technology 47 9.92
50 Telecommunications 9 1.9
55 Utilities 39 8.23
Total 474 100
Country Frequency Percentage (%)
Panel B: Distribution by country
ASX 200 65 13.71
FTSE 350 159 33.54
S&P 500 250 52.74
Total 474 100
GICS = Global Industry Classification Standard.
305 DataStream in millions of U.S. dollars. In addition, firms operating in carbon-emission-intensive sectors are subject to more cli-
306 mate change legislation and regulations and their carbon disclosure tends to be more open; thus, we controlled for the sector
307 fixed effect. The variable INTENSIVE was assigned a value of 1 if a firm operated in a carbon-intensive sector (materials, energy,
308 or utilities) and 0 otherwise. The country of origin was also considered in the previous literature to be an important determinant
309 of the level and type of corporate social and environmental disclosure (Luo et al., 2012, 2013; Meek et al., 1995; Van der Laan
310 Smith et al., 2005). In our context, the three surveyed countries are likely to differ in their climate change laws and carbon insti-
311 tutions, which would potentially have different effects on disclosure. Thus, two country dummy variables (UK and US) were
312 included in the equations to control for the country effect. In model (1), CD1 and CP1 represent total carbon disclosures and
313 total carbon intensity of emissions; thus, we included nine sector dummy variables to control for sector effects, whereas in
314 models (2)–(4), CD2, CP2, CP3, and CP4 incorporate adjustments for sector effect and thus the models include only one dummy
315 variable, INTENSIVE.
316 3.4. Sample selection
317 Companies from the United States, the United Kingdom, and Australia were selected based on the following criteria:
318 inclusion in the list of the CDP S&P 500, FTSE 350, and ASX 200 reports in 2010;
319 published carbon disclosure score for 2010;
320 published carbon emissions data in both 2009 and 2010; and
321 availability of complete data for the other independent variables (SIZE, ROA, LEV).
322
323 Our initial sample included 1050 U.S., U.K., and Australian companies that received a 2010 participation request from the
324 CDP (See company list in CDP Australian & New Zealand Report, 2010, pp. 39–45; CDP FTSE 350 Report, 2010, pp. 42–52; CDP
325 S&P 500 Report, 2010, pp. 39–49). The final sample comprised a total of 474 observations that met all the selection criteria.
326 Table 1, Panel A, shows the sector distribution of the sample. The financials, consumer discretionary, and industrial sectors
327 made up the largest proportion of firms (19.2%, 14.35%, and 13.08%, respectively), whereas the telecommunications, health
328 care, and energy sectors made up the smallest proportions (1.9%, 6.54%, and 6.96%, respectively). Table 1, Panel B, presents
329 the sample distribution by country. S&P 500 companies accounted for a majority of our sample (52.74%), followed by FTSE
330 350 companies (33.54%) and ASX 200 companies (13.71%).
331 4. Empirical results
332 4.1. Descriptive statistics
333 Table 2 reports the key descriptive statistics for the sample. The average and median carbon disclosure scores (CD1) are
334 65.286 and 66, respectively. The mean CP1 value is 432.643, showing that our sample firms emit approximately 433 tonnes
335 GHG per million dollars of sales on average. The mean CP3 is 4.513 out of a maximum score of 8. The distribution of carbon
5 September 2014
Table 2
Descriptive analysis of complete sample (N = 474).
Variable Mean Standard deviation Median Min Max

