Signaling without
Certification: The Critical
Role of Civil Society Scrutiny
Susan A. Kayser
John W. Maxwell
Michael W. Toffel
Working Paper
15-009
July 19, 2015
Copyright © 2014, 2015 by Susan A. Kayser, John W. Maxwell, and Michael W. Toffel
Working papers are in draft form. This working paper is distributed for purposes of comment and discussion
only. It may not be reproduced without permission of the copyright holder. Copies of working papers are
available from the author.
Signaling without Certification: The Critical Role of Civil Society Scrutiny
Susan A. Kayser
Post-Doctoral Fellow
Erb Institute for Global Sustainable Enterprise & Graham Sustainability Institute
University of Michigan
440 Church Street
Ann Arbor, MI 48109
skayser@umich.edu
John W. Maxwell
W. George Pinnell Professor
Kelley School of Business
Indiana University
1309 E. 10th St.
Bloomington, IN 47405
jwmax@indiana.edu
Michael W. Toffel
Professor of Business Administration
Harvard Business School
Morgan Hall 415
Boston, MA 02163
mtoffel@hbs.edu
July 19, 2015
In response to stakeholders’ growing concerns, firms are joining self-regulatory environmental
programs to signal their superior environmental management capabilities. In contrast to the
literature’s focus on programs featuring third-party certification, we theorize that programs lacking
certification can nonetheless sometimes serve as credible signals by instead promoting civil
society scrutiny, thus mitigating adverse selection. We hypothesize that (a) institutional
environments that support civil society scrutiny and (b) organizational characteristics that increase
the impact of that scrutiny enhance the credibility of the signal by making participation more
costly for firms with inferior capabilities. Our theory is supported by examining the decisions by
2,604 firms in 44 countries whether to participate in the United Nations Global Compact.
INTRODUCTION
Increased demand for corporate environmentalism has created a need for companies to
communicate their ability to manage environmental risks and impacts to a broad range of
stakeholders. A recent global survey of over 4,000 large companies indicated that 71 percent
issued a corporate responsibility report in 2013 (KPMG, 2013). But the survey authors
concluded that these reports “are often not that easy to read” and advised companies “to
communicate the information in a more digestible and engaging way” (KPMG, 2013: 9). One
1
way companies can bridge this information asymmetry is to use a signal: an action that
succinctly conveys information that may be too complex to convey directly (Spence, 1973).
Studies based on signaling theory have demonstrated that several self-regulatory
programs that require third-party certification are credible indicators of a firm’s environmental
capabilities (King, Lenox, and Terlaak, 2005; Montiel, Husted, and Christmann, 2012; Potoski
and Prakash, 2005; Rivera, 2002). Such programs serve as credible signals when meeting their
standards is sufficiently more costly for companies with inferior management capabilities such
that they forego participating (Darnall and Carmin, 2005; Darnall and Edwards, 2006; Terlaak,
2007). Programs featuring third-party certification impose two types of costs on participants: (1)
the managerial effort, technology, and equipment required meet the program’s standards, and (2)
a fee paid to a third-party auditor to attest that the standards have been met. The audit fee, which
can be tens of thousands of dollars, is unrelated to a firm’s environmental capabilities, degrades
the program’s signaling value if it deters participation by firms that could readily meet the
requirements but choose to avoid incurring the fee.1
In this paper, we examine a different type of self-regulatory program that encourages
scrutiny from civil society by requiring participants to publicly commit to the program’s goals
and principles. The ability of these programs to serve as signals does not rely on participants
paying for third-party certification, but on monitoring and sanctions provided by civil society.
We propose that civil society scrutiny holds participants accountable to the program’s goals by
exposing them to the risk of reputation damage if they fall short of their commitments. The
higher cost that lower-capability firms face to meet such commitments increases the risk that
they will fall short, and the resulting heightened risk of reputational damage deters their
1
This is akin to Spence’s (1973) observation that tuition costs might deter individuals who are highly productive but
poor from pursuing a college degree.
2
participation. Because meeting such commitments is more costly (in expectation) for lowercapability firms, they are less and therefore deters their participation. Herein lies the signaling
credibility of such programs.
We hypothesize that two institutional factors and one organizational attribute will
contribute to this effect. First, we hypothesize that institutional environments more likely to
impose civil society scrutiny will be more effective in deterring firms with inferior capabilities
from participating in self-regulatory programs. Second, we hypothesize that institutional
environments with greater expectations for corporate socially responsible behavior—and
therefore the threat of worse consequences for failing to meet those expectations—will also be
especially effective in deterring companies with inferior capabilities from participating. Third,
we hypothesize that smaller companies with inferior capabilities will be more deterred from
participating than larger companies with inferior capabilities. This results from the heightened
scrutiny that participation elicits having a greater impact on smaller firms, larger firms being
already more visible and exposed to scrutiny.
We test our theory by empirically analyzing companies’ decisions whether or not to join
the United Nations (UN) Global Compact, a self-regulatory program that espouses a set of
environmental and social principles that participants pledge to uphold, but that eschews thirdparty certification that could attest that they actually do. The Global Compact imposes no direct
costs of joining. Instead, its secretariat publicizes the names of participating companies and their
annual progress reports, which invites civil society scrutiny (UN Global Compact, 2015a).
To study whether Global Compact participation can serve as a credible signal of a
company’s capability to manage environmental risks and impacts, we examine the conditions
under which companies with higher-quality environmental disclosures—that is, disclosures that
3
reveal more quantitative information on the company’s most meaningful environmental risks and
impacts—are more likely to participate. Quantitative disclosures are a compelling measure of
environmental capabilities because prior research has shown that companies that issue highquality environmental disclosures have superior environmental performance (Al-Tuwaijri,
Christensen, and Hughes, 2004; Clarkson et al., 2008) and are able to better manage
environmental risks (Blacconiere and Northcut, 1997; Blacconiere and Patten, 1994).2
Using a sample of 2,604 companies (with headquarters in 44 countries), we find evidence
to support our three hypotheses. We find that companies with lower-quality disclosures are
particularly deterred from participating in the Global Compact in institutional settings in which
(a) the risk of civil society scrutiny is greater, and (b) the expectations for corporate socially
responsible behavior are higher. Finally, we find that smaller companies with lower-quality
disclosures are more deterred than larger companies with comparably low-quality disclosures.
After presenting our theory, analysis, and results, we conclude by discussing our study’s
implications for scholarship on signaling theory and self-regulation.
THEORY AND HYPOTHESES
Firms know more about their capabilities than do their external stakeholders. This information
asymmetry creates a transaction cost of identifying companies with desirable characteristics
(Akerlof, 1970; Williamson, 1985). Thus, it is in the best interest of firms with characteristics
that are desirable to their stakeholders to reduce information asymmetries by sending a signal—
an action that succinctly conveys information that may be too complex to convey directly
2
Some stakeholder groups, such as nongovernmental organizations (NGOs), may be able to assess a company’s
capabilities directly from its disclosures and will not need to rely on self-regulatory programs as signals. Other
stakeholders, such as potential joint venture partners, employees, and consumers, will often find it too difficult or
time-consuming to obtain or digest complex environmental disclosures. We argue that participation in a selfregulatory program can be a simple way to convey such information to these stakeholders.
4
(Spence, 1973). A signal will be credible if firms with inferior management capabilities find the
costs of sending the signal to exceed the benefits, whereas firms with superior management
capabilities find the benefits to exceed the costs.
Several studies have examined how self-regulatory programs that require third-party
certification can credibly signal superior capabilities to manage quality, health and safety, or
environmental issues (Anderson, Daly, and Johnson, 1999; King, Lenox, and Terlaak, 2005;
Levine and Toffel, 2010; Potoski and Prakash, 2005). These papers theorize that firms with
inferior management capabilities will forgo participation because the costs of implementing the
policies and procedures required to pass the certification audit exceed the anticipated benefits. In
contrast, firms with superior management capabilities will participate because the anticipated
benefits of meeting the program’s requirements exceed the costs (Terlaak, 2007). Some have
argued that self-regulatory programs that lack third-party auditing requirements cannot be
credible signals as they allow for free-riding because it would be no costlier for firms with
inferior capabilities to join than for firms with superior capabilities (Darnall and Carmin, 2005;
Lenox and Nash, 2003). Others have argued that it is too costly for external stakeholders to
scrutinize the participants to prevent such free-riding (Montiel, Husted, and Christmann, 2012).
We note, however, that some self-regulatory programs promote scrutiny from civil
society actors by requiring participants to publicly commit to the program’s principles. Programs
such as the chemical industry’s Responsible Care, the ski industry’s Sustainable Slopes, the
United Nations Global Compact (the focus of this study), and the Environmental Protection
Agency’s (EPA) WasteWise programs do not require third-party certification, but do require
participants to commit to their principles and publish participants’ names on their websites.
However, previous studies have not found evidence that participation in such programs can serve
5
as a credible signal (Delmas and Montes-Sancho, 2010; King and Lenox, 2000; Rivera and de
Leon, 2004; Rivera, de Leon, and Koerber, 2006). In contrast to this prior work, we explore
whether or how a company’s institutional or organizational context might influence the
perceived likelihood or consequences of civil society scrutiny in such a way that participation
might indeed serve as a credible signal.
We propose that scrutiny from civil society actors can, under certain conditions, impose a
higher cost of participation in self-regulatory programs featuring scrutiny from civil society for
firms with inferior capabilities, so that participation does serve as a credible signal. Our work
thus addresses Connelly et al.'s (2011: 61) call for research on how penalties inflicted by third
parties for sending “false signals” might act as a substitute for more formal signaling costs, such
as those imposed by third-party certification auditors.
