Description
Using a direct measure of environmental legitimacy, we explore the impact of annual report environmental disclosures
and environmental press releases as legitimation tools. The sample comprises North American firms (Canada and
the United States). The results obtained show that environmental legitimacy is significantly and positively affected by
the quality of the economic-based segments of annual report environmental disclosures and by reactive environmental
press releases, but not by proactive press releases
Media legitimacy and corporate
environmental communication
Walter Aerts
a,
*
, Denis Cormier
b,1
a
Faculty of Applied Economics, University of Antwerpen, Prinsstraat 13, 2000 Antwerpen, Belgium
b
E
´
cole des Sciences de la Gestion, UQAM, C.P. 8888, Succ. Centre-Ville, Montreal, Quebec, Canada H3C 1P8
Abstract
Using a direct measure of environmental legitimacy, we explore the impact of annual report environmental disclo-
sures and environmental press releases as legitimation tools. The sample comprises North American ?rms (Canada and
the United States). The results obtained show that environmental legitimacy is signi?cantly and positively a?ected by
the quality of the economic-based segments of annual report environmental disclosures and by reactive environmental
press releases, but not by proactive press releases. Moreover, our results suggest that negative media legitimacy is a
driver of environmental press releases but not of annual report environmental disclosures.
Ó 2008 Elsevier Ltd. All rights reserved.
Introduction
We analyse the relationship between corporate
environmental communication through annual
report disclosures and press releases and environ-
mental legitimacy. In concert with the strategic
approach to organizational legitimacy, we contend
that ?rms use corporate communication media
(such as annual report disclosures and press
releases) to manage perceived environmental legit-
imacy by signalling to relevant publics that their
behavior is appropriate and desirable and, at the
same time, to react to public pressures by adapting
the level, content and quality of their environmen-
tal information dissemination processes.
Legitimacy is mainly about perception. Central
to the concept of legitimacy are the perceptions
held by relevant publics and by society at large.
Contrary to previous research on corporate envi-
ronmental communication, we implement a direct
measure of environmental legitimacy. We use pub-
lic media data to assess generalized perceptions
of a ?rm’s environmental legitimacy. We focus on
0361-3682/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aos.2008.02.005
*
Corresponding author. Tel.: +32 32204110; fax: +32
2204064.
E-mail addresses: [email protected] (W. Aerts), cormier.
[email protected] (D. Cormier).
1
Tel.: +1 514 987 3000x8358.
www.elsevier.com/locate/aos
Available online at www.sciencedirect.com
Accounting, Organizations and Society 34 (2009) 1–27
legitimacy as transpiring through media coverage
and media evaluations as one of the most salient
and prominent sources of societal legitimacy per-
ceptions. In addition, by measuring environmental
legitimacy as transpiring through the media we
are able to directly assess the e?ectiveness of cor-
porate environmental communication as an orga-
nizational perception management tool (Elsbach,
2003) to enhance its perceived environmental
legitimacy.
Credibility is an important factor in any impres-
sion management process and will have a sig-
ni?cant impact on its e?ectiveness (Leary &
Kowalski, 1990). Within a communication context,
credibility refers to the congruence between the
source’s verbal claims and the corresponding acts
and events. Credibility issues with regard to corpo-
rate environmental communication have been
hinted at in a number of environmental reporting
studies (e.g. Patten, 2002a, 2002b), especially with
regard to ?rms operating in environmentally-sensi-
tive industries. Environmental legitimacy percep-
tions at the industry level may well have a
signi?cant e?ect on the ex ante believability of
any corporate environmental communication act.
In that vein, we integrate industry legitimacy as a
factor moderating the e?ectiveness of a ?rm’s envi-
ronmental communication initiatives.
Previous environmental reporting research
neglected the complementary roles of di?erent
communication channels in environmental legiti-
macy management processes. Communication
channels such as annual reports and press releases
di?er in their capacity to process timely, focused
and sensitive information cues. Press releases typ-
ically cover more fragmented issues and are mainly
short-term, tactically oriented, whereas annual
reports, by their nature, are more comprehensive
in their approach and focus on more enduring
issues. The more timely and ?exible use of environ-
mental press releases may well a?ect the impact of
annual report disclosures on environmental legiti-
macy. By analyzing both annual report disclosures
and press releases, taking into account their
distinctive features, we explore the relative impor-
tance and roles of these two common information
dissemination channels as perception management
tools.
Our results show that environmental legitimacy
is signi?cantly and positively a?ected by the extent
and quality of annual report environmental disclo-
sures and by reactive environmental press releases.
Press releases with a proactive environmental con-
tent do not seem to a?ect media legitimacy. The
e?ect of annual report environmental disclosures
seems to be driven by the extent and quality of
the economic-based information segments. More-
over, the impact of annual report disclosures on
media legitimacy is negatively a?ected if the ?rm
belongs to an environmentally-sensitive industry.
The impact of annual report environmental disclo-
sures on media legitimacy is not a?ected by the use
of proactive press releases, while it appears that
the use of reactive press releases substitutes for
the e?ect of annual environmental disclosures to
a signi?cant extent. Moreover, our results docu-
ment that, in concert with the main assumption
of legitimacy theory, negative prior environmen-
tal legitimacy is a signi?cant driver of environmen-
tal press releases. Such a relationship can,
however, not be demonstrated for annual report
environmental disclosures, suggesting a more stra-
tegic orientation of annual report environmental
communication.
The remainder of the paper is organized as fol-
lows. Section 2 presents a theoretical framework
for analyzing the interplay of environmental legit-
imacy and corporate environmental communica-
tion as well as hypotheses. The study’s empirical
models and sample are described in Section 3.
Findings are reported in Section 4. Finally, Section
5 provides a discussion regarding the potential
implications of our ?ndings.
Theoretical background and development of
hypotheses
Corporate environmental communication and media
legitimacy
The legitimacy concept, with its roots in institu-
tional theory and socio-political research, has been
a highly in?uential theoretical perspective within
the domain of corporate environmental reporting
research. Strategic legitimacy theory suggests that
2 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
legitimacy is, to a certain extent, controllable by
organizations (Ashforth & Gibbs, 1990; Oliver,
1991). It contends that organizations are able to
make strategic choices to alter their legitimacy sta-
tus and to cultivate the resource through corporate
actions, by adapting their activities and changing
perceptions. Organizations might take various
action to ensure that their behavior is perceived
to be legitimate. One of the main options is to
attempt, through communication, to become iden-
ti?ed with symbols, values, and methods of opera-
tion with institutions, values, or outputs that are
strongly believed to be legitimate, and, as such,
to demonstrate congruence between its organiza-
tional practices and the values professed by its
social environment (Dowling & Pfe?er, 1975;
Lindblom, 1994; Patterson & Allen, 1997).
This strategic legitimacy perspective has been
quite popular in environmental disclosure research
in recent years and several studies explicitly adopt
such a perspective (Bansal & Clelland, 2004;
Brown & Deegan, 1998; Buhr, 1998; Cormier &
Gordon, 2001; Deegan & Gordon, 1996; Deegan
& Rankin, 1996; Deegan, Rankin, & Tobin,
2002; Hogner, 1982; Neu, Warsame, & Pedwell,
1998; O’Donovan, 2002; O’Dwyer, 2002; Patten,
1991, 1992, 2005; Savage, Rowlands, & Cataldo,
1999; Walden & Schwartz, 1997; Wilmshurst &
Frost, 2000). However, a shortcoming of extant
research related to the legitimacy theory perspec-
tive on corporate environmental reporting is that
the legitimacy construct is seldom measured
directly. As legitimacy is not directly observable,
it is common for researchers to infer legitimation
processes and e?ects by examining relationships
between observable corporate performance attri-
butes (like measures of environmental perfor-
mance or environmental disasters in the oil and
gas industry) or third-party actions (e.g. lawsuits
for environmental matters) and environmental
reporting measures. In this study, we use a direct
measure of organizational legitimacy, grounded
in the role of public media in social perception
construction processes.
With its roots in beliefs and perceptions, legiti-
macy can best be conceived as a social assessment
or appraisal of acceptance, appropriateness and/or
desirability (Fombrun, 1996; Rao, 1994; Suchman,
1995; Zimmerman & Zeitz, 2002). It refers to a col-
lective awareness and recognition of an organiza-
tion in its organizational ?eld as appropriate and
acceptable. In the same vein, legitimation refers
to ‘‘the characteristic of being legitimized by being
placed within a framework through which some-
thing is viewed as right and proper” (Tyler, 2006,
p. 376). As institutional intermediaries specializing
in disseminating information about organizations
or in evaluating their outputs, public media play
an important role in such legitimation processes
(Fombrun, 1996; Rao, 1998). They are believed
to have superior ability to access and disseminate
information by virtue of their institutional roles
or structural positions (Rao, 1998, 2001) and are
closely followed because of their perceived superi-
ority in accessing and evaluating ?rms (Fanelli &
Misangyi, 2006). Although di?erent relevant pub-
lics may hold di?erent perceptions of what good
or bad environmental management is, these per-
ceptions are ?ltered through the media lens to a
kind of common impression. The information
and evaluations provided by public media tend
to be distributed more broadly than the opinions
of the average stakeholder. As a result, they are
likely to have a high degree of in?uence on which
organizations and organizational properties
become prominent in the minds of stakeholders
(Fanelli & Misangyi, 2006).
Media research within the agenda-setting para-
digm (Caroll & McCombs, 2003; McCombs &
Shaw, 1972) demonstrates a close alignment
between the content of public media and public
opinion (or the degree of salience that di?erent
topics have for the general public),
2
corroborating
the assertion that public media are actively
involved in public impression construction pro-
cesses (Gamson, Croteau, Hoynes, & Sasson,
2
The agenda-setting hypothesis on media behavior posits
that media do not mirror public concerns, but actively in?uence
them, particularly through the transfer of salience from the
media agenda to the public agenda (Caroll & McCombs, 2003;
McCombs & Shaw, 1972). This indicates that the impact of
public media on stakeholders’ perceptions derives primarily
from their ability to focus public attention on the issues and
entities that they select to report on (Deephouse, 2000; Pollock
& Rindova, 2003).
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 3
1992). Media research shows that the agenda-set-
ting e?ect is especially strong for unobtrusive issues
or issues with which individuals have little personal
contact and for which they rely on the media as the
primary (and sometimes only) source of informa-
tion. Environmental issues are generally perceived
as unobtrusive (Ader, 1995), reinforcing the rele-
vance of a media proxy for environmental legiti-
macy. Moreover, from management’s point of
view, media evaluations may be the only persistent
and long-term proxy for collective legitimacy
impressions on which it can benchmark and model
its environmental communication strategy. In the
same vein, Neu et al. (1998) argue that environ-
mental disclosures focus on relevant publics that
are dominant rather than peripheral and ‘critical’
ones.
Previous research in corporate environmental
reporting (Bewley & Li, 2000; Brown & Deegan,
1998; Deegan, Rankin, & Voght, 2000; Patten,
2002a) elaborates on the e?ect of media exposure
on environmental disclosures, mainly establishing
that higher public media coverage of environmen-
tal issues puts public pressure on ?rms to increase
environmental disclosures. In addition, Brown and
Deegan (1998) show a strong association between
the industry-wide number of environmental arti-
cles that were negative in tone and the amount
of ‘‘positive” corporate environmental disclosures.
These studies mainly focus on the level of ?rm-spe-
ci?c media coverage as a proxy for public concerns
regarding environmental issues. In this paper, we
extend previous environmental reporting research
by incorporating media exposure not only as an
antecedent but also as a consequence of environ-
mental communication acts. Moreover, we elabo-
rate on the ?rm-speci?c evaluative content of
media exposure in measuring corporate environ-
mental (media) legitimacy.
Given the potential relevance of corporate envi-
ronmental communication as a legitimation tool,
we posit the following hypothesis:
Hypothesis 1. Enhanced corporate environmental
communication (annual report dis-
closures and press releases) is asso-
ciated with higher environmental
(media) legitimacy.
E?ect of industry
Credibility is an important factor in any
impression management process and will have a
signi?cant impact on its e?ectiveness (Leary &
Kowalski, 1990; Shapiro, 1991). Perceived message
credibility will depend on the perceptions of source
competence, of actual or expected (incentives) bias
behavior and of the characteristics of the message
itself (Barton & Mercer, 2005; Birnbaum & Steg-
ner, 1979; Mercer, 2004). More speci?cally, Tede-
schi and Melburg (1984) distinguish between
credibility and believability, with believability
referring to the subjective probability the target
assigns to any communication from a source.
While a source’s credibility can be expected to
have a direct e?ect on believability, other percep-
tions of the source and of the organizational ?eld
in which it functions may have a strong e?ect on
the believability of its communications as well. In
that respect, entire industries can have more or less
legitimacy that can be conferred upon the ?rms
operating within them (Aldrich & Fiol, 1994;
Scott, 1995; Suchman, 1995). Industry-level legiti-
macy has to do with the degree to which the oper-
ations and business processes of ?rms in a given
industry, and their products and services o?ered,
are accepted as appropriate and useful by broader
publics (Hannan & Freeman, 1989; Scott, 1995).
Prior accounting-related studies document that
?nancial reporting users appear to consider the
consistency between a ?rm’s reporting incentives,
which are to a large degree circumstantial, and
its actual disclosures when assessing the credibility
of the disclosures. Especially when the level of
discretion in the reporting environment is high,
users tend to discount the credibility (believability)
of incentive-consistent disclosures (Anderson,
Kadous, & Koonce, 2004; Hirst, Koonce, & Sim-
ko, 1995; Hodge, Hopkins, & Pratt, 2006). In this
vein, industry legitimacy may be directly related to
the believability of its constituents’ communica-
tions, especially in the ?eld of voluntary environ-
mental communication where there is a
straightforward reporting incentive for manage-
ment. Even if the objective credibility of the source
is high, low industry legitimacy may undermine the
believability of environmental reporting and thus
4 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
the e?ectiveness of a ?rm’s communicative legiti-
mation e?orts.
Prior research suggests that the believability of
environmental reporting may be problematic.
The believability of corporate environmental com-
munication e?orts would generally be validated
through a positive association between environ-
mental performance and environmental communi-
cation content, but previous research fails to
clearly demonstrate such an association (Freed-
man & Wasley, 1990; Hughes, Anderson, &
Golden, 2001; Ingram & Frazier, 1980; Wiseman,
1982). Some minor evidence of a signi?cant posi-
tive relationship could be established for very spe-
ci?c pollution disclosures (e.g. Al-Tuwaijri,
Christensen, & Hughes, 2004; Patten, 2002b), but
surely not for more comprehensive disclosure mea-
sures. Moreover, Patten (2002b) documents a
lower relationship between environmental perfor-
mance and environmental disclosure within more
environmentally-sensitive industries, suggesting
that environmental disclosures may be perceived
as less credible for ?rms from industries with high
environmental exposure. Belonging to an environ-
mentally-sensitive industry (with its negative con-
notations) may be the ?rst observation one has
about a ?rm, leading to a discounting of later posi-
tive messages coming from that ?rm and impeding
the e?ectiveness of its legitimacy-enhancement
e?orts. Therefore, we hypothesize:
Hypothesis 2. The association between the level of
corporate environmental communi-
cation and environmental (media)
legitimacy will be lower for those
?rms operating in more environ-
mentally-sensitive industries.
Tactical verbal impression management
Press releases are among the most common and
widespread communication vehicles used by public
?rms to disseminate voluntary information and
their content is highly discretionary. Further, they
di?er from annual report disclosures in their
capacity to process rich information and, conse-
quently, may ful?l complementary communication
goals. Rich information media
3
will generally be
more e?ective in directing attention, establishing
prominence and changing impressions (Daft &
Lengel, 1986). For our purposes, the main infor-
mation richness criteria to di?erentiate annual
report disclosures and press releases are timeliness
(and related feedback capability), topical focus
and language variety. Press releases can be used
in a more timely, more elaborate and focused fash-
ion and with more expressive language than
annual report disclosures to confront sensitive
environmental issues. They may also be more e?ec-
tive in signalling commitment. In this vein, their
use will be more tactical than that of the more
comprehensive and longer-term characteristic of
annual report disclosures.
The use and e?ectiveness of information rich
media will be more pronounced when timeliness
(reaction speed) is important from a self-presenta-
tional perspective (Sheer & Chen, 2004). From a
tactical impression management perspective, press
releases can be used in a proactive or in a reactive
fashion (Gardner & Martinko, 1988; Tedeschi &
Melburg, 1984). Proactive (or acclaiming) verbal
behavior aims at establishing a positive identity
for an audience, most likely through verbal state-
ments of accomplishments. It is not merely initi-
ated as a reaction to situational demands. On the
other hand, reactive self-presentational verbal
behavior is mainly defensive and it is typically ini-
tiated as a response to a controversial situation in
which negative and undesirable qualities may be
attributed to the ?rm. Proactive (or acclaiming)
content would stress the importance, relevance
and scope of positive environmental outcomes or
actions, whereas more reactive content would
focus on downplaying the signi?cance of nega-
tively perceived or controversial events related to
the natural environment. In this vein, proactive
3
Daft and Lengel (1984, 1986) de?ne information richness as
the ability of communication media to change understanding
within a time interval, with face-to-face communication having
the highest richness and periodic statistical documents posi-
tioned at the lower end. Information richness depends on the
medium’s capability of immediate feedback (the most impor-
tant aspect), the number of cues and channels utilized,
personalisation and language variety.
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 5
environmental press releases would express com-
mitment to environmental concerns and generally
focus on the positive aspects of corporate environ-
mental performance. On the other hand, reactive
press releases would tackle more equivocal con-
cerns, with equivocality referring to the presence
of multiple and possibly con?icting interpretations
of the available information.
From an information richness perspective, per-
formance improves if richer media are used for
more equivocal tasks and leaner media are used
for non-equivocal tasks (Daft & Lengel, 1986).
In that sense, reactive press releases would be more
e?ective than proactive press releases. By enabling
equivocal information to be conveyed quickly,
timely reactive press releases can be used as a nor-
malizing account, separating a negative environ-
mental event or outcome from larger assessments
of the ?rm as a whole and, in the end, putting a
positive spin on events (Suchman, 1995).
Traditional forms of defensive verbal impres-
sion management tactics, such as excuses, justi?ca-
tions, defences of innocence and apologies, have
been shown to be e?ective as perception manage-
ment devices in numerous research settings (Barton
& Mercer, 2005; Elsbach, 2003; Kim, Dirks, Coo-
per, & Ferrin, 2006; Wood & Mitchell, 1981). In
?eld settings, defensive impression management
tactics are frequently accompanied by selective cat-
egorizations that stress and make salient other
legitimate and competing dimensions along which
a ?rm’s actions and performance should be
assessed. By selectively directing and focusing
attention on performance dimensions with positive
rami?cations, categorization processes relieve the
dissonance related to a single dimension (e.g. a neg-
atively perceived environmental outcome or event)
and tend to bolster the believability of more
straightforward defensive verbal tactics (Elsbach
& Kramer, 1996). Defenses of innocence and justi-
?cations typically shift focus away from the nega-
tive towards the more positive aspects of an event
or outcome and, in that vein, set the stage for
acclaiming tactics with regard to the progress made
toward environmentally desirable goals (Elsbach &
Sutton, 1992). By its nature, timeliness will be more
important for reactive impression management tac-
tics relative to proactive impression management
and is thus expected to promote the relative e?ec-
tiveness of reactive environmental press releases.
Moreover, previous research suggested that speci?c
occurrences of defensive impression management
are more e?ective than assertive verbal behaviors
in shaping evaluative perceptions of an external
audience (Suchman, 1995). In this vein, Crant
and Bateman (1993) documented that defensive
causal accounts actually diminished observers’
assessment of blame for failure, whereas acclaim-
ing causal accounts did not a?ect the assignment
of credit for success.
Given the importance of response timeliness in
the face of equivocal environmental concerns and
the relative e?ectiveness of defensive verbal
impression management tactics, we hypothesize
Hypothesis 3. The association between reactive
environmental press releases and
environmental (media) legitimacy
will be higher relative to the associa-
tion between proactive environmen-
tal press releases and environmental
(media) legitimacy.
Sample selection and empirical models
Sample
The initial sample is comprised of 623 non-
?nancial North American ?rms (205 from S&P/
TSX300 for Canada and 418 from S&P500 for
the US). From the sample of 623 ?rms, we ?nd
environmental news exposure for 165 ?rms. Out
of these 165 ?rms, 158 reported pollutants released
in the environment (Toxics release inventory).
These 158 ?rms (119 US, 39 Canadian) constitute
our ?nal sample. Environmental disclosure is col-
lected from corporate annual reports on Internet
sites and is coded for the year 2002. Financial data
is collected from Worldscope and from ?rms’
Internet sites. The sample ?rms operate in the fol-
lowing industries (S&P classi?cation):
Consumer goods and services;
Energy;
Chemicals and drugs;
6 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
Industrials;
Information technology, Telecom and media;
Mining and resources;
Utilities.
Energy; chemicals and drugs; mining and
resources (e.g. pulp and paper), and utilities are
considered to be environmentally-sensitive indus-
tries (cf. Cho & Patten, 2007).
Empirical models
Based on Al-Tuwaijri et al. (2004), we posit that
a ?rm’s environmental communication strategy
simultaneously a?ects perceived environmental
legitimacy, the extent of environmental news cov-
erage and the content of environmental communi-
cation. As such, we endogenize the related
variables in simultaneous equations models. The
following two sets of structural equations models
summarize the approach adopted in the empirical
analysis. The ?rst set integrates a ?rm’s annual
report environmental disclosure while the second
set integrates a ?rm’s press releases concerning
environmental matters:
Environmental legitimacy
itþ1
¼ ðb
0
Annual report environmental disclosure
þ b
1
Annual report environmental disclosure
ÂEnvironmentally-sensitive industries
þ b
2
Environmentally-sensitive industries
þ b
3
Legitimacy þ b
4
Size
þ b
5
Environmental performanceÞ
it
ð1:1Þ
Annual report environmental disclosure
it
¼ ðb
0
Lag Legitimacy
þ b
1
Environmentally-sensitive industries
þ b
2
Leverage þ b
3
Return on assets
þ b
4
Environmental news exposure
þ b
5
Environmental performance
þ b
5
CountryÞ
it
ð1:2Þ
Environmental news exposure
it
¼ ðb
0
Return on assets þ b
1
Foreign listing
þ b
2
Size þ b
3
Environmental performance
þ b
4
Industry þ b
5
CountryÞ
it
ð1:3Þ
Environmental legitimacy
itþ1
¼ ðb
0
Press releases þ b
1
Press releases
 Environmentally-sensitive industries
þ b
2
Environmentally-sensitive industries
þ b
3
Legitimacy þ b
4
Size
þ b
5
Environmental performanceÞ
it
ð2:1Þ
Press releases
it
¼ ðb
0
Lag Legitimacy
þ b
1
Environmentally-sensitive industries
þ b
2
Leverage þ b
3
Return on assets
þ b
4
Environmental news exposure
þ b
5
Environmental performance
þ b
6
CountryÞ
it
ð2:2Þ
Environmental news exposure
it
¼ ðb
0
Return on assets
þ b
1
Foreign listing þ b
2
Size
þ b
3
Environmental performance
þ b
4
Industry þ b
5
CountryÞ
it
ð2:3Þ
Description of dependent variables
Environmental legitimacy
To assess corporate environmental legitimacy,
we rely on a content analysis of press media cover-
age for corporate environmental issues. Institu-
tionalists suggest that content analysis of press
media sources may be particularly useful in study-
ing legitimation processes, as detailed archives of
media coverage exist for many industries and ana-
lyzing the content of those public sources o?ers a
potentially powerful technique for operationaliz-
ing legitimation (Baum & Powell, 1995).
News media content is extracted from the ABI/
Inform Global database and from three distinct
sources:
(1) Business, economics: local and regional busi-
ness publications (local and regional business
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 7
news coverage of large corporations, privately
held companies, local start-ups, executive pro-
?les, marketing, ?nance, and industry news.