CD1 65.286 15.904 66 11 96
CD2 0 15.768 1.254 56.262 32.818
CP1 432.643 1154.114 46.129 0.661 7616.73
CP2 0 892.2 32.821 2742.48 7379.336
CP3 4.513 1.977 5 0 8
CP4 2 0.917 2 0 4
SIZE 9.421 1.689 9.284 5.961 14.583
ROA 0.058 0.056 0.051 0.106 0.239
LEV 0.241 0.157 0.222 0 0.699
INTENSIVE 0.243 0.429 0 0 1
Notes: N represents the number of observations. All variables were Winsorised at the 1st and 99th percentiles, with the exception of the indicator variables.
Financial data are in millions of U.S. dollars. CD1 is a proxy for the total level of carbon disclosure and is scored using an index developed by the CDP. CD2 is a
sector-adjusted carbon disclosure proxy. CP1, CP2, CP3, and CP4 are four different proxies for carbon performance, with CP1 and CP2 based on the carbon
intensity of emissions and CP3 and CP4 focusing on carbon reduction outcomes. SIZE is measured as the natural logarithm of market capitalisation at the end
of fiscal year 2009. ROA is return on assets, which is used as a proxy for a firm’s profitability and measured as net income before extraordinary items/
preferred dividends divided by total assets. LEV is calculated as total debts divided by total assets at the end of fiscal year 2009. INTENSIVE is a dummy
variable that equals 1 if a firm belongs to the energy, materials, or utility sectors and 0 otherwise.
336 performance (CP3) in our sample (not shown) is as follows: 17 of the 474 firms (3.59%) had 0 points for CP3; 4 firms (0.84%)
337 had 1 point; 76 firms (16.03%) had 2 points; 49 firms (10.34%) had 3 points; 43 firms (9.07%) had 4 points; 172 firms (36.29%)
338 had 5 points; 16 firms (3.38%) and 68 firms (14.35%) had 6 and 7 points, respectively; and 29 firms (6.12%) had 8 points. The
339 mean CP4 value is 2, and the distribution is as follows: 17 of 474 firms (3.59%) had 0 points for CP4, 119 firms (25.11%) had 1
340 point, 214 firms (45.15%) had 2 points, 95 firms (20.04%) had 3 points, and 29 firms (6.12%) had 4 points. Table 2 also pre-
341 sents the mean value of firm size proxy, which is 9.421 (equivalent of U.S. $12 billion), showing that our sample included
342 relatively larger firms. The average values of LEV and ROA are 24.1% and 5.8%, respectively, which are comparable with pre-
343 vious studies (Luo et al., 2012, 2013).
Table 3
Descriptive analysis by sector and country.
GICS code GICS sector CD1 CD2 CP1 CP2 CP3 CP4 SIZE ROA LEV
Panel A: Descriptive analysis by GICS sector
10 Energy 61.182 0 850.515 0 3.909 1.667 9.705 0.048 0.183
15 Materials 66.512 0 632.376 0 4.535 1.953 8.957 0.062 0.223
20 Industrials 64.808 0 175.973 0 4.823 2.161 8.909 0.048 0.266
25 Consumer Discretionary 63.478 0 198.103 0 4.588 2.015 8.527 0.073 0.245
30 Consumer Staples 67.143 0 237.395 0 5.078 2.314 9.423 0.081 0.312
35 Health Care 64.015 0 27.561 0 3.871 1.71 9.635 0.077 0.207
40 Financials 67.218 0 34.221 0 4.341 1.945 10.51 0.03 0.217
45 Information Technology 62.872 0 38.961 0 4.83 2.149 8.77 0.098 0.121
50 Telecommunications 69.778 0 60.352 0 3.556 1.444 10.665 0.036 0.41
55 Utilities 67.262 0 2743.137 0 4.385 1.949 9.847 0.03 0.366
Total 65.286 0 432.643 0 4.513 2 9.421 0.058 0.241
Variable ASX 200 (N = 65) FTSE 350 (N = 159) S&P 500 (N = 250)
Panel B: Descriptive analysis by country
CD1 71.729 63.447 64.78
CD2 6.1 1.878 0.392
CP1 390.062 269.204 547.662
CP2 23.072 72.158 51.891
CP3 4.231 4.409 4.652
CP4 1.846 1.981 2.052
SIZE 8.896 8.765 9.975
ROA 0.037 0.058 0.064
LEV 0.214 0.247 0.245
INTENSIVE 0.354 0.182 0.252
Notes: The table presents the mean values of all variables. N represents the number of observations. GICS = Global Industry Classification Standard. All
variables were Winsorised at the 1st and 99th percentiles, with the exception of the indicator variables. Financial data are in millions of U.S. dollars. CD1 is a
proxy for the total level of carbon disclosure and is scored using an index developed by the CDP. CD2 is a sector-adjusted carbon disclosure proxy. CP1, CP2,
CP3, and CP4 are four different proxies for carbon performance, with CP1 and CP2 based on the carbon intensity of emissions and CP3 and CP4 focusing on
carbon reduction outcomes. SIZE is measured as the natural logarithm of market capitalisation at the end of fiscal year 2009. ROA is return on assets, which
is used as a proxy for a firm’s profitability and is measured as net income before extraordinary items/preferred dividends divided by total assets. LEV is
calculated as total debts divided by total assets at the end of fiscal year 2009. INTENSIVE is a dummy variable that equals 1 if a firm belongs to the energy,
materials, or utility sectors and 0 otherwise.
5 September 2014
344 Table 3 Panels A and B, reports the averages of all the variables by GICS sector and country. Panel A shows that all GICS
345 sectors have an average total carbon disclosure in excess of 60 out of 100, with the telecommunications sector having the
346 highest and the energy sector the lowest. In addition, utilities, energy, and materials have the highest total carbon intensity
347 of emissions. These three sectors are classified as carbon-intensive sectors, so these firms will inherently emit more than
348 other sectors. In contrast, the consumer staples and information technology sectors have the highest carbon mitigation per-
349 formance, suggesting that they have made efforts to reduce emissions, whereas the telecommunications, health care, and
350 energy sectors lag slightly behind in carbon mitigation activities. In Panel B of Table 3, one-way analysis of variance (ANOVA)
351 tests show that there is a significant variation in carbon disclosure (CD1 and CD2) among the three countries. Specifically,
352 Bonferroni multiple comparison tests indicate that the means of CD1 and CD2 for the ASX 200 companies are significantly
353 higher than the means for the FTSE 350 and S&P 500 companies, suggesting that Australian firms are more transparent than
354 their U.S. and U.K. counterparts. Regarding carbon performance, a higher value of CP1 or CP2 and a lower value of CP3 or CP4
355 suggest lower performance. Panel B in Table 3 shows insignificant differences in CP3 and CP4 between the U.S., the U.K., and
356 Australia. However, CP1 and CP2 for U.K. and Australian firms appear significantly lower than for U.S. firms, implying that the
357 ASX 200 and FTSE 350 firms have carbon performance that is superior to that of U.S. firms. In addition, the S&P 500 group is
358 larger and more profitable than its FTSE 350 and ASX 200 counterparts.
359 4.2. Univariate analysis
360 Table 4 reports both parametric and nonparametric correlation coefficients for the variables used in our tests. CD1 and
361 CD2 are significantly correlated. This is expected, because the firms with a high score for disclosure in the whole sample
362 are likely to be the firms with a high score in their respective sectors. However, the coefficients between proxies of carbon
363 disclosure and carbon performance vary significantly. CP3 and CP4 are consistently and positively related to both CD1 and
364 CD2, suggesting that good carbon mitigation performance increases carbon disclosure. However, only the Pearson correlation
365 coefficients between CP2 and CD1 (CD2) are significant and negative, as expected. In sum, the correlation matrix shows that
366 the values of any pairs of independent variables are well below the critical value of 0.8, thus indicating no evidence of
367 multicollinearity.
368 4.3. Multivariate analysis
369 4.3.1. Regression results for the pooled sample

370 Table 5 presents the OLS regression results. The coefficient for CP1 in model (1) is 0.00175 and is significant at the
371 p < 0.05 level, suggesting that a firm with a lower level of total carbon emissions tends to disclose more comprehensive car-
372 bon information. CP2 measures a firm’s carbon emissions relative to its sector counterparts. The coefficient of CP2 in model
373 (2) is negative and significant (0.00182, p < 0.05), indicating that good carbon emission performers within the sector have a
374 greater incentive to report increased carbon information than poor performers in the same sector. In other words, the result
375 at the sector level using model (2) reinforced the result found with model (1). In addition, model (3) used CP3 and model (4)
Table 4
Q5 Correlation matrix (N = 474).
Variable CD1 CD2 CP1 CP2 CP3 CP4 SIZE ROA LEV INTENSIVE
CD1 1.000 0.991*** 0.067 0.024 0.138*** 0.228*** 0.220*** 0.027 0.112** 0.005
CD2 0.992*** 1.000 0.069 0.016 0.136*** 0.225*** 0.199*** 0.001 0.078* 0.008
CP1 0.076 0.091** 1.000 0.204*** 0.139*** 0.172*** 0.028 0.016 0.314*** 0.633***
CP2 0.117** 0.118** 0.773*** 1.000 0.300*** 0.333*** 0.097** 0.113** 0.030 0.144***
CP3 0.161*** 0.161*** 0.175*** 0.203*** 1.000 0.947*** 0.083* 0.003 0.030 0.061
CP4 0.238*** 0.237*** 0.194*** 0.223*** 0.954*** 1.000 0.099** 0.011 0.017 0.095**
SIZE 0.234*** 0.204*** 0.001 0.038 0.095** 0.122*** 1.000 0.202*** 0.047 0.057
ROA 0.017 0.009 0.111** 0.022 2–0.0093 0.003 0.175*** 1.000 0.114** 0.090*
LEV 0.075 0.042 0.157*** 0.021 0.022 0.017 0.027 0.123*** 1.000 0.087*
INTENSIVE 0.002 0.000 0.480*** 0.000 0.060 0.081* 0.018 0.113** 0.067 1.000
CD1 is a proxy for the total level of carbon disclosure and is scored using an index developed by the CDP. CD2 is a sector-adjusted carbon disclosure proxy.
CP1, CP2, CP3, and CP4 are four different proxies for carbon performance, with CP1 and CP2 based on the carbon intensity of emissions and CP3 and CP4
focusing on carbon reduction outcomes. SIZE is measured as the natural logarithm of market capitalisation at the end of fiscal year 2009. ROA is return on
assets, which is used as a proxy for a firm’s profitability and is measured as net income before extraordinary items/preferred dividends divided by total
assets. LEV is calculated as total debts divided by total assets at the end of fiscal year 2009. INTENSIVE is a dummy variable that equals 1 if a firm belongs to
the energy, materials, or utility sectors and 0 otherwise.
Pearson (Spearman) correlation coefficients are below (above) the diagonal. All variables were Winsorised at the 1st and 99th percentiles, with the
exception of the indicator variables. Financial data are in millions of U.S. dollars.
***
Represent significance (two-tailed) at the 1% level.
**
*
5 September 2014
Table 5
Regression results for the pooled sample.
Variable (1) (2) (3) (4)