By identifying how organizational and institutional factors interact to influence
participation in a self-regulatory program, our work builds on prior research that has examined
other organizational and institutional factors that influence participation. For example, others
have found that participation is more prevalent among firms that are larger (e.g., Arora and
Cason, 1996), more export-oriented (Christmann and Taylor, 2001, 2006; King, Lenox, and
Terlaak, 2005), and that spend more on R&D (Terlaak and King, 2006). Research has found that
institutional features such as regulatory scrutiny (Rivera, 2002) and corruption also influence
firms’ propensity to participate in self-regulatory programs (Montiel, Husted, and Christmann,
2012). We theorize that institutional and organizational factors can impose differentially greater
costs on companies with inferior capabilities such that participation in a self-regulatory program
featuring civil society scrutiny—but not third-party certification—can serve as a credible signal.
6
The critical role of civil society scrutiny
For firms, as for individuals, the decision to signal information to uninformed parties hinges on
the perceived costs and benefits of that action. The benefits of signaling superior environmental
capabilities are the additional profits the company expects once its superior capabilities are
known in the marketplace. These profits may come from new business generated from buyers
with environmental preferences (Arora and Gangopadhyay, 1995), reduced environmental
oversight from regulators in the form of shorter permitting delays (Decker, 2003) and fewer
inspections (Toffel and Short, 2011), or from the company’s ability to attract more highly
motivated employees (Brekke and Nyborg, 2008).
Some have argued that self-regulatory programs that lack third-party certification
requirements risk allowing firms to receive such benefits without being held accountable for
their commitments (Darnall and Carmin, 2005; Delmas and Montes-Sancho, 2010; King and
Lenox, 2000). However, firms that participate in a self-regulatory program that instead features
scrutiny from civil society risk being criticized if such scrutiny reveals their failure to uphold
their commitments. For example, Greenpeace criticized Dow Chemical, a Responsible Care
participant, for lobbying to raise the legal allowance of a pollutant, dioxin, instead of reducing its
use. Greenpeace noted that this lobbying effort was in direct contrast to its alleged Responsible
Care commitment “to lead in the development of responsible laws, regulations and standards that
safeguard the community, workplace and environment.”3 In line with this logic, researchers have
argued that the threat of such criticism from civil society might actually cause firms to “clam up”
(Lyon and Maxwell, 2011: 21) and dissuade them from publicly committing to making
environmental improvements (Bansal and Clelland, 2004; Bowen, 2014).
3
“Dodgy deals and irresponsible care,” March 16, 2004, Greenpeace website,
http://www.greenpeace.org/international/en/news/features/dodgy-deals-and-irresponsible/, accessed June 2015.
7
Studies examining the impacts of social movements have shown that such negative press
resulting from civil society scrutiny can impose costly reputational damage (e.g., Bartley and
Child, 2011; King and Soule, 2007). However, even when no lasting reputational damage occurs,
such exposure is likely to negate the benefits of participation because it can make the firm’s
participation seem merely symbolic. For example, the Business and Human Rights Centre, which
maintains a database of articles regarding corporate human rights abuses, also lists whether the
company in question is a participant in the UN Global Compact. It is less likely that participants
exposed as symbolic participants will benefit from their membership in the program, especially
given the public’s growing skepticism of symbolic environmental commitments, sometimes
referred to as greenwash (Lyon and Montgomery, 2015).
Because it is more costly for firms with inferior capabilities than for firms with superior
capabilities to uphold the program’s public commitments, we posit that when the probability of
being caught by civil society is sufficient, participation in a self-regulatory program featuring
scrutiny can serve as a signal. Clearly, however, the credibility of the signal will depend on the
extent to which a participant’s institutional context supports civil society oversight.4 As a result,
we predict that:
H1: Firms with inferior capabilities to manage their environmental risks and impacts will
be less likely to participate in a self-regulatory program featuring scrutiny from civil
society, especially in institutional contexts that provide greater support for such scrutiny.
A firm’s decision to join a self-regulatory program featuring scrutiny from civil society
will depend not only on the likelihood that scrutiny will result in sanctions, but also on the
4
When such oversight is effective, participation in such self-regulatory programs serves as a purer signal than thirdparty certification programs. This is because third-party certification also comes with certification fees that are
independent of managerial capabilities and therefore may deter firms with superior managerial capabilities for
financial reasons. Self-regulatory programs featuring scrutiny from civil society lack these certification fees and
allow sorting to be based solely on managerial capabilities.
8
impact of those sanctions. This impact will vary with the extent to which stakeholders expect
firms to behave in a socially responsible manner, which varies considerably across countries
(e.g., Chapple and Moon, 2005; Doh and Guay, 2006; Matten and Moon, 2008). Firms
headquartered in institutional contexts that have greater expectations for socially responsible
behavior are held to higher standards (Rowley and Moldoveanu, 2003) and face greater
punishments if they fall short (Campbell, 2007; Peloza and Papania, 2008). Thus, firms in such
contexts that fail to uphold their commitment to a self-regulatory program’s objectives are at
greater risk of provoking negative attention. We propose that firms with inferior capabilities to
manage their environmental risks and impacts will be especially deterred from participating in a
program featuring scrutiny from civil society when they face these heightened expectations. We
therefore predict that participation in such a program will serve as a more credible signal
amongst firms in this institutional context.
H2: Firms with inferior capabilities to manage their environmental risks and impacts will
be less likely to participate in a self-regulatory program featuring scrutiny from civil
society, especially in institutional contexts that have greater expectations for socially
responsible behavior.
In addition to a firm’s institutional context, organizational characteristics can also affect
its perceived costs of participating in a self-regulatory program featuring civil society scrutiny.
We propose that because the public is less likely to have strong opinions about smaller firms
(Fombrun, Gardberg, and Barnett, 2000), any new information regarding corporate behavior will
likely have a greater reputational impact for a smaller firm than it would have for a larger firm.5
The potential cost is less likely to be meaningful for larger firms since they are already more
visible to their stakeholders and tend to attract more scrutiny than smaller firms (Greening and
5
The literature on Bayesian updating indicates that as prior beliefs are less certain, new pieces of information will
be more influential in updating expectations (Ericson, 1970).
9
Gray, 1994; Salancik and Pfeffer, 1978; Scott, 1992). Among smaller firms, those with inferior
capabilities, or those that would find upholding the program’s commitments more difficult, will
be less keen to invite the spotlight of stakeholder scrutiny. We therefore predict that smaller
firms with inferior capabilities will be especially deterred from participating.
H3: Firms with inferior capabilities to manage their environmental risks and impacts will
be less likely to participate in a self-regulatory program featuring scrutiny from civil
society when they are smaller.
DATA AND METHODOLOGY
Empirical context
To test our hypotheses, we examined participation in the United Nations Global Compact, a selfregulatory program that has a set of environmental and social standards that participants publicly
pledge to uphold, but that lacks an auditing requirement to ensure that they do. The Global
Compact requires two things of a participant: it must issue a “commitment letter” publicly
declaring its support for the 10 Global Compact principles on environmental and social issues6
and it must commit to submitting an annual public report on its progress in implementing the
principles, with the first report due one year after the company joins the Global Compact. Unlike
programs that require third-party certification, there are no mandatory fees or certification costs
required to participate in the Global Compact. Instead, it publicizes a member’s participation and
progress reports, which encourages monitoring from external stakeholders such as NGOs.
6
The 10 principles are: “Human Rights: (1) Businesses should support and respect the protection of internationally
proclaimed human rights; and (2) make sure that they are not complicit in human rights abuses. Labour: (3)
Businesses should uphold the freedom of association and the effective recognition of the right to collective
bargaining; (4) the elimination of all forms of forced and compulsory labour; (5) the effective abolition of child
labour; and (6) the elimination of discrimination in respect of employment and occupation. Environment: (7)
Businesses should support a precautionary approach to environmental challenges; (8) undertake initiatives to
promote greater environmental responsibility; and (9) encourage the development and diffusion of environmentally
friendly technologies. Anti-corruption: (10) Businesses should work against corruption in all its forms, including
extortion and bribery” (UN Global Compact, 2015a).
10
These minimal requirements have raised concerns that the Global Compact has allowed
some firms to participate that are not truly committed to its principles and do not intend to
implement them (Bigge, 2004; Deva, 2006; Lim and Tsutsui, 2012; Nolan, 2005; Waddock and
McIntosh, 2011). However, NGOs do monitor Global Compact participants and speak up when
they perceive firms to not be adhering to the principles. For example, SOMO7—a self-described
corporate watchdog—maintains the “Global Compact Critics” blog, which publicly shames
participants that fall short of their commitments.8 Moreover, the “Global Compact Compliance
Watch” blog highlights negative news regarding participants’ corporate social performance. For
example, one blog post reports that Antamina (a copper-zinc mine owned by four mining
companies that are all Global Compact participants) was fined by the Peruvian government for a
toxic spill.9
The Global Compact also allows NGOs to participate, both to work with the business
participants and to hold them accountable to their commitments by voicing concerns when they
find that participants are not abiding by the principles (Rasche, 2009: 520; UN Global Compact,
2015b). For example, the Business and Human Rights Resource Centre, which joined the Global
Compact in 2004, lists whether the targets of its reports on human rights or labor abuses are also
Global Compact participants.10
7
SOMO stands for Stichting Onderzoek Multinationale Ondernemingen [Center for Research on Multinational
Corporations].
