If one lie doesn’t work, lie again? A sequential decoupling act in
response to corporate misconduct
ABSTRACT
Corporate misconduct damages the social evaluations of a firm. Since firms depend on positive evaluations from corporate audiences to survive, a substantial amount of research has directed its efforts to examining how firms influence their social evaluations. This work, however, has mainly focused on the effects of individual tactics. In practice, firms often combine multiple actions over an extended period of time. To advance this line of research, we examine how firms combine multiple impression management tools to revert negative assessments from the public following misconduct. We test our theory in the context of environmental misconduct by publicly traded US firms from 2009 to 2018. Results from the empirical analyses confirm our hypothesis. We find that when firms combine multiple impression management tools in a sequential and interrelated way, they can revert negative social evaluations even if the actions announced are largely symbolic. These results provide important implications for both theory and practice.
Keywords: Impression management, social evaluations, reputation, corporate misconduct, environmental misconduct, greenwashing, top management, information disclosure, environmental disclosure.
Guido Berends University of Groningen
S2709457
Groningen, August 2019
prof. dr. Jordi Surroca University of Groningen
INTRODUCTION
Corporate misconduct damages the social evaluations of a firm (Devers, Dewett, Mishina &
Belsito, 2009; Gomulya & Mishina, 2017; Hersel, Helmuth, Zorn, Shropshire & Ridge, 2019;
Mishina, Dykes, Block & Pollock, 2010). Firms depend on positive social evaluations from
corporate audiences to survive and generate performance benefits (Rao, 1994; Rindova,
Petkova & Sever, 2005). Therefore, they must adequately respond in the aftermath of
misconduct (Carlos & Lewis, 2018; Gomulya & Mishina, 2017; Hersel et al., 2019; Lamin &
Zaheer, 2012).
Firms can attempt to influence their evaluations by issuing verbal accounts through the
media (Bolino, Kacmar, Turnley & Gilstrap, 2008; Elsbach, 1994; Pfarrer, Decelles, Smith & Taylor, 2008; McDonell & King, 2013). The goal being to return an actor that engaged in controversial activities to social acceptability and thereby avoiding negative outcomes (Davidson & Friedman, 1998). The effectiveness of verbal accounts depends on the actions the firm announces (Coombs, 2007; Mishina, Block & Mannor, 2012; Pfarrer et al., 2008). These
actions should send a signal that the firm is substantively addressing the misconduct to allay concerns that they are not purely symbolized and decoupled from actual changes (Elsbach, 1994; Mishina et al., 2010; Westphal & Zajac, 2001). Research shows that firms dedicating effort and resources to address the misconducts as soon as possible are able to revert negative assessments (Elsbach, 1994; Mishina et al., 2012). In contrast, attempts to influence evaluations with highly visible symbolic actions consistent with social expectations may backfire, as the negative behavior by the firm creates a situation where subsequent actions are met with more suspicion, distrust and scepticism (Ashforth & Gibbs, 1990; Zavyalova, Pfarrer, Reger &
Shapiro, 2012).
focused on the effects of individual tools on social evaluations (e.g., McDonnell & King, 2013; Zajac & Westphal, 1995), the use of multiple tools is rarely studied. This is surprising considering that in practice, firms almost always use more than one type of action in response
to corporate misconduct (Hersel et al., 2019). Consequently, we know very little about the
perceived efficacy of the combined use of multiple responses. Anecdotal evidence already
suggests that the combined use of different impression management tools may revert negative evaluations (Elsbach, 2003). For instance, Elsbach (2003) argues that in 1996, the National Rifle Association (NRA), while at a low-point in terms of public perceptions and support, substantially improved its perceptions by first appointing a new leader, then communicating about the NRA’s role in mainstream life and finally publicized its links to other legitimate organizations. This combined use of multiple tools seems especially salient when corporations are responding to controversial events (Arndt & Bigelow, 2000; Elsbach, 2003). Although research provides us with accounts of this phenomenon, the literature has been silent about the characteristics that these multiple tools should possess to be effective.
Grounded in existing research, we propose that firms have develop more sophisticated
strategies to influence social evaluations following misconduct (Elsbach, 2003; McDonnell &
King, 2013). They do so by combining multiple symbolic impression management tools.
Namely, following misconduct, firms first respond by adjusting their top management to signal
forthcoming change (Coombs, 2007; Pfarrer et al., 2008). Secondly, firms confirm that the
changes in top management amounted into significant improvements on an operational level
by communicating about their efforts (Aragón-Correa, Marcus & Hurtado-Torres, 2016). The
effectiveness of this process is dependent on three characteristics, namely the multiplicity of
symbols, the sequence in which they are used and their relatedness.
We test this theory in the context of environmental misconduct by publicly traded US
major social concern (Berrone & Gomez-Mejia, 2009). For instance, Volkswagen’s consumer
perception level in the U.S. took a dramatic fall in September 2015 as a result of an emission
scandal that was under extensive scrutiny by the media and governmental agencies (Marzille,
2015). Among other repercussions, this damage in social evaluation resulted in a 25% drop of
sales in November 2015 compared to the same month that previous year (Kasperkevic, 2015).
Our empirical results indicate that firms responding to environmental misconduct by first
adjusting the top management to signal environmental commitments and then exaggerate
environmental performance successfully revert negative assessments. Furthermore, the results
show that the effectiveness of these symbolic actions is dependent on whether the three
characteristics multiplicity, sequence, and relatedness are satisfied in a firm’s impression
management strategy. Finally, in line with existing research, we find that simple forms of
decoupling using only individual tools are ineffective in reverting negative assessments in the
long run.
Based on the findings of this study, we extend corporate misconduct and impression
management literature by examining how firms combine multiple impression management
tools in response to misconduct. We show that the effectiveness of these strategies in
influencing social evaluations is dependent on three characteristics. When the tools are
multiple, sequential and related, they can revert negative social evaluations even if the actions
announced are mainly symbolic. By doing so, we challenge the assumption held by both
impression management and environmental management scholars that the actions the firm
announces following misconduct should be substantive (Berrone et al., 2017; Coombs, 2007;
Pfarrer et al., 2008; Walker & Wan, 2012; Zavyalova et al., 2012). Rather, firms still
Furthermore, we also contribute to the discussion on greenwashing. Greenwashing
refers to the practice of making unsubstantiated claims about a firm’s environmental practices
or impact on the natural environment (Laufer, 2003). Here, we introduce greening the top
management as an alternative symbolic action used by firms to signal environmental
commitments. More importantly, whilst this research studied the effects of individual symbolic
actions, such as symbolically increasing environmental disclosure, and deemed them
ineffective (Berrone et al., 2017; Morales-Raya, Martín-Tapia & Ortiz-de-Mandojana, 2018;
Walker & Wan, 2012), we show that when these actions are combined with changes in the top
management of a firm, they may still be effective in reverting negative social evaluations.