CD1 CD2 CD2 CD2
CP1 0.00175**
(2.239)
CP2 0.00182**
(2.331)
CP3 1.229***
(3.495)
CP4 3.871***
(5.166)
SIZE 3.007*** 2.343*** 2.307*** 2.167***
(5.992) (5.193) (5.150) (4.894)
ROA 17.08 25.20* 27.06** 26.04**
(1.234) (1.915) (2.072) (2.024)
LEV 7.716 6.418 6.621 6.639
(1.583) (1.436) (1.492) (1.518)
UK 9.385*** 8.615*** 8.732*** 8.982***
(4.107) (3.808) (3.887) (4.056)
US 11.73*** 9.824*** 10.46*** 10.53***
(5.011) (4.431) (4.749) (4.856)
INTENSIVE 0.633 0.266 0.0311
(0.385) (0.163) (0.0193)
Constant 39.57*** 16.86*** 21.94*** 22.72***
(7.017) (3.578) (4.533) (4.822)
Sector effect YES NO NO NO
Observations 474 474 474 474
VIF 2.17 1.47 1.47 1.48
F-value 4.554*** 7.142*** 8.202*** 10.46***
Adjusted R2 0.101 0.083 0.096 0.123
Notes: t statistics are shown in parentheses. All variables were Winsorised at the 1st and 99th percentiles, with the exception of the indicator variables.
Financial data are in millions of U.S. dollars. CD1 is a proxy for the total level of carbon disclosure and is scored using an index developed by the CDP. CD2 is a
sector-adjusted carbon disclosure proxy. CP1, CP2, CP3, and CP4 are four different proxies for carbon performance, with CP1 and CP2 based on the carbon
intensity of emissions and CP3 and CP4 focusing on carbon reduction outcomes. SIZE is measured as the natural logarithm of market capitalisation at the
end of fiscal year 2009. ROA is return on assets, which is used as a proxy for a firm’s profitability and is measured as net income before extraordinary items/
preferred dividends divided by total assets. LEV is calculated as total debts divided by total assets at the end of fiscal year 2009. UK is a dummy variable and
equals 1 if a firm is listed in the FTSE 350 index and 0 otherwise. US is a dummy variable and equals 1 if a firm listed in the S&P 500 index and 0 otherwise.
INTENSIVE is a dummy variable that equals 1 if a firm belongs to the energy, materials, or utility sectors and 0 otherwise. VIF refers to the variance inflation
factor.
***
p < 0.01.
**
p < 0.05.
*
p < 0.1.
376 used CP4 to measure carbon mitigation performance (based on a four-item index). The coefficient of CP3 is 1.229 in model (3)
377 and that of CP4 is 3.871 in model (4); both are positive and significant at the 1% level.
378 All the results presented in models (1)–(4) suggest that good carbon performers, in particular companies with a large car-
379 bon emission reduction, disclose more overall carbon information than poor performers, consistent with our hypothesis.
380 Good performers tend to be more transparent so as to distinguish themselves from poor performers via sufficient disclosure
381 to outsiders, whereas poor performers are unable to imitate this behaviour and will send a false signal due to higher disclo-
382 sure costs and/or litigation risks. This finding contrasts with legitimacy theory, which predicts an inverse relationship. This
383 evidence implies that the standard CDP report allows for less flexibility to manipulate the outcome, meaning that the carbon
384 information that it provides is indeed useful for decision makers.
385 Thus, our study has made an incremental contribution beyond the existing literature. For example, Clarkson et al. (2008)
386 argued that previous empirical evidence provided mixed results regarding the relationship between corporate environmen-
387 tal performance and the level of environmental disclosure. They revisited this relationship by testing competing predictions
388 from economics-based and socio-political theories of voluntary disclosure and found a positive association. Their result is
389 consistent with economics disclosure theory but inconsistent with socio-political theories. However, very limited empirical
390 studies so far have attempted to validate their findings and these theories in the context of carbon emissions. Furthermore,
391 Luo et al. (2012) explored the empirical evidence on managerial incentives for voluntary carbon disclosure, but they did not
392 consider or include carbon emissions in their test model as a determinant for decision making. We found that our sample
393 firms differ in the level of emissions, in the extent of disclosure, and in the degree of carbon performance across sectors
394 and countries. For example, Table 5 shows that the coefficients of US and UK are all negative and significant, implying that
395 the quantity and comprehensiveness of voluntary carbon disclosure for U.S. and U.K. companies are significantly lower than
5 September 2014
396 those seen for Australian companies.12 This evidence is consistent with that reported in Panel B of Table 3. In addition, a prior
397 study (e.g. Luo et al., 2013) showed that different sectors and countries have different carbon legislations and systems. Notwith-
398 standing these differences, our evidence suggests that the pattern of the association between carbon disclosure and perfor-
399 mance shows the same tendency. That is, better carbon performers tend to be more ecologically transparent, whereas
400 poorer performers are inclined to follow a strategy that keeps their carbon profiles opaque. The implication of the findings is
401 that if a company is reluctant to unveil its emission profile, it is likely to have a poor climate change record.
402 4.3.2. Regression results by country