ABI Inform provides access to business infor-
mation not typically found in national news
sources. It contains news and analysis, informa-
tion on local markets, and more data gathered
from major business tabloids, magazines, daily
newspapers, wire services, and city, state, and
regional business publications);
(2) Business, ?nance, economics: journals, com-
pany pro?les, Wall Street Journal (most schol-
arly and comprehensive way to explore and
understand business research topics. It includes
nearly 1800 worldwide business periodicals for
in-depth coverage of business and economic
conditions, management techniques, theory,
and business practices, advertising, marketing,
economics, human resources, ?nance, taxation,
computers, and more. It constitutes extensive
international coverage with quick access to
information on more than 60,000 companies
with business and executive pro?les);
(3) Canadian newsstand, which o?ers unparal-
leled access to the full text of Canadian newspa-
pers (Montreal Gazette, National Post and
Toronto Star). We extracted articles using a
?rm’s name and the following keywords: ‘‘envi-
ronment”, ‘‘sustainable development”, ‘‘recy-
cling”, ‘‘pollution”, ‘‘toxic”, ‘‘ISO14000”,
‘‘conservation”, ‘‘remediation”, ‘‘spills”, ‘‘waste
management”, ‘‘energy”, ‘‘awards”, ‘‘environ-
mental audit”.
The legitimacy measure is computed for 2003,
2002 (lag measure in legitimacy models) and
2001 (lag measure in disclosure models). In the
summer of 2005, two research assistants found
413 articles: 226 provided good news, 172 provided
bad news and 25 provided neutral news. Each arti-
cle was coded in terms of its impact on the ?rm’s
environmental legitimacy, i.e. neutral, negative,
or positive (See Appendix 1). Good news stories
are those that convey environmental commitment
and that emphasize the positive aspects of a ?rm’s
activities. Examples of good news stories include
investment in facilities that will reduce energy con-
sumption, or the reduction of greenhouse gas
emissions. The legitimacy score is computed based
on this coding. The two coders agreed on 81%
good news, 87% bad news and 96% neutral news.
Internal consistency estimates (Cronbach’s alpha)
computed over the 2001 and 2003 period show
that the variance between the two coders’ scores
is quite systematic (alpha = 0.888 for good news,
0.926 for bad news and 0.864 for neutral news).
This suggests a high level of intercoder reliability
(Weber, 1990). A researcher reconciled all coding
disagreements between the two coders.
Annual legitimacy measures were calculated
using the Janis–Fadner coe?cient of imbalance
(Bansal & Clelland, 2004; Janis & Fadner, 1965).
The Janis–Fadner coe?cient ranges from À1.0 to
+1.0; a high presence of favourable articles in a
given year yields a value closer to +1.0, and a high
presence of unfavourable articles yields a value
closer to À1.0. The formula is as follows:
Janis–Fadner coefficient ¼
ðe
2
À ecÞ
t
2
if e > c
ðec À c
2
Þ
t
2
if c > e
where e is the number of favourable environmental
articles in a given year, c is the number of unfa-
vourable environmental articles in a given year
and t is the sum of e and c.
Annual report environmental disclosure
Environmental disclosure is measured using a
coding instrument in a manner similar to Wiseman
(1982), Cormier and Magnan (2003), and Al-
Tuwaijri et al. (2004). The grid comprises 39 items
measuring environmental disclosure quality where
the items are grouped into six categories as fol-
lows: expenditures and risk; laws and regulations;
pollution abatement; sustainable development;
land remediation; and environmental manage-
ment. The rating is based on a score from one to
three. Three points are awarded for an item
described in monetary or quantitative terms, two
are awarded when an item is described speci?cally,
and one is awarded for an item discussed in gen-
eral. The information is coded according to the
grid presented in Appendix 1.
We believe that the use of a coding scale to
qualify a ?rm’s environmental disclosure is appro-
8 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
priate for the following reasons. First, it allows for
some integration of di?erent types of information
into a single ?gure that is comparable across ?rms
in terms of relevance. Second, while other disclo-
sure studies rely on word counts to measure
environmental disclosure (e.g. Neu et al., 1998;
Williams & Ho Wern Pei, 1999), a qualitative scale
allows for the researcher’s judgment to be utilized
in rating the value or quality of the disclosures
made by a ?rm. While this process is more subjec-
tive, it ensures that irrelevant or redundant gener-
alities are not considered strategic environmental
disclosures.
To ensure consistency across ?rms, two persons
reviewed all individual scores independently. All
disagreements were subsequently reviewed by one
of the co-researchers.
4
Press releases
We collected press releases related to environ-
mental information from ?rms’ web sites in 2002
for 144 ?rms. Since the information was not avail-
able on the web page for 21 ?rms, we completed
the data collection using the Lexis/Nexis database.
We searched for press releases using the same key-
words used for articles pertaining to environmen-
tal legitimacy. Two research assistants identi?ed
236 press releases, 153 classi?ed as proactive
accounts and 83 as reactive accounts. Proactive
accounts tend to bolster the positive environmen-
tal image of a ?rm through descriptive statements
that describe various aspects of a ?rm’s environ-
mental policy and accomplishments that re?ect
positively upon the ?rm and through the use of
acclaiming tactics, including enhancements
intended to boost the apparent desirability of spe-
ci?c environmental events or of ?rm-speci?c envi-
ronmental actions and entitlements intended to
heighten the ?rm’s perceived responsibility for
the events or actions (Gardner & Martinko,
1988). Reactive accounts are verbal remedial
tactics and are a reaction to environmental predic-
aments, i.e. situations with undesirable implica-
tions for a ?rm’s environmental image. They are
mainly defensive in nature and tend to attenuate
the negative meaning of events or outcomes.
Excuses, justi?cations and causality denials are
the more traditional forms of defensive impression
management tactics. To put a positive spin on
events, these traditional forms of impression man-
agement are frequently accompanied by selective
categorizations highlighting alternate (positive)
attributes of a ?rm’s environmental position and
related entitlement and enhancement tactics. The
coders agreed on 77% of proactive news and 89%
of reactive news. Internal consistency estimates
(Cronbach’s alpha) show that the variance
between the two coders’ scores is quite systematic
(alpha = 0.825 for proactive news and 0.834 for
reactive news). A researcher reconciled all coding
disagreements between the two coders. Smaller
disagreements were resolved by the two coders
themselves.
Environmental news exposure
The importance of news exposure in determin-
ing environmental disclosure indicates that ?rms
need to achieve social legitimacy with their envi-
ronmental management, i.e. their ultimate intent
is strategic.
4
A coding manual documenting coding instructions as well
as standardized coding worksheets were prepared beforehand.
Each coder then applied the following coding sequence: (1)
independent identi?cation of the occurrence of items relative to
the di?erent coding categories; (2) independent coding of the
items according to quality level of content and (3) timed
reconciliation on a subset of company reports. The coders were
intensively trained in applying coding instructions and in using
the coding worksheets. They were unaware of the research
hypotheses. Initial di?erences in identifying grid items
accounted for, on average, 6% of the maximum number of
items identi?ed. Of the information quality level coding, less
than 10% had to be discussed for reconciliation. Disagreement
between coders mostly occurred at the beginning of the coding
process (essentially the ?rst 20 ?rms in each country and the
?rst 5 ?rms by industry). A researcher reconciled coding
disagreements exceeding 5% of the highest total score between
the two coders. Smaller disagreements were resolved by the two
coders themselves. Overall, we think that this coding process
provides a reliable measure of environmental reporting. Inter-
nal consistency estimates (Cronbach’s alpha on score compo-
nents) show that the variance is quite systematic
(alpha = 0.803). This is higher than Botosan (1997), who ?nds
an alpha of 0.64 for an index including ?ve categories of
disclosure in annual reports. Cronbach’s alpha estimates the
proportion of variance in the test scores that can be attributed
to a true score variance. It can range from 0 (if no variance is
consistent) to 1.00 (if all variances are consistent). According to
Nunnaly (1978), a score of 0.70 is acceptable.
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 9
The concerns of government and local commu-
nities are di?cult to ascertain directly. However,
prior work does suggest that environmental news
exposure is an appropriate proxy for community
concerns (Deegan & Rankin, 1996). A ?rm’s envi-
ronmental news exposure is computed by taking
the average number of articles concerning environ-
mental issues for 1998 through 2002, as contained
in the ABI/Inform Global database. We searched
for articles using the keywords mentioned above.
The reason for this choice is that disclosure this
year (2002) may be a?ected by the number and
types of articles that have been published about
a ?rm in the recent past. A total of 764 relevant
stories were identi?ed over the period. We expect
that as environmental news exposure increases,
the ?rm will increase its environmental disclosure.
Hence, a positive relationship is expected between
environmental media coverage and environmental
disclosure.
Determinants of environmental legitimacy
Size
Size has been shown to be an antecedent of
legitimacy (Baum & Oliver, 1991; Deephouse &
Carter, 2005). Firm size will a?ect the ?rm’s visi-
bility to the general public and tends to engender
increased public scrutiny. Firm size, measured as
ln(Assets), is introduced as a control variable, with
no directional prediction.
Prior environmental legitimacy
Reputation and legitimacy issues have been
argued to be largely sticky (Schultz, Mouritsen,
& Gabrielsen, 2001). Like reputation, legitimacy
can be inertial or durable and have the tendency
to reproduce itself over time. Hence, the lagged
environmental legitimacy variable is introduced
to capture the inertia factor. Adding the lagged
dependent variable also implies control for omit-
ted ?rm characteristics, including the fact that spe-
ci?c environmental norms and value expectations
may vary from industry to industry. Alternatively,
we could have expressed the dependent variable as
a change variable. We chose not to do so because
such a procedure constrains the coe?cient of the
lagged variable to equal one and we preferred to
model the inertia factor as a theoretically relevant
determinant.
Environmental performance
Documentary evidence of a ?rm’s polluting
activities might be directly discounted in how the
media perceive a ?rm’s environmental posture.
Environmental performance is proxied by the Tox-
ics Release Inventory (TRI), a public database
available from the Environmental Protection
Agency (EPA) in the US and from Environment
Canada. These databases contain information on
toxic chemical releases and other waste manage-
ment activities reported annually by manufactur-
ing facilities. The Toxics Release Inventory is the
sum of all chemicals released in the air and water
and on land in 2002. Consistent with Clarkson,
Li, Richardson, and Vasvari (2006), the measure
is computed by summing all facilities for an indi-
vidual company in pounds de?ated by $1000 of
sales. Higher values of the variable imply a worse
environmental performance.
Determinants of corporate environmental
communication
Prior environmental legitimacy
In line with strategic legitimacy theory argu-
ments, we expect a negative relationship between
prior media legitimacy and corporate environmen-
tal communication.
Leverage
It is expected that ?rms able to withstand
potential proprietary costs from the disclosure of
environmental information and bene?tting from
more open disclosure (?rms in good ?nancial con-
dition) are likely to outweigh the costs from the
disclosure of environmental information. By
widely disseminating information about their envi-
ronmental management and showing their ability
to shoulder environmental obligations, these ?rms
establish their credibility as a reliable and socially
responsible partner. Roberts (1992) and Richard-
son and Welker (2001) ?nd a positive relationship
between leverage and social disclosure while Eliji-
do-Ten (2004) does not ?nd any signi?cant rela-
tionship between leverage and environmental
10 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
disclosure. Conversely, Cormier and Magnan
(2003) document a negative relationship between
leverage and environmental disclosure.
5
We mea-
sure leverage by the ratio of long-term ?nancial
debt over equity (Long-term ?nancial debt/
equity). Since the actual impact of leverage on
environmental disclosure is unclear, no directional
predictions are made for the variable.
Return on assets
Many studies document a positive association
between a ?rm’s level of disclosure and its ?nancial
performance (Cochran & Wood, 1984; Cormier &
Magnan, 1999, 2003; McGuire, Sundgren, &
Schneeweis, 1988; Mills & Gardner, 1984). Firms
with superior earnings performance have a higher
propensity to reveal their ‘‘good news”. Hence,
Murray, Sinclair, Power, and Gray (2006) docu-
ment that ?rms with consistently higher returns
tend to have higher levels of total and voluntary
social and environmental disclosure. In this vein,
we expect a positive relationship between pro?t-
ability, as measured by return on assets, and envi-
ronmental disclosure.
Media exposure
A number of studies document that higher lev-
els of media exposure relative to environmental
issues increase public concerns and thus public
policy pressure, to which companies react through
greater environmental disclosure (Bewley & Li,
2000; Brown & Deegan, 1998; Deegan et al.,
2000; Li, Richardson, & Thornton, 1997; Patten,
2002a).
Environmental performance
Many authors examine the association between
environmental disclosure in annual reports and a
?rm’s environmental performance. Results are
mixed. Ingram and Frazier (1980), Jaggi and
Freedman (1982), Wiseman (1982), Rockness
(1985), Freedman and Wasley (1990), Fekrat,
Inclan, and Petroni (1996) do not ?nd a signi?cant
association between environmental disclosure (in
the annual report or in the 10K report) and the
CEP index of environmental performance while
Patten (2002b) establishes a negative relationship.
More recent works document a positive associa-
tion between environmental performance and the
extent of discretionary environmental disclosures
(Al-Tuwaijri et al., 2004; Clarkson et al., 2006).
According to Al-Tuwaijri et al. (2004), a positive
relationship conjectures that prior literature’s
mixed results describing their interrelations may
be attributable to the fact that researchers have
not jointly considered Environmental disclosure,
Environmental performance, and Economic per-
formance. Since the actual impact of environmen-
tal performance on environmental disclosure is
unclear, no directional predictions are made.
Environmentally-sensitive industries
Companies in environmentally-sensitive indus-
tries are generally subject to greater environmental
scrutiny than other companies (Cowen, Ferreri, &
Parker, 1987; Hackston & Milne, 1996; Patten,
1991) and have been documented to exhibit higher
levels of environmental disclosure.
Determinants of environmental news exposure
We introduce ?ve variables that determine a
?rm’s exposure to environmental news and its
environmental ‘‘riskiness”: Firm size; Return on
assets; Foreign listings; Capital intensity, and
Environmental performance.
Firm size: Prior evidence is consistent in show-
ing a positive relationship between the extent of
media coverage and ?rm size (e.g. Caroll &
McCombs, 2003; Deephouse, 2000; Schultz et al.,
2001). We predict a positive relationship between
size and environmental news exposure.
Return on assets: Since ?rms with consistently
higher returns tend to have higher levels of envi-
ronmental disclosure, we expect that such disclo-
sure will attract environmental news media.
Foreign listings: The degree to which ?rms are
listed internationally can in?uence news coverage
and public awareness. It may also be an indicator
of diversi?cation. Based on Hope’s study (2003), a
5
An explanation for the inverse relationship (positive asso-
ciation for social disclosure and negative association for
environmental disclosure) could be that social disclosure is
more likely to be good news than environmental disclosure.
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 11
listing on a domestic exchange and on foreign
exchanges (except U.S. and London Stock
Exchange listings) are given a weight of 1 per list-
ing; London Stock Exchange and U.S. listings are
given a weight of 1.5 because of their importance.
The score for each ?rm is summed. We expect a
positive association between the variable stock
exchange listings and the level of environmental
news exposure.
Environmental performance: A ?rm’s risk of
unfavourable environmental exposure is primarily
determined by the relative perceived environmen-
tal risk of its activities. The poorer a ?rm’s
environmental performance, the greater its envi-
ronmental exposure and the more this will attract
media attention.
Results
Univariate results
As illustrated in Table 1, the level of annual
report environmental disclosure varies from a
mean score of 13.05 for Technology, Telecom
and Media to 76.80 for Utilities. Among the seven
industries, the four industries for which ?rms’
activities are more likely to a?ect the environment
exhibit the higher environmental disclosure scores:
Utilities 76.80; Energy 72.50; Mining and
resources 67.54; and to a lesser extent Chemicals
and drugs 36.44. This result is consistent with Pat-
ten (2002b), who ?nds that those ?rms operating
in environmentally-sensitive industries report
more environmental information. On average,
US ?rms publish more environmental information
than Canadian ?rms except for sustainable devel-
opment and environmental management.
Table 2 provides some descriptive statistics
regarding the sample ?rms’ dependent and inde-
pendent variables. The 2003 mean legitimacy score
is positive at 0.135. The 2002 mean environmental
news exposure approaches one article per ?rm. As
expected, there are twice as many proactive press
releases compared to reactive ones. Firm size, on
average, is quite large with mean total asset for
the sample at 27.5 billion Canadian dollars. Aver-
age ?rm size is larger for US ?rms and Canadian T
a
b
l
e
1
E
n
v
i
r
o
n
m
e
n
t
a
l
l
e
g
i
t
i
m
a
c
y
a
n
d
a
n
n
u
a
l
r
e
p
o
r
t
e
n
v
i
r
o
n
m
e
n
t
a
l
d
i
s
c
l
o
s
u
r
e
m
e
a
n
s
c
o
r
e
s
b
y
i
n
d
u
s
t
r
y
S
a
m
p
l
e
L
e
g
i
t
i
m
a
c
y
E
x
p
e
n
d
i
t
u
r
e
s
a
n
d
r
i
s
k
s
L
a
w
s
a
n
d
r
e
g
u
l
a
t
i
o
n
s
c
o
n
f
o
r
m
i
t
y
P
o
l
l
u
t
i
o
n
a
b
a
t
e
m
e
n
t
S
u
s
t
a
i
n
a
b
l
e
d
e
v
e
l
o
p
m
e
n
t
L
a
n
d
r
e
m
e
d
i
a
t
i
o
n
a
n
d
c
o
n
t
a
m
i
n
a
t
i
o
n
E
n
v
i
r
o
n
m
e
n
t
a
l
m
a
n
a
g
e
m
e
n
t
T
o
t
a
l
C
o
n
s
u
m
e
r
g
o
o
d
s
a
n
d
s
e
r
v
i
c
e
s
3
4
0
.
0
0
0
5
.
8
9
3
.
2
8
7
.
0
0
2
.
1
0
3
.
1
4
4
.
2
8
2
5
.
6
9
E
n
e
r
g
y
2
0
0
.
0
8
1
1
8
.
6
6
9
.
3
6
1
2
.
6
7
2
.
9
4
1
8
.
8
1
1
0
.
0
3
7
2
.
5
0
C
h
e
m
i
c
a
l
s
a
n
d
d
r
u
g
s
9
0
.
0
0
0
8
.
4
1
4
.
7
7
8
.
0
1
2
.
1
1
8
.
6
8
4
.
4
7
3
6
.
4
4
I
n
d
u
s
t
r
i
a
l
s
1
3
0
.
5
3
8
1
4
.
1
4
8
.
4
1
1
1
.
9
0
3
.
4
9
1
8
.
0
0
4
.
6
8
6
0
.
6
3
I
n
f
o
r
m
a
t
i
o
n
t
e
c
h
n
o
l
o
g
y
,
T
e
l
e
c
o
m
a
n
d
M
e
d
i
a
1
5
0
.
2
0
0
0
.
1
6
1
.
0
3
4
.
0
3
1
.
7
3
1
.
5
3
4
.
5
9
1
3
.
0
5
M
i
n
i
n
g
a
n
d
r
e
s
o
u
r
c
e
s
4
8
0
.
1
0
6
1
7
.
0
4
8
.
5
3
1
1
.
3
4
4
.
7
4
1
2
.
0
6
1
3
.
8
1
6
7
.
5
4
U
t
i
l
i
t
i
e
s
1
9
0
.
3
0
9
1
6
.
8
5
1
4
.
0
4
1
9
.
3
8
5
.
5
2
1
4
.
9
5
6
.
0
5
7
6
.
8
0
T
o
t
a
l
1
5
8
0
.
1
4
1
1
2
.
5
6
7
.
2
7
1
0
.
7
5
3
.
5
0
1
0
.
8
1
8
.
1
6
5
3
.
0
6
C
a
n
a
d
a
3
9
0
.
0
8
0
1
0
.
8
0
2
.
7
5
1
0
.
7
5
3
.
8
7
5
.
9
6
1
1
.
0
8
4
5
.
2
1
U
S
1
1
9
0
.
1
5
9
1
3
.
1
7
8
.
8
3
1
0
.
7
5
3
.
3
8
1
2
.
4
6
7
.
1
5
5
5
.
7
6
12 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
?rms show, on average, lower systematic risk.
Media exposure and number of press releases are
much higher for US ?rms while Canadian ?rms
show, on average, higher levels of pollution and
lower media legitimacy.
Table 3 presents correlations for the legitimacy
model, the disclosure model and the news exposure
model. Reactive press releases (0.179) and Environ-
mental performance (À0.109) are signi?cantly
correlated with media legitimacy. Annual report
environmental disclosure is correlated with
Environmentally-sensitive industries (0.267), Envi-
ronmental news exposure (0.318), and Environ-
mental performance (0.171). As expected,
environmental communication is negatively
correlated with lagged media legitimacy, especially
for press releases (À0.214 for total press releases,
À0.171 for proactive press releases and À0.223
for reactive press releases). Finally, environmental
news exposure is correlated with size (0.393), For-
eign listing (0.401), and Environmental perfor-
mance (0.129).
Multivariate results
Selection bias
From the sample of 623 ?rms, we ?nd environ-
mental media coverage for 165 ?rms. Because
legitimacy is only measured for ?rms covered by
the media, there might be a problem in terms of
selection bias (Heckman, 1979). To correct this
potential bias, Heckman’s two-step procedure
was used. In the Heckman procedure (Heckman,
1979; Lee, 1983), the residuals of the selection
Table 3A
Correlations – legitimacy model
2 3 4 5 6 7 8 9 10
1 Legitimacy 0.220
*
0.028 À0.044 À0.001 À0.156
*
0.012 0.073 0.179
*
À0.109
*
2 Lag legitimacy 1 0.073 À0.097 À0.087 À0.092 À0.159
*
0.025 0.011 À0.044
3 Size 1 0.110
*
0.149
*
À0.039 À0.275
*
0.358
*
0.219
*
0.004
4 Total disclosure 1 0.974
*
0.743
*
0.267
*
0.125
*
0.072 0.171
*
5 Economic-based disclosure 1 0.572
*
0.248
*
0.144
*
0.074 0.109
*
6 Social-based disclosure 1 0.237
*
0.027 0.035 0.295
*
7 Sensitive industry 1 À0.083 À0.076 0.196
*
8 Press release – proactive 1 0.643
*
0.003
9 Press releases – reactive 1 0.008
10 Environmental performance 1
*
Signi?cant at 0.10 two-tailed.
Table 2
Descriptive statistics – ?nancial, legitimacy, and environmental news exposure variables
Min. Max. Mean Std. Dev. Mean
US sample
Mean
Canadian sample
Legitimacy À1 1 0.14 0.47 0.16 0.08
Legitimacy – highly sensitive industries À1 1 0.11 0.48 0.13 0.06
Environmental news exposure (year average) 0.17 7.33 0.68 1.12 0.72 0.57
Press releases – proactive 0 18.00 0.96 2.24 1.09 0.56
Press releases – reactive 0 18.00 0.52 1.80 0.60 0.28
Environmental performance 0 17.29 1.03 2.72 0.58 2.67
Toxic release inventory (pounds per $1000 of sales)
Total Assets (million Can $) 349 575 000 27 500 59 800 32 700 8835
Leverage (debt/assets) 0 0.64 0.24 0.13 0.24 0.22
Pro?tability (ROA) À0.49 0.20 0.03 0.08 0.04 0.02
Foreign listing 0 4 0.97 1.27 0.94 1.08
N 158 119 39
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 13
equation in a probit/logit analysis (news exposure/
no news exposure) are used to construct a selec-
tion bias control factor, i.e. the inverse mills
ratio:
6
Expected value of news exposure=no news exposure
¼aþa
1
Return on assets þa
2
Foreign listing
þa
3
Firmsize þa
4
Capital intensity
þa
5
Countryþa
6À10
Industry
Table 4 shows descriptive statistics for the con-
trol variables relating to the selected sample and
the rest of the population with no media coverage
over the selected period. We ?nd that all the vari-
ables exhibit larger mean values for the ?rms
selected and the di?erences are all statistically sig-
ni?cant. The third column in Table 5 gives the
results of the ?rst-stage Logit regression (industry
dummies not shown).