THEORETICAL BACKGROUND
Social evaluations are judgements of an organization by the stakeholders and society. These social judgements form the basis of perceptions targeted at the organization (George, Dahlander, Graffin & Sim, 2016). Subsequently, organizations depend on the approval of social actors to survive and generate performance benefits (Devers et al., 2009; Pfarrer, Pollock
& Rindova, 2010; Rao, 1994; Rindova, Petkova & Sever, 2005). Various forms of social evaluation are discussed in the literature. As this study examines the effects of signals on social evaluations, we use the concept of reputation (Devers et al., 2009; Love & Kraatz, 2009; Rao, 1994). Reputation is defined as “a collective social judgment regarding the quality or
capabilities of a focal actor within a specific domain” (Boivie, Graffin & Gentry, 2016: 188).
Symbolic impression management
Research on impression management has established that in order to influence the judgements of corporate audiences, organizations send performance and quality signals that meet the expectations of stakeholders (Coombs, 2007; Devers et al., 2009; Elsbach, 1994; Marcus & Goodman, 1991). The actions the firm signals can be symbolic, substantive, or both.
aspects (Gomulya & Mishina, 2017). By anchoring itself in the strategic tradition and adopting
a managerial perspective, this research examines how organizations instrumentally manipulate
and deploy evocative symbols to attain support from corporate audiences (e.g., Ashforth &
Gibbs, 1990; Dowling & Pfeffer, 1975; Pfeffer, 1981; Pfeffer & Salancik, 1978; Suchman,
1995). This includes symbolic compliance with institutional requests (Westphal & Zajac,
1994), how accounts and justifications are used (e.g., Elsbach, 1994; Zajac & Westphal, 1995)
and the release of confounding information to obfuscate or dilute negative assessments (e.g.,
Graffin, Carpenter & Boivie, 2011; McDonnell & King, 2013).
The use of these types of impression management tactics is likely when pressure from
outside stakeholders’ conflicts with the interests of actors that hold power in the organization
(Delmas & Montes-Sancho, 2010; Meyer & Rowan, 1977; Westphal & Zajac, 2001).
Consequently, when responding to pressures for responsible behavior, corporations often use
symbolic responses in order to preserve resources and managerial flexibility (Ashforth &
Gibbs, 1990; Meyer & Rowan, 1977; Weaver, Trevino & Cochran, 1999). However, recent
research argues that decoupling using symbolic actions in response to misconduct is ineffective
and can even end up further reducing social acceptance (Berrone et al., 2017; Carlos & Lewis,
2018; Pfarrer et al., 2008; Walker & Wan, 2012; Zavyalova et al., 2012).
Following misconduct, signals regarding these controversial activities will have little credibility (Aerts & Cormier, 2009). Credibility refers to the congruence between the source’s
claims and the corresponding acts and events (Aearts & Cormier, 2009). Thus, when firms
make positive claims that are inconsistent with recent actions or behaviors, there is little
congruence between them and firms can be perceived as hypocritical (Carlos & Lewis, 2018).
This perceived hypocrisy means that the signals produced by these actors are less likely to be
other communications from the actor being met with scepticism from certain audiences
(Rosenfeld, Giacalone & Riordan, 1995).
To prevent being perceived as hypocritical, firms may signal a determination to purge
negative influences and focus on renewal by taking internal actions (Pfarrer et al., 2008). A
salient internal action that sends a strong signal of change is adjusting the management
(Elsbach, 2003; Gomulya & Mishina, 2017; Hersel et al., 2019). Adjusting the management,
however, does not necessarily imply substantive changes (Gangloff, Connelly & Shook, 2016;
Westphal & Graebner, 2010). Rather, these changes may be largely symbolic and aimed at
signalling intentions for implementing future change (Hersel et al., 2019).
Whilst this internal action may yield some positive responses from corporate audiences (Gangloff et al., 2016), when applied in isolation, it is unlikely to return an actor to social acceptability if the firm does not create the perception that the changes amounted into actual improvements (Pfarrer et al., 2008). Firms may therefore communicate about their actions to the public following the adjustments in the management to create consistency between their actions (Elsbach, 1994; McDonnell & King, 2013; Seeger & Ulmer, 2002). The combined use of these different symbolic impression management tools can revert negative evaluations (Elsbach, 2003). In this paper, we argue that the effectiveness of these strategies depends on
the multiplicity of tools, the sequence in which they are administered and the relatedness among
them. The context of this study are symbolic impression management tools used in response to
corporate environmental misconduct.
Greening the top management
Following misconduct, firms often attempt to signal forthcoming change as they cannot
implement large-scale corrective actions quickly (Hersel et al., 2019). By altering the
management, firms try to signal that they are accommodating stakeholder demands (Gangloff
as a signal of a firm’s commitment to change. While CEOs are often the target, other members
of the top management team such as C-level executives and board members can also be used
as signals (Hersel et al., 2019). For instance, Westphal and Graebner (2010) argue that changes
in board composition influences stakeholder’s perception regarding a range of corporate
policies and strategies. They found that increases in formal board independence, by increasing
the amount of (seemingly) independent directors, is associated with more positive subsequent
analyst appraisals of a firm.
As the highest entity, the top management is responsible for setting the values and
standards within the organization through their decisions regarding strategy, incentives and
internal control systems (Finkelstein, Hambrick & Cannella, 2009). Accordingly, they play an
important role in shaping perceptions about the firm (Mizruchi, 1996). Following
environmental misconduct firms may therefore use their top management to signal
environmental commitments and change. Similar to the argument of Westphal and Graebner
(2010), where firms change board composition to symbolically manage their relations with
financial stakeholders, firms may adjust the top management to include an individual with
green credentials. This signals conformity to the expectations of corporate audiences by
creating the perception that the environmental performance of the firm will increase as result
of the change in top management (De Villiers, Naiker & Van Staden, 2011). In the following
parts of the paper we refer to such an action as greening the top management.
While greening the top management signals stakeholders that, following misconduct,
the organization will change its behavior so to meet perceptions of acceptable behavior (Coombs, 2007; Hersel et al., 2019). Research has already established that firms are unlikely to revert negative assessments directly through strategic restructuring (Suchman, 1995).