403 Our cross-country research design allows for a consideration of the impact of national institutional difference on the rela-
404 tionship between carbon performance and disclosure. It should be noted that the significant negative coefficients on US and
405 UK for Model (1) to Model (4) in Table 5 mean that carbon disclosure is less in the U.S. and U.K. relative to Australia. However,
406 this does not necessarily mean that the association between CD and CP in the U.K. and U.S. is different from that in Australia,
407 because the association is affected by both CD and CP, not just one of these variables. It could be argued that the relationship
408 is affected by institutional factors of these three countries in terms of carbon emission mitigation and climate change adap-
409 tation policy and other relevant regulations. However, this is an empirical issue.
410 In order to test the association between CD and CP across these three countries and their relative strength, we reran the
411 regression by country and present the results in Table 6. The empirical evidence in Table 6 shows that although not all the
412 coefficients are statistically significant, the results are qualitatively the same as those in Table 5. The coefficients of both CP1
413 and CP2 are negative and CP3 and CP4 are positive as predicted. However, it appears that the relationship between CD and CP
414 is not sensitive to national institutional difference. For example, CD1 is not sensitive to CP1 in all three countries, whereas
415 CD2 is significantly sensitive to CP4 in all three nations. In addition, although CD2 is strongly correlated with CP2 in the
416 U.K., it is not in Australia and the U.S. Thus, this evidence is insufficient to conclude that the relationship varies greatly across
417 nations.
418 In contrast, the results suggest that the sensitivity of the association between CD and CP seems related to how carbon
419 performance is measured. That is, the sector-adjusted measure of CP2 is better than the unadjusted measure CP1, because
420 CP2 has a significant coefficient at least in the U.K., but CP1 is significant in none of the three nations. Furthermore, CP4 uses
421 equal weight for all four components of the performance index, whereas CP3 has unequal and somewhat arbitrary weights.
422 CP4 is significant in all three nations. In contrast, CP3 is only significant in the U.S. Thus, the results indicate that the asso-
423 ciation between CD and CP is more significantly affected by sector differences than country characteristics.
424 4.4. Robustness checks
425 We conducted several tests for robustness to determine whether our findings would be subject to alternative research
426 designs. First, we used Tobit regression instead of OLS regression, because our dependent variables, CD1 and CD2, only range
427 from 0 to 100. When we reran the regression, the results (not tabulated) were consistent with the main results presented in
428 Table 5.
429 Second, we added two relevant control variables identified in the previous literature. First, investment in newer technol-
430 ogy would result in cleaner and more energy-efficient equipment, so firms tend to disclose this good news to stakeholders
431 (Clarkson et al., 2008). Hence, we created a control variable, CAPSPEND, which was calculated as capital spending divided by
432 total sales. Our second additional variable, TOBINQ, controlled for management capability. Firms exhibiting higher levels of
433 management capability and innovation are expected to disclose more information. TOBINQ was calculated as the total mar-
434 ket value based on the year-end price and the number of shares outstanding, plus preferred shares, book value of long-term
435 debt, and current liabilities, divided by the book value of total assets. The results (not tabulated) remained qualitatively the
436 same.
437 Third, Clarkson et al. (2013) argued that a firm’s general predisposition toward disclosure should also be considered an
438 important factor. Thus, we used managerial earnings forecast (MEF), Big 4 auditor, and growth potential as proxies for firms’
439 general predisposition toward disclosure. First, the disclosure of earnings forecast may suggest that the manager is willing to
440 disclose firm-specific information as part of an overall corporate transparency strategy. Such a firm is also more likely to
441 show a higher degree of general disclosure propensity (Clarkson et al., 2013; Coller and Yohn, 1997; De Franco et al.,
442 2013; Marquardt and Wiedman, 1998; Rogers et al., 2009; Rogers and Van Buskirk, 2009). Thus, we created a dummy var-
443 iable MEF that was equal to 1 if the firm made an earnings forecast and 0 otherwise (Clarkson et al., 2013). Second, auditor
444 size is associated with a firm’s disclosure tendency because firms hire large auditors to signal their commitment to transpar-
445 ency. Brand-name auditors are perceived to relate to high-quality reporting, lowering financing costs (Beatty, 1989) and rais-
446 ing the company’s value (Becker et al., 1998; Chaney and Philipich, 2002; Krishnan, 2003; Mansi et al., 2004).So auditor firm
447 size was a control variable measured as an indicator variable, Big 4. Third, the previous literature suggests that firm growth
448 opportunity is correlated with disclosure behaviour because of higher expectations from stakeholders and the market for
12
We replaced US and UK with another dummy variable, ETS, and examined the effect of a national ETS on a firm’s level of voluntary carbon disclosure. ETS
equals 1 if a firm is listed in a country that has implemented an emission trading scheme and 0 otherwise. The results demonstrate that the coefficient on ETS is
not significant, suggesting that there is no significant association between the presence of a national ETS and the quantity and comprehensiveness of voluntary
carbon disclosure.
12
5 September 2014
JCAE 57
Table 6
Regression results by country.
Variable U.S. U.K. AUS

(1) (2) (3) (4) (1) (2) (3) (4) (1) (2) (3) (4)
CD1 CD2 CD2 CD2 CD1 CD2 CD2 CD2 CD1 CD2 CD2 CD2
CP1 0.000926 0.00207 0.00247
L. Luo, Q. Tang / Journal of Contemporary Accounting & Economics xxx (2014) xxx–xxx
(0.788) (1.632) (0.668)
CP2 0.000933 0.00244** 0.00205
(0.810) (1.976) (0.702)
CP3 1.736*** 0.535 1.565
(3.489) (0.911) (1.582)
CP4 4.619*** 2.739** 4.731**
(4.375) (2.137) (2.418)
SIZE 2.638*** 1.998*** 1.918*** 1.686** 3.925*** 3.052*** 3.158*** 3.090*** 0.755 0.337 0.0114 0.218
(3.097) (2.673) (2.625) (2.327) (5.246) (4.588) (4.719) (4.668) (0.532) (0.288) (0.010) (0.191)
ROA 12.87 26.86 31.08* 28.07 19.36 20.68 18.09 18.27 46.69 38.96 36.46 36.84
(0.605) (1.408) (1.668) (1.529) (0.812) (0.883) (0.766) (0.783) (1.356) (1.211) (1.150) (1.195)
LEV 7.198 3.426 2.519 2.557 5.559 6.478 8.800 9.277 16.16 19.98 17.56 17.10
(0.943) (0.511) (0.387) (0.398) (0.768) (0.956) (1.289) (1.379) (0.964) (1.295) (1.150) (1.154)
INTENSIVE 0.392 0.551 0.686 0.0672 1.585 1.726 4.404 4.996 4.614
(0.169) (0.243) (0.308) (0.0211) (0.510) (0.562) (1.112) (1.275) (1.216)
Constant 31.38*** 22.73*** 30.34*** 29.28*** 20.35** 31.62*** 35.43*** 38.05*** 58.19*** 1.102 3.959 4.076
(3.074) (2.741) (3.616) (3.600) (2.179) (4.728) (4.970) (5.437) (4.352) (0.101) (0.364) (0.388)
Sector effect YES NO NO NO YES NO NO NO YES NO NO NO
Observations 250 250 250 250 159 159 159 159 65 65 65 65
F 1.18 1.72 4.1*** 5.54*** 3.25*** 5.93*** 5.22*** 6.09*** 1.36 1.19 1.63 2.36*
Adjusted R2 0.010 0.014 0.059 0.084 0.156 0.135 0.118 0.139 0.063 0.015 0.047 0.096
Notes: t statistics are shown in parentheses. All variables were Winsorised at the 1st and 99th percentiles, with the exception of the indicator variables. Financial data are in millions of U.S. dollars. CD1 is a proxy
for the total level of carbon disclosure and is scored using an index developed by the CDP. CD2 is a sector-adjusted carbon disclosure proxy. CP1, CP2, CP3, and CP4 are four different proxies for carbon performance,
with CP1 and CP2 based on the carbon intensity of emissions and CP3 and CP4 focusing on carbon reduction outcomes. SIZE is measured as the natural logarithm of market capitalisation at the end of fiscal year
No. of Pages 15, Model 3G