7
The model is well speci?ed
with a pseudo-R-square of 37.0% and a classi?ca-
Table 3B
Correlations – disclosure model
1 2 3 4 5 6 7 8 9 10 11 12
1 Total disclosure 1 0.974
*
0.744
*
0.111
*
0.125
*
0.070 À0.054 0.076 À0.052 0.267
*
0.318
*
0.171
*
2 Economic-based
disclosure
1 0.571
*
0.125
*
0.144
*
0.074 À0.013 0.093 À0.057 0.248
*
0.294
*
0.109
*
3 Social-based
disclosure
1 0.034 0.027 0.035 À0.142
*
0.001 À0.020 0.238
*
0.280
*
0.295
*
4 Total Press
release
1 0.927
*
0.884
*
À0.214
*
0.127
*
À0.017 À0.088 0.407
*
0.006
5 Press release –
proactive
1 0.643
*
À0.171
*
0.075 0.032 À0.083 0.338
*
0.004
6 Press releases –
reactive
1 À0.223
*
0.157
*
À0.075 À0.076 0.409
*
0.008
7 Lag legitimacy 1 À0.092 0.115
*
À0.169
*
À0.131
*
0.091
8 Leverage 1 À0.071 À0.081 0.168
*
À0.048
9 Return on assets 1 À0.057 À0.089 0.018
10 Sensitive industry 1 0.039 0.196
*
11 Environmental
news exposure
1 0.129
*
12 Environmental
performance
1
*
Signi?cant at 0.10 two-tailed.
Table 3C
Correlations – news exposure model
1 2 3 4 5
1 News exposure 1 À0.089 0.393
*
0.401
*
0.129
*
2 Return on assets 1 0.133
*
À0.111
*
0.018
3 Size 1 0.244
*
0.004
4 Foreign listing 1 À0.030
5 Environmental
performance
1
*
Signi?cant at 0.10 two-tailed.
6
Since we did not collect data on environmental performance
for the remaining sample, we replace that variable by capital
intensity. Physical plant and equipment makes a ?rm much
more visible to the public and the community at large.
Moreover, a ?rm’s capital intensity is likely to be related to
polluting activities.
7
A disadvantage of the PROBIT procedure is that this
procedure cannot directly compute predicted values. Lee (1983)
has developed a method to estimate the selection model with a
logit analysis that o?ers a less labour-intensive alternative for
computing LAMBDA. Hence, we compute LAMBDAbased on
the following procedure: (1) save predicted probabilities using
LOGIT regression (IKL); (2) using the inverse cumulative
distribution function of the normal distribution, these individual
probabilities are translated into the form they would have had
when they would have been computed based on a probit model
(IPS = probit (IKL)); (3) the IPS variable nowincludes the quasi-
probit scores and can be used to compute LAMBDAin the same
way as when using a probit selection model: LAMBDA = ((1/
sqrt(2 Ã 3.141592654)) Ã (exp (ÀIPS Ã IPSÃ0.5)))/cdfnorm(IPS).
14 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
Table 5
LOGIT and 3SLS regressions on the determinants of environmental legitimacy, environmental news exposure and environmental
disclosure
LOGIT 3SLS
Predicted sign Environmental
news exposure 1/0
Legitimacy Disclosure Environmental
news exposure
Legitimacy
Disclosure + 0.005
**
Disclosure  Environmentally-sensitive industries À À0.005
**
Environmentally-sensitive industries ± 0.127
Environmental performance À À0.006
***
Lag Legitimacy + 0.144
***
Size ± À0.045
Disclosure
Lag legitimacy À 7.174
Environmentally-sensitive industries + 46.986
***
Leverage ± À9.641
Return on assets + À8.086
Environmental news exposure + 32.773
***
Inverse Mills Ratio ± 2.810
Environmental performance ± 0.130
Canada ± À19.076
**
Environmental news exposure
Return on assets + 0.775 À0.569
Foreign listing + 0.194
**
0.390
***
Firm size + 0.881
***
0.317
***
Environmental performance + 0.009
**
Capital intensity ± 0.487
*
Canada ± À0.718
***
À0.346
*
Nagelkerke R-square
Chi-square = 182.24 (0.000)
Overall classi?cation rate = 80.1% 37.0%
R-Square 24.3% 13.6% 32.8%
Chi-square p-value 0.000 0.000 0.000
N 616 158 158 158
One-tailed if there is a predicted sign, two-tailed otherwise. Coe?cients for industry-speci?c dummies not reported.
*
p < 0.10.
**
p < 0.05.
***
p < 0.01.
Table 4
Test of self-selection bias – control variable mean values
Selected sample
(with media coverage)
Remaining sample
(without media coverage)
Mean di?erence
p-value
Return on assets 0.032 0.009 0.088
Foreign listing 0.857 0.697 0.013
Firm size (lnassets) 23.323 22.031 0.000
Capital intensity 0.444 0.341 0.000
N 165 458
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 15
tion rate of 80.1%. All coe?cients except for
Return on assets are statistically signi?cant. This
might be an indication of the presence of selection
bias.
In the second step of the Heckman procedure,
based on the expected probability value, we use
the selection bias control factor (Inverse Mills
Ratio–Lambda) as an additional independent
variable to control for the selection bias in corpo-
rate communication regressions.
Simultaneous test of legitimacy, environmental
disclosure, and environmental news exposure
Since we posit that a ?rm’s communication
strategy a?ects environmental legitimacy, environ-
mental communication, and environmental media
exposure, we ?rst assess whether or not endogene-
ity exists between these variables using the Haus-
man test. Results con?rm endogeneity between
Environmental legitimacy and Annual report envi-
ronmental disclosure (t = 1.839; p < 0.068). We
also reject the null hypothesis of no endogeneity
with respect to Annual report environmental dis-
closure and Media exposure (t = 1.620; p <
0.100). Therefore, it is important to control for
?rms’ incentives to disclose environmental infor-
mation as well as the characteristics of ?rms facing
environmental media exposure in assessing the
determinants of a ?rm’s environmental legitimacy.
Table 5 provides evidence regarding the simul-
taneous test of environmental legitimacy, (Eq.
(1.1)), total annual report environmental disclo-
sure (Eq. (1.2)) and environmental news exposure
(Eq. (1.3)). Concerning the determinants of envi-
ronmental legitimacy, consistent with Hypothesis
1, there is a positive relationship between environ-
mental disclosure and environmental legitimacy
(0.005, p < 0.05). Consistent with Hypothesis 2,
the interaction term ‘‘Annual report disclosure Â
Environmentally-sensitive industries” is negative
and signi?cant (À0.005; p < 0.05) suggesting that
environmental disclosure has a lower impact on
legitimacy for those ?rms operating in more envi-
ronmentally-sensitive industries. Moreover, as
expected, Environmental performance (À0.006;
p < 0.01) is signi?cantly and negatively related to
a ?rm’s media legitimacy. Corroborating the sticky
character of media legitimacy, a ?rm’s prior legit-
imacy is signi?cantly related to its current level of
legitimacy (0.144; p < 0.01)
8
.
As for the determinants of environmental news
exposure, results show that a ?rm’s potential envi-
ronmental visibility, as measured by Foreign list-
ing (0.390; p < 0.01), Environmental performance
(0.009; p < 0.05) and ?rm size (0.317; p < 0.01)
lead to more environmental news. Concerning
the determinants of Annual report environmental
disclosure, self-selection bias related to news
exposure does not appear to be an issue in this
regression model. There is a signi?cant relation-
ship between Environmentally-sensitive industries
(46.986; p < 0.01), Environmental news exposure
(32.773; p < 0.01) and the total Annual report
environmental disclosure score. Furthermore,
total annual report environmental disclosure does
not seem to be in?uenced by prior media legiti-
macy, or by the level of ?rm’s polluting activities
as measured by the environmental performance
variable. The absence of a signi?cant relationship
between Environmental disclosure and Environ-
mental performance is not consistent with prevail-
ing legitimacy theory arguments, but could be
related to the content of the disclosures (Patten,
2002). In this respect, Cho and Patten (2007)
argue that the legitimizing nature of di?erent
types of annual report environmental disclosures
is not identical and that it is important to distin-
guish di?erent types of information when
assessing legitimation e?ects. They primarily dis-
tinguish between litigation related and non-litiga-
tion related disclosure. It is important to note that
they do not distinguish between such types of
information based on the information’s direct
legitimizing e?ect, but rather on the association
of environmental disclosures and environmental
performance. In that vein, their results indicate
that non-litigation related disclosure is higher
for ?rms with lower environmental performance
and for ?rms from environmentally-sensitive
industries.
8
As a sensitivity analysis, we added the Market-to-Book and
Return on assets variables to the model separately. None of the
coe?cients are signi?cant and they do not a?ect our results.
16 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
In our direct approach to measuring legitima-
tion e?ects, we believe the association of environ-
mental disclosure and environmental performance
is not so much an indicator of the legitimizing
nature of the information types, but a proxy for
the information’s intrinsic believability. The
higher the association between the type of envi-
ronmental disclosure and environmental perfor-
mance, the higher the message’s believability and
the greater its potential to a?ect legitimacy per-
ceptions. Although litigation related disclosures
are to a large extent mandatory, their counter-
parts (non-litigation related disclosures) qualify
more easily as incentive-consistent disclosures,
the believability of which users tend to discount
(Hirst et al., 1995; Hodge et al., 2006). Although
our coding grid does not incorporate a clear dis-
tinction between litigation related and non-litiga-
tion related information, we distinguish between
economic-based and social-based environmental
information, with social-based environmental dis-
closure being more incentive-consistent than eco-
nomic-based disclosure. This dichotomous split
of disclosure types resembles the distinction made
by Clarkson et al. (2006) between soft and hard
environmental disclosures, with hard disclosures
re?ecting factual, objective information that can-
not easily be mimicked by poor environmental
performers.
Economic-based types of information focus on
the ?nancial, legal and operational consequences
of corporate environmental activities, most of
which are required disclosures in 10-Ks for public
companies listed with the SEC. This type of
information is mainly comprised within the fol-
lowing four components of our content grid:
expenditure and risk; compliance with laws and
regulations; pollution abatement; and land reme-
diation and contamination, whereas social-based
information relates to the ‘sustainable develop-
ment’ and ‘environmental management’ grid cap-
tions. Disclosure about sustainable development
and environmental management is likely to be
more discretionary, less factual and objective,
and easier to imitate even without substance to
support the claims made. In that vein, social-
based environmental information is likely to be
more incentive-consistent and, thus, we might
expect this type of disclosure to have a lower
impact on a ?rm’s environmental legitimacy than
the more objective, economic-based disclosures.
In our sample, environmental disclosure is mostly
composed of economic-based information (78%
of total disclosures) while more social-related
information is reported by environmentally-sensi-
tive industries (28% of total disclosures for Min-
ing and resources). We estimate our model
separately for economic-based and social-based
disclosures.
The results presented in Table 6 suggest that
economic-based annual report environmental dis-
closure signi?cantly a?ects a ?rm’s media legiti-
macy (0.006; p < 0.05), but that ‘‘social-related”
environmental information segments (0.020;
p < 0.26 one-tailed) do not.
9
For economic-based
environmental disclosure, consistent with Hypoth-
esis 2, the interaction term ‘‘Economic-based
annual report disclosure  Environmentally-sensi-
tive industries” is signi?cant (À0.005; p < 0.05),
suggesting that environmental disclosure has a
marginally di?erent impact on legitimacy for those
?rms operating in more environmentally-sensitive
industries. Moreover, as expected, a bad environ-
mental performance negatively a?ects a ?rm’s
legitimacy (À0.006; p < 0.01). Concerning the
determinants of annual report environmental dis-
closure, a striking di?erence between the types of
content disclosed relates to the association of
Environmental performance and Annual report
disclosure score. Worse environmental perfor-
mance is associated with more elaborate social-
based environmental disclosures but not with the
economic-based variants.
9
The disclosure component Compliance with laws and regu-
lation comprises some items that could communicate negative
environmental image (e.g. ?nes, order to conform, incidents).
These litigation related disclosures are more mandated in nature
and less likely to be used as a legitimizing tool (Patten, 2002).
Therefore, the economic-based disclosure model is re-estimated
dropping the component Compliance with laws and regulation
from the disclosure score. Results remain similar as those
reported in Table 6 (Coe?cient for economic-based disclosure =
0.007; p < 0.050, and Economic-based disclosure  Environ-
mentally-sensitive industries = À0.006; p < 0.050).
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 17
Since visibility factors are likely to di?er across
US and Canadian ?rms, we test for potential inter-
active country di?erences of the impact of disclosure
on legitimacy. As a sensitivity analysis (results not
tabulated), we add two interaction terms to the
legitimacy model ‘‘Economic-based annual
report disclosure  Canada” and ‘‘Economic-based
annual report disclosure  Environmentally-sensi-
tive industries  Canada”. Coe?cients for these
two interaction terms are not signi?cant (À0.001;
p < 0.775 two-tailed, and -0.0008; p < 0.875 two-
tailed), suggesting that mixing US and Canadian
data does not bias our results.
Simultaneous test of legitimacy, environmental
press releases, and environmental news exposure
Table 7 provides evidence regarding the simul-
taneous test of environmental legitimacy, (Eq.
(2.1)), environmental press releases (Eq. (2.2))
Table 6
3SLS regressions on the determinants of environmental legitimacy, environmental news exposure and environmental disclosure
Predicted
sign
Economic-based (expenditures/
remediation/pollution abatement and
norms)
Sustainable development and
environmental management
Legitimacy Disclosure Environmental
news exposure
Legitimacy Disclosure Environmental
news exposure
Legitimacy
Disclosure + 0.006
**
0.020
Disclosure  Environmentally-
sensitive industries
À À0.005
**
À0.021
Environmentally-sensitive
industries
± 0.091 0.124
Environmental performance À À0.006
***
À0.006
***
Lag legitimacy + 0.151
***
0.130
**
Size ± À0.043 À0.049
Disclosure
Lag Legitimacy À 8.035 À0.866
Environmentally-sensitive
industries
+ 39.029
***
7.927
**
Leverage ± À0.725 À7.471
Return on assets + À14.821 5.846
Environmental news exposure + 23.918
***
8.630
***
Inverse Mills ratio ± À7.195 9.814
Environmental performance + 0.035 0.101
*
Canada ± À20.844
***
1.537
Environmental news exposure
Return on assets + À0.666 À0.505
Foreign listing + 0.379
***
0.426
***
Firm size + 0.342
***
0.285
***
Environmental performance + 0.009
**
0.009
**
Canada ± À0.322
*
À0.389
**
R-Square 27.3% 19.7% 34.8% 18.2% 6.3% 32.8%
F-statistic p-value 0.000 0.000 0.000 0.000 0.000 0.000
N 158 158 158 158 158 158
One-tailed if there is a predicted sign, two-tailed otherwise.
Coe?cients for industry-speci?c dummies not reported.
*
p < 0.10.
**
p < 0.05.
***
p < 0.01.
18 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
and environmental news exposure (Eq. (2.3)).
Total press releases are not associated with a ?rm’s
legitimacy (0.006; p < 0.306 one-tailed). In Table 8,
we present separate regressions for proactive and
reactive press releases. Consistent with Hypothesis
3, our results show that only reactive press releases
increase a ?rm’s perceived environmental legiti-
macy (0.029; p < 0.05) while there is no impact
for proactive press releases (À0.009; p < 0.652
two-tailed). However, the interaction term
‘‘Reactive Press releases  Environmentally-sensi-
tive industries” is negative but not signi?cant
(À0.001; p < 0.494 one-tailed), suggesting that the
impact of environmental press releases on legiti-
macy does not di?er for those ?rms operating in
more environmentally-sensitive industries. Con-
cerning the determinants of environmental press
releases, as expected, results show a negative
relationship between prior media legitimacy and
the use of environmental press releases (À0.617;
p < 0.05 for proactive press releases and À0.600;
p < 0.01 for reactive press releases), suggesting
tactical impression management motives behind
the use of press releases with the objective of
enhancing corporate environmental legitimacy
perceptions.
In order to explore whether annual report envi-
ronmental disclosures and environmental press
releases interfere in their e?ect on media legiti-
macy, we include both variables in simultaneous
Table 7
3SLS regressions on the determinants of environmental legitimacy, environmental news exposure and environmental press releases
Predicted
sign
Legitimacy Environmental
press releases
Environmental
news exposure
Legitimacy
Press releases + 0.006
Press releases  Environmentally-sensitive industries À 0.001
Environmentally-sensitive industries ± À0.010
Environmental performance À À0.006
***
Lag Legitimacy + 0.164
***
Size ± À0.027
Press releases
Lag Legitimacy À À1.216
***
Environmentally-sensitive industries + À0.161
*
Leverage ± 1.300
Return on assets + 1.652
Environmental news exposure + 1.367
***
Inverse Mills Ratio ± À2.025
Environmental performance ± 0.007
Canada ± À0.863
Environmental news exposure
Return on assets + À1.012
Foreign listing + 0.412
***
Firm size + 0.374
***
Environmental performance + 0.010
**
Canada ± À0.836
R-square 28.9% 20.7% 33.4%
F-statistic p-value 0.000 0.000 0.000
N 158 158 158
One-tailed if there is a predicted sign, two-tailed otherwise.
Coe?cients for industry-speci?c dummies not reported.
*
p < 0.10.
**
p < 0.05.
***
p < 0.01.
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 19
regressions of legitimacy, environmental press
releases and environmental disclosure. We restrict
the analysis to economic-based environmental dis-
closures. Including both economic-based annual
report disclosures and proactive press releases
(Table 9), we ?nd that proactive press releases do
not add to the legitimacy-enhancement e?ect of
annual report environmental disclosures, suggest-
ing that proactive press releases do not provide
incremental information to the other with respect
to environmental legitimacy enhancement. On
the other hand, including both economic-based
annual report disclosures and reactive press
releases (Table 9), the e?ect of annual report
environmental disclosure becomes insigni?cant,
while the coe?cient of reactive press releases
remains positive and signi?cant, suggesting a sub-
stitutional e?ect of reactive press releases with
regard to annual environmental disclosures. In line
with earlier results concerning the determinants of
environmental communication, we ?nd a negative
relationship between prior legitimacy and
press releases (À0.629; p < 0.05 for proactive press
releases and À0.598; p < 0.01 for reactive press
releases) while there is no signi?cant association
between prior media legitimacy and current annual
report economic-based disclosure.
Discussion and conclusion
Symbolic management theory (Pfe?er & Sala-
ncik, 1978) and related impression management
perspectives (Ginzel, Kramer, & Sutton, 1993;
Suchman, 1995) suggest that a ?rm’s management
is expected not only to manage the ?rm’s perfor-
mance (including its environmental performance),
but also perceptions of its (environmental) perfor-
mance. In that sense, a ?rm’s environmental
communication e?orts through environmental dis-
closures in annual reports and environmental press
releases represent predictable opportunities for
impression management and legitimation for the
?rm’s environmental activities and performance.
However, studying legitimation and legitimacy
issues empirically has its methodological chal-
lenges, such as the practical problems of assessing
subjective perceptions and beliefs of relevant pub-
lics. In this study, we investigated the impact of a
?rm’s environmental communication e?orts on
media legitimacy as a direct measure of perceived
legitimacy. Public media content captures the per-
spective of the general public and has been used as
a proxy for normative legitimacy issues (Bansal &
Clelland, 2004; Deephouse, 1996; Elsbach & Sut-
ton, 1992).
Previous environmental reporting studies
mainly focused on one communication channel
(annual report disclosures). However, in most
cases, annual report disclosures constitute only
one part of the ?rm’s communication repertoire
and the use of one channel may well a?ect the
use and e?ectiveness of other channels. Moreover,
di?erent channels may not be equally e?cient or
e?ective in reaching speci?c communication goals
(such as legitimacy enhancement). In this study,
we looked at both annual report environmental
disclosures and environmentally-related press
releases which may function in complementary or
substitutional roles.
The results obtained in this paper suggest that
perceived environmental legitimacy is positively
a?ected by the extent and quality of economic-
based segments of environmental disclosures in
annual reports and by reactive environmental
press releases. Consistent with Hypothesis 1, our
results document a signi?cant association between
the extent of annual report environmental disclo-
sures and reactive press releases on the one hand
and environmental (media) legitimacy on the
other. The association between annual report envi-
ronmental disclosures and media legitimacy is
essentially driven by the economic-based segments
of the annual report disclosures. We did not ?nd
evidence to support Hypothesis 1 with regard to
the extent of proactive press releases. The di?eren-
tial results with regard to the association of reac-
tive versus proactive press releases and media
legitimacy is consistent with Hypothesis 3 that pre-
dicts a stronger association for reactive environ-
mental press releases relative to proactive ones.
Relatedly, the interactive e?ect of corporate envi-
ronmental communication and environmentally-
sensitive industries (predicted in Hypothesis 2)
can only be demonstrated for the economic-based
segments of annual report environmental disclo-
20 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
sures. This suggests that industry-level legitimacy
functions as a constraining a priori impression
for annual report economic-based disclosures and
diminishes the legitimation e?ectiveness of these
communication e?orts, but not so of reactive press
releases.
Although not explicitly hypothesized, our
results shed some light on the complementary roles
of annual report environmental disclosures and
environmental press releases as they suggest that
proactive press releases do not interfere with the
e?ectiveness of annual report environmental dis-
closures, while the opposite is true for reactive
press releases. Moreover, in concert with the main
theme of the strategic legitimacy perspective on
environmental reporting, our results suggest that
Table 8
3SLS regressions on the determinants of environmental legitimacy, environmental news exposure and environmental press releases
Predicted
sign
Proactive press releases Reactive press releases
Legitimacy Environmental
press releases
Environmental
news exposure
Legitimacy Environmental
press releases
Environmental
news exposure
Legitimacy
Press releases + À0.009 0.029
**
Press releases Â
Environmentally-
sensitive industries
À 0.010 À0.001
Environmentally-
sensitive industries
± À0.017 À0.011
Environmental
performance
À À0.006
***
À0.006
***
Lag legitimacy + 0.165
***
0.165
***
Size ± À0.019 À0.032
Press releases
Lag Legitimacy À À0.617
**
À0.600
***
Environmentally-
sensitive industries
+ À0.038 À0.124
Leverage ± 0.076 1.205
Return on Assets + 2.118 À0.466
Environmental news
exposure
+ 0.819
***
0.546
***
Inverse mills ratio ± À0.139 À1.887
*
Environmental
performance
± 0.001 0.007
Canada ± À0.482 À0.355
Environmental news exposure
Return on assets + À0.906 À0.905
Foreign listing + 0.392
***
0.405
***
Firm size + 0.393
***
0.366
***
Environmental
performance
+ 0.008
**
0.008
**
Canada ± À0.295 À0.320
*
R-square 28.6% 13.8% 33.3% 30.1% 22.5% 33.5%
F-statistic p-value 0.000 0.000 0.000 0.000 0.000 0.000
N 158 158 158 158 158 158
One-tailed if there is a predicted sign, two-tailed otherwise.
Coe?cients for industry-speci?c dummies not reported.
*
p < 0.10.
**
p < 0.05.
***
p < 0.01.
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 21
negative media legitimacy is a signi?cant driver of
environmental press releases. This is, however, not
the case for annual report environmental disclo-
sures. Finally, environmental news exposure is
associated with both annual report environmental
disclosures and environmental press releases.