8
The blog can be found at http://globalcompactcritics.blogspot.com/. Accessed February 2015.
9
“Peru fines Antamina’s owners US$77,000 for toxic spill,” Global Compact Compliance Watch website, July 6,
2013, http://globalcompactcompliance.blogspot.com/2013/07/peru-fines-antaminas-owners-us77000-for.html,
accessed May 2015.
10
For example, see the Business and Human Rights Resource Centre’s page on Barrick Gold at http://businesshumanrights.org/en/barrick-gold. Accessed May 2015.
11
Sample
Our sampling frame was determined by the coverage of Trucost Plc, a company that produces
and sells corporate environmental profiles to socially responsible investors and is the source of
several of our key variables. Trucost collected data on all 4,819 public companies that were
listed on any of the following major stock indices during 2004 through 2008: ASX 200, Dow
Jones STOXX Europe 600, FTSE All Share, MSCI Asia ex Japan, MSCI World, Nikkei 225,
Russell 1000, and S&P 500. We omit from our sample companies in industries for which direct
environmental impacts are typically inconsequential, as others have done (e.g., Cho and Patten,
2007; Clarkson et al., 2008). In particular, we omit the 1,901 companies in service-related
industries and focus on the remaining 2,918 companies that are in manufacturing, resources, and
other industries listed in Table 1.11
Our outcome of interest is participation in the UN Global Compact. By 2008, 2,811
companies from 107 countries and a wide array of industries had joined. Of the 730 publicly
traded companies that had joined by 2008, 264 were in industrial-related industries and in the
Trucost sample. This includes 145 that joined during our sample period (2004–2008) and 119
companies that had joined before our sample period and that we therefore omitted.
Linking the remaining 2,799 companies to our other data sources—the United Nations,
the World Bank, the World Economic Forum, the World Values Survey, Worldscope, and the
Yearbook of International Organizations—resulted in a loss of 195 more companies. In sum,
restricting our sample to the firms within the Trucost universe that were in industrial-related
industries, had not joined the Global Compact before 2004, and were linked to our other data
sources yielded an estimation sample of 2,604 companies headquartered in 44 countries.
11
We include the following Industry Classification Benchmark (ICB) supersectors: automobiles and parts, basic
resources, chemicals, construction and materials, food and beverages, industrial goods and services, oil and gas,
personal and household goods, technology, telecommunications, and utilities.
12
Tables 1 and 2 report the industry and headquarter country distributions of the estimation
sample.
[Insert Tables 1 and 2 here]
Variables
We measure Global Compact participation as an annual binary variable coded 1 the year a
company initially participated in the Global Compact and 0 otherwise.12 For nonparticipants, it is
always coded 0. We obtained a comprehensive list of all companies that participated in the
Global Compact, including the date they began participating, from the Global Compact
secretariat.
To proxy for a company’s capabilities to manage its environmental risks and impacts, we
rely on the quality of its environmental disclosures. Research has shown that companies that
issue higher-quality environmental disclosures are better able to manage environmental risks
(Blacconiere and Northcut, 1997; Blacconiere and Patten, 1994) and have lower environmental
impacts (Al-Tuwaijri, Christensen, and Hughes, 2004; Clarkson et al., 2008). Our use of this
proxy is consistent with the accounting literature’s discretionary disclosure theory, which posits
that companies with superior environmental capabilities will disclose objective information
regarding their environmental impacts to convey those capabilities to stakeholders (Dye, 1985;
Verrecchia, 1983). Because it is difficult for companies with inferior capabilities to reveal
equally positive information, stakeholders can use disclosures as an indicator of a company’s
environmental capabilities.
We measure quality of environmental disclosure by relying on Trucost’s weighted
disclosure ratio (Trucost Plc, 2008). This ratio is the proportion of a company’s relevant
12
As described below, our analysis omits Global Compact participants’ observations in the years after it joined.
13
environmental indicators for which it disclosed worldwide quantitative figures, each indicator
being weighted according to the cost of the environmental damage associated with it. To
construct this ratio, Trucost first identified the subset of 464 industries from which each
company derived revenues each year, based on the FactSet Fundamentals database, financial
disclosures, and company feedback. From a comprehensive list of environmental indicators
(such as sulfur dioxide emissions and hydrofluorocarbons), Trucost identified the subset that it
deemed relevant to each of these industries, based on lifecycle assessment and economic inputoutput tables.
Second, Trucost reviewed each company’s annual reports, sustainability reports, and
websites to determine for which of the relevant indicators the company publicly disclosed
quantitative figures. To appropriately weight these disclosures, Trucost multiplied each indicator
by a damage cost factor; for example, $31 per ton of greenhouse gas emitted. Trucost obtains
these damage cost factors from the environmental economics literature. The sum of these
products is the numerator of the weighted disclosure ratio.
To construct the denominator, Trucost estimated the environmental damage that would
have been reported had the company made a complete disclosure; the denominator of the
weighted disclosure ratio is therefore also referred to as the company’s environmental damage.
To do so, Trucost estimated the relevant indicators which the company did not disclose, based on
economic input-output data and lifecycle assessment data from various sources. It multiplied
each of those indicators by its damage cost factor and added the sum of those products to the
numerator. The weighted disclosure ratio captures the extent to which a company disclosed its
more environmentally damaging or less environmentally damaging indicators.13
13
For example, if a company disclosed quantitative data for 10 of its 20 relevant indicators but those 10 indicators
collectively represented a mere 10% of the company’s overall environmental damage, then its quality of
14
We measured civil society scrutiny as the number of global NGOs in a company’s
headquarters country that were participants in the Global Compact in a given year, or NGO
participants. According to the Global Compact, global NGOs are nonprofit entities that pursue
positive social and environmental changes both within and beyond their country’s borders. The
role of global NGO participants is to hold business participants accountable to their
commitments to implement the Global Compact’s principles (UN Global Compact, 2015b). Of
all civil society organizations, these NGOs are most likely to understand the Global Compact’s
principles and objectives and to scrutinize its participants.14 We obtained data from the Global
Compact’s website on the number of participating NGOs.15 Examples include Amnesty
International, CERES, Conservation International, the Indian chapter of Leadership for
Environment and Development (LEAD), and the Vietnamese chapter of ActionAid.
We measured stakeholder expectations of responsible behavior, or ethical context, based
on data from the World Economic Forum’s annual Executive Opinion Survey, which is a part of
the World Competitiveness Report. Data from these surveys have been used by many other
scholars examining international datasets (e.g., Aggarwal and Goodell, 2014; Johnson,
Kaufmann, and Zoido-Lobatón, 1998; van Stel, Carree, and Thurik, 2005; Yuan, Low, and Tang,
environmental disclosure would be 0.1. If those 10 disclosed indicators instead collectively represented 90% of the
company’s overall environmental damage, then its quality of environmental disclosure would be a much more
impressive 0.9.
14
We focus on NGOs in each company’s headquarters country because is it easier for NGOs to research companies
that operate (and release information) in the same language and because NGOs can more easily educate the local
community and engage with the local regulators (Billings, 1971; Sine and Lee, 2009). The Appendix describes a
robustness test that uses Internet usage as an alternative measure of civil society scrutiny, based on the premise that
greater Internet penetration increases the threat of scrutiny from NGOs based domestically or elsewhere. Our
regression estimates based on this alternative measure yield the same inferences that our primary model do.
15
As a robustness test, we also run our regressions with the number of NGO participants in the Global Compact as a
percentage of the total number of NGOs in that country. The total number of NGOs is gathered from the Yearbook
of International Organizations. The results in Model 2 remain largely similar in terms of effect sizes and
significance levels. In Model 3, however, the interaction between the percentage of NGOs in the Global Compact
and the quality of environmental disclosure becomes significant only at the 10% level (p=0.074).
15
2010).16 Business leaders were asked, “In your country, how would you rate the corporate ethics
of companies (ethical behavior in interactions with public officials, politicians, and other
firms)?” Responses ranged from 1 (“extremely poor—among the worst in the world”) to 7
(“excellent—among the best in the world”). This question was not asked in the 2003–2005
surveys, which prevents us from using annual figures. Instead, we use average country values
based on surveys conducted in 2006–2008.17 We assign these country average values to
companies based on their headquarters location. A company’s key stakeholders—including
employees, regulators, and investors—tend to be concentrated in its headquarters country, which
makes this institutional environment especially influential (Guler, Guillén, and Macpherson
2002).18
We measure company size using its annual sales, a common approach to capture a
company’s visibility to its stakeholders (e.g., Cho and Patten, 2007; Elsayed and Hoque, 2010;
Hackston and Milne, 1996; Patten, 2002). We measure sales in millions of U.S. dollars based on
data from Worldscope and we standardize sales by country to account for the fact that the mean
and variation in company sales differ substantially by country.19 To adjust for outliers within
each country, we winsorize at the 99th percentile.
16
These surveys are available at http://www.weforum.org/issues/global-competitiveness (accessed February 2015).
As a robustness test, we construct an alternative measure that uses 2006 values for 2003–2006, 2007 values for
2007, and 2008 values for 2008. The results remain similar in terms of effect sizes and significance levels.
18
As a robustness test, we instead measure ethical context as the Corruption Perception Index (CPI) of a company’s
headquarters country, obtained from Transparency International (available at
http://www.transparency.org/research/cpi/overview; accessed May 2015). When CPI is reversed so that larger scores
represent more corruption, this measure has a correlation of 0.95 with the ethical context measure. When using the
reversed CPI score in our models, our results remain largely unchanged in terms of effect sizes and significance
levels, except the coefficient on the interaction between CPI and quality of environmental disclosure becomes only
weakly significant (p=0.072).