2009. ROA is return on assets, which is used as a proxy for a firm’s profitability and is measured as net income before extraordinary items/preferred dividends divided by total assets. LEV is calculated as total debts
divided by total assets at the end of fiscal year 2009. UK is a dummy variable and equals 1 if a firm is listed in the FTSE 350 index and 0 otherwise. US is a dummy variable and equals 1 if a firm listed in the S&P 500
index and 0 otherwise. INTENSIVE is a dummy variable that equals 1 if a firm belongs to the energy, materials, or utility sectors and 0 otherwise.
***
p < 0.01.
**
p < 0.05.
*
p < 0.1.
5 September 2014
449 fast-growth companies (Skinner and Sloan, 2002). So sales growth was used as a control variable that was 1-year, 3-year, or
450 5-year annual sales growth. We reran the test model, and the results (not tabulated) were qualitatively the same as those
451 reported in Table 5, suggesting that using earnings forecast, Big 4 auditor, and growth potential to proxy for general disclo-
452 sure predisposition does not alter our main results.
453 Fourth, we examined the emissions intensity variable data and found that the distribution of the data was skewed in our
454 sample firms, which could have possibly led to incorrect conclusions. Thus, we employed the natural logarithm transforma-
455 tion of the ratio of total Scope 1 and Scope 2 GHG emissions to total sales to measure CP1 (Al-Tuwaijri et al., 2004; Clarkson
456 et al., 2011b). The use of the log form of CP1 did not alter our results.
457 Finally, in our main tests, we restricted our sample firms to those that voluntarily participated in the CDP. This sample
458 might have introduced a self-selection bias in favour of rejecting a nonassociation when, in fact, the null scenario was true.
459 To address this issue, we adopted the Heckman (1979) two-stage estimation procedure. In the first stage, we modelled the
460 firm’s voluntary carbon disclosure decision (i.e., the decision to participate in the CDP) using firm size, profitability and lever-
461 age, capital spending, and Tobin’s Q as well as sector dummies because the relationship is sector sensitive (Bewley and Li,
462 2000; Clarkson et al., 2008; Luo et al., 2012). The results (not tabulated) suggested that our results were qualitatively
463 unchanged when we controlled for sampling bias.
464 5. Summary and conclusion
465 GHG information is increasingly important. However, the usefulness of voluntarily disclosed information from corpora-
466 tions is an issue, and there is a lack of empirical evidence for the correspondence between the level of carbon information
467 disclosed and the underlying genuine carbon performance. We examined the association and found that good performers are
468 more forthright in reporting their commitment to carbon control using the CDP disclosure code. Our results are generally
469 consistent with those of Clarkson et al. (2008) and Al-Tuwaijri et al. (2004), suggesting that firms with good performance
470 are likely to disclose more to distinguish themselves for investors and other stakeholders. Our findings were robust when
471 we controlled for other influences on carbon disclosure and applied different and complementary measures for carbon per-
472 formance. In particular, we used three research settings, namely, the United States, the United Kingdom, and Australia, and
473 the results were consistent. Thus, our empirical evidence should enhance stakeholders’ confidence in CDP reports as a mea-
474 sure of a firm’s carbon performance. The policy implication of the study is that voluntary disclosure should be regulated with
475 a standard format and contents so that carbon disclosure can be made in a consistent manner, even when the disclosure is
476 voluntary. Ideally, a stand-alone GHG statement is preferred, because carbon information is very complex, and disclosure in
477 other forms (such as in annual reports or sustainability statements) is likely to be brief and not sufficient for users to make an
478 informed decision.
479 The present study has some limitations. First, we relied on CDP reports for analysis; thus, it is probably inappropriate to
480 generalise our results to information disclosed through other communication channels. Second, although the quality of CDP
481 reports is improving with updated guidelines, there will always be potential problems with the adoption of a generic set of
482 rules given the diversity of issues covered and the complex nature of firms (Frost et al., 2005). Third, we used large firms, and
483 our sample was relatively small; therefore, caution should be exercised when generalising the findings to other firms. Finally,
484 our analysis is merely a snapshot of reporting practices over a single year. Thus, further research is necessary to provide a
485 more complete picture of the changing nature of carbon reporting by extending the study over multiple periods. Studies cov-
486 ering multiple periods would be particularly useful if in a given year firms adopt a proactive strategy that may affect future
487 performance (Clarkson et al., 2011a). Carbon mitigation is but one possible measure of a firm’s carbon performance. For
488 example, some studies (Krishnan, 2003) have examined the valuation relevance of a firm’s carbon emissions. Future studies
489 may consider using the impact of carbon activities on share price as a measure of a firm’s carbon performance (He et al.,
490 2013).
491 Because carbon disclosure has already moved from a voluntary to a mandatory requirement in many jurisdictions, the
492 format and content of CDP reports might be considered to develop a formal GHG statement. Based on our results, we
493 observed some room for improvement in the current version of the CDP report. For instance, we believe that there should
494 be industry-specific disclosure guidelines, and more disclosure should be made at the project level. In addition, we identified
495 some weaknesses in the carbon management systems in our sample firms. For example, there is no separate climate com-