Table 9
3SLS regressions on the determinants of environmental legitimacy, environmental press releases, and economic-based environmental
disclosure
Predicted sign Proactive press releases Reactive press releases
Legitimacy Environmental
press releases
Environmental
disclosure
Legitimacy Environmental
press releases
Environmental
disclosure
Legitimacy
Disclosure + 0.006
**
0.001
Press releases + À0.048 0.035
**
Lag legitimacy + 0.176
***
0.166
***
Environmentally-
sensitive industries
± À0.193
*
À0.022
Environmental
performance
À0.007
***
À0.006
***
Size ± À0.035 À0.035
Press releases
Lag Legitimacy À À0.629
**
0.598
***
Environmentally-
sensitive
industries
+ À0.059 À0.121
Leverage ± 0.155 1.199
Return on Assets + 1.900 À0.440
Environmental news
exposure
+ 0.629
***
0.567
***
Inverse mills ratio ± À0.133 À1.885
*
Environmental
performance
± 0.003 0.009
Canada ± À0.529 À0.202
Environmental disclosure
Lag legitimacy À 5.836 6.312
Environmentally-
sensitive
industries
+ 36.984
***
37.150
***
Leverage ± 6.247 8.094
Return on assets + À27.299 À29.128
Lag environmental
news exposure
+ 10.572
***
10.638
***
Inverse mills ratio ± À13.035 À12.013
Environmental
performance
0.226 0.226
Canada À24.132
***
À24.494
***
R-square 28.4% 14.6% 29.1% 29.9% 22.5% 29.1%
F-statistic p-value 0.000 0.000 0.000 0.000 0.015 0.000
N 158 158 158 158 158 158
One-tailed if there is a predicted sign, two-tailed otherwise.
Coe?cients for industry-speci?c dummies not reported.
*
p < 0.10.
**
p < 0.05.
***
p < 0.01.
22 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
The present research especially highlights the
role of economic-based environmental informa-
tion in annual reports and of reactive press
releases as perception management tools. The
more subjective social-based environmental dis-
closure in annual reports does not a?ect media
legitimacy, whereas the more objective, harder
part of annual report environmental disclosure
does. Moreover, reactive environmental press
releases, through a mixture of verbal remedial tac-
tics, are found to be e?ective in transforming
events that seem at ?rst to be image-threatening
into messages that ultimately protect and even
enhance a ?rm’s environmental media legitimacy.
This is in line with previous impression manage-
ment research that suggested that di?erent forms
of verbal accounts a?ect legitimacy by attenuating
organizational responsibility for controversial
events and by accentuating the positive aspects
of such events. In particular, accounts with refer-
ences to widely institutionalized structures, pro-
grams and procedures (such as formal company
policies, budgetary constraints) are taken to be
e?ective in protecting legitimacy, because they
improve the adequacy of the verbal claims by pro-
viding evidence to bolster the verbal accounts
(Elsbach, 1994; Elsbach & Elofson, 2000). Future
research might look into the composition of (dif-
ferent sets and sequences of) verbal remedial tac-
tics, their form and content to identify the most
e?ective patterns.
The lack of impact of proactive press releases
and of the more subjective, social-based annual
report environmental disclosures on media legiti-
macy suggests that public media seem to discount
transparent self-promotional behavior, recogniz-
ing that ?rms tend to exaggerate their speci?c mer-
its in the environmental management domain. It
may be another occurrence of the ‘‘self-promoter’s
paradox”: as actors increase claims of competence,
audiences become more sceptical as competent
actors often downplay their successes (Jones &
Pittman, 1982). Alternatively, the lack of signi?-
cant association between the use of proactive press
releases and media legitimacy may also be
explained by the assertion that narcissistic behav-
ior is only what is generally expected of ?rms
and, as such, does not add to the assignment of
credit. This result is also consistent with the fact
that the softer part of annual report environmental
disclosures is not associated with enhanced media
legitimacy, whereas higher-quality economic-
based environmental disclosures do a?ect media
legitimacy, except when complemented by reactive
press releases.
Finally, the observation that mainly the extent
and not the evaluative propensity of public media
environmental exposure is associated with the
level and quality of annual report environmental
disclosure corroborates our initial assertion that
annual report environmental disclosures are
mainly driven by longer-term, strategic motives.
Annual report environmental disclosure does not
directly respond to the relatively current and tem-
porary legitimacy perceptions that transpire
through the public media. On the other hand,
the value as a legitimation device of environmen-
tal press releases is most evident in times of crisis
or controversy when environmental legitimacy is
challenged or threatened. Our results suggest that
in these circumstances both proactive and reactive
press releases are actively triggered, with only
reactive press releases being capable of signi?-
cantly mitigating legitimacy perceptions, while in
those circumstances the e?ect of annual report
environmental disclosures vanishes. These obser-
vations suggest that future research in corporate
environmental reporting might fruitfully distin-
guish between more enduring reputational e?ects
and relatively short-lived legitimacy e?ects based
on speci?c perceptions of environmental issues
(Elsbach, 2003) and of environmental communi-
cation devices. Substantiating such di?erential
e?ects would add to the corporate environmental
reporting literature.
Our measure of legitimacy might raise questions
about the generalizability of our ?ndings since it lar-
gely depends on the extent of comprehensiveness of
the media coverage database and the reliability of
measurements for that coverage. Editorial bias that
is systematically related to the independent vari-
ables of our models might also a?ect our results.
Despite these limitations, it is encouraging to ?nd
that with the di?culty in measuring data, we are
able to document direct legitimation e?ects of envi-
ronmental communication practices.
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 23
Acknowledgements
The authors acknowledge the ?nancial support
of the Social Sciences and Humanities Research
Council of Canada, the Fonds que´be´cois de la
recherche sur la socie´te´ et la culture (FQRSC),
the National Bank of Belgium and the Autorite´
des Marche´s Financiers (Que´bec) and the Corpo-
rate Reporting Chair (UQAM).
Appendix 1
Environmental Disclosure Grid
Expenditures and risks Sustainable development
Investments Natural resource conservation
Operation costs Recycling
Future investments Life cycle information
Future operating costs Land remediation and contamination
Financing for investments Sites
Environmental debts Remediation e?orts
Risk provisions Potential liability-remediation
Risk litigation Implicit liability
Provision for future expenditures Spills (number, nature, reduction e?orts )
Compliance with laws and regulations Environmental management
Litigation, actual and potential Environmental policies or company concern for the environment
Fines Environmental management system
Orders to comply Environmental auditing
Corrective action Goals and targets
Incidents Awards
Future legislation and regulations Department, group, service assigned to the environment
Pollution abatement ISO 14000
Emission of pollutants Involvement of the ?rm to develop environmental standards
Discharges Involvement of environmental organizations (industry
committees, etc.)
Waste management Joint environmental management projects with other ?rms
Installation and process controls
Compliance status of facilities
Noise and odours
Rating scale: 3: Item described in monetary or quantitative terms; 2: Item described speci?cally; 1: Item discussed in general.
References
Ader, C. R. (1995). A longitudinal study of agenda setting for
the issue of environmental pollution. Journalism and Mass
Communication Quarterly, 72(2), 300–311.
Aldrich, H. E., & Fiol, C. M. (1994). Fools rush in? The
institutional context of industry creation. Academy of
Management Review, 19(4), 645–670.
Al-Tuwaijri, S., Christensen, T. E., & Hughes, K. E. II, (2004).
The relations among environmental disclosure, environ-
mental performance, and economic performance: A simul-
taneous equations approach. Accounting, Organizations and
Society, 29, 447–471.
Anderson, U., Kadous, K., & Koonce, L. (2004). The role of
incentives to manage earnings and quanti?cation in audi-
tors’ evaluations of management-provided information.
Auditing: A Journal of Practice and Theory, 23(1), 11–27.
Ashforth, B. E., & Gibbs, B. W. (1990). The double-edge of
organizational legitimation. Organization Science, 1,
177–194.
24 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
Bansal, P., & Clelland, I. (2004). Talking trash: Legitimacy,
impression management, and unsystematic risk in the
context of the natural environment. Academy of Manage-
ment and Business Research, 29(1), 21–41.
Barton, J., & Mercer, M. (2005). To blame or not to blame:
Analysts’ reactions to external explanations for poor
?nancial performance. Journal of Accounting and Econom-
ics, 39(3), 509–533.
Baum, J. A. C., & Oliver, C. (1991). Institutional linkages and
organizational mortality. Administrative Science Quarterly,
36, 187–218.
Baum, Joel A. C., & Powell, W. W. (1995). Cultivating an
institutional ecology of organizations: comment on Han-
nan, Carroll, Dundon, and Torres. American Sociological
Review, 60, 529–539.
Bewley, K., & Li, Y. (2000). Disclosure of environmental
information by Canadian manufacturing companies: A
voluntary disclosure perspective. Advances in Environmental
Accounting and Management, 1, 201–226.
Birnbaum, M., & Stegner, S. (1979). Source credibility in social
judgment: Bias, expertise, and the judge’s point of view.
Journal of Personality and Social Psychology, 48–74.
Botosan, C. (1997). Disclosure level and the cost of equity
capital. The Accounting Review, 72, 323–349.
Brown, N., & Deegan, C. (1998). The public disclosure of
environmental performance information – A dual test of
media agenda setting theory and legitimacy theory.
Accounting and Business Research, 29(1), 21–41.
Buhr, N. (1998). Environmental performance, legislation and
annual report disclosure: The case of acid rain and
falconbridge. Accounting, Auditing and Accountability Jour-
nal, 163–190.
Caroll, C. E., & McCombs, M. (2003). Agenda-setting e?ects of
business news on the public’s images and opinions about
major corporations. Corporate Reputation Review, 6(1),
36–46.
Cho, C. H., & Patten, D. M. (2007). The role of environmental
disclosures as tools of legitimacy: A research note. Account-
ing, Organizations and Society, 32, 639–647.
Clarkson, P., Li, Y., Richardson, G. D., & Vasvari, F. P.
(2006). Revisiting the relation between environmental
performance and environmental disclosure: An empirical
analysis. In CAAA 2006 annual conference paper, April 5,
2006. Available at SSRN: .
Cochran, P., & Wood, R. (1984). Corporate social responsibil-
ity and ?nancial performance. Academy of Management
Journal, 42–56.
Cormier, D., & Gordon, I. M. (2001). An examination of social
and environmental reporting strategies. Accounting, Audit-
ing and Accountability Journal, 14(5), 587–616.
Cormier, D., & Magnan, M. (1999). Corporate environmental
disclosure strategies: Determinants, costs and bene?ts.
Journal of Accounting, Auditing and Finance, 14(3), 429–451.
Cormier, D., & Magnan, M. (2003). Environmental reporting
management: A european perspective. Journal of Accounting
and Public Policy, 22, 43–62.
Cowen, S. S., Ferreri, L. B., & Parker, L. D. (1987). The impact
of corporate characteristics on social responsibility disclo-
sure: A typology and frequency-based analysis. Accounting,
Organizations and Society, 111–122.
Crant, J. M., & Bateman, T. S. (1993). Assignment of credit and
blame for performance outcomes. Academy of Management
Journal, 36(1), 7–27.
Daft, R. L., & Lengel, R. H. (1984). Information richness: A
new approach to managerial behavior and organizational
design. Research in Organizational Behavior, 6, 191–233.
Daft, R. L., & Lengel, R. H. (1986). Organizational informa-
tion requirements, media richness and structural design.
Management Science, 32(5), 554–571.
Deegan, C., & Gordon, B. (1996). A study of the environmental
disclosure practices of Australian corporations. Accounting
and Business Research, 26(3), 187–199.
Deegan, C., & Rankin, M. (1996). Do australian companies
report environmental disclosures by ?rms prosecuted suc-
cessfully by environmental protection authority. Account-
ing, Auditing and Accountability Journal, 9(2), 50–67.
Deegan, C., Rankin, M., & Tobin, J. (2002). An examination
of the corporate social and environmental disclosures of
BHP from 1983 to 1997 – A test of legitimacy theory.
Accounting, Auditing and Accountability Journal, 15(3),
312–343.
Deegan, C., Rankin, M., & Voght, P. (2000). Firms’ disclosure
reactions to major social incidents: Australian evidence.
Accounting Forum, 24(1), 101–130.
Deephouse, D. L. (1996). Does isomorphism legitimate?
Academy of Management Journal, 39, 1024–1039.
Deephouse, D. L. (2000). Media reputation as a strategic
resource: An integration of mass communication and
resource-based theories. Journal of Management, 26(6),
1091–1112.
Deephouse, D. L., & Carter, S. M. (2005). An examination of
di?erences between organizational legitimacy and organiza-
tional reputation. Journal of Management Studies, 42(2),
329–360.
Dowling, J., & Pfe?er, J. (1975). Organizational legitimacy:
Social values and organizational behavior. Paci?c Sociolog-
ical Review, 18(1), 122–136.
Elijido-Ten, E. (2004). Determinants of environmental disclo-
sures in a developing country: An application of the
stakeholder theory. In Fourth Asia Paci?c interdisciplinary
research in accounting conference, Singapore.
Elsbach, K. D. (1994). Managing organizational legitimacy in
the california cattle industry: The construction and e?ec-
tiveness of verbal accounts. Administrative Science Quar-
terly, 39, 57–88.
Elsbach, K. D. (2003). Organizational perception management.
In L. L. Cummings & B. M. Staw (Eds.). Research in
Organizational Behavior (25, pp. 297–332).
Elsbach, K. D., & Elofson, G. (2000). How the packaging of
decision explanations a?ects perceptions of trustworthiness.
Academy of Management Journal, 43(1), 80–89.
Elsbach, K. D., & Kramer, R. M. (1996). Members’ responses
to organizational identity threats: Encountering and
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 25
countering the business week rankings. Administrative
Science Quarterly, 4.
Elsbach, K. D., & Sutton, R. (1992). Acquiring organizational
legitimacy through illegitimate actions: A marriage of
institutional and impression management theories. Academy
of Management Journal, 35, 699–739.
Fanelli, A., & Misangyi, V. F. (2006). Bringing out charisma:
CEO charisma and external stakeholders. Academy of
Management Review, 31(4), 1049–1061.
Fekrat, M. A., Inclan, C., & Petroni, D. (1996). Corporate
environmental disclosures: Competitive disclosure hypoth-
esis using 1991 annual report data. The International Journal
of Accounting, 31(2), 175–195.
Fombrun, C. (1996). Reputation: Realizing value from the
corporate image. Boston, MA: Harvard Business School
Press.
Freedman, M., & Wasley, C. (1990). The association between
environmental performance and environmental disclosure in
annual reports and 10Ks. Advances in Public Interest
Accounting, 183–193.
Gamson, W. A., Croteau, D., Hoynes, W., & Sasson, T. (1992).
Media images and the social construction of reality. Annual
Review of Sociology, 18, 373–393.
Gardner, W. L., & Martinko, M. J. (1988). Impression
management in organizations. Journal of Management,
14(2), 321–338.
Ginzel, L. E., Kramer, R. M., & Sutton, R. I. (1993).
Organizational Impression Management as a Reciprocal
In?uence Process: The Neglected Role of the Organiza-
tional Audience. In L. L. Cummings & B. M. Staw
(Eds.). Research in Organizational Behavior (15, pp. 227–
266).
Hackston, D., & Milne, M. J. (1996). Some determinants of
social and environmental disclosures in New Zealand
companies. Accounting, Auditing and Accountability Journal,
25, 77–108.
Hannan, M. T., & Freeman, J. (1989). Organizational ecology.
Cambridge, MA: Harvard University Press.
Heckman, J. J. (1979). Sample selection bias as a speci?cation
error. Econometrica, 47(1), 153–162.
Hirst, D., Koonce, L., & Simko, P. (1995). Investor reactions to
?nancial analysts’ research reports. Journal of Accounting
Research, 33(2), 335–351.
Hodge, F., Hopkins, P. E., & Pratt, J. (2006). Management
reporting incentives and classi?cation credibility: The e?ects
of reporting discretion and reputation. Accounting, Organi-
zations and Society, 31(7), 623–634.
Hogner, R. H. (1982). Corporate social reporting: Eight
decades of development at US steel. Research in Corporate
Performance and Policy, 4, 243–250.
Hughes, S. B., Anderson, A., & Golden, S. (2001). Corporate
environmental disclosures: Are they useful in determining
environmental performance?. Journal of Accounting and
Public Policy 20, 217–240.
Ingram, R. W., & Frazier, K. B. (1980). Environmental
performance and corporate disclosure. Journal of Account-
ing Research, 18, 615–622.
Jaggi, B., & Freedman, M. (1982). An analysis of the
informational content of pollution disclosures. The Finan-
cial Review, 142–152.
Janis, I. L., & Fadner, R. (1965). The Coe?cient of Imbalance.
In H. Lasswell, N. Leites, & Associates (Eds.), Language of
politics (pp. 153–169). Cambridge, MA: MIT Press.
Jones, E. E., & Pittman, T. S. (1982). Toward a theory of
strategic self-presentation. In J. Suls (Ed.), Psychological
perspectives on the self (pp. 231–262). Hillsdale, NJ:
Erlbaum.
Kim, P. H., Dirks, KT., Cooper, C. D., & Ferrin, D. L.
(2006). When more blame is better than less: The
implications of internal vs. external attributions for the
repair of trust after a competence- vs. integrity-based trust
violation. Organizational Behavior and Human Decision
Processes, 99, 49–65.
Leary, M. R., & Kowalski, R. M. (1990). Impression manage-
ment: A literature review and two-component model.
Psychological Bulletin, 107(1), 34–47.
Lee, L. F. (1983). Generalized econometric models with
selectivity. Econometrica, 51(2), 507–513.
Lindblom, C. K. (1994). The implications of organizational
legitimacy for corporate social performance and disclosure.
In Paper presented at the critical perspectives on accounting
conference, New York, NY.
Li, Y., Richardson, G. D., & Thornton, D. B. (1997).
Corporate disclosure of environmental liability information:
Theory and evidence. Contemporary Accounting Research,
14(3), 435–474.
McCombs, M. E., & Shaw, D. L. (1972). The agenda setting
function of the mass media. Public Opinion Quarterly, 36,
176–187.
McGuire, J., Sundgren, A., & Schneeweis, T. (1988). Corporate
social responsibility and ?rm ?nancial performance. Acad-
emy of Management Journal(December), 854–872.
Mercer, M. (2004). How do investors assess the credibility of
management disclosures?. Accounting Horizons 18(3),
185–196.
Mills, D., & Gardner, M. (1984). Financial pro?les and the
disclosure of expenditures for socially responsible purposes.
Journal of Business Research(December), 407–424.
Murray, A., Sinclair, D., Power, D., & Gray, R. (2006). Do
?nancial markets care about social and environmental
disclosure? Accounting, Auditing and Accountability Journal,
19(2), 228–255.
Neu, D., Warsame, H., & Pedwell, K. (1998). Managing public
impressions: Environmental disclosures in annual reports.
Accounting, Organizations and Society, 23(3), 265–282.
Nunnaly, J. (1978). Psychometric theory (end ed.). New York:
McGraw Hill.
O’Donovan, G. (2002). Environmental disclosures in annual
report – Extending the applicability and power of legitimacy
theory. Accounting, Auditing and Accountability Journal,
15(3), 344–371.
O’Dwyer, B. (2002). Managerial perceptions of corporate social
disclosure – An Irish story. Accounting, Auditing and
Accountability Journal, 15(3), 406–436.
26 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
Oliver, C. (1991). Strategic responses to institutional processes.
Academy of Management Review, 16(1), 145–179.
Patten, D. M. (1991). Exposure, legitimacy, and social disclo-
sure. Journal of Accounting and Public Policy, 10, 297–308.
Patten, D. M. (1992). Intra-industry environmental disclosures
in response to the alaskan oil spill: A note on legitimacy
theory. Accounting, Organizations and Society, 17(5),
471–475.
Patten, D. M. (2002a). Media exposure, public policy pressure,
and environmental disclosure: An examination of the
impact of tri data availability. Accounting Forum, 26(2),
153–171.
Patten, D. M. (2002b). The relation between environmental
performance and environmental disclosure: A research note.
Accounting, Organizations and Society, 27, 763–773.
Patten, D. M. (2005). The accuracy of ?nancial reports
projections of future environmental capital expenditures:
A research note. Accounting, Organizations and Society, 30,
457–468.
Patterson, J. D., & Allen, M. W. (1997). Accounting for your
actions: How stakeholders respond to the strategic commu-
nication of environmental activist organizations. Journal of
Applied Communication Research, 25, 293–316.
Pfe?er, J., & Salancik, G. R. (1978). The external control of
organizations. New York: Harper and Row.
Pollock, T., & Rindova, V. P. (2003). Media legitimation e?ects
in the market for initial public o?erings. Academy of
Management Journal, 46(5), 631–642.
Rao, H. (1994). The social construction of reputation: Certi-
?cation contests, legitimation and the survival of organiza-
tions in the American automobile industry: 1985–1912.
Strategic Management Journal, 15, 29–44.
Rao, H. (1998). Caveat emptor: The construction of nonpro?t
consumer watchdog organizations. American Journal of
Sociology, 103(4), 912–961.
Rao, H. (2001). In Joel A. C. Baum (Ed.), Interorganizational
ecology, companion to organizations (pp. 541–556). London:
Basil Blackwell.
Richardson, A. J., & Welker, M. (2001). Social disclosure,
?nancial disclosure and the cost of capital. Accounting,
Organizations and Society, 26(7/8), 597–616.
Roberts, R. W. (1992). Determinants of corporate social
responsibility disclosure: An application of stakeholder
theory. Accounting, Organizations and Society, 17(6),
595–612.
Rockness, J. W. (1985). An assessment of the relationship
between US corporate environmental performance and
disclosure. Journal of Business Finance and Accounting,
12(3), 339–354.
Savage, A., Rowlands, J. & Cataldo, A. J. (1999). Environmen-
tal disclosure in annual reports: A legitimacy theory frame-
work. Working Paper Oakland University.
Schultz, M., Mouritsen, J., & Gabrielsen, G. (2001). Sticky
reputation: Analyzing a ranking system. Corporate Reputa-
tion Review, 4(1), 24–41.
Scott, W. R. (1995). Institutions and organizations. Thousand
Oaks, CA: Sage.
Shapiro, D. L. (1991). The e?ects of explanations on negative
reactions of deceit. Administrative Science Quarterly, 36,
614–630.
Sheer, V. C., & Chen, L. (2004). Improving media richness
theory. Management Communication Quarterly, 18(1),
76–93.
Suchman, M. C. (1995). Managing legitimacy: Strategic and
institutional approaches. Academy of Management Review,
20, 571–610.
Tedeschi, J. T., & Melburg, V. (1984). Impression management
and in?uence in the organization. In S. B. Bacharach & E. J.
Lawler (Eds.). Research in the Sociology of Organizations
(3, pp. 31–58).
Tyler, T. R. (2006). Legitimacy and legitimation. Annual Review
of Psychology, 57, 375–400.
Walden, D. W., & Schwartz, B. N. (1997). Environmental
disclosures and public policy pressure. Journal of Accounting
and Public Policy, 16, 125–154.
Weber, R. P. (1990). Basic content analysis (second ed.).
Newbury Park, CA: Sage.
Williams, S. M., & Ho Wern Pei, C. A. (1999). Corporate social
disclosures by listed companies on their web sites: An
international comparison. The International Journal of
Accounting, 34(3), 389–419.
Wilmshurst, T. D., & Frost, G. R. (2000). Corporate environ-
mental reporting: A test of legitimacy theory. Accounting,
Auditing and Accountability Journal, 13(1), 10–26.
Wiseman, J. (1982). An evaluation of environmental disclosures
made in corporate annual reports. Accounting, Organiza-
tions and Society, 7(1), 53–63.
Wood, R. E., & Mitchell, T. R. (1981). Manager behavior in a
social context: The impact of impression management on
attributions and disciplinary actions. Organizational Behav-
ior and Human Decision Processes, 28, 356–378.
Zimmerman, M. A., & Zeitz, G. J. (2002). Beyond survival:
Achieving new venture growth by building legitimacy.
Academy of Management Review, 27, 414–431.