Rather, adjusting corporate structures to conform to stakeholders’ expectations sets the stage
impression management literature found that organizations support their verbal accounts by
referring to organizational structures and procedures in order to influence the social evaluations
of an organization in response to a controversial event (Elsbach, 1994). Consequently, a context
of scepticism surrounding green credentials and environmental misconduct may require
additional signals showing that the changes in the management amounted into significant
improvements on an operational level (Lyon & Montgomery, 2015; Pfarrer et al., 2008). Firms
can accomplish this by communicating their actions to the public.
Exaggerating environmental performance
Communication about social efforts helps corporations build a reputation and restore it after
negative publicity (Vanhamme & Grobben, 2009). Brown and Deegan (1998) show that when
concerns about a firm’s environmental performance are raised by the media, they increase the
disclosure of environmental information. Environmental disclosure is strategically important
since investors, customers, regulators and non-governmental organizations act upon this type
of corporate communication (Lewis, Walls & Dowell, 2014). More specifically, it is proposed
that environmental disclosures can be used as a tool to change stakeholder’s perception
regarding the environmental performance of a firm (Clarkson, Richardson & Vasvari, 2008).
While firms may use environmental disclosure to signal substantive improvements on an operational level, recently, scholars argue that environmental disclosure and actual
performance are different dimensions (Aragón-Correa, Marcus & Hurtado-Torres, 2016; Cho
et al., 2012). Environmental disclosure can remove stakeholder pressure while providing little
benefit to the environment (Aragón-Correa, Marcus & Hurtado-Torres, 2016). Research even shows that in some industries, extensive environmental disclosure mediates the potential
negative effects of poorer environmental performance on environmental reputation (Cho et al.,
2012). Hence, environmental disclosure is used as a symbolic tool to influence social
The effectiveness of this impression management tool, however, depends on the
credibility of the source (Aerts & Cormier, 2009). When the firm's actions are criticized by the
media, credibility is compromised (Gomulya & Mishina, 2017). Accordingly, the signals
produced by these firms less are likely to provoke a positive response from corporate audiences
(Aerts & Cormier, 2009; Gomulya & Mishina, 2017). In a context of increased scepticism
surrounding green credentials and environmental misconduct, firms are, therefore, unlikely to
be able to revert negative social evaluations by only exaggerating environmental performance
through increasing the extent of environmental disclosure.
Sequential decoupling
Despite the fact that in isolation, greening the top management and exaggerating environmental
performance are unlikely to yield their intended outcome, when applied together they may lead
to positive reactions from corporate audience (Elsbach, 1994). Based on accounts from the
literature, we propose that firms revert negative assessments from the public by engaging in a
two-stage, or sequential, decoupling process. Namely, firms first signal a change by altering
the top management. Secondly, firms confirm the signal of change by communicating about
their efforts.
This two-stage decoupling is focused on leveraging the synergistic value between
signals produced by the symbolic actions. The effectiveness of this process is dependent on
three characteristics, namely the multiplicity of symbols, their sequence and relatedness.
Multiplicity. When organizations want to signal or change a perception, they can
develop strategies that combine multiple impression management tools or symbolic behaviors
(Elsbach, 2003). For instance, Chen and Meindl (1991) established in their investigation of the
image of the founder of People Express that his image evolved over time through the use of
various perception management tactics. These perception management tactics included press
accounts combining acknowledgments of controversies with references to institutionalized characteristics, such as new organizational structures, are more effective in managing the perceptions of organizational legitimacy than accounts with only one of these elements.
The power of this employment of multiple tools possibly lies in the credibility that is
created by the consistency of the signals these symbolic actions send (Chen & Meindl, 1991;
Crant, 1996). This means that organizations may successfully influence a perception by
engaging in multiple and different types of perception management tactics all aimed at that
single perception (Elsbach, 2003). As greening the top management and exaggerating
environmental performance both send signals aimed at positively influencing environmental
reputation, we posit that firms combine these actions in order to successfully revert negative
assessments. However, even though the signals these different symbolic actions send both
relate to environmental performance, their role in reverting negative assessments differs. Due
to the different roles they play, the sequence in which these actions are executed substantially
influences their effectiveness (Elsbach & Sutto, 1992).
Sequence. Elsbach and Sutton (1992) found that after illegitimate actions, the social
As discussed above, this paper focuses on two different types of impression management tools, each with a distinct role. Greening the top management is a visible change in the highest body of an organization. This action signals conformity to stakeholder
expectations by creating the perception that the environmental performance of the firm will
increase as result of the change (De Villiers, Naiker & Van Staden, 2011). In contrast,
exaggerating environmental performance, is a form of corporate communication aimed at changing the perception of the actual environmental performance (Aragón-Correa, Marcus & Hurtado-Torres, 2016).
When in a situation where a firm need to defend its reputation, the most effective way of responding seems to be first show conformity to societal expectations by adjusting existing structures and signalling change (Elsbach & Sutton, 1992; Hersel et al., 2019). Adjusting these
existing structures to represent socially legitimate structures can enhance the credibility and lend credence to the actor’s subsequent claims (Westphal & Graebner, 2010). Following this logic, firms may first respond to environmental misconduct by altering the top management to signal change. Next, they confirm that adjusting the top management amounted into improvements by exaggerating their environmental performance. While this research demonstrates the importance of multiple tools and their sequence (Chen & Meindl, 1991;
Elsbach, 1994; Elsbach, 2003; Elsbach & Sutton, 1992; Westphal & Graebner, 2010), a final
critical aspect is relatedness.
Relatedness. Relatedness refers to the first signal creating an expectation that is confirmed by the second one (e.g. Gilmore & Ferris, 1989). The first action, greening the top
management, aims to create a perception that the firm will adjust its environmental
performance to align with societal expectations. The second signal, exaggerating
environmental performance, confirms that the firm has, in the eyes of the stakeholder, in fact
Clarkson, Richardson & Vasvari, 2008). Together, these two related signals send a strong
message to stakeholders that, following the misconduct, the firm addressed their concerns
regarding environmental performance.
In existing research, one can find instances of symbols that complement one another
when related. For example, MacLeod and Urquiola (2015) investigate a model in which the
labour market infers the skill of an individual based on two signals, namely graduation test and
college identity, as it is argued to never be directly observed. These signals are related and
dependent on each other with regards to their effectiveness, since a high graduation score from
a college with a relatively ‘bad’ reputation will not lead to a positive inference of the skills of
an individual. Only when this high graduation score is combined with a college possessing a
‘good’ reputation is their effect present in terms of the positive inference regarding the skill of
an individual. In a similar vein, research puts forth the argument that impression management
tactics in an employment interview interact with the applicant’s credentials (Gilmore & Ferris,
1989). So that candidates with strong credentials using impression management tactics will be
perceived as better prospects when compared to those that don’t. The rationale behind this
argument is that applicants with poor qualifications for a job, employing impression
management tactics may be viewed as overdoing. Similarly, when an applicant with strong
credentials does not engage in tactical impression management behaviors they may be seen as
a poor prospect (Gilmore & Ferris, 1989). Thus, while these credentials can create the
credibility necessary for impression management tactics (Aerts & Cormier, 2009), alone they
are not enough in effectively influence perceptions. Impression management tactics specifically aimed at endorsing these credentials are necessary.