19
For example, a $500-million-a-year company in Malaysia is large relative to its peers and will attract more
scrutiny from its stakeholders than a company that size in the United States would do. To standardize by country, we
take the mean for each country in the sample and mean-center each observation by country. Then we divide each
observation by the country’s standard deviation. The observations within each country then have a mean of zero and
a standard deviation of one.
17
16
We measure the extent to which a company’s activities result in environmental damage
using the previously described metric provided by Trucost, which is reported in millions of U.S.
dollars. As with sales, we standardize environmental damage by country to account for the fact
that the mean and variation of environmental damage differs substantially by country. We also
winsorize at the 99th percentile to account for outliers within each country.
We created two variables based on data from the Business and Human Rights Research
Centre’s (BHRRC) library of global news articles about companies and human rights. Each
article in this database, which is freely available on the organization’s website (http://businesshumanrights.org/), contains a summary of its content as well as BHRRC’s tags, which include
the companies named and the specific human rights issues discussed. We measure the extent to
which a company’s activities result in any human rights concerns based on whether the company
had been tagged in any article in the BHRRC database in a given year that contained an “abuse”
tag such as beatings and violence, denial of freedom of association, and sexual harassment.20 We
likewise coded any labor rights concerns based on whether the BHRRC database contained an
article in a given year that tagged a company’s name and either a “labor condition” issue such as
child labor or forced labor or a “discrimination” issue such as pregnancy or racial/ethical/caste
origin.21 Because only four percent of the companies in our sample were tagged in more than one
such article, we measure any human rights concerns and any labor rights concerns as dummy
variables at the company-year level rather than as counts, in order to avoid undue influence of
outliers, which is akin to top-coding at the 96th percentile.
20
BHRRC tags for articles related to labor abuse referred to abduction, arbitrary detention, beatings and violence,
complicity, deaths, death penalty, death threats, denial of freedom of association, denial of freedom of expression,
denial of freedom of movement, disappearances, displacement, genocide, injuries, intimidation and threats, killings,
rape and sexual abuse, sexual harassment, slavery, torture and ill-treatment, and unfair trial.
21
BHRRC tags for articles related to labor conditions were child labor, export processing zones (which are
sometimes exempt from legal labor protections), forced labor, labor (general), living wage, and prison labor.
BHRRC discrimination tags referred to discrimination based on age, disability, diversity, gender, HIV/AIDS,
marital status, political opinion, pregnancy, racial/ethnic/caste origin, religious, and sexual orientation.
17
We calculated peer participation for each company as the number of companies sharing
its headquarters country and industry classification (that is, Industry Classification Benchmark
[ICB] supersector) that were already participating in the Global Compact the prior year. We also
include several country-level variables. We measure the annual total number of NGOs in a
company’s headquarters country based on data from the annual Yearbook of International
Organizations.22 We measured the extent to which the population of each company’s
headquarters country generally supports the United Nations and its mission by obtaining data
from the following World Values Survey question: “Could you tell me how much confidence
you have in [the United Nations]: is it a great deal of confidence, quite a lot of confidence, not
very much confidence, or none at all?”23 UN support is the percentage of World Values Survey
respondents from each country that responded “a great deal” or “quite a lot.” Because it takes the
World Values Survey several years to gather data from over 100 countries, the surveys are
compiled into “waves” that use the same set of survey questions. We use data from the 2005–
2008 survey wave, which provides us with a single average value per country.24
We measure ISO 14001 popularity as the number of ISO 14001 certifications received
annually in each company’s headquarters country per million population. This is gathered from
the ISO Surveys for 2005 and 2007, which provide historic annual data on the number of ISO
14001 certifications per country for the years 2004 to 2007.25 Finally, we measure the wealth of a
22
The Yearbook of International Organizations is available at http://www.uia.org/yearbook (accessed August 2013).
We used the 2003 through 2007 editions because we lag all independent and control variables one year.
23
Responses range from 1 (“a great deal of confidence”) to 4 (“none at all”). The World Values Survey is available
at http://www.worldvaluessurvey.org/WVSDocumentationWV5.jsp (accessed May 2014).
24
Because the World Values Survey does not collect data from four countries in our sample (Hong Kong, Israel,
Singapore, and Sri Lanka), we recode these missing values of UN support to 0 and include a corresponding dummy
variable coded 1 to denote these recoded observations (Greene, 2007: 62; Maddala, 1977: 202). Our results are
nearly identical when, as a robustness test, we estimate our models on the slightly smaller sample that excludes the
330 company-year observations from the 102 companies in these four countries.
25
The ISO Surveys can be found at http://www.iso.org/iso/survey2005.pdf and
http://www.iso.org/iso/survey2007.pdf. Accessed May 2015.
18
company’s headquarters country, or GDP per capita, as its gross domestic product (GDP) per
thousand people based on annual data from the World Bank.26
Summary statistics and correlations are reported in Table 3.
[Insert Table 3 about here]
EMPIRICAL MODEL AND RESULTS
Specification
We estimate the following model:
, , ,
F
where
, , ,
X1 ,
,
,
X2 ,
,
,
X3 ,
,
,
X4
,
,
X5 , ,
,
,, ,
),
refers to whether company i in industry j headquartered in country c participated in
the UN Global Compact in year t (Global Compact participation). Because our model predicts a
company’s decision to initially participate in the Global Compact, we omit from the sample
observations in the years after a participant joined. The function F(·) refers to the logistic
function. The term X1 ,
,
refers to our key explanatory variables measured at the firm or
country level, lagged one year: quality of environmental disclosure, NGO participants, ethical
context, and sales.
The term X2 ,
,
refers to several annual firm-level controls that might also affect a
company’s decision to participate in the Global Compact. Because performance in a selfregulatory program’s domain has been found to be an important predictor of participation in that
program (Arora and Cason, 1995; King and Lenox, 2000; Short and Toffel, 2008), we control for
environmental damage, any human rights concerns, and any labor rights concerns to reflect
performance dimensions associated with the Global Compact’s social and environmental
26
Data on GDP per capita can be found at http://data.worldbank.org/indicator/NY.GDP.PCAP.CD. Accessed May
2015.
19
principles. The term X3 ,
,
refers to our controlling for peer participation, calculated based on
the focal firm’s industry and country (and observation year), to account for the possibility that
peer pressure or mimetic institutional forces might influence a company’s decision to participate
in the Global Compact (Bennie, Bernhagen, and Mitchell, 2007; Perez-Batres, Miller, and Pisani,
2011).
X4
,
refers to our annual country-level controls. Like Berliner and Prakash (2012), we
control for a country’s total number of NGOs to account for general scrutiny from NGOs
including the vast majority that do not participate in the Global Compact. We control for ISO
14001 popularity because a company headquartered in a country where there are greater rates of
environmental certification might face similar pressures to participate in the Global Compact
(Berliner and Prakash, 2012). This term also includes GDP per capita because a company whose
headquarters country is wealthier might benefit less from participating in the Global Compact.
The term X5 refers to our time-invariant country-level measure of UN support. We control for
this because the value of Global Compact participation might be greater in countries that are
more supportive of the United Nations (Berliner and Prakash, 2012), where the program might
be more recognizable and where stakeholders might view membership more favorably (Bennie,
Bernhagen, and Mitchell, 2007).
The term
represents industry dummies (ICB supersector) to control for time-invariant
industry differences in the propensity to participate in the Global Compact. Year dummies ( )
control for overall temporal trends that might affect participation, such as the Global Compact
becoming more recognizable over time. Finally, the term
20
,, ,
represents the error term.
Results
Like other studies that have tested signaling models (e.g., King and Lenox, 2000; Lenox and
Nash, 2003; Montiel, Husted, and Christmann, 2012), we use logistic regression to predict the
likelihood that a company will decide to take the signaling action of participating in the Global
Compact. We lag all independent variables, moderators, and control variables by one year to
avoid reverse-causality concerns.27 To ease interpretation, all interacted variables are
standardized and sales and environmental damage are standardized by country, as noted above.
Because our explanatory variables are measured at the company or country level, we
report standard errors clustered by country (Petersen, 2009). Our diagnostic check for
multicollinearity yielded no cause for concern, as the maximum variance inflation factor was 5.1,
well below the conventional threshold of 10 (Belsley, Kuh, and Welsch, 1980).
Regression results are reported in Table 4, both as coefficients and as average marginal
effects (AME). Model 1 includes only control variables. Our results indicate that overall,
companies with higher-quality disclosures are significantly more likely to participate in the
Global Compact (Model 1: β = 0.35; p < 0.01).28
[Insert Table 4 and Figure 1 about here]
Model 2 adds the interaction between quality of environmental disclosure and NGO
participants. The positive significant coefficient on this interaction term (β = 0.11; p < 0.05)
indicates that companies with lower-quality disclosures are especially unlikely to participate
when headquartered in countries with more civil society engagement in the Global Compact.
27
The one exception is UN support, for which we have a single average value per country during our sample period.
To interpret the magnitude of this effect, we note that the baseline probability of Global Compact participation is
1.74% (calculated as the 145 observation-years for which a company joined the Global Compact, divided by the
8,338 total observation-years). The average marginal effect of 0.58% indicates that a one-standard-deviation
increase in quality of environmental disclosure corresponds to a 25% increase in the probability of joining the
Global Compact from the 1.74% baseline to 2.32%.