496 mittee with the sole responsibility for addressing issues related to global warming risks and opportunities. Currently, this
497 function is often performed by a social, environmental, or sustainability committee. In addition, firms require an upgraded
498 accounting system to meet the needs of a low-carbon economy (Ratnatunga et al., 2011), and thus there is growing demand
499 for courses in carbon accounting, auditing, and management. Accounting academics will need to develop more teaching
500 material to meet the needs of the accounting students who will practice in a carbon-restricted business environment.
501 Acknowledgement
502 We thank all the participants at the Postgraduate Research Student Colloquium at the University of Western Sydney. We
503 also gratefully acknowledge constructive comments by anonymous reviewers and the editor. Qingliang Tang gratefully
504 acknowledges financial support from the School of Business, University of Western Sydney. Le Luo acknowledges financial
5 September 2014
505 Q3 support from the National Natural Science Foundation of China (Project No. 71272237) and the National Social Science Foun-
Q4
506 dation of China (Project No. 13BJY069).
507 References
508 Aerts, W., Cormier, D., 2009. Media legitimacy and corporate environmental communication. Account. Organ. Soc. 34, 1–27.
509 Al-Tuwaijri, S.A., Christensen, T.E., Hughes, K.E., 2004. The relations among environmental disclosure, environmental performance, and economic
510 performance: a simultaneous equations approach. Account. Organ. Soc. 29, 447–471.
511 Andrew, J., Cortese, C., 2010. Greenhouse gas reporting and the Carbon Disclosure Project, Paper Presented at the Sixth Asia Pacific Interdisciplinary
512 Research in Accounting (APIRA) Conference, Sydney, Australia, 12 and 13 July <http://apira2010.econ.usyd.edu.au/conference_proceedings/APIRA-
513 2010-080-Andrew-Greenhouse-gas-reporting-and-the-carbon-disclosure-project.pdf>.
514 Becker, C.L., DeFond, M.L., Jiambalvo, J., Subramanyam, K., 1998. The effect of audit quality on earnings management⁄. Contemp. Account. Res. 15, 1–24.
515 Belkaoui, A., Karpik, P.G., 1989. Determinants of the corporate decision to disclose social information. Account. Audit. Accountab. J. 2, 36–51.
516 Bewley, K., Li, Y., 2000. Disclosure of environmental information by Canadian manufacturing companies: a voluntary disclosure perspective. Adv. Environ.
517 Account. Manage. 1, 201–226.
518 CDP Australian & New Zealand Report, 2010. Australia and New Zealand Report, Carbon Disclosure Project <https://www.cdproject.net/en-US/Results/
519 Pages/All-Investor-Reports.aspx> (accessed 22.03.12).
520 CDP FTSE 350 Report, 2010. Carbon Disclosure Project 2010 FTSE 350 Report.
521 CDP Rating Methodology, 2010. CDP 2010 Rating Methodology, Carbon Disclosure Project <https://www.cdproject.net/en-US/Results/Pages/CDP-2010-
522 performance-scores.aspx> (accessed 02.09.11).
523 CDP S&P 500 Report, 2010. Carbon Disclosure Project 2010 S&P 500 Report.
524 CDP Scoring Methodology, 2010. CDP 2010 Scoring Methodology, Carbon Disclosure Project <https://www.cdproject.net/en-US/Results/Pages/leadership-
525 index-2010.aspx> (accessed 09.05.13).
526 Chaney, P.K., Philipich, K.L., 2002. Shredded reputation: the cost of audit failure. J. Account. Res. 40, 1221–1245.
527 Chapple, L., Clarkson, P.M., Gold, D.L., 2013. The cost of carbon: capital market effects of the proposed emission trading scheme. Abacus 49, 1–33.
528 Chatterji, A.K., Levine, D.I., Toffel, M.W., 2009. How well do social ratings actually measure corporate social responsibility? J. Econ. Manage. Strategy 18,
529 125–169.
530 Cho, C., Patten, D., 2007. The role of environmental disclosures as tools of legitimacy: a research note. Account. Organ. Soc. 32, 639–647.
531 Clarkson, P.M., Li, Y., Richardson, G.D., Vasvari, F.P., 2008. Revisiting the relation between environmental performance and environmental disclosure: an
532 empirical analysis. Account. Organ. Soc. 33, 303–327.
533 Clarkson, P.M., Li, Y., Richardson, G.D., Vasvari, F.P., 2011a. Does it really pay to be green? Determinants and consequences of proactive environmental
534 strategies. J. Account. Public Policy 30, 122–144.
535 Clarkson, P.M., Overell, M.B., Chapple, L., 2011b. Environmental reporting and its relation to corporate environmental performance. Abacus 47, 27–60.
536 Clarkson, P.M., Fang, X., Li, Y., Richardson, G., 2013. The relevance of environmental disclosures: are such disclosures incrementally informative? J. Account.
537 Public Policy 32, 410–431.
538 Coller, M., Yohn, T.L., 1997. Management forecasts and information asymmetry: an examination of bid-ask spreads. J. Account. Res. 35, 181–191.
539 Cormier, D., Magnan, M., 1999. Corporate environmental disclosure strategies: determinants, costs and benefits. J. Account. Audit. Finance 14, 429–452.
540 Cormier, D., Magnan, M., Van Velthoven, B., 2005. Environmental disclosure quality in large German companies: economic incentives, public pressures or
541 institutional conditions? Euro. Account. Rev. 14, 3–39.
542 Cotter, J., Najah, M.M., 2012. Institutional investor influence on global climate change disclosure practices. Austr. J. Manage. 37, 169–187.
543 Cowan, S., Gadenne, D., 2005. Australian corporate environmental reporting: a comparative analysis of disclosure practices across voluntary and mandatory
544 disclosure systems. J. Account. Organ. Change 1, 165–179.
545 De Franco, G., Hope, O.-K., Larocque, S., 2013. The effect of disclosure on the pay-performance relation. J. Account. Public Policy 32, 319–341.
546 Deegan, C., 2002. Introduction: the legitimising effect of social and environmental disclosures – a theoretical foundation. Account. Audit. Accountab. J. 15,
547 282–311.
548 Deegan, C., Gordon, B., 1996. A study of the environmental disclosure practices of Australian corporations. Account. Bus. Res. 26, 187–199.
549 Deegan, C., Rankin, M., 1996. Do Australian companies report environmental news objectively? Account. Audit. Accountab. J. 9, 50–67.
550 Dhaliwal, D.S., Li, O.Z., Tsang, A., Yang, Y.G., 2011. Voluntary nonfinancial disclosure and the cost of equity capital: the initiation of corporate social