Based on this discussion, we suggest that companies may try to manage impressions
using multidimensional, sequential and interrelated symbolic tools. In this process, firms
to protecting the natural environment, giving stakeholders the perception that the firm will
become more sensitive to environmental issues and subsequently change their environmental performance. Next, exaggerating environmental performance by increasing the extent of
environmental disclosure confirms the expectation that the firm has changed as a consequence
of the ‘green’ stance. This confirmation sends a strong signal toward stakeholders that the firm
improved its corporate environmental performance to be in line with perceptions of acceptable
firm behavior (Chen & Meindl, 1991; Coombs, 2007; Pfarrer et al., 2008). Thus, the following
hypothesis is proposed:
Hypothesis 1: For firms involved in corporate environmental misconduct, the decision of first, greening the top management and, then, exaggerating environmental performance will revert negative social evaluations.
METHODS Data collection
To test the proposed hypothesis, we developed a panel dataset using multiple sources spanning from 2009 to 2018. The sample of the study is based on the list of all publicly traded US firms, 2,508 firms in total, from Thomson Reuters ASSET4 database. A single nation sample is chosen to avoid any differences in cultural or reporting requirements (Pfarrer et al., 2008).
which a news article is positive or negative. We overcome these issues by introducing a novel and more precise way for research on social evaluations to measure corporate reputation. Namely, we use the Event Sentiment Scores, which is a granular score between 0 and 100 that represents the sentiment of a news story, from RavenPack News Analytics to measure reputation. While management research has yet to adopt it, research in finance and accounting are increasingly using RavenPack (e.g., Akbas, Boehmer, Erturk & Sorescu, 2017; Warren & Sorescu, 2018). RavenPack automatically processes hundreds of thousands of articles a day from leading publishers and web aggregators covering more than 68,000 companies. The Web Edition that was used for this study monitors over 19,000 sources, including national and local news, blog sites and industry and business publishers. For each news story, RavenPack provides a score between 0-100 that indicates how strongly related the entity is to the underlying news story. In our sample we only included news stories with a “relevance score” of 100, which means that the firm played a key role in the news story and is considered highly relevant. Using this method, we collected data on more than 74 million news stories related to 2,508 firms in our sample from 2009 till 2018. In the next part, we will discuss how this data is used to measure a firm’s reputation.
Data on corporate environmental misconduct was collected by combining information from Thomson Reuters ASSET4 database, which tracks environmental controversies in the media related to resource impact, product impact and spills or pollution, with the violation tracker of the Corporate Research Project, which is the first wide-ranging database on corporate misconduct. 1 Thompson Reuters identified 204 firms in the sample that were involved in one
or more environmental controversy covered by the media. We combined the data from
1 The Corporate Research Project is a non-profit center that helps community, environmental and labour
organizations in researching companies and industries. The violation tracker covers banking, consumer protection, false claims, environmental, wage & hour, unfair labour practice, health, safety, employment discrimination, price-fixing, bribery and other cases resolved by more than 40 federal regulatory agencies and all parts of the Justice Department since 2000. More information can be found on:
ASSET4 with that of the Violation Tracker based on a point raised by McDonnel and King (2013), which is that we prevent just focusing on those accounts of misconduct reported by the news. Currently, the violation tracker of the Corporate Research Project covers 41,619 environmental violations as reported by the US federal regulatory agencies and the US Justice
Department. Using the data from the violation tracker we matched 1,709 violations to the firms in our sample. On overage, the penalty amount of these violations was 3,9 million US dollars. Combining the violations with monthly aggregated ESS data from RavenPack, giving us a monthly reputation score, showed that 917 out of the 1,709 violations caused a drop in monthly reputation as compared to the three-month prior average reputation. By merging the data from both sources, we were able to identify 379 firms that were involved in environmental misconduct from 2009 to 2018.
Data on greening the top management, the first impression management tactic, was gathered using BoardEx. To construct this data, we first collect information on the first-tier affiliations of the directors and senior managers for all 2,508 firms from 2009 till 2018, 740,269 observations in total, from BoardEx.2 Next, a list of Environmental Non-Governmental
Organizations (ENGOs) was created based on United Nations (UN) accredited ENGOs.3 We
limited the list of ENGOs to those primarily concerned with environmental protection, as the UN list also includes organizations such as the International Law Association. The final list included 303 ENGOs. Next, this list was matched with the first-tier affiliations in order to identify directors or senior managers with green credentials. Of the 2,508 firms in the sample, 686 firms greened their top management. We observe that of the 379 firms involved in environmental misconduct, 199 engaged in greening the top management. Hence, 52,5% of firms involved in environmental misconduct greened their top managements in the period of
2 BoardEx provides information on the affiliations of senior managers (board and non-board) and directors.
Hence, our sample includes both senior managers and directors.
https://www.unenvironment.org/civil-2009 till 2018 as compared to only 22,6% of the firms not involved in environmental misconducted.
Data on the exaggeration of environmental performance, the second impression management tactic, was collected using environmental disclosure information from Thomson Reuters ASSET4. The validity of ASSET4 in assessing the extent of environmental disclosure of firms has already been established by prior Corporate Social Responsibility (CSR) research (e.g., Hawn & Ioannou, 2016). The database provides CSR information collected by trained research analysts based on around 900 evaluations points per firm. The information provided is based on primary data that are objective and publicly available, sources include annual reports, sustainability reports, nongovernmental websites. We observe that 72,8% of firms involved in environmental misconduct exaggerate their environmental performance, while only 31.3% of firms not involved in misconduct exaggerate environmental performance.