28
21
This supports H1. Figure 1 depicts the average predicted probability of participating in the
Global Compact along varying levels of quality of environmental disclosure and as if the
companies were headquartered in countries with low or high levels of civil society scrutiny—that
is, at the 5th or 95th percentile of the sample distribution of NGO participants—with all other
variables retaining their original values. The increasing slope of both lines indicates a greater
propensity to participate in the Global Compact among companies with higher-quality
environmental disclosures. The dashed line’s steeper slope indicates that this positive
relationship is more intensive for companies headquartered in countries with more civil society
scrutiny. 29 Our model predicts that when companies face low civil society scrutiny, the
probability of participating in the Global Compact approximately doubles when comparing
companies with no environmental disclosures to those with very high-quality disclosure. That is,
estimates based on setting NGO participants at the 5th percentile of the sample distribution
indicate that the predicted probability of participation increases from 1.9 to 4.2 percent (a factor
of 2.2) as quality of environmental disclosure increases from the 5th to the 95th percentile of our
sample distribution. In contrast, the corresponding increase when companies face high civil
society scrutiny is nearly fourfold. That is, estimates based on setting NGO participants at the
95th percentile indicate that the predicted probability of participation increases from 0.8 to 3.0
percent (a factor of 3.8). The greater increase in the predicted probability of participation when
calculated as if the companies were headquartered in countries with strong civil society scrutiny
suggests that participation provides a more credible signal in such countries.
29
The dashed line refers to estimates as if the companies were all headquartered in countries whose civil society
scrutiny was at the 95th percentile of our sample; that is, countries in which seven global NGOs participated in the
Global Compact, such as the United Kingdom and United States. The solid line refers to average predicted values
estimated as if the companies were all headquartered in countries whose civil society scrutiny was at the 5th
percentile of our sample distribution; that is, countries in which no global NGOs participated in the Global Compact,
such as in Japan and Australia.
22
Model 3 includes the interaction between quality of environmental disclosure and ethical
context. The significant positive coefficient on this interaction term (β = 0.10; p < 0.05) indicates
that companies with lower-quality disclosures are especially unlikely to participate when
headquartered in countries with greater expectations for corporate ethical behavior. This supports
H2. Figure 2 depicts the average predicted probability of participating in the Global Compact
along varying levels of quality of environmental disclosure and as if the companies were
headquartered in countries with weak and strong ethical contexts; that is, at the 5th or 95th
percentile of the sample distribution of ethical context.30 The nearly flat solid line in Figure 2
indicates that for companies headquartered in countries with very weak ethical contexts, the data
reveal little relationship between quality of environmental disclosure and the average predicted
probability of participating. Our model predicts that when companies are headquartered in
countries with weak ethical contexts, the probability of a company with very high-quality
disclosure participating in the Global Compact is only 1.15 times the probability for a company
making no environmental disclosures. That is, estimates based on setting ethical context at the
5th percentile of the sample distribution indicate that the predicted probability of participation
increases from 3.3 to 3.8 percent (a factor of 1.15) as quality of environmental disclosure
increases from the 5th to the 95th percentile. In contrast, the corresponding increase when
companies face high civil society scrutiny is fourfold. That is, estimates based on setting ethical
context at the 95th percentile of the sample distribution indicate that the predicted probability of
participation increases from 0.8 to 3.2 percent (a factor of 4.0) as quality of environmental
disclosure increases from the 5th to the 95th percentile. The greater increase in the predicted
30
The dashed line refers to estimates as if the companies were all headquartered in a country whose ethical context
was at the 95th percentile of our sample, such as Denmark, Finland, or New Zealand. The solid line refers to average
predicted values estimated as if the companies were all headquartered in a country whose ethical context was at the
fifth percentile of our sample distribution, such as in Argentina, Indonesia, the Philippines, or Russia.
23
probability of participation when calculated as if the companies were headquartered in countries
with a strong ethical context suggests that participation provides a more credible signal in these
countries.
[Insert Figures 2 and 3 about here]
Model 4 includes the interaction between quality of environmental disclosure and sales.
The negative significant coefficient on this interaction term (β = -0.09; p = 0. 08) indicates that
companies with low quality of environmental disclosure are especially deterred from
participation in the Global Compact when they are less visible, which weakly supports H3. This
relationship is depicted in Figure 3. The increasing slope of both lines indicates a greater
propensity to participate in the Global Compact among companies with higher-quality
environmental disclosures. The solid line’s steeper slope indicates that this positive relationship
is more intensive for smaller companies—those at the 5th percentile of our sample distribution.
Our model predicts that when companies are larger, the probability of a company with very highquality disclosure participating in the Global Compact is twice that of a company with no
environmental disclosures. That is, estimates based on setting sales at the 95th percentile indicate
that the predicted probability of participation increases from 2.4 to 4.8 percent (a factor of 2.0) as
quality of environmental disclosure increases from the 5th to the 95th percentile. In contrast, the
corresponding increase for smaller companies more than triples. That is, estimates based on
setting sales at the 5th percentile of the sample distribution indicate that the predicted probability
of participation increases from 0.9 to 3.1 percent (a factor of 3.4). The greater increase in the
predicted probability of participation when calculated as if the companies are smaller suggests
that participation among these companies provides a more credible signal.
24
DISCUSSION
Our analysis reveals that self-regulatory programs lacking certification can sometimes serve as
credible signals by promoting civil society scrutiny of the participants. Below, we discuss how
our study contributes to the literature, provide implications for managers, acknowledge
limitations of our study, and offer suggestions for future research.
Contributions
The central contribution of our work is to the literature that examines whether participation in a
self-regulatory program can serve as a credible signal of a company’s superior capabilities
related to the program’s principles. Prior studies have concluded that a self-regulatory program
that lack third-party certification cannot prevent adverse selection and therefore cannot serve as a
credible signal (Darnall and Carmin, 2005; King and Lenox, 2000; Lenox and Nash, 2003;
Rivera and de Leon, 2004). We argue, however, that self-regulatory programs that instead
feature scrutiny from civil society—by requiring participants to publicly commit to the
program’s principles—can, under certain conditions, mitigate adverse selection. By exploring
institutional and organizational factors that support and enhance the impact of scrutiny, ours is
the first study to find that such a program can sometimes serve as a credible signal.
Our research also contributes to the management literature that has applied signaling
theory to a variety of other contexts. Several researchers have noted that the literature using
signaling theory is mainly focused on ex ante signaling costs, such as the cost of achieving thirdparty certification, but largely ignores potential ex post penalty costs such as being vulnerable to
lawsuits or reputation damage (Connelly et al., 2011: 61; Payne et al., 2013: 232). Ours is among
the first studies to examine whether the risk of a penalty, in the form of criticism from civil
society, can effectively substitute for ex ante signaling costs. In what may be the closest study to
25
ours, Lin, Prabhala, and Viswanathan (2013) examine people seeking peer-to-peer loans and find
that the number of online “friendships” a borrower possesses serves as a credible signal of her
likelihood of defaulting on the loan. They theorize that more “friends” increases the risk of social
stigma should the borrower default. In other words, this potential penalty cost prevents adverse
selection. Our research goes beyond this analysis by applying the concept of ex post penalty
costs to a business setting and by identifying several conditions under which the penalty risk is
especially likely to make a signal credible.
Implications for managers
Self-regulatory programs that require third-party certification have been proposed as a
supplemental governance mechanism for companies operating in institutional contexts with lax
regulatory enforcement and little social pressure for corporate social responsibility (e.g.,
Christmann and Taylor, 2006). Our study provides evidence that participation in a self-regulatory
program that lacks such certification requirements, but instead features scrutiny from civil
society, can provide a credible signal of superior capabilities when there is institutional support
for such scrutiny. Our results therefore suggest that managers can screen suppliers in countries
that provide institutional support for scrutiny based in part on which suppliers choose to
participate in self-regulation programs featuring civil society scrutiny.
Furthermore, participation in self-regulatory programs that feature civil society scrutiny
might provide a better signal of superior capabilities than participation in programs requiring
third-party certification does. In his seminal “Job Market Signaling” paper, Spence (1973) notes
that the cost of an education might prevent highly capable people from earning a degree.
Similarly, third-party certification fees can be substantial (e.g., Darnall and Edwards, 2006) and
might prevent highly capable companies from pursuing certification. Self-regulatory programs
26
that instead feature civil society scrutiny lack such certification costs that are unrelated to the
quality of the company’s capabilities and thus do not similarly dissuade highly capable
companies from participating.
Limitations and future work
Our work has several limitations. Because we lack a precise measure of a company’s capabilities
to manage environmental risks and impact, we instead used the quality of the company’s
environmental disclosures. While the number of environmental accidents might have provided a
better proxy, we lack such a measure for our global sample, which spans 44 countries. The
accounting literature has compellingly demonstrated that companies with better disclosures are
also more capable of managing environmental risks (Blacconiere and Northcut, 1997;
Blacconiere and Patten, 1994) and impacts (Al-Tuwaijri, Christensen, and Hughes, 2004;
Clarkson et al., 2008). Of course, this prompts a second concern that rather than using
participation in a self-regulatory program as a signal of capabilities to manage environmental
risks and impacts, stakeholders could instead simply assess each company’s environmental
disclosures. We argue that because environmental disclosures can be difficult to comprehend,
companies often need a simpler means—such as participation in a self-regulatory program—to
convey such information.