551 responsibility reporting. Account. Rev. 86, 59.
552 Dowling, J., Pfeffer, J., 1975. Organizational legitimacy: social values and organizational behavior. Pac. Sociol. Rev. 18, 122–136.
553 Freedman, M., Jaggi, B., 1982. Pollution disclosures, pollution performance and economic performance. Omega 10, 167–176.
554 Freedman, M., Jaggi, B., 2004. Carbon dioxide emissions and disclosures by electric utilities. Adv. Publ. Interest Account. 10, 105–129.
555 Freedman, M., Jaggi, B., 2005. Global warming, commitment to the Kyoto Protocol, and accounting disclosures by the largest global public firms from
556 polluting industries. Int. J. Account. 40, 215–232.
557 Freedman, M., Jaggi, B., 2010. Global warming and corporate disclosures: a comparative analysis of companies from the European Union, Japan and Canada.
558 Adv. Environ. Account. Manage. 4, 129–160.
559 Freedman, M., Jaggi, B., 2011. Global warming disclosures: impact of Kyoto Protocol across countries. J. Int. Fin. Manage. Account. 22, 46–90.
560 Freedman, M., Wasley, C., 1990. The association between environmental performance and environmental disclosure in annual reports and 10Ks. Adv. Publ.
561 Interest Account. 3, 183–193.
562 Frost, G., Jones, S., Loftus, J., Laan, S., 2005. A survey of sustainability reporting practices of Australian reporting entities. Austr. Account. Rev. 15, 89–96.
563 Gibson, K., O’Donovan, G., 2007. Corporate governance and environmental reporting: an Australian study. Corp. Gov. Int. Rev. 15, 944–956.
564 Gray, R., Kouhy, R., Lavers, S., 1995. Corporate social and environmental reporting: a review of the literature and a longitudinal study of UK disclosure.
565 Account. Audit. Account. J. 8, 47–77.
566 Griffin, P.A., Lont, D.H., Sun, Y., 2012. The Relevance to Investors of Greenhouse Gas Emission Disclosures. Working Paper, University of California, University
567 of Otago, 3 October 2011.
568 He, Y., Tang, Q., Wang, K., 2013. Carbon disclosure, carbon performance, and cost of capital. China J. Account. Stud. 1, 190–220.
569 Heckman, J.J., 1979. Sample selection bias as a specification error. Econometrica 47, 153–161.
570 Hoffmann, V.H., Busch, T., 2008. Corporate carbon performance indicators. J. Ind. Ecol. 12, 505–520.
571 Hughes, S., Sander, J., Reier, J., 2000. Do environmental disclosures in US annual reports differ by environmental performance? Adv. Environ. Account.
572 Manage. 1, 141–161.
573 Hughes, S.B., Anderson, A., Golden, S., 2001. Corporate environmental disclosures: are they useful in determining environmental performance? J. Account.
574 Public Policy 20, 217–240.
575 Ingram, R., Frazier, K., 1980. Environmental performance and corporate disclosure. J. Account. Res. 18, 614–622.
576 International Union for the Conservation of Nature and Natural Resources, 1980. World Conservation Strategy: Living Resource Conservation for Sustainable
577 Development (Gland, Switzerland).
578 IPCC, 2013. Climate Change 2013: The Physical Science Basis, Intergovernmental Panel on Climate Change <http://www.climatechange2013.org/images/
579 uploads/WGIAR5_WGI-12Doc2b_FinalDraft_All.pdf> (accessed 20.12.13).
5 September 2014
580 Johnston, D., Sefcik, S., Soderstrom, N., 2008. The value relevance of greenhouse gas emissions allowances: an exploratory study in the related United States
581 SO2 market. Euro. Account. Rev. 17, 747–764.
582 Kim, E.H., Lyon, T.P., 2011. Strategic environmental disclosure: evidence from the DOE’s voluntary greenhouse gas registry. J. Environ. Econ. Manage. 61,
583 311–326.
584 Klassen, R.D., McLaughlin, C.P., 1996. The impact of environmental management on firm performance. Manage. Sci. 42, 1199–1214.
585 Kolk, A., Levy, D., Pinkse, J., 2008. Corporate responses in an emerging climate regime: the institutionalization and commensuration of carbon disclosure.
586 Euro. Account. Rev. 17, 719–745.
587 Krishnan, G.V., 2003. Audit quality and the pricing of discretionary accruals. Audit. J. Practice Theory 22, 109–126.
588 Lash, J., Wellington, F., 2007. Competitive advantage on a warming planet. Harvard Bus. Rev. 85, 94.
589 Leighton Holdings CDP Report, 2010. Investor CDP 2010 Information Request-Leighton Holdings Ltd, the Carbon Disclosure Project <https://
590 www.cdproject.net/en-US/Pages/HomePage.aspx> (accessed 03.08.12).
591 Lend Lease CDP Report, 2010. Investor CDP 2010 Information Request-Lend Lease Corporation Ltd, the Carbon Disclosure Project <https://
592 www.cdproject.net/en-US/Pages/HomePage.aspx> (accessed 03.08.12).
593 Li, Y., McConomy, B., 1999. An empirical examination of factors affecting the timing of environmental accounting standard adoption and the impact on
594 corporate valuation. J. Account. Audit. Fin. 14, 279–313.
595 Li, Y., Richardson, G.D., Thornton, D.B., 1997. Corporate disclosure of environmental liability information: theory and evidence. Contemp. Account. Res. 14,
596 435–474.
597 Liao, L., Luo, L., Tang, Q., 2014. Gender diversity, board independence, environmental committee and greenhouse gas disclosure. British Account. Rev.
598 (forthcoming).
599 Luo, L., Tang, Q., 2014. Carbon tax, corporate carbon profile and financial return. Pac. Account. Rev. (forthcoming).
600 Luo, L., Lan, Y.-C., Tang, Q., 2012. Corporate incentives to disclose carbon information: evidence from the CDP global 500 report. J. Int. Finan. Manage.
601 Account. 23, 93–120.
602 Luo, L., Tang, Q., Lan, Y.-C., 2013. Comparison of propensity for carbon disclosure between developing and developed countries: a resource constraint
603 perspective. Account. Res. J. 26, 6–34.
604 Mansi, S.A., Maxwell, W.F., Miller, D.P., 2004. Does auditor quality and tenure matter to investors? Evidence from the bond market. J. Account. Res. 42, 755–
605 793.
606 Marquardt, C., Wiedman, C.I., 1998. Voluntary disclosure, information asymmetry, and insider selling through secondary equity offerings⁄. Contemp.
607 Account. Res. 15, 505–537.
608 Matsumura, E.M., Prakash, R., Vera-Muñoz, S.C., 2014. Firm-value effects of carbon emissions and carbon disclosures. Account. Rev. 89, 695–724.