Sample
Merging the data from the above-mentioned sources resulted in an initial sample of 2,508 US firms with 25,080 observations. This sample was used to create subsamples of firms engaging in different impression management tactics following environmental misconduct. We analyse the impact of impression management strategies on reputation over three years. Accordingly, our subsamples include three yearly observations per firm. In our timeframe, t0 is the year in
which the environmental misconduct occurs, t+1 the year in which the first impression
management tactics occurs and t+2 the second impression management tactic (for firms using
strategy 4 or 5, see Table 1). The three-year observation window is because of empirical considerations and limitations of the data.4
Subsamples were created by matching firms that responded to environmental misconduct using different impression management strategies with firms taking no action. Creating subsamples was required since we propose that firms, following environmental misconduct, first green the top managements and, then, exaggerate environmental performance to revert negative social evaluations. The argument here is that the effectiveness does not only rest upon the multiplicity of tactics, but also on the sequence and relatedness. In order to empirically test this, we develop the subsamples based the different strategies that are summarized in Table 1.
-- TABLE 1 --
no significant differences (p > 0.05) between the control groups and treatment groups in each of the five samples. An overview of the matched subsamples is provided in Table 2. Note here that we also created a fifth subsample in which double decoupling is compared to sequential decoupling to assess the effect of the sequence of impression management tactics.
-- TABLE 2 --
Dependent variable
Our dependent variable is firm reputation, which is measured as the tenor of media coverage about a focal firm. To capture the tenor of media coverage we use the Event Sentiment Scores (ESS) from RavenPack (Warren & Sorescu, 2018). The ESS is a score between 0 and 100 that represents the sentiment of a single news event. This score is determined by matching stories that are typically categorized by experts with extensive professional backgrounds in finance and economics as having a positive or negative impact. These experts evaluated more than 2,000 types of corporate events for their content. The ratings from this training sample are used in an algorithm that then evaluates the content of each news event and generates a score ranging from 0-100, where 50 indicates neutral sentiment, values above 50 indicate positive sentiment and values below 50 show negative sentiment. The algorithm dynamically assigns an ESS score based on score ranges assigned by the experts and by performing analysis and computation when factors such as magnitudes, comparative values or ratings are disclosed in the story5. In line with research using media coverage to assess reputation, we measure yearly firm reputation (e.g., Deephouse, 2000). Hence, for each firm we averaged the ESS of all news stories having a relevance score of 100 in a given year. On average, RavenPack analysed 3,243 news stories
5 In their manual, RavenPack states that the algorithm is capable of interpreting actual figures, estimates, ratings,
for each firm in the sample per year. The average yearly sentiment score for these firms is 53.48 with a standard deviation of 3.25.
Independent variables
Environmental misconduct. We consider environmental misconduct as actions by a firm that violates environmental laws, or be within the bounds of law, but perceived by many
in society to be unethical, such as hydraulic fracturing (Hersel et al., 2019). Environmental
misconduct is measured as a binary variable that was coded “1” if a firm was involved in one or more accounts of environmental misconduct that covered by the media or a fine was imposed
by a US federal regulatory agency in a given year, and “0” otherwise.
Greening the top management. Greening the top management is considered an action
in which either a top-level manager or director affiliated with an ENGOs is appointed or, whilst already appointed, starts to engage with an ENGO. We measure greening the top management as a binary variable that was coded “1” if a firm greened their top management in a given year, and “0” otherwise.
Exaggerating environmental performance. We measure exaggerating environmental
environmental, social and governance items, we solely focus on environmental items from ASSET4. Guided by their validated items (Cronbach’s alpha of 0.74 for external and 0.75 for internal), we coded 25 items as external actions and 20 items as internal actions. Appendix A provides an overview of the items that were coded as internal or external CSR actions. Since each item can only take a value of 0 (“no”) or 1 (“yes”) we sum the items in both categories and measure the absolute difference between internal and external actions. The absolute difference is then compared to that of the previous year in order to assess an increase (or decrease) in divergence between external CSR actions and internal CSR actions. Since we are interested in an exaggeration of environmental performance, we focus only on cases in which the sum of external actions increases more (or decreases less) than internal actions.
Control variables
Industry. Industries differ in terms of how sensitive they are to environmental issues
(Patten, 2002). For instance, in polluting industries, the impression management tactics may
be met with greater skepticism (Walker & Wan, 2012). To control for these industry effects,
we include industry dummies based on Thomson Reuters economic sector codes.
Industry reputation. Fluctuations in firm reputation may be in part due to increases or
decreases in the general reputation of an industry. For example, the misconduct of an individual firm may cast an entire industry into a bad light just as positive publicity may enhance the industry’s reputation (Winn, MacDonald & Zietsma 2008). To account for such a possibility, all models control for the average yearly reputation of the industry, which is calculated by averaging the yearly firm reputation per economic sector for the 2,508 firms. The industries are defined using the Thomson Reuters economic sectors.
Reputation volatility. Firms in the sample differed considerably in terms of how much
volatility of firm reputation over 10 year. Reputation volatility is measured as the standard deviation of the reputation over 10 years.
Media coverage. We control for the extent of media coverage per firm in a given year
by including the natural log of the total amount of news stories published each year (Zavyalova et al., 2012).
Environmental performance. The focus of this study is on symbolic actions firms use
to revert negative assessments in the media. To control for the possibility that increases in reputation are due to increases in actual environmental performance, we include the natural log of the environmental score provided by Thompson Reuters ASSET4 (Cheng, Ioannou & Serafeim, 2014). Thompson Reuters provides an environmental pillar score based on 101 datapoints. Next to the environmental disclosure items included, these datapoints also cover for instance absolute emission levels, energy usage and water usage.
Firm size. To control for firm size, we include the both the natural logs of the market
capitalization (Fiss & Zajac, 2004) and sales of a firm (Boivie et al., 2016; Hawn & Ioannou, 2016; Pfarrer et al., 2010).
Subsequent misconduct. A number of firms in the sample were involved in corporate
misconduct in consecutive years. To account for the effects of environmental misconduct happing in t+1 or t+2 we include a dummy taking a value of “1” if a firm was involved in
environmental misconduct in one of these years of observation, and “0” otherwise.
Financial performance. Research shows that financial performance has a substantial
impact on firm reputation (Roberts & Dowling, 2002). To control the effects of financial performance, we include the return on assets of a firm (e.g., Basdeo et al. 2006).
Non-board role. With greening the top management, we focus on both the top
effectiveness between the board and non-board roles, we include a dummy variable taking a
value of “0” if it is a board member and “1” if it is a non-board member.
Type of greening. Greening the top management is considered an action in which either a top-level manager or director affiliated with an ENGOs is appointed or, whilst already appointed, starts to engage with an ENGO. Although they send similar signals, they are two distinct types of actions. For instance, appointing an individual with green credentials to the top management requires considerable investments and may be perceived as more credible (Connelly, Certo, Ireland & Reutzel, 2011). To account for the difference between these two actions we include a dummy variable taking a value of “0” if an individual from the top
management starts engaging with an ENGO and “1” if an individual with ENGO affiliations is
appointed to the top management.