We theorize that the threat of scrutiny from civil society can, under certain conditions, be
a sufficient cost to companies with inferior capabilities to keep them from participating in
programs featuring such scrutiny. However, while the requirement to make a public commitment
to program principles is a feature of many self-regulatory programs featuring such scrutiny—
such as the EPA’s 33/50 and WasteWise programs, the chemical industry’s Responsible Care
program, and the alpine ski industry’s Sustainable Slopes program—these programs also differ
27
in ways that might call into question the generalizability of our results to all of them. For
example, some programs, including the Global Compact, seek participants from all industries,
but others, such as Responsible Care and Sustainable Slopes, focus on one industry. Programs
also differ based on whether they are operated by a multigovernment agency, a regulator, an
industry association, or a multi-stakeholder consortium. These factors might influence the extent
to which civil society scrutinizes the program’s participants. Furthermore, the Global Compact is
a particularly large and well-known self-regulatory program launched and operated by an
unusually prominent organization; it has also attracted considerable criticism from researchers,
the media, and activists (e.g., Bigge, 2004; Deva, 2006; Nolan, 2005; Waddock and McIntosh,
2011). Future research is needed to confirm how different programs might differ in the degree to
which participation serve as a credible signal.
Another question of generalizability arises from our data being constrained to relatively
large, publicly listed companies, whereas many self-regulatory programs also attract relatively
small and privately held companies. In terms of firm size, we find that smaller companies with
inferior capabilities to manage environmental risks and impacts are especially deterred from
participating in a self-regulatory program that features scrutiny from civil society, which we
attribute to the greater risk that criticism would damage their reputations. This suggests that our
results might also apply to even smaller companies that were not included in our analysis. Very
small companies, however, might attract no additional scrutiny at all upon joining a selfregulatory program, which would imply that the mechanisms driving our results might not apply
to them. It also remains unclear whether our results apply to private companies, which are not
subject to important sources of institutional pressure such as investors and socially-responsibleinvestment rating agencies. Upon joining self-regulatory programs featuring scrutiny from civil
28
society, private companies might not experience a sufficient increase in scrutiny to prevent
adverse selection. Future work is therefore needed on the extent to which our hypothesized
relationships apply to smaller companies and privately held companies.
Finally, Etzion and Pe’er (2014) argue that the credibility of a signaling mechanism can
change over time because the conditions that influence its costs and benefits can change. It is
therefore plausible that the signaling credibility of the Global Compact might change over time.
Because our data on the quality of companies’ environmental disclosures is limited to the period
of 2003 to 2007, we cannot determine whether our results would hold for participants that joined
either before 2003 (the Global Compact was initiated in 2000) or after 2007. Participation in the
Global Compact has increased rapidly and it has become the largest self-regulatory program,
with over 10,000 companies participating (Rasche, Waddock, and McIntosh, 2013). This rapid
growth might be overwhelming civil society’s ability to scrutinize participants, which could
gradually reduce the program’s signaling credibility.
Conclusion
Our study aimed to reveal the circumstances under which self-regulatory programs that lack
third-party certification requirements, but instead require participants to publicly commit to the
program’s principles, can still serve as a credible signal. We extend theory of signaling
mechanisms in the management literature by demonstrating how the threat of a penalty, in the
form of criticism from civil society, can prevent companies with inferior capabilities to manage
environmental risks and impacts from participating in a self-regulatory program featuring such
scrutiny.
29
REFERENCES
Aggarwal R, Goodell JW. 2014. Cross-national differences in access to finance: influence of
culture and institutional environments. Research in International Business and Finance 31:
193–211.
Akerlof GA. 1970. The market for "lemons": quality uncertainty and the market mechanism.
Quarterly Journal of Economics 84(3): 488–500.
Al-Tuwaijri SA, Christensen TE, Hughes KE. 2004. The relations among environmental
disclosure, environmental performance, and economic performance: a simultaneous
equations approach. Accounting, Organizations and Society 29(5): 447–471.
Anderson SW, Daly JD, Johnson MF. 1999. Why firms seek ISO 9000 certification: regulatory
compliance or competitive advantage? Production and Operations Management 8(1): 28–
43.
Arora S, Cason TN. 1995. An experiment in voluntary environmental regulation: Participation in
EPA’s 33/50 program. Journal of Environmental Economics and Management 28(3): 271–
286.
Arora S, Cason TN. 1996. Why do firms volunteer to exceed environmental regulations?
Understanding participation in EPA’s 33/50 program. Land Economics 72(4): 413–432.
Arora S, Gangopadhyay S. 1995. Toward a theoretical model of voluntary overcompliance.
Journal of Economic Behavior & Organization 28(3): 289–309.
Bansal P, Clelland I. 2004. Talking trash: legitimacy, impression management, and unsystematic
risk in the context of the natural environment. Academy of Management Journal 47(1): 93–
103.
Bartley T, Child C. 2011. Movements, markets and fields: the effects of anti-sweatshop
campaigns on US firms, 1993–2000. Social Forces 90(2): 425–451.
Belsley DA, Kuh E, Welsch RE. 1980. Regression Diagnostics: Identifying Influential
Observations and Sources of Collinearity. Wiley: New York.
Bennie L, Bernhagen P, Mitchell NJ. 2007. The logic of transnational action: the good
corporation and the Global Compact. Political Studies 55(4): 733–753.
Berliner D, Prakash A. 2012. From norms to programs: the United Nations Global Compact and
global governance. Regulation and Governance 6(2): 149–166.
BHRRC. 2015. About us: Business and Human Rights Research Centre, http://businesshumanrights.org/en/about-us [May 2015].
Billings L. 1971. Sierra Club 1970: New chapters, groups, growth and change. Sierra 56: 20–22.
Bigge DM. 2004. Bring on the bluewash: a social constructivist argument against using Nike v.
Kasky to attack the UN Global Compact. International Legal Perspectives 14(6): 6–21.
Blacconiere WG, Northcut WD. 1997. Environmental information and market reactions to
environmental legislation. Journal of Accounting, Auditing & Finance 12(2): 149–178.
Blacconiere WG, Patten DM. 1994. Environmental disclosures, regulatory costs, and changes in
firm value. Journal of Accounting and Economics 18(3): 357–377.
Bowen F. 2014. After Greenwashing: Symbolic Corporate Environmentalism and Society.
Cambridge University Press: Cambridge, U.K.
Brekke KA, Nyborg K. 2008. Attracting responsible employees: green production as labor
market screening. Resource and Energy Economics 30(4): 509–526.
30
Campbell JL. 2007. Why would corporations behave in socially responsible ways? An
institutional theory of corporate social responsibility. Academy of Management Review
32(3): 946–967.
Chapple W, Moon J. 2005. Corporate social responsibility (CSR) in Asia: a seven-country study
of CSR web site reporting. Business & Society 44(4): 415–441.
Cho CH, Patten DM. 2007. The role of environmental disclosures as tools of legitimacy: a
research note. Accounting, Organizations and Society 32(7): 639–647.
Christmann P, Taylor G. 2001. Globalization and the environment: Determinants of firm selfregulation in China. Journal of International Business Studies (3): 439–458.
Christmann P, Taylor G. 2006. Firm self-regulation through international certifiable standards:
determinants of symbolic versus substantive implementation. Journal of International
Business Studies 37(6): 863–878.
Clarkson PM, Li Y, Richardson GD, Vasvari, FP. 2008. Revisiting the relation between
environmental performance and environmental disclosure: an empirical analysis.
Accounting, Organizations and Society 33(4): 303–327.
Connelly BL, Certo ST, Ireland RD, Reutzel CR. 2011. Signaling theory: a review and
assessment. Journal of Management 37(1): 39–67.
Darnall N, Carmin J. 2005. Greener and cleaner? The signaling accuracy of US voluntary
environmental programs. Policy Sciences 38(2–3): 71–90.
Darnall N, Edwards D. 2006. Predicting the cost of environmental management system adoption:
the role of capabilities, resources and ownership structure. Strategic Management Journal
27(4): 301–320.
Decker CS. 2003. Corporate environmentalism and environmental statutory permitting. Journal
of Law and Economics 46(1): 103–129.
Delmas MA, Montes‐Sancho MJ. 2010. Voluntary agreements to improve environmental
quality: symbolic and substantive cooperation. Strategic Management Journal 31(6): 575–
601.
Deva S. 2006. Global Compact: a critique of the UN’s public-private partnership for promoting
corporate citizenship. Syracuse Journal of International Law and Commerce 34(1): 107–
152.
Doh JP, Guay TR. 2006. Corporate social responsibility, public policy, and NGO activism in
Europe and the United States: an institutional‐stakeholder perspective. Journal of
Management Studies 43(1): 47–73.
Dye RA. 1985. Disclosure of non-proprietary information. Journal of Accounting Research
23(1): 123–145.
Elsayed MO, Hoque Z. 2010. Perceived international environmental factors and corporate
voluntary disclosure practices: an empirical study. British Accounting Review 42(1): 17–35.
Ericson WA. 1970. On the posterior mean and variance of a population mean. Journal of the
American Statistical Association 65(330): 649–652.
Etzion D, Pe'er A. 2014. Mixed signals: a dynamic analysis of warranty provision in the
automotive industry, 1960–2008. Strategic Management Journal 35(11): 1605–1625.
Fombrun CJ, Gardberg GA, Barnett ML. 2000. Opportunity platforms and safety nets: corporate
citizenship and reputational risk. Business and Society Review 105(1): 85–106.