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 27
doc_331361131.pdf
Using a direct measure of environmental legitimacy, we explore the impact of annual report environmental disclosures
and environmental press releases as legitimation tools. The sample comprises North American firms (Canada and
the United States). The results obtained show that environmental legitimacy is significantly and positively affected by
the quality of the economic-based segments of annual report environmental disclosures and by reactive environmental
press releases, but not by proactive press releases
Media legitimacy and corporate
environmental communication
Walter Aerts
a,
*
, Denis Cormier
b,1
a
Faculty of Applied Economics, University of Antwerpen, Prinsstraat 13, 2000 Antwerpen, Belgium
b
E
´
cole des Sciences de la Gestion, UQAM, C.P. 8888, Succ. Centre-Ville, Montreal, Quebec, Canada H3C 1P8
Abstract
Using a direct measure of environmental legitimacy, we explore the impact of annual report environmental disclo-
sures and environmental press releases as legitimation tools. The sample comprises North American ?rms (Canada and
the United States). The results obtained show that environmental legitimacy is signi?cantly and positively a?ected by
the quality of the economic-based segments of annual report environmental disclosures and by reactive environmental
press releases, but not by proactive press releases. Moreover, our results suggest that negative media legitimacy is a
driver of environmental press releases but not of annual report environmental disclosures.
Ó 2008 Elsevier Ltd. All rights reserved.
Introduction
We analyse the relationship between corporate
environmental communication through annual
report disclosures and press releases and environ-
mental legitimacy. In concert with the strategic
approach to organizational legitimacy, we contend
that ?rms use corporate communication media
(such as annual report disclosures and press
releases) to manage perceived environmental legit-
imacy by signalling to relevant publics that their
behavior is appropriate and desirable and, at the
same time, to react to public pressures by adapting
the level, content and quality of their environmen-
tal information dissemination processes.
Legitimacy is mainly about perception. Central
to the concept of legitimacy are the perceptions
held by relevant publics and by society at large.
Contrary to previous research on corporate envi-
ronmental communication, we implement a direct
measure of environmental legitimacy. We use pub-
lic media data to assess generalized perceptions
of a ?rm’s environmental legitimacy. We focus on
0361-3682/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aos.2008.02.005
*
Corresponding author. Tel.: +32 32204110; fax: +32
2204064.
E-mail addresses: [email protected] (W. Aerts), cormier.
[email protected] (D. Cormier).
1
Tel.: +1 514 987 3000x8358.
www.elsevier.com/locate/aos
Available online at www.sciencedirect.com
Accounting, Organizations and Society 34 (2009) 1–27
legitimacy as transpiring through media coverage
and media evaluations as one of the most salient
and prominent sources of societal legitimacy per-
ceptions. In addition, by measuring environmental
legitimacy as transpiring through the media we
are able to directly assess the e?ectiveness of cor-
porate environmental communication as an orga-
nizational perception management tool (Elsbach,
2003) to enhance its perceived environmental
legitimacy.
Credibility is an important factor in any impres-
sion management process and will have a sig-
ni?cant impact on its e?ectiveness (Leary &
Kowalski, 1990). Within a communication context,
credibility refers to the congruence between the
source’s verbal claims and the corresponding acts
and events. Credibility issues with regard to corpo-
rate environmental communication have been
hinted at in a number of environmental reporting
studies (e.g. Patten, 2002a, 2002b), especially with
regard to ?rms operating in environmentally-sensi-
tive industries. Environmental legitimacy percep-
tions at the industry level may well have a
signi?cant e?ect on the ex ante believability of
any corporate environmental communication act.
In that vein, we integrate industry legitimacy as a
factor moderating the e?ectiveness of a ?rm’s envi-
ronmental communication initiatives.
Previous environmental reporting research
neglected the complementary roles of di?erent
communication channels in environmental legiti-
macy management processes. Communication
channels such as annual reports and press releases
di?er in their capacity to process timely, focused
and sensitive information cues. Press releases typ-
ically cover more fragmented issues and are mainly
short-term, tactically oriented, whereas annual
reports, by their nature, are more comprehensive
in their approach and focus on more enduring
issues. The more timely and ?exible use of environ-
mental press releases may well a?ect the impact of
annual report disclosures on environmental legiti-
macy. By analyzing both annual report disclosures
and press releases, taking into account their
distinctive features, we explore the relative impor-
tance and roles of these two common information
dissemination channels as perception management
tools.
Our results show that environmental legitimacy
is signi?cantly and positively a?ected by the extent
and quality of annual report environmental disclo-
sures and by reactive environmental press releases.
Press releases with a proactive environmental con-
tent do not seem to a?ect media legitimacy. The
e?ect of annual report environmental disclosures
seems to be driven by the extent and quality of
the economic-based information segments. More-
over, the impact of annual report disclosures on
media legitimacy is negatively a?ected if the ?rm
belongs to an environmentally-sensitive industry.
The impact of annual report environmental disclo-
sures on media legitimacy is not a?ected by the use
of proactive press releases, while it appears that
the use of reactive press releases substitutes for
the e?ect of annual environmental disclosures to
a signi?cant extent. Moreover, our results docu-
ment that, in concert with the main assumption
of legitimacy theory, negative prior environmen-
tal legitimacy is a signi?cant driver of environmen-
tal press releases. Such a relationship can,
however, not be demonstrated for annual report
environmental disclosures, suggesting a more stra-
tegic orientation of annual report environmental
communication.
The remainder of the paper is organized as fol-
lows. Section 2 presents a theoretical framework
for analyzing the interplay of environmental legit-
imacy and corporate environmental communica-
tion as well as hypotheses. The study’s empirical
models and sample are described in Section 3.
Findings are reported in Section 4. Finally, Section
5 provides a discussion regarding the potential
implications of our ?ndings.
Theoretical background and development of
hypotheses
Corporate environmental communication and media
legitimacy
The legitimacy concept, with its roots in institu-
tional theory and socio-political research, has been
a highly in?uential theoretical perspective within
the domain of corporate environmental reporting
research. Strategic legitimacy theory suggests that
2 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
legitimacy is, to a certain extent, controllable by
organizations (Ashforth & Gibbs, 1990; Oliver,
1991). It contends that organizations are able to
make strategic choices to alter their legitimacy sta-
tus and to cultivate the resource through corporate
actions, by adapting their activities and changing
perceptions. Organizations might take various
action to ensure that their behavior is perceived
to be legitimate. One of the main options is to
attempt, through communication, to become iden-
ti?ed with symbols, values, and methods of opera-
tion with institutions, values, or outputs that are
strongly believed to be legitimate, and, as such,
to demonstrate congruence between its organiza-
tional practices and the values professed by its
social environment (Dowling & Pfe?er, 1975;
Lindblom, 1994; Patterson & Allen, 1997).
This strategic legitimacy perspective has been
quite popular in environmental disclosure research
in recent years and several studies explicitly adopt
such a perspective (Bansal & Clelland, 2004;
Brown & Deegan, 1998; Buhr, 1998; Cormier &
Gordon, 2001; Deegan & Gordon, 1996; Deegan
& Rankin, 1996; Deegan, Rankin, & Tobin,
2002; Hogner, 1982; Neu, Warsame, & Pedwell,
1998; O’Donovan, 2002; O’Dwyer, 2002; Patten,
1991, 1992, 2005; Savage, Rowlands, & Cataldo,
1999; Walden & Schwartz, 1997; Wilmshurst &
Frost, 2000). However, a shortcoming of extant
research related to the legitimacy theory perspec-
tive on corporate environmental reporting is that
the legitimacy construct is seldom measured
directly. As legitimacy is not directly observable,
it is common for researchers to infer legitimation
processes and e?ects by examining relationships
between observable corporate performance attri-
butes (like measures of environmental perfor-
mance or environmental disasters in the oil and
gas industry) or third-party actions (e.g. lawsuits
for environmental matters) and environmental
reporting measures. In this study, we use a direct
measure of organizational legitimacy, grounded
in the role of public media in social perception
construction processes.
With its roots in beliefs and perceptions, legiti-
macy can best be conceived as a social assessment
or appraisal of acceptance, appropriateness and/or
desirability (Fombrun, 1996; Rao, 1994; Suchman,
1995; Zimmerman & Zeitz, 2002). It refers to a col-
lective awareness and recognition of an organiza-
tion in its organizational ?eld as appropriate and
acceptable. In the same vein, legitimation refers
to ‘‘the characteristic of being legitimized by being
placed within a framework through which some-
thing is viewed as right and proper” (Tyler, 2006,
p. 376). As institutional intermediaries specializing
in disseminating information about organizations
or in evaluating their outputs, public media play
an important role in such legitimation processes
(Fombrun, 1996; Rao, 1998). They are believed
to have superior ability to access and disseminate
information by virtue of their institutional roles
or structural positions (Rao, 1998, 2001) and are
closely followed because of their perceived superi-
ority in accessing and evaluating ?rms (Fanelli &
Misangyi, 2006). Although di?erent relevant pub-
lics may hold di?erent perceptions of what good
or bad environmental management is, these per-
ceptions are ?ltered through the media lens to a
kind of common impression. The information
and evaluations provided by public media tend
to be distributed more broadly than the opinions
of the average stakeholder. As a result, they are
likely to have a high degree of in?uence on which
organizations and organizational properties
become prominent in the minds of stakeholders
(Fanelli & Misangyi, 2006).
Media research within the agenda-setting para-
digm (Caroll & McCombs, 2003; McCombs &
Shaw, 1972) demonstrates a close alignment
between the content of public media and public
opinion (or the degree of salience that di?erent
topics have for the general public),
2
corroborating
the assertion that public media are actively
involved in public impression construction pro-
cesses (Gamson, Croteau, Hoynes, & Sasson,
2
The agenda-setting hypothesis on media behavior posits
that media do not mirror public concerns, but actively in?uence
them, particularly through the transfer of salience from the
media agenda to the public agenda (Caroll & McCombs, 2003;
McCombs & Shaw, 1972). This indicates that the impact of
public media on stakeholders’ perceptions derives primarily
from their ability to focus public attention on the issues and
entities that they select to report on (Deephouse, 2000; Pollock
& Rindova, 2003).
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 3
1992). Media research shows that the agenda-set-
ting e?ect is especially strong for unobtrusive issues
or issues with which individuals have little personal
contact and for which they rely on the media as the
primary (and sometimes only) source of informa-
tion. Environmental issues are generally perceived
as unobtrusive (Ader, 1995), reinforcing the rele-
vance of a media proxy for environmental legiti-
macy. Moreover, from management’s point of
view, media evaluations may be the only persistent
and long-term proxy for collective legitimacy
impressions on which it can benchmark and model
its environmental communication strategy. In the
same vein, Neu et al. (1998) argue that environ-
mental disclosures focus on relevant publics that
are dominant rather than peripheral and ‘critical’
ones.
Previous research in corporate environmental
reporting (Bewley & Li, 2000; Brown & Deegan,
1998; Deegan, Rankin, & Voght, 2000; Patten,
2002a) elaborates on the e?ect of media exposure
on environmental disclosures, mainly establishing
that higher public media coverage of environmen-
tal issues puts public pressure on ?rms to increase
environmental disclosures. In addition, Brown and
Deegan (1998) show a strong association between
the industry-wide number of environmental arti-
cles that were negative in tone and the amount
of ‘‘positive” corporate environmental disclosures.
These studies mainly focus on the level of ?rm-spe-
ci?c media coverage as a proxy for public concerns
regarding environmental issues. In this paper, we
extend previous environmental reporting research
by incorporating media exposure not only as an
antecedent but also as a consequence of environ-
mental communication acts. Moreover, we elabo-
rate on the ?rm-speci?c evaluative content of
media exposure in measuring corporate environ-
mental (media) legitimacy.
Given the potential relevance of corporate envi-
ronmental communication as a legitimation tool,
we posit the following hypothesis:
Hypothesis 1. Enhanced corporate environmental
communication (annual report dis-
closures and press releases) is asso-
ciated with higher environmental
(media) legitimacy.
E?ect of industry
Credibility is an important factor in any
impression management process and will have a
signi?cant impact on its e?ectiveness (Leary &
Kowalski, 1990; Shapiro, 1991). Perceived message
credibility will depend on the perceptions of source
competence, of actual or expected (incentives) bias
behavior and of the characteristics of the message
itself (Barton & Mercer, 2005; Birnbaum & Steg-
ner, 1979; Mercer, 2004). More speci?cally, Tede-
schi and Melburg (1984) distinguish between
credibility and believability, with believability
referring to the subjective probability the target
assigns to any communication from a source.
While a source’s credibility can be expected to
have a direct e?ect on believability, other percep-
tions of the source and of the organizational ?eld
in which it functions may have a strong e?ect on
the believability of its communications as well. In
that respect, entire industries can have more or less
legitimacy that can be conferred upon the ?rms
operating within them (Aldrich & Fiol, 1994;
Scott, 1995; Suchman, 1995). Industry-level legiti-
macy has to do with the degree to which the oper-
ations and business processes of ?rms in a given
industry, and their products and services o?ered,
are accepted as appropriate and useful by broader
publics (Hannan & Freeman, 1989; Scott, 1995).
Prior accounting-related studies document that
?nancial reporting users appear to consider the
consistency between a ?rm’s reporting incentives,
which are to a large degree circumstantial, and
its actual disclosures when assessing the credibility
of the disclosures. Especially when the level of
discretion in the reporting environment is high,
users tend to discount the credibility (believability)
of incentive-consistent disclosures (Anderson,
Kadous, & Koonce, 2004; Hirst, Koonce, & Sim-
ko, 1995; Hodge, Hopkins, & Pratt, 2006). In this
vein, industry legitimacy may be directly related to
the believability of its constituents’ communica-
tions, especially in the ?eld of voluntary environ-
mental communication where there is a
straightforward reporting incentive for manage-
ment. Even if the objective credibility of the source
is high, low industry legitimacy may undermine the
believability of environmental reporting and thus
4 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
the e?ectiveness of a ?rm’s communicative legiti-
mation e?orts.
Prior research suggests that the believability of
environmental reporting may be problematic.
The believability of corporate environmental com-
munication e?orts would generally be validated
through a positive association between environ-
mental performance and environmental communi-
cation content, but previous research fails to
clearly demonstrate such an association (Freed-
man & Wasley, 1990; Hughes, Anderson, &
Golden, 2001; Ingram & Frazier, 1980; Wiseman,
1982). Some minor evidence of a signi?cant posi-
tive relationship could be established for very spe-
ci?c pollution disclosures (e.g. Al-Tuwaijri,
Christensen, & Hughes, 2004; Patten, 2002b), but
surely not for more comprehensive disclosure mea-
sures. Moreover, Patten (2002b) documents a
lower relationship between environmental perfor-
mance and environmental disclosure within more
environmentally-sensitive industries, suggesting
that environmental disclosures may be perceived
as less credible for ?rms from industries with high
environmental exposure. Belonging to an environ-
mentally-sensitive industry (with its negative con-
notations) may be the ?rst observation one has
about a ?rm, leading to a discounting of later posi-
tive messages coming from that ?rm and impeding
the e?ectiveness of its legitimacy-enhancement
e?orts. Therefore, we hypothesize:
Hypothesis 2. The association between the level of
corporate environmental communi-
cation and environmental (media)
legitimacy will be lower for those
?rms operating in more environ-
mentally-sensitive industries.
Tactical verbal impression management
Press releases are among the most common and
widespread communication vehicles used by public
?rms to disseminate voluntary information and
their content is highly discretionary. Further, they
di?er from annual report disclosures in their
capacity to process rich information and, conse-
quently, may ful?l complementary communication
goals. Rich information media
3
will generally be
more e?ective in directing attention, establishing
prominence and changing impressions (Daft &
Lengel, 1986). For our purposes, the main infor-
mation richness criteria to di?erentiate annual
report disclosures and press releases are timeliness
(and related feedback capability), topical focus
and language variety. Press releases can be used
in a more timely, more elaborate and focused fash-
ion and with more expressive language than
annual report disclosures to confront sensitive
environmental issues. They may also be more e?ec-
tive in signalling commitment. In this vein, their
use will be more tactical than that of the more
comprehensive and longer-term characteristic of
annual report disclosures.
The use and e?ectiveness of information rich
media will be more pronounced when timeliness
(reaction speed) is important from a self-presenta-
tional perspective (Sheer & Chen, 2004). From a
tactical impression management perspective, press
releases can be used in a proactive or in a reactive
fashion (Gardner & Martinko, 1988; Tedeschi &
Melburg, 1984). Proactive (or acclaiming) verbal
behavior aims at establishing a positive identity
for an audience, most likely through verbal state-
ments of accomplishments. It is not merely initi-
ated as a reaction to situational demands. On the
other hand, reactive self-presentational verbal
behavior is mainly defensive and it is typically ini-
tiated as a response to a controversial situation in
which negative and undesirable qualities may be
attributed to the ?rm. Proactive (or acclaiming)
content would stress the importance, relevance
and scope of positive environmental outcomes or
actions, whereas more reactive content would
focus on downplaying the signi?cance of nega-
tively perceived or controversial events related to
the natural environment. In this vein, proactive
3
Daft and Lengel (1984, 1986) de?ne information richness as
the ability of communication media to change understanding
within a time interval, with face-to-face communication having
the highest richness and periodic statistical documents posi-
tioned at the lower end. Information richness depends on the
medium’s capability of immediate feedback (the most impor-
tant aspect), the number of cues and channels utilized,
personalisation and language variety.
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 5
environmental press releases would express com-
mitment to environmental concerns and generally
focus on the positive aspects of corporate environ-
mental performance. On the other hand, reactive
press releases would tackle more equivocal con-
cerns, with equivocality referring to the presence
of multiple and possibly con?icting interpretations
of the available information.
From an information richness perspective, per-
formance improves if richer media are used for
more equivocal tasks and leaner media are used
for non-equivocal tasks (Daft & Lengel, 1986).
In that sense, reactive press releases would be more
e?ective than proactive press releases. By enabling
equivocal information to be conveyed quickly,
timely reactive press releases can be used as a nor-
malizing account, separating a negative environ-
mental event or outcome from larger assessments
of the ?rm as a whole and, in the end, putting a
positive spin on events (Suchman, 1995).
Traditional forms of defensive verbal impres-
sion management tactics, such as excuses, justi?ca-
tions, defences of innocence and apologies, have
been shown to be e?ective as perception manage-
ment devices in numerous research settings (Barton
& Mercer, 2005; Elsbach, 2003; Kim, Dirks, Coo-
per, & Ferrin, 2006; Wood & Mitchell, 1981). In
?eld settings, defensive impression management
tactics are frequently accompanied by selective cat-
egorizations that stress and make salient other
legitimate and competing dimensions along which
a ?rm’s actions and performance should be
assessed. By selectively directing and focusing
attention on performance dimensions with positive
rami?cations, categorization processes relieve the
dissonance related to a single dimension (e.g. a neg-
atively perceived environmental outcome or event)
and tend to bolster the believability of more
straightforward defensive verbal tactics (Elsbach
& Kramer, 1996). Defenses of innocence and justi-
?cations typically shift focus away from the nega-
tive towards the more positive aspects of an event
or outcome and, in that vein, set the stage for
acclaiming tactics with regard to the progress made
toward environmentally desirable goals (Elsbach &
Sutton, 1992). By its nature, timeliness will be more
important for reactive impression management tac-
tics relative to proactive impression management
and is thus expected to promote the relative e?ec-
tiveness of reactive environmental press releases.
Moreover, previous research suggested that speci?c
occurrences of defensive impression management
are more e?ective than assertive verbal behaviors
in shaping evaluative perceptions of an external
audience (Suchman, 1995). In this vein, Crant
and Bateman (1993) documented that defensive
causal accounts actually diminished observers’
assessment of blame for failure, whereas acclaim-
ing causal accounts did not a?ect the assignment
of credit for success.
Given the importance of response timeliness in
the face of equivocal environmental concerns and
the relative e?ectiveness of defensive verbal
impression management tactics, we hypothesize
Hypothesis 3. The association between reactive
environmental press releases and
environmental (media) legitimacy
will be higher relative to the associa-
tion between proactive environmen-
tal press releases and environmental
(media) legitimacy.
Sample selection and empirical models
Sample
The initial sample is comprised of 623 non-
?nancial North American ?rms (205 from S&P/
TSX300 for Canada and 418 from S&P500 for
the US). From the sample of 623 ?rms, we ?nd
environmental news exposure for 165 ?rms. Out
of these 165 ?rms, 158 reported pollutants released
in the environment (Toxics release inventory).
These 158 ?rms (119 US, 39 Canadian) constitute
our ?nal sample. Environmental disclosure is col-
lected from corporate annual reports on Internet
sites and is coded for the year 2002. Financial data
is collected from Worldscope and from ?rms’
Internet sites. The sample ?rms operate in the fol-
lowing industries (S&P classi?cation):
Consumer goods and services;
Energy;
Chemicals and drugs;
6 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
Industrials;
Information technology, Telecom and media;
Mining and resources;
Utilities.
Energy; chemicals and drugs; mining and
resources (e.g. pulp and paper), and utilities are
considered to be environmentally-sensitive indus-
tries (cf. Cho & Patten, 2007).
Empirical models
Based on Al-Tuwaijri et al. (2004), we posit that
a ?rm’s environmental communication strategy
simultaneously a?ects perceived environmental
legitimacy, the extent of environmental news cov-
erage and the content of environmental communi-
cation. As such, we endogenize the related
variables in simultaneous equations models. The
following two sets of structural equations models
summarize the approach adopted in the empirical
analysis. The ?rst set integrates a ?rm’s annual
report environmental disclosure while the second
set integrates a ?rm’s press releases concerning
environmental matters:
Environmental legitimacy
itþ1
¼ ðb
0
Annual report environmental disclosure
þ b
1
Annual report environmental disclosure
ÂEnvironmentally-sensitive industries
þ b
2
Environmentally-sensitive industries
þ b
3
Legitimacy þ b
4
Size
þ b
5
Environmental performanceÞ
it
ð1:1Þ
Annual report environmental disclosure
it
¼ ðb
0
Lag Legitimacy
þ b
1
Environmentally-sensitive industries
þ b
2
Leverage þ b
3
Return on assets
þ b
4
Environmental news exposure
þ b
5
Environmental performance
þ b
5
CountryÞ
it
ð1:2Þ
Environmental news exposure
it
¼ ðb
0
Return on assets þ b
1
Foreign listing
þ b
2
Size þ b
3
Environmental performance
þ b
4
Industry þ b
5
CountryÞ
it
ð1:3Þ
Environmental legitimacy
itþ1
¼ ðb
0
Press releases þ b
1
Press releases
 Environmentally-sensitive industries
þ b
2
Environmentally-sensitive industries
þ b
3
Legitimacy þ b
4
Size
þ b
5
Environmental performanceÞ
it
ð2:1Þ
Press releases
it
¼ ðb
0
Lag Legitimacy
þ b
1
Environmentally-sensitive industries
þ b
2
Leverage þ b
3
Return on assets
þ b
4
Environmental news exposure
þ b
5
Environmental performance
þ b
6
CountryÞ
it
ð2:2Þ
Environmental news exposure
it
¼ ðb
0
Return on assets
þ b
1
Foreign listing þ b
2
Size
þ b
3
Environmental performance
þ b
4
Industry þ b
5
CountryÞ
it
ð2:3Þ
Description of dependent variables
Environmental legitimacy
To assess corporate environmental legitimacy,
we rely on a content analysis of press media cover-
age for corporate environmental issues. Institu-
tionalists suggest that content analysis of press
media sources may be particularly useful in study-
ing legitimation processes, as detailed archives of
media coverage exist for many industries and ana-
lyzing the content of those public sources o?ers a
potentially powerful technique for operationaliz-
ing legitimation (Baum & Powell, 1995).
News media content is extracted from the ABI/
Inform Global database and from three distinct
sources:
(1) Business, economics: local and regional busi-
ness publications (local and regional business
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 7
news coverage of large corporations, privately
held companies, local start-ups, executive pro-
?les, marketing, ?nance, and industry news.