Analysis
than 50, we therefore analyse this sample using both MLE and REML estimators. The results remain unchanged when using REML estimation.
RESULTS
Table 3 reports the descriptive statistics and pairwise correlations for all variables. All control
variables, except the type of greening, are significantly correlated with the dependent variable.
-- TABLE 3 --
Notably, environmental misconduct is negatively correlated (p < 0.05) with reputation and
sequential decoupling is the only impression management strategy significant and positively
correlated with reputation. All the variance inflation factors (VIF) were well below five, with
an average VIF of 1.47 across all models. From this, we can conclude that multicollinearity is
not a concern (Chatterjee & Price, 1991).
The results of the mixed-effects linear regressions for subsamples 1-4 are presented in
Table 4. Models 1, 3, 5 and 7 are the baselines, containing control variables only. Models 2, 4,
6 and 8 introduce the main effects, namely time and group (control vs treatment) and the
interaction between them. The latter once’s are the full models, so serve the basis for our
-- TABLE 4 --
interpretation. Model 2 shows that greening the top management does not significantly impact
the reputation in the two years after environmental misconduct. This finding is in line with
expectations that individual symbolic impression management tactics do not successfully
revert negative assessments. Model 4 reports that exaggerating environmental performance the
year following environmental misconduct does positively and significantly (β = 0.831, p <
0.05) affect reputation. This effect, however, becomes insignificant in the second year
only short lived. Model 6 indicates that double decoupling, or just using multiple impression
management tactics, over time does not significantly impact reputation. Similar to model 4,
firms in the treatment group in model 6 engage in exaggerating environmental performance in
the year following misconduct. While we observe a similar coefficient for these models in the
year firm exaggerate their environmental performance, it is insignificant in model 6. This is
very likely due to the lower number of observations in model 6 (218 observations) as compared
to model 4 (436 observations). Finally, model 8 shows that, in line with model 2, greening the
top management in the year following misconduct does not significantly impact reputation.
However, when firms confirm this signal send by greening the top management with
exaggerating environmental performance in the second year after misconduct, they are able to
revert negative assessments as the second year shows a positive and significant effect on
reputation (β = 1.585, p < 0.05). Thus, the results from models 2, 4, 6 and 8 provide strong
support for our hypothesis arguing that the effectiveness of reverting negative assessments
from the public rests upon the multiplicity, sequence and relatedness of impression
management tactics. An illustration of the effect of sequential decoupling on firm reputation,
as compared to no action following misconduct, is provided by Figure 1. We included the year prior to the misconduct (t-1 year) in this figure to show that the reputation of firms taking no action
and firms engaging in sequential decoupling were on similar levels the year prior to the misconduct and experienced a similar drop in reputation as a result of the misconduct.
-- FIGURE 1 --
regressions. Since in this sample the J-q-1 is lower than 50 (Snijders & Bosker, 2012), this table provide the results using both MLE and REML estimators. First of all, no substantive differences exist between the models estimated using MLE (models 1 and 2) and the models estimated using REML (models 3 and 4). The full models, model 2 and 4, show no significant differences between firms that first green the top management or firms that first exaggerate their environmental performance in the year following misconduct. However, in the second year after environmental misconduct, we observe a significant difference between using a double decoupling strategy and sequential decoupling strategy (β = 1.585, p < 0.05). This result
confirms the evidence from table 4 and provides further support for our hypothesis.
-- TABLE 5 –
Among the control variables, average yearly industry reputation has a positive and
significant effect across all models (p < 0.01). Firm performance, measured as return on assets,
also shows a significant (p < 0.05) and positive effect in all models (p < 0.10 for models 7 and
8). The type of role, board or non-board role, shows a highly significant and negative effect (p
< 0.01) on reputation when firms only green the top management (model 2). However, when
firms engage in sequential decoupling (model 8) this effect is highly significant and positive (p
< 0.01). This shows that when firms green the top management using the top management team
the year following misconduct but do not engage in subsequent impression management the
next year, it is perceived as worse than when the greening involves a member of the board. On
the other hand, when firms do confirm that the changes in the top management amounted into
seeming increases in environmental performance, it may be perceived as more sincere and thus
has a stronger and positive effect on reputation. The time variable shows a negative coefficient
in all models (both control and treatment group), with a significant negative effect of time 2
that corporate misconduct damages the reputation of a firm, holds. In the next section we
provide additional support confirming the negative effect of corporate misconduct.
Additional analyses
In addition to the results in Table 3, 4 and 5, we perform supplementary analyses to address
potential concerns with the data and model specifications.
First, we started this paper off by stating that corporate misconduct damages the social
evaluations of a firm. While the previous results already show a significant drop in reputation
the years following misconduct (e.g., Figure 1), to provide further evidence that environmental
misconduct negatively impacts reputation we performed a t-test using the sample of 319 firms
with 3,190 observations. The t-test showed a significantly lower reputation (p < 0.01) for years
in which environmental misconduct was discovered by the media or a fine was imposed by a
US federal regulatory agency (mean reputation of 53.18) as opposed to years in which this was
not the case (mean reputation of 53.78). To further explore the effects of the penalty amount of
the violation on reputation, we performed a t-test using monthly reputation data (ESS scores
aggregated to a monthly average media tenor score). Here, we compared violations that caused
a drop in reputation in the month the fine was administered as opposed to the three month
average prior reputation to violations that did not cause a drop. This t-test revealed a significant
difference (p < 0.05) in the penalty amounts of violations that caused a drop in reputation (5.5 million US dollars on average) as compared to those violations that did not cause a drop (2.1 million US dollars on average).
Second, greening the top management is considered an action in which either a top-level
Third, while research on social evaluations to date focused on a limited sample of sources
from which articles are used to calculate media tenor, we use the full universe of sources from
RavenPack. A concern about the data then is how reliable a reputation score based on 19,000
different sources is. Figure 2 provides a comparison of the reputation score based ESS of
articles from a few major news outlets from 2009 till 2015. These are: The Washington Post,
Forbes, Fortune, The New York Times and The Wall Street Journal. As shown in this figure,
both reputation scores follow the same general trend. However, when limiting the articles to
those only published by major news sources, there is a substantial reduction in observations
and articles per year (on which reputation score is based). For example, using all outlets we
were able to generate a reputation score for 88% of the 25,080 observations, while with major
news sources this drops to 39%. Further examining the differences between the
reputation scores shows that the median number of articles used in the reputation score from
the five major news outlets is 10, whereas the median is 557 articles when using all news
outlets. For 25% of the reputations scores from the five major news outlets, the number of
articles used is three or less. Finally, the standard deviation or volatility of the reputation scores
also differs, while the standard deviation of the reputation score from all news outlets is 3.54,
relying on the major outlet raises the standard deviation to 8.23. Hence, the reputation score
based on the major news outlets is not only more volatile but also much less precise.