Greene WH. 2007. Econometric Analysis (6th ed.). Pearson/Prentice Hall: Upper Saddle River,
NJ.
31
Greening DW, Gray B. 1994. Testing a model of organizational response to social and political
issues. Academy of Management Journal 37(3): 467–498.
Guler I, Guillén MF, Macpherson JM. 2002. Global competition, institutions, and the diffusion
of organizational practices: the international spread of ISO 9000 Quality Certificates.
Administrative Science Quarterly 47(2): 207–232.
Hackston D, Milne MJ. 1996. Some determinants of social and environmental disclosures in
New Zealand companies. Accounting, Auditing and Accountability Journal 9(1): 77–108.
Johnson SD, Kaufmann D, Zoido-Lobatón P. 1998. Regulatory discretion and the unofficial
economy. American Economic Review 88(2): 387–392.
King AA, Lenox MJ. 2000. Industry self-regulation without sanctions: the chemical industry’s
Responsible Care program. Academy of Management Journal 43(4): 698–716.
King AA, Lenox MJ, Terlaak A. 2005. The strategic use of decentralized institutions: exploring
certification with the ISO 14001 management standard. Academy of Management Journal
48(6): 1091–1106.
King, BG, Soule SA. 2007. Social movements as extra-institutional entrepreneurs: the effect of
protests on stock price returns. Administrative Science Quarterly 52(3): 413–442.
KPMG. 2013. The KPMG Survey of Corporate Responsibility Reporting 2013,
https://www.kpmg.com/Global/en/IssuesAndInsights/ArticlesPublications/corporateresponsibility/Documents/corporate-responsibility-reporting-survey-2013-execsummary.pdf [June 2015].
Lenox MJ, Nash J. 2003. Industry self‐regulation and adverse selection: a comparison across
four trade association programs. Business Strategy and the Environment 12(6): 343–356.
Levine DI, Toffel MW. 2010. Quality management and job quality: how the ISO 9001 standard
for quality management systems affects employees and employers. Management Science
56(6): 978–996.
Lim A, Tsutsui K. 2012. Globalization and commitment in corporate social responsibility: crossnational analyses of institutional and political-economy effects. American Sociological
Review 77(1): 69–98.
Lin, M, Prabhala NR, Viswanathan S. 2013. Judging borrowers by the company they keep:
friendship networks and information asymmetry in online peer-to-peer lending.
Management Science 59(1): 17–35.
Lyon TP, Maxwell JW. 2011. Greenwash: corporate environmental disclosure under threat of
audit. Journal of Economics and Management Strategy 20(1): 3–41.
Lyon TP, Montgomery AW. 2015. The means and end of greenwash. Organization and
Environment 28(2): 223–249.
Maddala GS. 1977. Econometrics. McGraw-Hill: New York.
Matten D, Moon J. 2008. “Implicit” and “explicit” CSR: a conceptual framework for a
comparative understanding of corporate social responsibility. Academy of management
Review 33(2): 404–424.
Montiel I, Husted BW, Christmann P. 2012. Using private management standard certification to
reduce information asymmetries in corrupt environments. Strategic Management Journal
33(9): 1103–1113.
Nolan J. 2005. The United Nations’ compact with business: hindering or helping the protection
of human rights? University of Queensland Law Journal 24(2): 445–466.
Patten, DM. 2002. The relation between environmental performance and environmental
disclosure: a research note. Accounting, Organizations and Society 27(8): 763–773.
32
Payne GT, Moore CB, Bell RG, Zachary MA. 2013. Signaling organizational virtue: an
examination of virtue rhetoric, country‐level corruption, and performance of foreign IPOs
from emerging and developed economies. Strategic Entrepreneurship Journal 7(3): 230–
251.
Peloza J, Papania L. 2008. The missing link between corporate social responsibility and financial
performance: stakeholder salience and identification. Corporate Reputation Review 11(2):
169–181.
Perez-Batres LA, Miller VV, Pisani MJ. 2011. Institutionalizing sustainability: an empirical
study of corporate registration and commitment to the United Nations Global Compact
guidelines. Journal of Cleaner Production 19(8): 843–851.
Petersen MA. 2009. Estimating standard errors in finance panel data sets: comparing approaches.
Review of Financial Studies 22(1): 435–480.
Potoski M, Prakash A. 2005. Green clubs and voluntary governance: ISO 14001 and firms’
regulatory compliance. American Journal of Political Science 49(2): 235–248.
Rasche A. 2009. “A necessary supplement:” what the United Nations Global Compact is and is
not. Business & Society 48(4): 511–537.
Rasche AS, Waddock S, McIntosh M. 2013. The United Nations Global Compact: retrospect and
prospect. Business and Society 52(1): 6–30.
Rivera J. 2002. Assessing a voluntary environmental initiative in the developing world: the Costa
Rican Certification for Sustainable Tourism. Policy Science 35(4): 333–360.
Rivera J, de Leon P. 2004. Is greener whiter? Voluntary environmental performance of Western
ski areas. Policy Studies Journal 32(3): 417–437.
Rivera J, De Leon P, Koerber C. 2006. Is greener whiter yet? The sustainable slopes program
after five years. Policy Studies Journal 34(2): 195–221.
Rowley T, Moldoveanu M. 2003. When will stakeholder groups act? An interest- and identitybased model of stakeholder group mobilization. Academy of Management Review 28(2):
204–219.
Salancik GR, Pfeffer J. 1978. A social information processing approach to job attitudes and task
design. Administrative Science Quarterly 23(2): 224–253.
Scott WR. 1992. Organizations: Rational, Natural, and Open Systems. Prentice-Hall:
Englewood Cliffs, NJ.
Short JL, Toffel MW. 2008. Coerced confessions: self-policing in the shadow of the regulator.
Journal of Law, Economics & Organization 24(1): 45–71.
Sine WD, Lee BH. 2009. Tilting at windmills? The environmental movement and the emergence
of the US wind energy sector. Administrative Science Quarterly 54(1): 123-155.
Spence AM. 1973. Job market signaling. Quarterly Journal of Economics 87(3): 355–379.
Terlaak A. 2007. Order without law: the role of certified management standards in shaping
socially desired firm behaviors. Academy of Management Review 32(3): 968–985.
Terlaak A, King A. 2006. The effect of certification with the ISO 9000 Quality Management
Standard: a signaling approach. Journal of Economic Behavior and Organization 60(4):
579–602.
Toffel MW, Short JL. 2011. Coming clean and cleaning up: does voluntary self-reporting
indicate effective self-policing? Journal of Law and Economics 54(3): 609–649.
Trucost Plc. 2008. Trucost Methodology Overview: Measuring Company Environmental
Impacts. Trucost Plc: London.
33
UN Global Compact. 2015a. About us: the ten principles,
http://www.unglobalcompact.org/AboutTheGC/TheTenPrinciples/index.html [March 2015].
UN Global Compact. 2015b. How to participate: civil society,
http://www.unglobalcompact.org/HowToParticipate/civil_society/index.html [March 2015].
van Stel A, Carree M, Thurik R. 2005. The effect of entrepreneurial activity on national
economic growth. Small Business Economics 24(3): 311–321.
Verrecchia RE. 1983. Discretionary disclosure. Journal of Accounting & Economics 5(1): 179–
194.
Waddock S, McIntosh N. 2011. SEE Change: Making the Transition to a Sustainable Enterprise
Economy. Greenleaf: Sheffield, U.K.
Williamson OE. 1985. The Economic Institutions of Capitalism: Firms, Marketing, Relational
Contracting. Free Press: New York.
Yuan XM, Low JM, Tang LC. 2010. Roles of the airport and logistics services on the economic
outcomes of an air cargo supply chain. International Journal of Production Economics
127(2): 215–225.
34
Table 1. Industry composition of sample
ICB supersector
Automobiles & parts
Basic resources
Chemicals
Construction & materials
Food & beverage
Industrial goods & services
Oil & gas
Personal & household goods
Technology
Telecommunications
Utilities
Total
Firms Percent
68
2.6
211
8.1
138
5.3
164
6.3
164
6.3
715
27.5
285
10.9
220
8.4
342
13.1
112
4.3
185
7.1
2,604
Table 2. Headquarters composition of sample
Country
Australia
Austria
Belgium
Brazil
Canada
Chile
China
Czech Republic
Denmark
Egypt
Finland
France
Germany
Greece
Hong Kong
Hungary
India
Indonesia
Ireland
Israel
Italy
Japan
Firms Percent
121
4.6
13
0.5
14
0.5
31
1.2
98
3.8
5
0.2
61
2.3
2
0.1
18
0.7
2
0.1
34
1.3
37
1.4
58
2.2
13
0.5
61
2.3
2
0.1
50
1.9
19
0.7
15
0.6
18
0.7
28
1.1
324
12.4
Country
Luxembourg
Malaysia
Mexico
Netherlands
New Zealand
Norway
Pakistan
Philippines
Poland
Portugal
Russia
Singapore
South Africa
South Korea
Spain
Sri Lanka
Sweden
Switzerland
Thailand
Turkey
United Kingdom
United States
Total
35
Firms Percent
7
0.3
38
1.5
15
0.6
42
1.6
7
0.3
52
2.0
11
0.4
10
0.4
4
0.2
6
0.2
17
0.7
21
0.8
19
0.7
80
3.1
15
0.6
3
0.1
46
1.8
29
1.1
23
0.9
5
0.2
384
14.7
746
28.6
2,604
Table 3. Summary statistics
Variable
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
Global Compact participation
Environmental damage
Any human rights concerns
Any labor concerns
Peer participants per m pop
Quality of environmental disclosure
Sales
ISO14001 certifications per m pop
GDP per thousand pop
NGO participants
Total number of NGOs
UN support
Stakeholder expectations
Mean
SD
0.02
0.13
193
699
0.06
0.24
0.09
0.29
0.04
0.09
0.17
0.32
6,337 15,495
75.31 70.29
34.52 12.30
3.56
3.76
6,283 1,858
0.46
0.16
5.46
0.58
Min.