ABI Inform provides access to business infor-
mation not typically found in national news
sources. It contains news and analysis, informa-
tion on local markets, and more data gathered
from major business tabloids, magazines, daily
newspapers, wire services, and city, state, and
regional business publications);
(2) Business, ?nance, economics: journals, com-
pany pro?les, Wall Street Journal (most schol-
arly and comprehensive way to explore and
understand business research topics. It includes
nearly 1800 worldwide business periodicals for
in-depth coverage of business and economic
conditions, management techniques, theory,
and business practices, advertising, marketing,
economics, human resources, ?nance, taxation,
computers, and more. It constitutes extensive
international coverage with quick access to
information on more than 60,000 companies
with business and executive pro?les);
(3) Canadian newsstand, which o?ers unparal-
leled access to the full text of Canadian newspa-
pers (Montreal Gazette, National Post and
Toronto Star). We extracted articles using a
?rm’s name and the following keywords: ‘‘envi-
ronment”, ‘‘sustainable development”, ‘‘recy-
cling”, ‘‘pollution”, ‘‘toxic”, ‘‘ISO14000”,
‘‘conservation”, ‘‘remediation”, ‘‘spills”, ‘‘waste
management”, ‘‘energy”, ‘‘awards”, ‘‘environ-
mental audit”.
The legitimacy measure is computed for 2003,
2002 (lag measure in legitimacy models) and
2001 (lag measure in disclosure models). In the
summer of 2005, two research assistants found
413 articles: 226 provided good news, 172 provided
bad news and 25 provided neutral news. Each arti-
cle was coded in terms of its impact on the ?rm’s
environmental legitimacy, i.e. neutral, negative,
or positive (See Appendix 1). Good news stories
are those that convey environmental commitment
and that emphasize the positive aspects of a ?rm’s
activities. Examples of good news stories include
investment in facilities that will reduce energy con-
sumption, or the reduction of greenhouse gas
emissions. The legitimacy score is computed based
on this coding. The two coders agreed on 81%
good news, 87% bad news and 96% neutral news.
Internal consistency estimates (Cronbach’s alpha)
computed over the 2001 and 2003 period show
that the variance between the two coders’ scores
is quite systematic (alpha = 0.888 for good news,
0.926 for bad news and 0.864 for neutral news).
This suggests a high level of intercoder reliability
(Weber, 1990). A researcher reconciled all coding
disagreements between the two coders.
Annual legitimacy measures were calculated
using the Janis–Fadner coe?cient of imbalance
(Bansal & Clelland, 2004; Janis & Fadner, 1965).
The Janis–Fadner coe?cient ranges from À1.0 to
+1.0; a high presence of favourable articles in a
given year yields a value closer to +1.0, and a high
presence of unfavourable articles yields a value
closer to À1.0. The formula is as follows:
Janis–Fadner coefficient ¼
ðe
2
À ecÞ
t
2
if e > c
ðec À c
2
Þ
t
2
if c > e
where e is the number of favourable environmental
articles in a given year, c is the number of unfa-
vourable environmental articles in a given year
and t is the sum of e and c.
Annual report environmental disclosure
Environmental disclosure is measured using a
coding instrument in a manner similar to Wiseman
(1982), Cormier and Magnan (2003), and Al-
Tuwaijri et al. (2004). The grid comprises 39 items
measuring environmental disclosure quality where
the items are grouped into six categories as fol-
lows: expenditures and risk; laws and regulations;
pollution abatement; sustainable development;
land remediation; and environmental manage-
ment. The rating is based on a score from one to
three. Three points are awarded for an item
described in monetary or quantitative terms, two
are awarded when an item is described speci?cally,
and one is awarded for an item discussed in gen-
eral. The information is coded according to the
grid presented in Appendix 1.
We believe that the use of a coding scale to
qualify a ?rm’s environmental disclosure is appro-
8 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
priate for the following reasons. First, it allows for
some integration of di?erent types of information
into a single ?gure that is comparable across ?rms
in terms of relevance. Second, while other disclo-
sure studies rely on word counts to measure
environmental disclosure (e.g. Neu et al., 1998;
Williams & Ho Wern Pei, 1999), a qualitative scale
allows for the researcher’s judgment to be utilized
in rating the value or quality of the disclosures
made by a ?rm. While this process is more subjec-
tive, it ensures that irrelevant or redundant gener-
alities are not considered strategic environmental
disclosures.
To ensure consistency across ?rms, two persons
reviewed all individual scores independently. All
disagreements were subsequently reviewed by one
of the co-researchers.
4
Press releases
We collected press releases related to environ-
mental information from ?rms’ web sites in 2002
for 144 ?rms. Since the information was not avail-
able on the web page for 21 ?rms, we completed
the data collection using the Lexis/Nexis database.
We searched for press releases using the same key-
words used for articles pertaining to environmen-
tal legitimacy. Two research assistants identi?ed
236 press releases, 153 classi?ed as proactive
accounts and 83 as reactive accounts. Proactive
accounts tend to bolster the positive environmen-
tal image of a ?rm through descriptive statements
that describe various aspects of a ?rm’s environ-
mental policy and accomplishments that re?ect
positively upon the ?rm and through the use of
acclaiming tactics, including enhancements
intended to boost the apparent desirability of spe-
ci?c environmental events or of ?rm-speci?c envi-
ronmental actions and entitlements intended to
heighten the ?rm’s perceived responsibility for
the events or actions (Gardner & Martinko,
1988). Reactive accounts are verbal remedial
tactics and are a reaction to environmental predic-
aments, i.e. situations with undesirable implica-
tions for a ?rm’s environmental image. They are
mainly defensive in nature and tend to attenuate
the negative meaning of events or outcomes.
Excuses, justi?cations and causality denials are
the more traditional forms of defensive impression
management tactics. To put a positive spin on
events, these traditional forms of impression man-
agement are frequently accompanied by selective
categorizations highlighting alternate (positive)
attributes of a ?rm’s environmental position and
related entitlement and enhancement tactics. The
coders agreed on 77% of proactive news and 89%
of reactive news. Internal consistency estimates
(Cronbach’s alpha) show that the variance
between the two coders’ scores is quite systematic
(alpha = 0.825 for proactive news and 0.834 for
reactive news). A researcher reconciled all coding
disagreements between the two coders. Smaller
disagreements were resolved by the two coders
themselves.
Environmental news exposure
The importance of news exposure in determin-
ing environmental disclosure indicates that ?rms
need to achieve social legitimacy with their envi-
ronmental management, i.e. their ultimate intent
is strategic.
4
A coding manual documenting coding instructions as well
as standardized coding worksheets were prepared beforehand.
Each coder then applied the following coding sequence: (1)
independent identi?cation of the occurrence of items relative to
the di?erent coding categories; (2) independent coding of the
items according to quality level of content and (3) timed
reconciliation on a subset of company reports. The coders were
intensively trained in applying coding instructions and in using
the coding worksheets. They were unaware of the research
hypotheses. Initial di?erences in identifying grid items
accounted for, on average, 6% of the maximum number of
items identi?ed. Of the information quality level coding, less
than 10% had to be discussed for reconciliation. Disagreement
between coders mostly occurred at the beginning of the coding
process (essentially the ?rst 20 ?rms in each country and the
?rst 5 ?rms by industry). A researcher reconciled coding
disagreements exceeding 5% of the highest total score between
the two coders. Smaller disagreements were resolved by the two
coders themselves. Overall, we think that this coding process
provides a reliable measure of environmental reporting. Inter-
nal consistency estimates (Cronbach’s alpha on score compo-
nents) show that the variance is quite systematic
(alpha = 0.803). This is higher than Botosan (1997), who ?nds
an alpha of 0.64 for an index including ?ve categories of
disclosure in annual reports. Cronbach’s alpha estimates the
proportion of variance in the test scores that can be attributed
to a true score variance. It can range from 0 (if no variance is
consistent) to 1.00 (if all variances are consistent). According to
Nunnaly (1978), a score of 0.70 is acceptable.
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 9
The concerns of government and local commu-
nities are di?cult to ascertain directly. However,
prior work does suggest that environmental news
exposure is an appropriate proxy for community
concerns (Deegan & Rankin, 1996). A ?rm’s envi-
ronmental news exposure is computed by taking
the average number of articles concerning environ-
mental issues for 1998 through 2002, as contained
in the ABI/Inform Global database. We searched
for articles using the keywords mentioned above.
The reason for this choice is that disclosure this
year (2002) may be a?ected by the number and
types of articles that have been published about
a ?rm in the recent past. A total of 764 relevant
stories were identi?ed over the period. We expect
that as environmental news exposure increases,
the ?rm will increase its environmental disclosure.
Hence, a positive relationship is expected between
environmental media coverage and environmental
disclosure.
Determinants of environmental legitimacy
Size
Size has been shown to be an antecedent of
legitimacy (Baum & Oliver, 1991; Deephouse &
Carter, 2005). Firm size will a?ect the ?rm’s visi-
bility to the general public and tends to engender
increased public scrutiny. Firm size, measured as
ln(Assets), is introduced as a control variable, with
no directional prediction.
Prior environmental legitimacy
Reputation and legitimacy issues have been
argued to be largely sticky (Schultz, Mouritsen,
& Gabrielsen, 2001). Like reputation, legitimacy
can be inertial or durable and have the tendency
to reproduce itself over time. Hence, the lagged
environmental legitimacy variable is introduced
to capture the inertia factor. Adding the lagged
dependent variable also implies control for omit-
ted ?rm characteristics, including the fact that spe-
ci?c environmental norms and value expectations
may vary from industry to industry. Alternatively,
we could have expressed the dependent variable as
a change variable. We chose not to do so because
such a procedure constrains the coe?cient of the
lagged variable to equal one and we preferred to
model the inertia factor as a theoretically relevant
determinant.
Environmental performance
Documentary evidence of a ?rm’s polluting
activities might be directly discounted in how the
media perceive a ?rm’s environmental posture.
Environmental performance is proxied by the Tox-
ics Release Inventory (TRI), a public database
available from the Environmental Protection
Agency (EPA) in the US and from Environment
Canada. These databases contain information on
toxic chemical releases and other waste manage-
ment activities reported annually by manufactur-
ing facilities. The Toxics Release Inventory is the
sum of all chemicals released in the air and water
and on land in 2002. Consistent with Clarkson,
Li, Richardson, and Vasvari (2006), the measure
is computed by summing all facilities for an indi-
vidual company in pounds de?ated by $1000 of
sales. Higher values of the variable imply a worse
environmental performance.
Determinants of corporate environmental
communication
Prior environmental legitimacy
In line with strategic legitimacy theory argu-
ments, we expect a negative relationship between
prior media legitimacy and corporate environmen-
tal communication.
Leverage
It is expected that ?rms able to withstand
potential proprietary costs from the disclosure of
environmental information and bene?tting from
more open disclosure (?rms in good ?nancial con-
dition) are likely to outweigh the costs from the
disclosure of environmental information. By
widely disseminating information about their envi-
ronmental management and showing their ability
to shoulder environmental obligations, these ?rms
establish their credibility as a reliable and socially
responsible partner. Roberts (1992) and Richard-
son and Welker (2001) ?nd a positive relationship
between leverage and social disclosure while Eliji-
do-Ten (2004) does not ?nd any signi?cant rela-
tionship between leverage and environmental
10 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
disclosure. Conversely, Cormier and Magnan
(2003) document a negative relationship between
leverage and environmental disclosure.
5
We mea-
sure leverage by the ratio of long-term ?nancial
debt over equity (Long-term ?nancial debt/
equity). Since the actual impact of leverage on
environmental disclosure is unclear, no directional
predictions are made for the variable.
Return on assets
Many studies document a positive association
between a ?rm’s level of disclosure and its ?nancial
performance (Cochran & Wood, 1984; Cormier &
Magnan, 1999, 2003; McGuire, Sundgren, &
Schneeweis, 1988; Mills & Gardner, 1984). Firms
with superior earnings performance have a higher
propensity to reveal their ‘‘good news”. Hence,
Murray, Sinclair, Power, and Gray (2006) docu-
ment that ?rms with consistently higher returns
tend to have higher levels of total and voluntary
social and environmental disclosure. In this vein,
we expect a positive relationship between pro?t-
ability, as measured by return on assets, and envi-
ronmental disclosure.
Media exposure
A number of studies document that higher lev-
els of media exposure relative to environmental
issues increase public concerns and thus public
policy pressure, to which companies react through
greater environmental disclosure (Bewley & Li,
2000; Brown & Deegan, 1998; Deegan et al.,
2000; Li, Richardson, & Thornton, 1997; Patten,
2002a).
Environmental performance
Many authors examine the association between
environmental disclosure in annual reports and a
?rm’s environmental performance. Results are
mixed. Ingram and Frazier (1980), Jaggi and
Freedman (1982), Wiseman (1982), Rockness
(1985), Freedman and Wasley (1990), Fekrat,
Inclan, and Petroni (1996) do not ?nd a signi?cant
association between environmental disclosure (in
the annual report or in the 10K report) and the
CEP index of environmental performance while
Patten (2002b) establishes a negative relationship.
More recent works document a positive associa-
tion between environmental performance and the
extent of discretionary environmental disclosures
(Al-Tuwaijri et al., 2004; Clarkson et al., 2006).
According to Al-Tuwaijri et al. (2004), a positive
relationship conjectures that prior literature’s
mixed results describing their interrelations may
be attributable to the fact that researchers have
not jointly considered Environmental disclosure,
Environmental performance, and Economic per-
formance. Since the actual impact of environmen-
tal performance on environmental disclosure is
unclear, no directional predictions are made.
Environmentally-sensitive industries
Companies in environmentally-sensitive indus-
tries are generally subject to greater environmental
scrutiny than other companies (Cowen, Ferreri, &
Parker, 1987; Hackston & Milne, 1996; Patten,
1991) and have been documented to exhibit higher
levels of environmental disclosure.
Determinants of environmental news exposure
We introduce ?ve variables that determine a
?rm’s exposure to environmental news and its
environmental ‘‘riskiness”: Firm size; Return on
assets; Foreign listings; Capital intensity, and
Environmental performance.
Firm size: Prior evidence is consistent in show-
ing a positive relationship between the extent of
media coverage and ?rm size (e.g. Caroll &
McCombs, 2003; Deephouse, 2000; Schultz et al.,
2001). We predict a positive relationship between
size and environmental news exposure.
Return on assets: Since ?rms with consistently
higher returns tend to have higher levels of envi-
ronmental disclosure, we expect that such disclo-
sure will attract environmental news media.
Foreign listings: The degree to which ?rms are
listed internationally can in?uence news coverage
and public awareness. It may also be an indicator
of diversi?cation. Based on Hope’s study (2003), a
5
An explanation for the inverse relationship (positive asso-
ciation for social disclosure and negative association for
environmental disclosure) could be that social disclosure is
more likely to be good news than environmental disclosure.
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 11
listing on a domestic exchange and on foreign
exchanges (except U.S. and London Stock
Exchange listings) are given a weight of 1 per list-
ing; London Stock Exchange and U.S. listings are
given a weight of 1.5 because of their importance.
The score for each ?rm is summed. We expect a
positive association between the variable stock
exchange listings and the level of environmental
news exposure.
Environmental performance: A ?rm’s risk of
unfavourable environmental exposure is primarily
determined by the relative perceived environmen-
tal risk of its activities. The poorer a ?rm’s
environmental performance, the greater its envi-
ronmental exposure and the more this will attract
media attention.
Results
Univariate results
As illustrated in Table 1, the level of annual
report environmental disclosure varies from a
mean score of 13.05 for Technology, Telecom
and Media to 76.80 for Utilities. Among the seven
industries, the four industries for which ?rms’
activities are more likely to a?ect the environment
exhibit the higher environmental disclosure scores:
Utilities 76.80; Energy 72.50; Mining and
resources 67.54; and to a lesser extent Chemicals
and drugs 36.44. This result is consistent with Pat-
ten (2002b), who ?nds that those ?rms operating
in environmentally-sensitive industries report
more environmental information. On average,
US ?rms publish more environmental information
than Canadian ?rms except for sustainable devel-
opment and environmental management.
Table 2 provides some descriptive statistics
regarding the sample ?rms’ dependent and inde-
pendent variables. The 2003 mean legitimacy score
is positive at 0.135. The 2002 mean environmental
news exposure approaches one article per ?rm. As
expected, there are twice as many proactive press
releases compared to reactive ones. Firm size, on
average, is quite large with mean total asset for
the sample at 27.5 billion Canadian dollars. Aver-
age ?rm size is larger for US ?rms and Canadian T
a
b
l
e
1
E
n
v
i
r
o
n
m
e
n
t
a
l
l
e
g
i
t
i
m
a
c
y
a
n
d
a
n
n
u
a
l
r
e
p
o
r
t
e
n
v
i
r
o
n
m
e
n
t
a
l
d
i
s
c
l
o
s
u
r
e
m
e
a
n
s
c
o
r
e
s
b
y
i
n
d
u
s
t
r
y
S
a
m
p
l
e
L
e
g
i
t
i
m
a
c
y
E
x
p
e
n
d
i
t
u
r
e
s
a
n
d
r
i
s
k
s
L
a
w
s
a
n
d
r
e
g
u
l
a
t
i
o
n
s
c
o
n
f
o
r
m
i
t
y
P
o
l
l
u
t
i
o
n
a
b
a
t
e
m
e
n
t
S
u
s
t
a
i
n
a
b
l
e
d
e
v
e
l
o
p
m
e
n
t
L
a
n
d
r
e
m
e
d
i
a
t
i
o
n
a
n
d
c
o
n
t
a
m
i
n
a
t
i
o
n
E
n
v
i
r
o
n
m
e
n
t
a
l
m
a
n
a
g
e
m
e
n
t
T
o
t
a
l
C
o
n
s
u
m
e
r
g
o
o
d
s
a
n
d
s
e
r
v
i
c
e
s
3
4
0
.
0
0
0
5
.
8
9
3
.
2
8
7
.
0
0
2
.
1
0
3
.
1
4
4
.
2
8
2
5
.
6
9
E
n
e
r
g
y
2
0
0
.
0
8
1
1
8
.
6
6
9
.
3
6
1
2
.
6
7
2
.
9
4
1
8
.
8
1
1
0
.
0
3
7
2
.
5
0
C
h
e
m
i
c
a
l
s
a
n
d
d
r
u
g
s
9
0
.
0
0
0
8
.
4
1
4
.
7
7
8
.
0
1
2
.
1
1
8
.
6
8
4
.
4
7
3
6
.
4
4
I
n
d
u
s
t
r
i
a
l
s
1
3
0
.
5
3
8
1
4
.
1
4
8
.
4
1
1
1
.
9
0
3
.
4
9
1
8
.
0
0
4
.
6
8
6
0
.
6
3
I
n
f
o
r
m
a
t
i
o
n
t
e
c
h
n
o
l
o
g
y
,
T
e
l
e
c
o
m
a
n
d
M
e
d
i
a
1
5
0
.
2
0
0
0
.
1
6
1
.
0
3
4
.
0
3
1
.
7
3
1
.
5
3
4
.
5
9
1
3
.
0
5
M
i
n
i
n
g
a
n
d
r
e
s
o
u
r
c
e
s
4
8
0
.
1
0
6
1
7
.
0
4
8
.
5
3
1
1
.
3
4
4
.
7
4
1
2
.
0
6
1
3
.
8
1
6
7
.
5
4
U
t
i
l
i
t
i
e
s
1
9
0
.
3
0
9
1
6
.
8
5
1
4
.
0
4
1
9
.
3
8
5
.
5
2
1
4
.
9
5
6
.
0
5
7
6
.
8
0
T
o
t
a
l
1
5
8
0
.
1
4
1
1
2
.
5
6
7
.
2
7
1
0
.
7
5
3
.
5
0
1
0
.
8
1
8
.
1
6
5
3
.
0
6
C
a
n
a
d
a
3
9
0
.
0
8
0
1
0
.
8
0
2
.
7
5
1
0
.
7
5
3
.
8
7
5
.
9
6
1
1
.
0
8
4
5
.
2
1
U
S
1
1
9
0
.
1
5
9
1
3
.
1
7
8
.
8
3
1
0
.
7
5
3
.
3
8
1
2
.
4
6
7
.
1
5
5
5
.
7
6
12 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
?rms show, on average, lower systematic risk.
Media exposure and number of press releases are
much higher for US ?rms while Canadian ?rms
show, on average, higher levels of pollution and
lower media legitimacy.
Table 3 presents correlations for the legitimacy
model, the disclosure model and the news exposure
model. Reactive press releases (0.179) and Environ-
mental performance (À0.109) are signi?cantly
correlated with media legitimacy. Annual report
environmental disclosure is correlated with
Environmentally-sensitive industries (0.267), Envi-
ronmental news exposure (0.318), and Environ-
mental performance (0.171). As expected,
environmental communication is negatively
correlated with lagged media legitimacy, especially
for press releases (À0.214 for total press releases,
À0.171 for proactive press releases and À0.223
for reactive press releases). Finally, environmental
news exposure is correlated with size (0.393), For-
eign listing (0.401), and Environmental perfor-
mance (0.129).
Multivariate results
Selection bias
From the sample of 623 ?rms, we ?nd environ-
mental media coverage for 165 ?rms. Because
legitimacy is only measured for ?rms covered by
the media, there might be a problem in terms of
selection bias (Heckman, 1979). To correct this
potential bias, Heckman’s two-step procedure
was used. In the Heckman procedure (Heckman,
1979; Lee, 1983), the residuals of the selection
Table 3A
Correlations – legitimacy model
2 3 4 5 6 7 8 9 10
1 Legitimacy 0.220
*
0.028 À0.044 À0.001 À0.156
*
0.012 0.073 0.179
*
À0.109
*
2 Lag legitimacy 1 0.073 À0.097 À0.087 À0.092 À0.159
*
0.025 0.011 À0.044
3 Size 1 0.110
*
0.149
*
À0.039 À0.275
*
0.358
*
0.219
*
0.004
4 Total disclosure 1 0.974
*
0.743
*
0.267
*
0.125
*
0.072 0.171
*
5 Economic-based disclosure 1 0.572
*
0.248
*
0.144
*
0.074 0.109
*
6 Social-based disclosure 1 0.237
*
0.027 0.035 0.295
*
7 Sensitive industry 1 À0.083 À0.076 0.196
*
8 Press release – proactive 1 0.643
*
0.003
9 Press releases – reactive 1 0.008
10 Environmental performance 1
*
Signi?cant at 0.10 two-tailed.
Table 2
Descriptive statistics – ?nancial, legitimacy, and environmental news exposure variables
Min. Max. Mean Std. Dev. Mean
US sample
Mean
Canadian sample
Legitimacy À1 1 0.14 0.47 0.16 0.08
Legitimacy – highly sensitive industries À1 1 0.11 0.48 0.13 0.06
Environmental news exposure (year average) 0.17 7.33 0.68 1.12 0.72 0.57
Press releases – proactive 0 18.00 0.96 2.24 1.09 0.56
Press releases – reactive 0 18.00 0.52 1.80 0.60 0.28
Environmental performance 0 17.29 1.03 2.72 0.58 2.67
Toxic release inventory (pounds per $1000 of sales)
Total Assets (million Can $) 349 575 000 27 500 59 800 32 700 8835
Leverage (debt/assets) 0 0.64 0.24 0.13 0.24 0.22
Pro?tability (ROA) À0.49 0.20 0.03 0.08 0.04 0.02
Foreign listing 0 4 0.97 1.27 0.94 1.08
N 158 119 39
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 13
equation in a probit/logit analysis (news exposure/
no news exposure) are used to construct a selec-
tion bias control factor, i.e. the inverse mills
ratio:
6
Expected value of news exposure=no news exposure
¼aþa
1
Return on assets þa
2
Foreign listing
þa
3
Firmsize þa
4
Capital intensity
þa
5
Countryþa
6À10
Industry
Table 4 shows descriptive statistics for the con-
trol variables relating to the selected sample and
the rest of the population with no media coverage
over the selected period. We ?nd that all the vari-
ables exhibit larger mean values for the ?rms
selected and the di?erences are all statistically sig-
ni?cant. The third column in Table 5 gives the
results of the ?rst-stage Logit regression (industry
dummies not shown).