-- FIGURE 2 --
Yet, in keeping with the tradition of research on social evaluations (e.g., Chandler et
al., 2019), we ran the analyses from Table 4 restricting coverage to the major outlets specified
above. With the alternative specification of the dependent variable, we find a significant and
positive effect (β = 4.661, p < 0.05) for greening the top management the year following
year after misconduct. For exaggerating environmental performance, the effect found in Table
4 becomes insignificant. The results for double decoupling and sequential decoupling remain
fully robust. Namely, when firms confirm this signal send by greening the top management
with exaggerating environmental performance in the second year after misconduct, they are
able to revert negative assessments as the second year shows a positive and significant effect
on reputation (β = 5.175, p < 0.05).6 However, with the alternative specification we see a
substantial number of missing observations and reputation scores based on three or less news
articles.
Fourth, since we use the ESS to calculate reputation scores instead of the widely used
coefficient of imbalance (Janis & Fadner, 1965), we also run the analyses from Table 4 using
the coefficient of imbalance as an indicator of firm reputation. We use the following formula
from Deephouse (2000) to calculate media tenor:
(f2 – fu)/(total)2 if f > u
Coefficient of media tenor = 0 if f = u
(fu – u)2/(total)2 if u > f
where f = the number of favorable news stories, defined as articles with an ESS higher than 50,
in a giver year; u = the number of unfavorable news stories, defined as articles with an ESS
lower than 50, in a given year; and total = total number of news stories in a given year. For
both greening the top management and double decoupling, the main effects remain
substantively unchanged. While exaggerating environmental performance showed a similar
trend in coefficients for the main effects, the significant effect found in Table 4 for the first
year following misconduct increases from below 0.05 to a p-value of 0.071 (β = 0.026). In line
with the results from Table 4, the second year following misconduct the effect becomes
insignificant. Similar to this finding, the main effects for sequential decoupling, using the
coefficient of imbalance, show an unchanged trend in coefficients. However, as observed with
exaggerating environmental performance, the significant effect found in Table 4 increases from
below 0.05 to a p-value of 0.083 (β = 0.047). The increase in p-value can be explained by the
impact of the extent to which an article is negative or positive, or rather the absence of this
spectrum when using the coefficient of imbalance. Therefore, we conclude that these findings
can be used to argue for the robustness of the results as reported in Table 4.
DISCUSSION
Firms almost always use more than one type of action in response to corporate misconduct
(Elsbach, 2003; Hersel et al., 2019). Yet, research has rarely examined how multiple responses
work together (Hersel et al., 2019). Therefore, we know very little about the perceived efficacy
of the combined use of multiple responses. To further our understanding of this phenomenon,
we examined how the use of multiple symbolic impression management tools may revert
negative assessments from corporate audiences following misconduct.
To date, impression management research mainly focused on the effects of individual symbolic tools and deemed them ineffective (e.g. Ashforth & Gibbs, 1990; Zavyalova et al.,
2012). We show that the accumulation of different tools, if organized in a sequential, integrated
way, can revert negative assessments of corporate audiences even if they are largely symbolic.
Specifically, we found that firms revert negative social evaluations by first altering their formal
visible structures to signal forthcoming change, next they confirm that these changes amounted
into improvements on an operational level by communicating about their efforts. These
findings have important implications for both theory and practice.
Theoretical implications
Following corporate misconduct, adequate action must be taken to prevent serious harm to the
a range of tools employed by firms to favorably shape the interpretations and reactions from
corporate audiences (e.g. Elsbach, 1994; Graffin et al., 2011; McDonnell & King, 2013;
Westphal & Zajac, 1994). When these tools conform to legitimate practices, structures and
expectations, they can successfully revert negative assessments (Gomulya & Mishina, 2017).
Whilst this research highlights the importance of symbolic aspects of impression management
tools, when responding to misconduct, they argue that the tools should not be decoupled from
actual changes (Elsbach, 1994; Mishina et al., 2010; Pfarrer et al., 2008). Our results confirm this and show that when firms use individual symbolic tools in response to misconduct, they are unable to revert negative assessments in the long run (Carlos & Lewis, 2018; Pfarrer et al., 2008).
First, firms that only adjusted their formal structures following corporate misconduct experienced a similar drop in reputation as firms taking no action. An explanation here is that adjusting the top management is a signal of change, so if the firm does not follow up by showing what this change amounted into, it will not receive positive social evaluations from corporate audience with regards to the action (Pfarrer et al., 2008). This is in line with what our analysis shows. We observe that when firms do not take action following adjustments in the top management, they receive more negative social evaluations if the adjustment took place in the senior management instead of the board of directors. Senior management has a more direct impact on firms’ strategies and policies than the board of directors, thus a signal of change coming from the senior management may create stronger expectation (Baysinger & Hoskisson, 1990). Therefore, if the firm adjusts the senior management but does not show changes on an operational level it can experience a bigger backlash (Zavlyalova et al., 2012).
not help revert negative assessment from corporate audience because when the firm's actions
are criticized by the media, credibility is compromised (Gomulya & Mishina, 2017). Thus, the
claims of these firms are less likely to provoke a positive response (Aerts & Cormier, 2009).
However, this finding indicates that corporate audience may still find it difficult to separate
symbolism from substance, and that even largely symbolic communication about efforts can
remove stakeholder pressure in the short run (Aragón-Correa, Marcus & Hurtado-Torres, 2016.
Since this effect is not hold the year after, our results also show that firms using this type of
communication are at an increased risked of being exposed in the long run (Hawn & Ioannou,
2016).
Until now, we have followed existing research and focused on the effects of individual
impression management tactics. Yet, in practice, firms often combine multiple actions in
response to misconduct (Elsbach, 2003; Hersel et al., 2019). Accordingly, recent research has
called for investigations into combinations of actions by firms in response to corporate
misconduct (Hersel et al., 2019).