609 Meek, G.K., Roberts, C.B., Gray, S.J., 1995. Factors influencing voluntary annual report disclosures by U.S., U.K. and Continental European Multinational
610 Corporations. J. Int. Bus. Stud. 26, 555–572.
611 Mitchell, J., Percy, M., McKinlay, B., 2006. Voluntary environmental reporting practices: a further study of ‘poor’ environmental performers. Austr. J. Corp.
612 Law 19, 182–215.
613 Neu, D., Warsame, H., Pedwell, K., 1998. Managing public impressions: environmental disclosures in annual reports. Account. Organ. Soc. 23, 265–282.
614 Patten, D., 2002. The relation between environmental performance and environmental disclosure: a research note. Account. Organ. Soc. 27, 763–773.
615 Peters, G., Romi, A., 2009. Carbon Disclosure Incentives in a Global Setting: An Empirical Investigation. Working Paper, University of Arkansas, 4 September.
616 Pinkse, J., Kolk, A., 2009. International Business and Global Climate Change. Taylor & Francis.
617 Prado-Lorenzo, J.-M., Garcia-Sanchez, I.-M., 2010. The role of the board of directors in disseminating relevant information on greenhouse gases. J. Bus. Ethics
618 97, 391–424.
619 Prado-Lorenzo, J., Rodriguez-Dominguez, L., Gallego-Alvarez, I., Garcia-Sanchez, I., 2009. Factors influencing the disclosure of greenhouse gas emissions in
620 companies world-wide. Management 47, 1133–1157.
621 Ratnatunga, J., Jones, S., Balachandran, K.R., 2011. The valuation and reporting of organizational capability in carbon emissions management. Account.
622 Horizons 25, 127–147.
623 Reid, E., Toffel, M., 2009. Responding to public and private politics: corporate disclosure of climate change strategies. Strategic Manage. J. 30, 1157–1178.
624 Rogers, J.L., Van Buskirk, A., 2009. Shareholder litigation and changes in disclosure behavior. J. Account. Econ. 47, 136–156.
625 Rogers, J.L., Skinner, D.J., Van Buskirk, A., 2009. Earnings guidance and market uncertainty. J. Account. Econ. 48, 90–109.
626 Schaltegger, S., Figge, F., 1997. Environmental Shareholder Value. WWZ and Bank Sarasin, Basel.
627 Shane, P.B., Spicer, B.H., 1983. Market response to environmental information produced outside the firm. Account. Rev. 58, 521–538.
628 Skinner, D.J., Sloan, R.G., 2002. Earnings surprises, growth expectations, and stock returns or don’t let an earnings torpedo sink your portfolio. Rev. Account.
629 Stud. 7, 289–312.
630 Stanny, E., 2010. Voluntary Disclosures by US Firms to the Carbon Disclosure Project. Working paper, School of Business and Economics, Sonoma State
631 University, CA.
632 Stanny, E., Ely, K., 2008. Corporate environmental disclosures about the effects of climate change. Corp. Soc. Respons. Environ. Manage. 15, 338–348.
633 Steffen, W., Burbidge, A.A., Hughes, L., Kitching, R., Lindenmayer, D., Musgrave, W., Stafford, S.M., Werner, P.A., 2009. Australia’s Biodiversity and Climate
634 Change: A Strategic Assessment of the Vulnerability of Australia’s Biodiversity to Climate Change. CSIRO Publishing, <http://
635 www.climatechange.gov.au/sites/climatechange/files/documents/04_2013/biodiversity-vulnerability-assessment.pdf> (accessed 12.08.14).
636 Strike, V.M., Gao, J., Bansal, P., 2006. Being good while being bad: social responsibility and the international diversification of US firms. J. Int. Bus. Stud. 37,
637 850–862.
638 Sutantoputra, A.W., Lindorff, M., Johnson, E.P., 2012. The relationship between environmental performance and environmental disclosure. Austral. J.
639 Environ. Manage. 19, 51–65.
640 Tang, Q., Luo, L., 2011. Transparency of Corporate Carbon Disclosure: International Evidence. Working Paper, University of Western Sydney, Sydney, 13 July.
641 Tang, Q., Luo, L., 2014. Carbon management systems and carbon mitigation. Austr. Account. Rev. 24, 84–98.
642 Tilt, C., 2001. The content and disclosure of Australian corporate environmental policies. Account. Audit. Accountab. J. 14, 190–212.
643 Tinker, T., Gray, R., 2003. Beyond a critique of pure reason: from policy to politics to praxis in environmental and social research. Account. Audit. Accountab.
644 J. 16, 727–761.
645 Van der Laan Smith, J., Adhikari, A., Tondkar, R.H., 2005. Exploring differences in social disclosures internationally: a stakeholder perspective. J. Account.
646 Publ. Policy 24, 123–151.
647 Verrecchia, R.E., 1983. Discretionary disclosure. J. Account. Econ. 5, 179–194.
648 Walls, J.L., Phan, P.H., Berrone, P., 2011. Measuring environmental strategy: construct development, reliability, and validity. Bus. Soc. 50, 71–115.
649 Williams, R.J., Bradstock, R.A., Cary, G.J., Enright, N.J., Gill, A.M., Liedloff, A.C., Lucas, C., Whelan, R.J., Andersen, A.N., Bowman, D.M.J.S., Clarke, P.J., Cook, G.D.,
650 Hennessy, K.J., York, A., 2009. Interactions between climate change, fire regimes and biodiversity in Australia: a preliminary assessment, CSIRO Climate
651 Adaptation Flagship, available.
652 Wiseman, J., 1982. An evaluation of environmental disclosures made in corporate annual reports. Account. Organ. Soc. 7, 53–63.
653

Journal of Contemporary Accounting & Economics: Le Luo, Qingliang Tang

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Journal of Contemporary Accounting & Economics: Le Luo, Qingliang Tang

Uploaded by

Copyright:

Available Formats

JCAE 57 No.

of Pages 15, Model 3G

Contents lists available at ScienceDirect

3 Does voluntary carbon disclosure reﬂect underlying carbon

103 2. Literature review and hypothesis development

175 3. Research design

176 3.1. The measurement of carbon disclosure

CD2i ¼ CD1i CD1i

239 3.2. The measurement of carbon performance

CP2i ¼ CP1i CP1i ;

283 3.3. Empirical models

GICS code Sector Frequency Percentage (%)

GICS = Global Industry Classiﬁcation Standard.

316 3.4. Sample selection

331 4. Empirical results

332 4.1. Descriptive statistics

Variable Mean Standard deviation Median Min Max

359 4.2. Univariate analysis

368 4.3. Multivariate analysis

369 4.3.1. Regression results for the pooled sample

Variable (1) (2) (3) (4)

402 4.3.2. Regression results by country

424 4.4. Robustness checks

Variable U.S. U.K. AUS

No. of Pages 15, Model 3G

464 5. Summary and conclusion

You might also like