7
The model is well speci?ed
with a pseudo-R-square of 37.0% and a classi?ca-
Table 3B
Correlations – disclosure model
1 2 3 4 5 6 7 8 9 10 11 12
1 Total disclosure 1 0.974
*
0.744
*
0.111
*
0.125
*
0.070 À0.054 0.076 À0.052 0.267
*
0.318
*
0.171
*
2 Economic-based
disclosure
1 0.571
*
0.125
*
0.144
*
0.074 À0.013 0.093 À0.057 0.248
*
0.294
*
0.109
*
3 Social-based
disclosure
1 0.034 0.027 0.035 À0.142
*
0.001 À0.020 0.238
*
0.280
*
0.295
*
4 Total Press
release
1 0.927
*
0.884
*
À0.214
*
0.127
*
À0.017 À0.088 0.407
*
0.006
5 Press release –
proactive
1 0.643
*
À0.171
*
0.075 0.032 À0.083 0.338
*
0.004
6 Press releases –
reactive
1 À0.223
*
0.157
*
À0.075 À0.076 0.409
*
0.008
7 Lag legitimacy 1 À0.092 0.115
*
À0.169
*
À0.131
*
0.091
8 Leverage 1 À0.071 À0.081 0.168
*
À0.048
9 Return on assets 1 À0.057 À0.089 0.018
10 Sensitive industry 1 0.039 0.196
*
11 Environmental
news exposure
1 0.129
*
12 Environmental
performance
1
*
Signi?cant at 0.10 two-tailed.
Table 3C
Correlations – news exposure model
1 2 3 4 5
1 News exposure 1 À0.089 0.393
*
0.401
*
0.129
*
2 Return on assets 1 0.133
*
À0.111
*
0.018
3 Size 1 0.244
*
0.004
4 Foreign listing 1 À0.030
5 Environmental
performance
1
*
Signi?cant at 0.10 two-tailed.
6
Since we did not collect data on environmental performance
for the remaining sample, we replace that variable by capital
intensity. Physical plant and equipment makes a ?rm much
more visible to the public and the community at large.
Moreover, a ?rm’s capital intensity is likely to be related to
polluting activities.
7
A disadvantage of the PROBIT procedure is that this
procedure cannot directly compute predicted values. Lee (1983)
has developed a method to estimate the selection model with a
logit analysis that o?ers a less labour-intensive alternative for
computing LAMBDA. Hence, we compute LAMBDAbased on
the following procedure: (1) save predicted probabilities using
LOGIT regression (IKL); (2) using the inverse cumulative
distribution function of the normal distribution, these individual
probabilities are translated into the form they would have had
when they would have been computed based on a probit model
(IPS = probit (IKL)); (3) the IPS variable nowincludes the quasi-
probit scores and can be used to compute LAMBDAin the same
way as when using a probit selection model: LAMBDA = ((1/
sqrt(2 Ã 3.141592654)) Ã (exp (ÀIPS Ã IPSÃ0.5)))/cdfnorm(IPS).
14 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
Table 5
LOGIT and 3SLS regressions on the determinants of environmental legitimacy, environmental news exposure and environmental
disclosure
LOGIT 3SLS
Predicted sign Environmental
news exposure 1/0
Legitimacy Disclosure Environmental
news exposure
Legitimacy
Disclosure + 0.005
**
Disclosure  Environmentally-sensitive industries À À0.005
**
Environmentally-sensitive industries ± 0.127
Environmental performance À À0.006
***
Lag Legitimacy + 0.144
***
Size ± À0.045
Disclosure
Lag legitimacy À 7.174
Environmentally-sensitive industries + 46.986
***
Leverage ± À9.641
Return on assets + À8.086
Environmental news exposure + 32.773
***
Inverse Mills Ratio ± 2.810
Environmental performance ± 0.130
Canada ± À19.076
**
Environmental news exposure
Return on assets + 0.775 À0.569
Foreign listing + 0.194
**
0.390
***
Firm size + 0.881
***
0.317
***
Environmental performance + 0.009
**
Capital intensity ± 0.487
*
Canada ± À0.718
***
À0.346
*
Nagelkerke R-square
Chi-square = 182.24 (0.000)
Overall classi?cation rate = 80.1% 37.0%
R-Square 24.3% 13.6% 32.8%
Chi-square p-value 0.000 0.000 0.000
N 616 158 158 158
One-tailed if there is a predicted sign, two-tailed otherwise. Coe?cients for industry-speci?c dummies not reported.
*
p < 0.10.
**
p < 0.05.
***
p < 0.01.
Table 4
Test of self-selection bias – control variable mean values
Selected sample
(with media coverage)
Remaining sample
(without media coverage)
Mean di?erence
p-value
Return on assets 0.032 0.009 0.088
Foreign listing 0.857 0.697 0.013
Firm size (lnassets) 23.323 22.031 0.000
Capital intensity 0.444 0.341 0.000
N 165 458
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 15
tion rate of 80.1%. All coe?cients except for
Return on assets are statistically signi?cant. This
might be an indication of the presence of selection
bias.
In the second step of the Heckman procedure,
based on the expected probability value, we use
the selection bias control factor (Inverse Mills
Ratio–Lambda) as an additional independent
variable to control for the selection bias in corpo-
rate communication regressions.
Simultaneous test of legitimacy, environmental
disclosure, and environmental news exposure
Since we posit that a ?rm’s communication
strategy a?ects environmental legitimacy, environ-
mental communication, and environmental media
exposure, we ?rst assess whether or not endogene-
ity exists between these variables using the Haus-
man test. Results con?rm endogeneity between
Environmental legitimacy and Annual report envi-
ronmental disclosure (t = 1.839; p < 0.068). We
also reject the null hypothesis of no endogeneity
with respect to Annual report environmental dis-
closure and Media exposure (t = 1.620; p <
0.100). Therefore, it is important to control for
?rms’ incentives to disclose environmental infor-
mation as well as the characteristics of ?rms facing
environmental media exposure in assessing the
determinants of a ?rm’s environmental legitimacy.
Table 5 provides evidence regarding the simul-
taneous test of environmental legitimacy, (Eq.
(1.1)), total annual report environmental disclo-
sure (Eq. (1.2)) and environmental news exposure
(Eq. (1.3)). Concerning the determinants of envi-
ronmental legitimacy, consistent with Hypothesis
1, there is a positive relationship between environ-
mental disclosure and environmental legitimacy
(0.005, p < 0.05). Consistent with Hypothesis 2,
the interaction term ‘‘Annual report disclosure Â
Environmentally-sensitive industries” is negative
and signi?cant (À0.005; p < 0.05) suggesting that
environmental disclosure has a lower impact on
legitimacy for those ?rms operating in more envi-
ronmentally-sensitive industries. Moreover, as
expected, Environmental performance (À0.006;
p < 0.01) is signi?cantly and negatively related to
a ?rm’s media legitimacy. Corroborating the sticky
character of media legitimacy, a ?rm’s prior legit-
imacy is signi?cantly related to its current level of
legitimacy (0.144; p < 0.01)
8
.
As for the determinants of environmental news
exposure, results show that a ?rm’s potential envi-
ronmental visibility, as measured by Foreign list-
ing (0.390; p < 0.01), Environmental performance
(0.009; p < 0.05) and ?rm size (0.317; p < 0.01)
lead to more environmental news. Concerning
the determinants of Annual report environmental
disclosure, self-selection bias related to news
exposure does not appear to be an issue in this
regression model. There is a signi?cant relation-
ship between Environmentally-sensitive industries
(46.986; p < 0.01), Environmental news exposure
(32.773; p < 0.01) and the total Annual report
environmental disclosure score. Furthermore,
total annual report environmental disclosure does
not seem to be in?uenced by prior media legiti-
macy, or by the level of ?rm’s polluting activities
as measured by the environmental performance
variable. The absence of a signi?cant relationship
between Environmental disclosure and Environ-
mental performance is not consistent with prevail-
ing legitimacy theory arguments, but could be
related to the content of the disclosures (Patten,
2002). In this respect, Cho and Patten (2007)
argue that the legitimizing nature of di?erent
types of annual report environmental disclosures
is not identical and that it is important to distin-
guish di?erent types of information when
assessing legitimation e?ects. They primarily dis-
tinguish between litigation related and non-litiga-
tion related disclosure. It is important to note that
they do not distinguish between such types of
information based on the information’s direct
legitimizing e?ect, but rather on the association
of environmental disclosures and environmental
performance. In that vein, their results indicate
that non-litigation related disclosure is higher
for ?rms with lower environmental performance
and for ?rms from environmentally-sensitive
industries.
8
As a sensitivity analysis, we added the Market-to-Book and
Return on assets variables to the model separately. None of the
coe?cients are signi?cant and they do not a?ect our results.
16 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
In our direct approach to measuring legitima-
tion e?ects, we believe the association of environ-
mental disclosure and environmental performance
is not so much an indicator of the legitimizing
nature of the information types, but a proxy for
the information’s intrinsic believability. The
higher the association between the type of envi-
ronmental disclosure and environmental perfor-
mance, the higher the message’s believability and
the greater its potential to a?ect legitimacy per-
ceptions. Although litigation related disclosures
are to a large extent mandatory, their counter-
parts (non-litigation related disclosures) qualify
more easily as incentive-consistent disclosures,
the believability of which users tend to discount
(Hirst et al., 1995; Hodge et al., 2006). Although
our coding grid does not incorporate a clear dis-
tinction between litigation related and non-litiga-
tion related information, we distinguish between
economic-based and social-based environmental
information, with social-based environmental dis-
closure being more incentive-consistent than eco-
nomic-based disclosure. This dichotomous split
of disclosure types resembles the distinction made
by Clarkson et al. (2006) between soft and hard
environmental disclosures, with hard disclosures
re?ecting factual, objective information that can-
not easily be mimicked by poor environmental
performers.
Economic-based types of information focus on
the ?nancial, legal and operational consequences
of corporate environmental activities, most of
which are required disclosures in 10-Ks for public
companies listed with the SEC. This type of
information is mainly comprised within the fol-
lowing four components of our content grid:
expenditure and risk; compliance with laws and
regulations; pollution abatement; and land reme-
diation and contamination, whereas social-based
information relates to the ‘sustainable develop-
ment’ and ‘environmental management’ grid cap-
tions. Disclosure about sustainable development
and environmental management is likely to be
more discretionary, less factual and objective,
and easier to imitate even without substance to
support the claims made. In that vein, social-
based environmental information is likely to be
more incentive-consistent and, thus, we might
expect this type of disclosure to have a lower
impact on a ?rm’s environmental legitimacy than
the more objective, economic-based disclosures.
In our sample, environmental disclosure is mostly
composed of economic-based information (78%
of total disclosures) while more social-related
information is reported by environmentally-sensi-
tive industries (28% of total disclosures for Min-
ing and resources). We estimate our model
separately for economic-based and social-based
disclosures.
The results presented in Table 6 suggest that
economic-based annual report environmental dis-
closure signi?cantly a?ects a ?rm’s media legiti-
macy (0.006; p < 0.05), but that ‘‘social-related”
environmental information segments (0.020;
p < 0.26 one-tailed) do not.
9
For economic-based
environmental disclosure, consistent with Hypoth-
esis 2, the interaction term ‘‘Economic-based
annual report disclosure  Environmentally-sensi-
tive industries” is signi?cant (À0.005; p < 0.05),
suggesting that environmental disclosure has a
marginally di?erent impact on legitimacy for those
?rms operating in more environmentally-sensitive
industries. Moreover, as expected, a bad environ-
mental performance negatively a?ects a ?rm’s
legitimacy (À0.006; p < 0.01). Concerning the
determinants of annual report environmental dis-
closure, a striking di?erence between the types of
content disclosed relates to the association of
Environmental performance and Annual report
disclosure score. Worse environmental perfor-
mance is associated with more elaborate social-
based environmental disclosures but not with the
economic-based variants.
9
The disclosure component Compliance with laws and regu-
lation comprises some items that could communicate negative
environmental image (e.g. ?nes, order to conform, incidents).
These litigation related disclosures are more mandated in nature
and less likely to be used as a legitimizing tool (Patten, 2002).
Therefore, the economic-based disclosure model is re-estimated
dropping the component Compliance with laws and regulation
from the disclosure score. Results remain similar as those
reported in Table 6 (Coe?cient for economic-based disclosure =
0.007; p < 0.050, and Economic-based disclosure  Environ-
mentally-sensitive industries = À0.006; p < 0.050).
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 17
Since visibility factors are likely to di?er across
US and Canadian ?rms, we test for potential inter-
active country di?erences of the impact of disclosure
on legitimacy. As a sensitivity analysis (results not
tabulated), we add two interaction terms to the
legitimacy model ‘‘Economic-based annual
report disclosure  Canada” and ‘‘Economic-based
annual report disclosure  Environmentally-sensi-
tive industries  Canada”. Coe?cients for these
two interaction terms are not signi?cant (À0.001;
p < 0.775 two-tailed, and -0.0008; p < 0.875 two-
tailed), suggesting that mixing US and Canadian
data does not bias our results.
Simultaneous test of legitimacy, environmental
press releases, and environmental news exposure
Table 7 provides evidence regarding the simul-
taneous test of environmental legitimacy, (Eq.
(2.1)), environmental press releases (Eq. (2.2))
Table 6
3SLS regressions on the determinants of environmental legitimacy, environmental news exposure and environmental disclosure
Predicted
sign
Economic-based (expenditures/
remediation/pollution abatement and
norms)
Sustainable development and
environmental management
Legitimacy Disclosure Environmental
news exposure
Legitimacy Disclosure Environmental
news exposure
Legitimacy
Disclosure + 0.006
**
0.020
Disclosure  Environmentally-
sensitive industries
À À0.005
**
À0.021
Environmentally-sensitive
industries
± 0.091 0.124
Environmental performance À À0.006
***
À0.006
***
Lag legitimacy + 0.151
***
0.130
**
Size ± À0.043 À0.049
Disclosure
Lag Legitimacy À 8.035 À0.866
Environmentally-sensitive
industries
+ 39.029
***
7.927
**
Leverage ± À0.725 À7.471
Return on assets + À14.821 5.846
Environmental news exposure + 23.918
***
8.630
***
Inverse Mills ratio ± À7.195 9.814
Environmental performance + 0.035 0.101
*
Canada ± À20.844
***
1.537
Environmental news exposure
Return on assets + À0.666 À0.505
Foreign listing + 0.379
***
0.426
***
Firm size + 0.342
***
0.285
***
Environmental performance + 0.009
**
0.009
**
Canada ± À0.322
*
À0.389
**
R-Square 27.3% 19.7% 34.8% 18.2% 6.3% 32.8%
F-statistic p-value 0.000 0.000 0.000 0.000 0.000 0.000
N 158 158 158 158 158 158
One-tailed if there is a predicted sign, two-tailed otherwise.
Coe?cients for industry-speci?c dummies not reported.
*
p < 0.10.
**
p < 0.05.
***
p < 0.01.
18 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
and environmental news exposure (Eq. (2.3)).
Total press releases are not associated with a ?rm’s
legitimacy (0.006; p < 0.306 one-tailed). In Table 8,
we present separate regressions for proactive and
reactive press releases. Consistent with Hypothesis
3, our results show that only reactive press releases
increase a ?rm’s perceived environmental legiti-
macy (0.029; p < 0.05) while there is no impact
for proactive press releases (À0.009; p < 0.652
two-tailed). However, the interaction term
‘‘Reactive Press releases  Environmentally-sensi-
tive industries” is negative but not signi?cant
(À0.001; p < 0.494 one-tailed), suggesting that the
impact of environmental press releases on legiti-
macy does not di?er for those ?rms operating in
more environmentally-sensitive industries. Con-
cerning the determinants of environmental press
releases, as expected, results show a negative
relationship between prior media legitimacy and
the use of environmental press releases (À0.617;
p < 0.05 for proactive press releases and À0.600;
p < 0.01 for reactive press releases), suggesting
tactical impression management motives behind
the use of press releases with the objective of
enhancing corporate environmental legitimacy
perceptions.
In order to explore whether annual report envi-
ronmental disclosures and environmental press
releases interfere in their e?ect on media legiti-
macy, we include both variables in simultaneous
Table 7
3SLS regressions on the determinants of environmental legitimacy, environmental news exposure and environmental press releases
Predicted
sign
Legitimacy Environmental
press releases
Environmental
news exposure
Legitimacy
Press releases + 0.006
Press releases  Environmentally-sensitive industries À 0.001
Environmentally-sensitive industries ± À0.010
Environmental performance À À0.006
***
Lag Legitimacy + 0.164
***
Size ± À0.027
Press releases
Lag Legitimacy À À1.216
***
Environmentally-sensitive industries + À0.161
*
Leverage ± 1.300
Return on assets + 1.652
Environmental news exposure + 1.367
***
Inverse Mills Ratio ± À2.025
Environmental performance ± 0.007
Canada ± À0.863
Environmental news exposure
Return on assets + À1.012
Foreign listing + 0.412
***
Firm size + 0.374
***
Environmental performance + 0.010
**
Canada ± À0.836
R-square 28.9% 20.7% 33.4%
F-statistic p-value 0.000 0.000 0.000
N 158 158 158
One-tailed if there is a predicted sign, two-tailed otherwise.
Coe?cients for industry-speci?c dummies not reported.
*
p < 0.10.
**
p < 0.05.
***
p < 0.01.
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 19
regressions of legitimacy, environmental press
releases and environmental disclosure. We restrict
the analysis to economic-based environmental dis-
closures. Including both economic-based annual
report disclosures and proactive press releases
(Table 9), we ?nd that proactive press releases do
not add to the legitimacy-enhancement e?ect of
annual report environmental disclosures, suggest-
ing that proactive press releases do not provide
incremental information to the other with respect
to environmental legitimacy enhancement. On
the other hand, including both economic-based
annual report disclosures and reactive press
releases (Table 9), the e?ect of annual report
environmental disclosure becomes insigni?cant,
while the coe?cient of reactive press releases
remains positive and signi?cant, suggesting a sub-
stitutional e?ect of reactive press releases with
regard to annual environmental disclosures. In line
with earlier results concerning the determinants of
environmental communication, we ?nd a negative
relationship between prior legitimacy and
press releases (À0.629; p < 0.05 for proactive press
releases and À0.598; p < 0.01 for reactive press
releases) while there is no signi?cant association
between prior media legitimacy and current annual
report economic-based disclosure.
Discussion and conclusion
Symbolic management theory (Pfe?er & Sala-
ncik, 1978) and related impression management
perspectives (Ginzel, Kramer, & Sutton, 1993;
Suchman, 1995) suggest that a ?rm’s management
is expected not only to manage the ?rm’s perfor-
mance (including its environmental performance),
but also perceptions of its (environmental) perfor-
mance. In that sense, a ?rm’s environmental
communication e?orts through environmental dis-
closures in annual reports and environmental press
releases represent predictable opportunities for
impression management and legitimation for the
?rm’s environmental activities and performance.
However, studying legitimation and legitimacy
issues empirically has its methodological chal-
lenges, such as the practical problems of assessing
subjective perceptions and beliefs of relevant pub-
lics. In this study, we investigated the impact of a
?rm’s environmental communication e?orts on
media legitimacy as a direct measure of perceived
legitimacy. Public media content captures the per-
spective of the general public and has been used as
a proxy for normative legitimacy issues (Bansal &
Clelland, 2004; Deephouse, 1996; Elsbach & Sut-
ton, 1992).
Previous environmental reporting studies
mainly focused on one communication channel
(annual report disclosures). However, in most
cases, annual report disclosures constitute only
one part of the ?rm’s communication repertoire
and the use of one channel may well a?ect the
use and e?ectiveness of other channels. Moreover,
di?erent channels may not be equally e?cient or
e?ective in reaching speci?c communication goals
(such as legitimacy enhancement). In this study,
we looked at both annual report environmental
disclosures and environmentally-related press
releases which may function in complementary or
substitutional roles.
The results obtained in this paper suggest that
perceived environmental legitimacy is positively
a?ected by the extent and quality of economic-
based segments of environmental disclosures in
annual reports and by reactive environmental
press releases. Consistent with Hypothesis 1, our
results document a signi?cant association between
the extent of annual report environmental disclo-
sures and reactive press releases on the one hand
and environmental (media) legitimacy on the
other. The association between annual report envi-
ronmental disclosures and media legitimacy is
essentially driven by the economic-based segments
of the annual report disclosures. We did not ?nd
evidence to support Hypothesis 1 with regard to
the extent of proactive press releases. The di?eren-
tial results with regard to the association of reac-
tive versus proactive press releases and media
legitimacy is consistent with Hypothesis 3 that pre-
dicts a stronger association for reactive environ-
mental press releases relative to proactive ones.
Relatedly, the interactive e?ect of corporate envi-
ronmental communication and environmentally-
sensitive industries (predicted in Hypothesis 2)
can only be demonstrated for the economic-based
segments of annual report environmental disclo-
20 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
sures. This suggests that industry-level legitimacy
functions as a constraining a priori impression
for annual report economic-based disclosures and
diminishes the legitimation e?ectiveness of these
communication e?orts, but not so of reactive press
releases.
Although not explicitly hypothesized, our
results shed some light on the complementary roles
of annual report environmental disclosures and
environmental press releases as they suggest that
proactive press releases do not interfere with the
e?ectiveness of annual report environmental dis-
closures, while the opposite is true for reactive
press releases. Moreover, in concert with the main
theme of the strategic legitimacy perspective on
environmental reporting, our results suggest that
Table 8
3SLS regressions on the determinants of environmental legitimacy, environmental news exposure and environmental press releases
Predicted
sign
Proactive press releases Reactive press releases
Legitimacy Environmental
press releases
Environmental
news exposure
Legitimacy Environmental
press releases
Environmental
news exposure
Legitimacy
Press releases + À0.009 0.029
**
Press releases Â
Environmentally-
sensitive industries
À 0.010 À0.001
Environmentally-
sensitive industries
± À0.017 À0.011
Environmental
performance
À À0.006
***
À0.006
***
Lag legitimacy + 0.165
***
0.165
***
Size ± À0.019 À0.032
Press releases
Lag Legitimacy À À0.617
**
À0.600
***
Environmentally-
sensitive industries
+ À0.038 À0.124
Leverage ± 0.076 1.205
Return on Assets + 2.118 À0.466
Environmental news
exposure
+ 0.819
***
0.546
***
Inverse mills ratio ± À0.139 À1.887
*
Environmental
performance
± 0.001 0.007
Canada ± À0.482 À0.355
Environmental news exposure
Return on assets + À0.906 À0.905
Foreign listing + 0.392
***
0.405
***
Firm size + 0.393
***
0.366
***
Environmental
performance
+ 0.008
**
0.008
**
Canada ± À0.295 À0.320
*
R-square 28.6% 13.8% 33.3% 30.1% 22.5% 33.5%
F-statistic p-value 0.000 0.000 0.000 0.000 0.000 0.000
N 158 158 158 158 158 158
One-tailed if there is a predicted sign, two-tailed otherwise.
Coe?cients for industry-speci?c dummies not reported.
*
p < 0.10.
**
p < 0.05.
***
p < 0.01.
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 21
negative media legitimacy is a signi?cant driver of
environmental press releases. This is, however, not
the case for annual report environmental disclo-
sures. Finally, environmental news exposure is
associated with both annual report environmental
disclosures and environmental press releases.