Our findings show that firms use more sophisticated impression management strategies
than previously investigated to symbolically address corporate misconduct. Namely, firms use
impression management tactics to first signal a change and at a later point in time confirm that
change occurred with subsequent impression management. The results show that effective
strategies (1) combine different types of impression management tools, (2) employ these tools
in a specific sequence, and (3) use tools that are related. We observe that firms which are able
to successfully revert negative assessments first signal a change by adjusting the top
management. This signals to corporate audiences that the firms are accommodating their
demands (Gangloff et al., 2016; Westphal & Graebner, 2010). Firms follow this up by claiming
that they have made improvements on an operation level, thereby confirm the first signal of
practices to align with perceptions of acceptable firm behavior (Chen & Meindl, 1991;
Coombs, 2007; Pfarrer et al., 2008).
Whilst existing research already provides anecdotal evidence of firms combining
multiple impression management tactics, we advance impression management literature by
empirically testing the characteristics these strategies possess in order to be able to revert
negative assessments. These characteristics are multiplicity, sequence and relatedness. When
impression management strategies include these characteristics, they can revert negative
assessments of corporate audiences even if they are largely symbolic. By challenging the
assumption that the actions the firm announces following misconduct should be substantive to
revert negative assessments (Coombs, 2007; Pfarrer et al., 2008; Zavyalova et al., 2012), we
argue that impression management research should extend its focus to include investigations
into the effectiveness of combined actions (Hersel et al., 2019). Since rather than being the
outcome of individual actions, reputation is built by the accumulation of corporate actions over
an extended period of time (Rindova, Williamson, Petkova & Sever, 2005; Roberts & Dowling,
2002).
Previously we argued that when firms alter their senior management without engaging
in subsequent actions, they experience a bigger backlash as opposed to altering the board of
directors. Interestingly, when firms do follow up altering the top management with
communication about their efforts this effect reverses. Firms that first alter their senior
management to signal change and then communicate about their efforts are better able to revert
negative assessments than firms that alter their board of directors as a signal a change. This
supports our argument that a signal of change coming from the senior management creates a
advance the literature on corporate reputation. Most of the research on corporate reputation has neglected the role of leaders in reputational judgements of the firm (Love, Lim & Bednar, 2017). Yet, we have long known that leaders help shape how people make sense of firms (e.g.
Meindl, Ehrlich & Dukerich, 1985). We show that in a context of increased skepticism
surrounding a firm’s behavior, the top management team does not help to build or restore the reputation directly. Rather, top management credentials can help to legitimize subsequent actions or claims that can build or restore reputation (Cohen & Dean, 2005). Further investigation is necessary to see whether firms not involved in misconduct reap the same benefits from adjusting the top management or whether it can even help them build their reputation directly.
Finally, we also contribute to the discussion on greenwashing. Greenwashing refers to
the practice of making unsubstantiated claims about a firm’s environmental practices or impact
on the natural environment (Laufer, 2003). Similar to impression management literature, this
stream of research studied the effects of individual symbolic actions, such as symbolically
increasing environmental disclosure, and deemed them ineffective (Berrone et al., 2017;
Morales-Raya, Martín-Tapia & Ortiz-de-Mandojana, 2018; Walker & Wan, 2012). The reason
for this is the increased levels of skepticism surrounding corporate green credentials and more
vigilant monitoring nowadays (Lyon & Montgomery, 2015). Yet, as they require changes at an
operational level, substantive environmental actions involve high costs (Schons & Steinmeier,
2016). So, even though it seems to become more difficult for firms to decouple, the incentives
to do so are still present. Subsequently, firms may have developed more sophisticated strategies
of decoupling (McDonnell & King, 2013). This article highlights an alternative tool in the
strategic arsenal of a firm that is used in response to environmental misconduct. Namely, firms
adjust their top management to include an individual with green credentials. This signal can
change in top management (De Villiers, Naiker & Van Staden, 2011). A recent example of this
symbolic action comes from Exxon Mobil Corp, the world’s largest publicly traded oil
producer. In January 2017, after repeated attacks in the past year from environmental groups,
they named an environmentalist, former professor and head of Hole Oceanographic Institution
Susan Avery, to its board (Reuters Staff, 2017). Our results show that 52,5% of firms involved
in environmental misconduct adjusted their top management to reflect green credentials, as opposed to only 22.6 % of firms not involved in misconduct. Further investigation into this action confirmed our argument that these highly visible structural changes are mainly symbolic (Ashforth & Gibbs, 1990; Westphal & Graebner, 2010). Only 18% of the firms that greened
their top management subsequently improved their environmental performance. More
importantly, whilst previous research found symbolic communication about environmental
efforts to be ineffective (Berrone et al., 2017; Morales-Raya, Martín-Tapia &
Ortiz-de-Mandojana, 2018; Walker & Wan, 2012), we show that when these actions are combined with
visible structural changes, such as greening the top management, they may still be effective in
garnering societal support.
Practical implications
This research also provides important insights to practice. Recently, the UN Intergovernmental
Panel on Climate Change (IPCC) released a report showing that by 2030, carbon pollution has
to be cut by 45% to prevent catastrophic events (IPCC, 2018). Businesses play a central role in
reaching this goal. In order to motivate organizations to take substantive action and
significantly reduce their carbon footprint, it is crucial to understand how they deal with
pressures for responsible behavior. We add to this discussion by investigating how and why
organizations still get away with symbolic responses, or simply put “lies”, to calls for
responsible behavior. This paper shows that while, in line with recent findings (Berrone et al.,
management do not revert negative responses by corporate audiences, more refined strategies
still do. Since it is difficult for corporate audiences to observe firm operations, this research
shows us that even more vigilant monitoring of firms’ environmental practices by third parties
may necessary in order to separate symbol from substance.
Limitations and future research
This paper is also subject to a number of limitations. First of all, we posit that the impression
management tools in this paper are symbolic. While we focus on the effects of the symbolic
aspects as much as possible, they cannot be completely separated from substance. For instance,
exaggerating environmental performance is a divergence between internal (substantive) and
external action (symbolic). In our empirical analysis we therefore included firms that increased
their symbolic actions more than they increase their substantive actions from one year to the
next. Yet, in this situation there is still an increase in substantive actions that may affect social
evaluations. Future research may therefore, instead of focusing on a divergence between
symbolic and substantive actions, direct its efforts to specific types of disclosure or
communication by firms to further isolate the effects of symbols.
Second, our conclusions are drawn based on firms involved in environmental
misconduct. Since environmental issues are currently a major social concern, the impression
management tactics related to environmental performance may have a more substantial impact
on social evaluations than others (Berrone & Gomez-Mejia, 2009). Future research should test
the effectiveness of impression management strategies presented here in different contexts and
see whether the effects still hold. Furthermore, we only focus on two types of impression
management tactics, namely structural change and subsequent communication. Other
combinations of actions should also be studied to support the robustness of the mechanism