Max.
0
1
0.00 11,035
0
1
0
1
0
1.66
0
1.00
0.01 358,600
0.16
417
0.60
83
0
9
1,676
8,813
0
0.85
3.36
6.54
Correlations
[1] [2] [3]
1.00
0.06 1.00
0.03 0.12 1.00
0.04 0.14 0.61
0.06 -0.05 -0.01
0.08 0.34 0.12
0.06 0.28 0.36
0.03 -0.08 -0.04
-0.04 -0.06 -0.01
-0.05 0.01 0.03
-0.01 -0.04 0.00
0.04 -0.02 -0.01
-0.03 -0.10 -0.04
[4]
[5]
[6]
[7]
[8]
[9]
[10] [11] [12]
1.00
-0.03
0.13
0.34
-0.07
0.03
0.10
0.02
-0.05
-0.05
1.00
-0.02
-0.03
0.21
0.06
-0.02
0.09
0.03
0.20
1.00
0.20
0.10
0.02
-0.07
0.02
0.09
0.04
1.00
-0.01
0.03
0.03
-0.01
0.01
-0.06
1.00
0.18
-0.41
-0.08
0.52
0.41
1.00
0.47
0.61
-0.07
0.69
1.00
0.73 1.00
-0.49 -0.15 1.00
0.10 0.40 -0.04
This table reports the actual means and standard deviations, but in most cases standardized variables are used for the analysis, as
indicated in Table 4.
36
Table 4. Logistic regression results
Dependent variable: Global Compact participation
(1)
Coef.
H1
H2
H3
Quality of environmental disclosure† ×
NGO participants †
Quality of environmental disclosure† ×
Stakeholder expectations †
Quality of environmental disclosure† ×
Sales ◊
Quality of environmental disclosure †
AME
(2)
Coef.
0.113*
[0.054]
(3)
Coef.
(4)
Coef.
0.103*
[0.052]
Environmental damage ◊
Any human rights concerns
Any labor concerns
Peer participants per million population
Sales ◊
ISO 14001 certifications per million population
GDP per thousand population
NGO participants †
Total number of NGOs †
UN support
Stakeholder expectations †
Industry dummies
Year dummies
Observations (company-years)
Companies
Participant companies
Countries
0.349**
[0.089]
0.207*
[0.094]
-0.422
[0.343]
0.527**
[0.169]
2.505**
[0.669]
0.339**
[0.079]
0.118
[0.144]
-0.030*
[0.014]
-0.426*
[0.202]
0.563**
[0.122]
0.704
[1.345]
-0.224*
[0.095]
Yes
Yes
8,338
2,604
145
44
0.58%
0.34%
-0.70%
0.87%
4.16%
0.56%
0.20%
-0.05%
-0.71%
0.93%
1.17%
-0.37%
0.395**
[0.063]
0.205*
[0.093]
-0.427
[0.336]
0.504**
[0.171]
2.472**
[0.665]
0.338**
[0.076]
0.114
[0.142]
-0.028*
[0.014]
-0.502*
[0.221]
0.562**
[0.120]
0.680
[1.323]
-0.227*
[0.095]
Yes
Yes
8,338
2,604
145
44
0.367**
[0.088]
0.209*
[0.091]
-0.418
[0.341]
0.537**
[0.169]
2.532**
[0.673]
0.333**
[0.077]
0.123
[0.147]
-0.029*
[0.014]
-0.429*
[0.201]
0.574**
[0.125]
0.649
[1.345]
-0.287**
[0.098]
Yes
Yes
8,338
2,604
145
44
Logistic regression coefficients and average marginal effects (AME) with standard errors clustered by
country in brackets. ** p<0.01, * p<0.05, + p<0.10. † indicates variables that are standardized. ◊ indicates
standardized within each country. Environmental damage and sales are winsorized (top-coded) at the topone-percent level to account for outliers. All models also include a dummy variable denoting the 332
instances in which UN support was coded to 0 to replace missing values.
37
-0.087+
[0.050]
0.394**
[0.077]
0.217*
[0.091]
-0.407
[0.334]
0.535**
[0.161]
2.457**
[0.690]
0.415**
[0.096]
0.120
[0.147]
-0.030*
[0.014]
-0.425*
[0.200]
0.570**
[0.119]
0.734
[1.339]
-0.223*
[0.095]
Yes
Yes
8,338
2,604
145
44
0.08
Probability of
participation
0.06
0.04
0.02
0.00
-0.5
0.0
0.5
1.0
1.5
Quality of environmental disclosure †
No NGO participation
Figure 1.
2.0
2.5
High NGO participation
Companies with low-quality disclosures are especially deterred from participating in
the Global Compact when headquartered in countries with many NGO participants.
The solid line represents no NGO participation, corresponding to the 5th percentile of our sample. The dashed line
represents high NGO participation, corresponding to the 95th percentile (that is, 7 global NGOs participating in the
Global Compact). † indicates variables that are standardized
0.08
Probability of
participation
0.06
0.04
0.02
0.00
-0.5
0.0
0.5
1.0
1.5
Quality of environmental disclosure †
Weak ethical context
Figure 2.
2.0
2.5
Strong ethical context
Companies with low-quality disclosures are especially deterred from participating in
the Global Compact when headquartered in countries with strong ethical contexts.
The solid line represents weak ethical contexts, corresponding to the 5th percentile of our sample (that is, companies
with headquarters in Argentina, Indonesia, the Philippines, or Russia). The dashed line represents strong ethical
contexts, corresponding to the 95th percentile (that is, companies with headquarters in Denmark, Finland, or New
Zealand). † indicates variables that are standardized
38
0.08
Probability of
participation
0.06
0.04
0.02
0.00
-0.5
0.0
0.5
1.0
1.5
Quality of environmental disclosure †
Smaller companies
2.0
2.5
Larger companies
Figure 3. Companies with low-quality disclosures that are smaller in size are especially deterred
from participating in the Global Compact.
The solid line represents smaller companies: those with sales at the 5th percentile in their headquarters country. The
dashed line represents larger companies: those with sales at the 95th percentile in their headquarters country. †
indicates variables that are standardized
39
Appendix. An alternative measure of civil society scrutiny
To test H1, we measure civil society scrutiny as the number of global NGO participants in the
Global Compact in each company’s headquarters country in a given year. This approach is
predicated on the assumption that it is easier for NGOs to access information on domestic
companies and to disseminate information to domestic stakeholders that includes communities,
the media, and regulators. However, NGOs are increasingly using the Internet to obtain and
disseminate information across national boundaries. For example, Greenpeace is headquartered
in the Netherlands, but the organization targets companies all over the world via their websites
and blogs. We use a country’s number of Internet users as an alternative measure of civil society
scrutiny. This metric proxies for the extent to which a company’s stakeholders discover that it
was criticized for having failed to uphold the Global Compact’s principles. From the World
Bank, we obtained data on the number of Internet users per 100 people, which it collects from
the International Telecommunication Union. This number includes people who have Internet
access at home, work, or school.31
Table A1 reports results that use this alternative metric to test H2. As in Model 2 in Table
4, the positive significant coefficient on this interaction term (β = 0.10; p = 0.05) indicates that
companies with higher-quality disclosures are especially likely to participate when headquartered
in countries with greater Internet usage. Thus, Global Compact participation in this institutional
context provides a more credible signal of a company’s capability to manage environmental risks
and impacts.
31
Data on Internet users from the World Bank can be accessed at
http://data.worldbank.org/indicator/IT.NET.USER.P2. Accessed May 2015.
Appendix - 1
Table A1. Regression results using Internet usage as an alternative measure of civil society
scrutiny
Dependent variable: Global Compact participation
Quality of environmental disclosure† ×
Number of Internet users in hundreds †
Quality of environmental disclosure †
Environmental damage
Any human rights concerns
Any labor concerns
Peer participants per million population
Sales ◊
ISO 14001 certifications per million population
GDP per thousand population
NGO participants †
Number of Internet users in hundreds †
Total number of NGOs †
UN support
Stakeholder expectations †
Industry dummies
Year dummies
Observations (company-years)
Companies
Participant companies
Countries
(1)
0.096+
[0.052]
0.378**
[0.090]
0.214*
[0.091]
-0.404
[0.332]
0.513**
[0.178]
2.521**
[0.677]
0.333**
[0.076]
0.075
[0.158]
-0.035*
[0.016]
-0.462*
[0.200]
0.103
[0.171]
0.606**
[0.142]
0.788
[1.462]
-0.288*
[0.119]
Yes
Yes
8,167
2,594
144
44
Logistic regression coefficients with standard errors clustered by country in brackets. ** p<0.01, *
p<0.05, + p<0.10. † indicates variables that are standardized. ◊ indicates standardized within each
country. Environmental damage and sales are winsorized (top-coded) at the top-one-percent level to
account for outliers. All models also include a dummy variable denoting the 332 instances in which
missing values of UN support were recoded to 0.
Appendix - 2