Table 9
3SLS regressions on the determinants of environmental legitimacy, environmental press releases, and economic-based environmental
disclosure
Predicted sign Proactive press releases Reactive press releases
Legitimacy Environmental
press releases
Environmental
disclosure
Legitimacy Environmental
press releases
Environmental
disclosure
Legitimacy
Disclosure + 0.006
**
0.001
Press releases + À0.048 0.035
**
Lag legitimacy + 0.176
***
0.166
***
Environmentally-
sensitive industries
± À0.193
*
À0.022
Environmental
performance
À0.007
***
À0.006
***
Size ± À0.035 À0.035
Press releases
Lag Legitimacy À À0.629
**
0.598
***
Environmentally-
sensitive
industries
+ À0.059 À0.121
Leverage ± 0.155 1.199
Return on Assets + 1.900 À0.440
Environmental news
exposure
+ 0.629
***
0.567
***
Inverse mills ratio ± À0.133 À1.885
*
Environmental
performance
± 0.003 0.009
Canada ± À0.529 À0.202
Environmental disclosure
Lag legitimacy À 5.836 6.312
Environmentally-
sensitive
industries
+ 36.984
***
37.150
***
Leverage ± 6.247 8.094
Return on assets + À27.299 À29.128
Lag environmental
news exposure
+ 10.572
***
10.638
***
Inverse mills ratio ± À13.035 À12.013
Environmental
performance
0.226 0.226
Canada À24.132
***
À24.494
***
R-square 28.4% 14.6% 29.1% 29.9% 22.5% 29.1%
F-statistic p-value 0.000 0.000 0.000 0.000 0.015 0.000
N 158 158 158 158 158 158
One-tailed if there is a predicted sign, two-tailed otherwise.
Coe?cients for industry-speci?c dummies not reported.
*
p < 0.10.
**
p < 0.05.
***
p < 0.01.
22 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
The present research especially highlights the
role of economic-based environmental informa-
tion in annual reports and of reactive press
releases as perception management tools. The
more subjective social-based environmental dis-
closure in annual reports does not a?ect media
legitimacy, whereas the more objective, harder
part of annual report environmental disclosure
does. Moreover, reactive environmental press
releases, through a mixture of verbal remedial tac-
tics, are found to be e?ective in transforming
events that seem at ?rst to be image-threatening
into messages that ultimately protect and even
enhance a ?rm’s environmental media legitimacy.
This is in line with previous impression manage-
ment research that suggested that di?erent forms
of verbal accounts a?ect legitimacy by attenuating
organizational responsibility for controversial
events and by accentuating the positive aspects
of such events. In particular, accounts with refer-
ences to widely institutionalized structures, pro-
grams and procedures (such as formal company
policies, budgetary constraints) are taken to be
e?ective in protecting legitimacy, because they
improve the adequacy of the verbal claims by pro-
viding evidence to bolster the verbal accounts
(Elsbach, 1994; Elsbach & Elofson, 2000). Future
research might look into the composition of (dif-
ferent sets and sequences of) verbal remedial tac-
tics, their form and content to identify the most
e?ective patterns.
The lack of impact of proactive press releases
and of the more subjective, social-based annual
report environmental disclosures on media legiti-
macy suggests that public media seem to discount
transparent self-promotional behavior, recogniz-
ing that ?rms tend to exaggerate their speci?c mer-
its in the environmental management domain. It
may be another occurrence of the ‘‘self-promoter’s
paradox”: as actors increase claims of competence,
audiences become more sceptical as competent
actors often downplay their successes (Jones &
Pittman, 1982). Alternatively, the lack of signi?-
cant association between the use of proactive press
releases and media legitimacy may also be
explained by the assertion that narcissistic behav-
ior is only what is generally expected of ?rms
and, as such, does not add to the assignment of
credit. This result is also consistent with the fact
that the softer part of annual report environmental
disclosures is not associated with enhanced media
legitimacy, whereas higher-quality economic-
based environmental disclosures do a?ect media
legitimacy, except when complemented by reactive
press releases.
Finally, the observation that mainly the extent
and not the evaluative propensity of public media
environmental exposure is associated with the
level and quality of annual report environmental
disclosure corroborates our initial assertion that
annual report environmental disclosures are
mainly driven by longer-term, strategic motives.
Annual report environmental disclosure does not
directly respond to the relatively current and tem-
porary legitimacy perceptions that transpire
through the public media. On the other hand,
the value as a legitimation device of environmen-
tal press releases is most evident in times of crisis
or controversy when environmental legitimacy is
challenged or threatened. Our results suggest that
in these circumstances both proactive and reactive
press releases are actively triggered, with only
reactive press releases being capable of signi?-
cantly mitigating legitimacy perceptions, while in
those circumstances the e?ect of annual report
environmental disclosures vanishes. These obser-
vations suggest that future research in corporate
environmental reporting might fruitfully distin-
guish between more enduring reputational e?ects
and relatively short-lived legitimacy e?ects based
on speci?c perceptions of environmental issues
(Elsbach, 2003) and of environmental communi-
cation devices. Substantiating such di?erential
e?ects would add to the corporate environmental
reporting literature.
Our measure of legitimacy might raise questions
about the generalizability of our ?ndings since it lar-
gely depends on the extent of comprehensiveness of
the media coverage database and the reliability of
measurements for that coverage. Editorial bias that
is systematically related to the independent vari-
ables of our models might also a?ect our results.
Despite these limitations, it is encouraging to ?nd
that with the di?culty in measuring data, we are
able to document direct legitimation e?ects of envi-
ronmental communication practices.
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 23
Acknowledgements
The authors acknowledge the ?nancial support
of the Social Sciences and Humanities Research
Council of Canada, the Fonds que´be´cois de la
recherche sur la socie´te´ et la culture (FQRSC),
the National Bank of Belgium and the Autorite´
des Marche´s Financiers (Que´bec) and the Corpo-
rate Reporting Chair (UQAM).
Appendix 1
Environmental Disclosure Grid
Expenditures and risks Sustainable development
Investments Natural resource conservation
Operation costs Recycling
Future investments Life cycle information
Future operating costs Land remediation and contamination
Financing for investments Sites
Environmental debts Remediation e?orts
Risk provisions Potential liability-remediation
Risk litigation Implicit liability
Provision for future expenditures Spills (number, nature, reduction e?orts )
Compliance with laws and regulations Environmental management
Litigation, actual and potential Environmental policies or company concern for the environment
Fines Environmental management system
Orders to comply Environmental auditing
Corrective action Goals and targets
Incidents Awards
Future legislation and regulations Department, group, service assigned to the environment
Pollution abatement ISO 14000
Emission of pollutants Involvement of the ?rm to develop environmental standards
Discharges Involvement of environmental organizations (industry
committees, etc.)
Waste management Joint environmental management projects with other ?rms
Installation and process controls
Compliance status of facilities
Noise and odours
Rating scale: 3: Item described in monetary or quantitative terms; 2: Item described speci?cally; 1: Item discussed in general.
References
Ader, C. R. (1995). A longitudinal study of agenda setting for
the issue of environmental pollution. Journalism and Mass
Communication Quarterly, 72(2), 300–311.
Aldrich, H. E., & Fiol, C. M. (1994). Fools rush in? The
institutional context of industry creation. Academy of
Management Review, 19(4), 645–670.
Al-Tuwaijri, S., Christensen, T. E., & Hughes, K. E. II, (2004).
The relations among environmental disclosure, environ-
mental performance, and economic performance: A simul-
taneous equations approach. Accounting, Organizations and
Society, 29, 447–471.
Anderson, U., Kadous, K., & Koonce, L. (2004). The role of
incentives to manage earnings and quanti?cation in audi-
tors’ evaluations of management-provided information.
Auditing: A Journal of Practice and Theory, 23(1), 11–27.
Ashforth, B. E., & Gibbs, B. W. (1990). The double-edge of
organizational legitimation. Organization Science, 1,
177–194.
24 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
Bansal, P., & Clelland, I. (2004). Talking trash: Legitimacy,
impression management, and unsystematic risk in the
context of the natural environment. Academy of Manage-
ment and Business Research, 29(1), 21–41.
Barton, J., & Mercer, M. (2005). To blame or not to blame:
Analysts’ reactions to external explanations for poor
?nancial performance. Journal of Accounting and Econom-
ics, 39(3), 509–533.
Baum, J. A. C., & Oliver, C. (1991). Institutional linkages and
organizational mortality. Administrative Science Quarterly,
36, 187–218.
Baum, Joel A. C., & Powell, W. W. (1995). Cultivating an
institutional ecology of organizations: comment on Han-
nan, Carroll, Dundon, and Torres. American Sociological
Review, 60, 529–539.
Bewley, K., & Li, Y. (2000). Disclosure of environmental
information by Canadian manufacturing companies: A
voluntary disclosure perspective. Advances in Environmental
Accounting and Management, 1, 201–226.
Birnbaum, M., & Stegner, S. (1979). Source credibility in social
judgment: Bias, expertise, and the judge’s point of view.
Journal of Personality and Social Psychology, 48–74.
Botosan, C. (1997). Disclosure level and the cost of equity
capital. The Accounting Review, 72, 323–349.
Brown, N., & Deegan, C. (1998). The public disclosure of
environmental performance information – A dual test of
media agenda setting theory and legitimacy theory.
Accounting and Business Research, 29(1), 21–41.
Buhr, N. (1998). Environmental performance, legislation and
annual report disclosure: The case of acid rain and
falconbridge. Accounting, Auditing and Accountability Jour-
nal, 163–190.
Caroll, C. E., & McCombs, M. (2003). Agenda-setting e?ects of
business news on the public’s images and opinions about
major corporations. Corporate Reputation Review, 6(1),
36–46.
Cho, C. H., & Patten, D. M. (2007). The role of environmental
disclosures as tools of legitimacy: A research note. Account-
ing, Organizations and Society, 32, 639–647.
Clarkson, P., Li, Y., Richardson, G. D., & Vasvari, F. P.
(2006). Revisiting the relation between environmental
performance and environmental disclosure: An empirical
analysis. In CAAA 2006 annual conference paper, April 5,
2006. Available at SSRN: .
Cochran, P., & Wood, R. (1984). Corporate social responsibil-
ity and ?nancial performance. Academy of Management
Journal, 42–56.
Cormier, D., & Gordon, I. M. (2001). An examination of social
and environmental reporting strategies. Accounting, Audit-
ing and Accountability Journal, 14(5), 587–616.
Cormier, D., & Magnan, M. (1999). Corporate environmental
disclosure strategies: Determinants, costs and bene?ts.
Journal of Accounting, Auditing and Finance, 14(3), 429–451.
Cormier, D., & Magnan, M. (2003). Environmental reporting
management: A european perspective. Journal of Accounting
and Public Policy, 22, 43–62.
Cowen, S. S., Ferreri, L. B., & Parker, L. D. (1987). The impact
of corporate characteristics on social responsibility disclo-
sure: A typology and frequency-based analysis. Accounting,
Organizations and Society, 111–122.
Crant, J. M., & Bateman, T. S. (1993). Assignment of credit and
blame for performance outcomes. Academy of Management
Journal, 36(1), 7–27.
Daft, R. L., & Lengel, R. H. (1984). Information richness: A
new approach to managerial behavior and organizational
design. Research in Organizational Behavior, 6, 191–233.
Daft, R. L., & Lengel, R. H. (1986). Organizational informa-
tion requirements, media richness and structural design.
Management Science, 32(5), 554–571.
Deegan, C., & Gordon, B. (1996). A study of the environmental
disclosure practices of Australian corporations. Accounting
and Business Research, 26(3), 187–199.
Deegan, C., & Rankin, M. (1996). Do australian companies
report environmental disclosures by ?rms prosecuted suc-
cessfully by environmental protection authority. Account-
ing, Auditing and Accountability Journal, 9(2), 50–67.
Deegan, C., Rankin, M., & Tobin, J. (2002). An examination
of the corporate social and environmental disclosures of
BHP from 1983 to 1997 – A test of legitimacy theory.
Accounting, Auditing and Accountability Journal, 15(3),
312–343.
Deegan, C., Rankin, M., & Voght, P. (2000). Firms’ disclosure
reactions to major social incidents: Australian evidence.
Accounting Forum, 24(1), 101–130.
Deephouse, D. L. (1996). Does isomorphism legitimate?
Academy of Management Journal, 39, 1024–1039.
Deephouse, D. L. (2000). Media reputation as a strategic
resource: An integration of mass communication and
resource-based theories. Journal of Management, 26(6),
1091–1112.
Deephouse, D. L., & Carter, S. M. (2005). An examination of
di?erences between organizational legitimacy and organiza-
tional reputation. Journal of Management Studies, 42(2),
329–360.
Dowling, J., & Pfe?er, J. (1975). Organizational legitimacy:
Social values and organizational behavior. Paci?c Sociolog-
ical Review, 18(1), 122–136.
Elijido-Ten, E. (2004). Determinants of environmental disclo-
sures in a developing country: An application of the
stakeholder theory. In Fourth Asia Paci?c interdisciplinary
research in accounting conference, Singapore.
Elsbach, K. D. (1994). Managing organizational legitimacy in
the california cattle industry: The construction and e?ec-
tiveness of verbal accounts. Administrative Science Quar-
terly, 39, 57–88.
Elsbach, K. D. (2003). Organizational perception management.
In L. L. Cummings & B. M. Staw (Eds.). Research in
Organizational Behavior (25, pp. 297–332).
Elsbach, K. D., & Elofson, G. (2000). How the packaging of
decision explanations a?ects perceptions of trustworthiness.
Academy of Management Journal, 43(1), 80–89.
Elsbach, K. D., & Kramer, R. M. (1996). Members’ responses
to organizational identity threats: Encountering and
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 25
countering the business week rankings. Administrative
Science Quarterly, 4.
Elsbach, K. D., & Sutton, R. (1992). Acquiring organizational
legitimacy through illegitimate actions: A marriage of
institutional and impression management theories. Academy
of Management Journal, 35, 699–739.
Fanelli, A., & Misangyi, V. F. (2006). Bringing out charisma:
CEO charisma and external stakeholders. Academy of
Management Review, 31(4), 1049–1061.
Fekrat, M. A., Inclan, C., & Petroni, D. (1996). Corporate
environmental disclosures: Competitive disclosure hypoth-
esis using 1991 annual report data. The International Journal
of Accounting, 31(2), 175–195.
Fombrun, C. (1996). Reputation: Realizing value from the
corporate image. Boston, MA: Harvard Business School
Press.
Freedman, M., & Wasley, C. (1990). The association between
environmental performance and environmental disclosure in
annual reports and 10Ks. Advances in Public Interest
Accounting, 183–193.
Gamson, W. A., Croteau, D., Hoynes, W., & Sasson, T. (1992).
Media images and the social construction of reality. Annual
Review of Sociology, 18, 373–393.
Gardner, W. L., & Martinko, M. J. (1988). Impression
management in organizations. Journal of Management,
14(2), 321–338.
Ginzel, L. E., Kramer, R. M., & Sutton, R. I. (1993).
Organizational Impression Management as a Reciprocal
In?uence Process: The Neglected Role of the Organiza-
tional Audience. In L. L. Cummings & B. M. Staw
(Eds.). Research in Organizational Behavior (15, pp. 227–
266).
Hackston, D., & Milne, M. J. (1996). Some determinants of
social and environmental disclosures in New Zealand
companies. Accounting, Auditing and Accountability Journal,
25, 77–108.
Hannan, M. T., & Freeman, J. (1989). Organizational ecology.
Cambridge, MA: Harvard University Press.
Heckman, J. J. (1979). Sample selection bias as a speci?cation
error. Econometrica, 47(1), 153–162.
Hirst, D., Koonce, L., & Simko, P. (1995). Investor reactions to
?nancial analysts’ research reports. Journal of Accounting
Research, 33(2), 335–351.
Hodge, F., Hopkins, P. E., & Pratt, J. (2006). Management
reporting incentives and classi?cation credibility: The e?ects
of reporting discretion and reputation. Accounting, Organi-
zations and Society, 31(7), 623–634.
Hogner, R. H. (1982). Corporate social reporting: Eight
decades of development at US steel. Research in Corporate
Performance and Policy, 4, 243–250.
Hughes, S. B., Anderson, A., & Golden, S. (2001). Corporate
environmental disclosures: Are they useful in determining
environmental performance?. Journal of Accounting and
Public Policy 20, 217–240.
Ingram, R. W., & Frazier, K. B. (1980). Environmental
performance and corporate disclosure. Journal of Account-
ing Research, 18, 615–622.
Jaggi, B., & Freedman, M. (1982). An analysis of the
informational content of pollution disclosures. The Finan-
cial Review, 142–152.
Janis, I. L., & Fadner, R. (1965). The Coe?cient of Imbalance.
In H. Lasswell, N. Leites, & Associates (Eds.), Language of
politics (pp. 153–169). Cambridge, MA: MIT Press.
Jones, E. E., & Pittman, T. S. (1982). Toward a theory of
strategic self-presentation. In J. Suls (Ed.), Psychological
perspectives on the self (pp. 231–262). Hillsdale, NJ:
Erlbaum.
Kim, P. H., Dirks, KT., Cooper, C. D., & Ferrin, D. L.
(2006). When more blame is better than less: The
implications of internal vs. external attributions for the
repair of trust after a competence- vs. integrity-based trust
violation. Organizational Behavior and Human Decision
Processes, 99, 49–65.
Leary, M. R., & Kowalski, R. M. (1990). Impression manage-
ment: A literature review and two-component model.
Psychological Bulletin, 107(1), 34–47.
Lee, L. F. (1983). Generalized econometric models with
selectivity. Econometrica, 51(2), 507–513.
Lindblom, C. K. (1994). The implications of organizational
legitimacy for corporate social performance and disclosure.
In Paper presented at the critical perspectives on accounting
conference, New York, NY.
Li, Y., Richardson, G. D., & Thornton, D. B. (1997).
Corporate disclosure of environmental liability information:
Theory and evidence. Contemporary Accounting Research,
14(3), 435–474.
McCombs, M. E., & Shaw, D. L. (1972). The agenda setting
function of the mass media. Public Opinion Quarterly, 36,
176–187.
McGuire, J., Sundgren, A., & Schneeweis, T. (1988). Corporate
social responsibility and ?rm ?nancial performance. Acad-
emy of Management Journal(December), 854–872.
Mercer, M. (2004). How do investors assess the credibility of
management disclosures?. Accounting Horizons 18(3),
185–196.
Mills, D., & Gardner, M. (1984). Financial pro?les and the
disclosure of expenditures for socially responsible purposes.
Journal of Business Research(December), 407–424.
Murray, A., Sinclair, D., Power, D., & Gray, R. (2006). Do
?nancial markets care about social and environmental
disclosure? Accounting, Auditing and Accountability Journal,
19(2), 228–255.
Neu, D., Warsame, H., & Pedwell, K. (1998). Managing public
impressions: Environmental disclosures in annual reports.
Accounting, Organizations and Society, 23(3), 265–282.
Nunnaly, J. (1978). Psychometric theory (end ed.). New York:
McGraw Hill.
O’Donovan, G. (2002). Environmental disclosures in annual
report – Extending the applicability and power of legitimacy
theory. Accounting, Auditing and Accountability Journal,
15(3), 344–371.
O’Dwyer, B. (2002). Managerial perceptions of corporate social
disclosure – An Irish story. Accounting, Auditing and
Accountability Journal, 15(3), 406–436.
26 W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27
Oliver, C. (1991). Strategic responses to institutional processes.
Academy of Management Review, 16(1), 145–179.
Patten, D. M. (1991). Exposure, legitimacy, and social disclo-
sure. Journal of Accounting and Public Policy, 10, 297–308.
Patten, D. M. (1992). Intra-industry environmental disclosures
in response to the alaskan oil spill: A note on legitimacy
theory. Accounting, Organizations and Society, 17(5),
471–475.
Patten, D. M. (2002a). Media exposure, public policy pressure,
and environmental disclosure: An examination of the
impact of tri data availability. Accounting Forum, 26(2),
153–171.
Patten, D. M. (2002b). The relation between environmental
performance and environmental disclosure: A research note.
Accounting, Organizations and Society, 27, 763–773.
Patten, D. M. (2005). The accuracy of ?nancial reports
projections of future environmental capital expenditures:
A research note. Accounting, Organizations and Society, 30,
457–468.
Patterson, J. D., & Allen, M. W. (1997). Accounting for your
actions: How stakeholders respond to the strategic commu-
nication of environmental activist organizations. Journal of
Applied Communication Research, 25, 293–316.
Pfe?er, J., & Salancik, G. R. (1978). The external control of
organizations. New York: Harper and Row.
Pollock, T., & Rindova, V. P. (2003). Media legitimation e?ects
in the market for initial public o?erings. Academy of
Management Journal, 46(5), 631–642.
Rao, H. (1994). The social construction of reputation: Certi-
?cation contests, legitimation and the survival of organiza-
tions in the American automobile industry: 1985–1912.
Strategic Management Journal, 15, 29–44.
Rao, H. (1998). Caveat emptor: The construction of nonpro?t
consumer watchdog organizations. American Journal of
Sociology, 103(4), 912–961.
Rao, H. (2001). In Joel A. C. Baum (Ed.), Interorganizational
ecology, companion to organizations (pp. 541–556). London:
Basil Blackwell.
Richardson, A. J., & Welker, M. (2001). Social disclosure,
?nancial disclosure and the cost of capital. Accounting,
Organizations and Society, 26(7/8), 597–616.
Roberts, R. W. (1992). Determinants of corporate social
responsibility disclosure: An application of stakeholder
theory. Accounting, Organizations and Society, 17(6),
595–612.
Rockness, J. W. (1985). An assessment of the relationship
between US corporate environmental performance and
disclosure. Journal of Business Finance and Accounting,
12(3), 339–354.
Savage, A., Rowlands, J. & Cataldo, A. J. (1999). Environmen-
tal disclosure in annual reports: A legitimacy theory frame-
work. Working Paper Oakland University.
Schultz, M., Mouritsen, J., & Gabrielsen, G. (2001). Sticky
reputation: Analyzing a ranking system. Corporate Reputa-
tion Review, 4(1), 24–41.
Scott, W. R. (1995). Institutions and organizations. Thousand
Oaks, CA: Sage.
Shapiro, D. L. (1991). The e?ects of explanations on negative
reactions of deceit. Administrative Science Quarterly, 36,
614–630.
Sheer, V. C., & Chen, L. (2004). Improving media richness
theory. Management Communication Quarterly, 18(1),
76–93.
Suchman, M. C. (1995). Managing legitimacy: Strategic and
institutional approaches. Academy of Management Review,
20, 571–610.
Tedeschi, J. T., & Melburg, V. (1984). Impression management
and in?uence in the organization. In S. B. Bacharach & E. J.
Lawler (Eds.). Research in the Sociology of Organizations
(3, pp. 31–58).
Tyler, T. R. (2006). Legitimacy and legitimation. Annual Review
of Psychology, 57, 375–400.
Walden, D. W., & Schwartz, B. N. (1997). Environmental
disclosures and public policy pressure. Journal of Accounting
and Public Policy, 16, 125–154.
Weber, R. P. (1990). Basic content analysis (second ed.).
Newbury Park, CA: Sage.
Williams, S. M., & Ho Wern Pei, C. A. (1999). Corporate social
disclosures by listed companies on their web sites: An
international comparison. The International Journal of
Accounting, 34(3), 389–419.
Wilmshurst, T. D., & Frost, G. R. (2000). Corporate environ-
mental reporting: A test of legitimacy theory. Accounting,
Auditing and Accountability Journal, 13(1), 10–26.
Wiseman, J. (1982). An evaluation of environmental disclosures
made in corporate annual reports. Accounting, Organiza-
tions and Society, 7(1), 53–63.
Wood, R. E., & Mitchell, T. R. (1981). Manager behavior in a
social context: The impact of impression management on
attributions and disciplinary actions. Organizational Behav-
ior and Human Decision Processes, 28, 356–378.
Zimmerman, M. A., & Zeitz, G. J. (2002). Beyond survival:
Achieving new venture growth by building legitimacy.
Academy of Management Review, 27, 414–431.
W. Aerts, D. Cormier / Accounting, Organizations and Society 34 (2009) 1–27 27
doc_331361131.pdf