EC O L O G IC A L E C O N O M IC S 6 8 ( 2 0 09 ) 71 9 – 7 30

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

w w w. e l s e v i e r. c o m / l o c a t e / e c o l e c o n

ANALYSIS

Why do manufacturing facilities introduce environmental

management systems? Improving and/or
signaling performance☆

Nick Johnstone a,⁎, Julien Labonne b

a
National Policies Division, OECD Environment Directorate, 2 rue André-Pascal, 75775 Paris Cedex 16, France
b
Social Development Department, World Bank, 1818 H Street, NW Washington, DC 20433, USA

AR TIC LE I N FO ABS TR ACT

Article history: In this paper we explore the motivation for the introduction of environmental management
Received 13 July 2007 systems, and their certification. A distinction is drawn between their role in bringing about
Received in revised form 22 April 2008 better compliance or improved performance, and as external indicators of good
Accepted 3 June 2008 environmental practices to both other market participants and regulatory authorities.
Available online 3 July 2008 Drawing upon a database of approximately 4000 facilities in seven OECD countries,
empirical evidence is found for the role that both factors play in encouraging the adoption
Keywords: and certification of EMS's, but that the relative importance of different factors varies
Environmental management system according to facility size.
ISO14001 © 2008 Elsevier B.V. All rights reserved.
Signaling
Asymmetric information
Corporate social responsibility
Environmental policy

JEL classification:
D82 (asymmetric and private
information); L15 (information and
product quality); Q50 (environmental
economics); M14 (social responsibility)

1. Introduction lower cost (Potoski and Prakash, 2005a). Indeed, both regula-
tors and buyers are increasingly concerned about the quality
Environmental management systems (EMS's) are usually of facilities' environmental management. The most wide-
viewed as playing a causal role in bringing about either: a) spread EMS standard (ISO 14001) mandates the introduction of
improved environmental performance and compliance; or, b) an explicit corporate environmental policy, including: the
comparable environmental performance and compliance at identification of the environmental impacts of its activities;

☆
The authors would like to thank Paul Lanoie and Gilles Grolleau for helpful comments, as well as two anonymous referees. All
remaining errors are ours. The views expressed here do not necessarily reflect those of the World Bank, the OECD or their respective
member countries.
⁎ Corresponding author. Tel.: +33 1 45247922; fax: +33 1 45247876.
E-mail addresses: [Link]@[Link] (N. Johnstone), jlabonne@[Link] (J. Labonne).

0921-8009/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/[Link].2008.06.003
720 EC O LO GIC A L E CO N O M ICS 6 8 ( 2 00 9 ) 7 1 9 – 7 30

the identification of legal and other requirements relating to frameworks differ significantly by country, sector, and facility
its activities; the establishment of an audit and review system; size the sample allows for a more robust assessment of the
and procedures to rectify any shortcomings identified. Imple- impacts of policy variation. (See Johnstone et al., 2007a for a
mented effectively, such measures can help facilities improve comparison of policy frameworks across these vectors.) In
their environmental performance. For this reason examples of addition, data on market structure and other factors which
“management-based regulation” are increasingly common, influence EMS adoption and certification is included in the
with the implicit assumption that good management will lead analysis.
to good performance (Bennear, 2007). However, the most important advance on previous work is
However, the widespread view that EMS's are a reflection of the rich characterisation of the environmental policy frame-
“better” environmental stewardship has also raised the work, including general policy measures as well as policy
possibility that certified EMS's serve as a signaling device, incentives which are targeted directly at EMS adoption and
informing others that they are managing their environmental certification. In particular, relative to previous literature (King
impacts efficiently.1 Since the quality of environmental et al., 2006; Toffel, 2005; Terlaak and King, 2006; King et al.,
management is not readily observable, this is an area 2006) we are able to examine whether the EMS certification
characterized by strong information asymmetries.2 As such, decision serves as a signal to public regulators, and not just
it can be a good strategy for facilities to provide information on private sector supply chain partners. For instance, Potoski and
their environmental management practices to both regulators Prakash (2005a) estimate the impact of ISO certification on
and buyers. In this paper, we argue this is one of the reasons compliance with regulations, but not whether a reduction in
why facilities decide to implement an EMS.3 regulatory scrutiny is an incentive to certify an EMS. To the
In the event that the primary motivation for the introduc- extent that policymakers come to see EMS's as a reflection of
tion of an EMS relates to the role that it plays in bringing about good environmental performance, an understanding of the
improved environmental performance, the motivation is role of regulatory “signaling” as a motivation for their
internal to the facility. In the event that the primary adoption and certification becomes increasingly important.
motivation relates to the perceptions of others, the motivation Overall, our results provide support for the view that in
is external, with the certified EMS acting as a “signal” to others. addition to their perceived role in bringing about improved
It is important to note that unless EMS brings about improved environmental performance, facilities implement and certify
environmental performance, the latter motivation will not be EMS to signal to others in the market, particularly when there
sustainable. Indeed Spence (1973) argues that a strategy based is significant potential for asymmetry of information between
upon signaling which provides misleading information (i.e. the facility in question and those that they are trying to signal,
does not reflect superior environment management in this confirming findings reported in Terlaak and King (2006). In
case) will not be viable in the longer-term. addition, we find strong evidence that certification serves as a
In this paper we explore the role that certified EMS's play as a signal to regulatory authorities, although the intended reci-
signaling device toward regulators and potential buyers through pients of the signal appear to differ by facility size.
the use of a database covering over 4000 manufacturing facilities. The remainder of the paper is organized as follow. Section 2
The database is unique in that it provides information on those presents some theoretical arguments as of why firms decide
facilities which have not introduced EMS's, those which have to implement EMS. Section 3 describes the data set and
done so, but not certified them, and those which have done so provides descriptive statistics. Section 4 presents our estima-
and certified them by third parties. The distinction between these tion strategy provides the results. The last section concludes
three groups is important since the motivations for implement- and presents some policy implications.
ing an EMS may be largely internal, while the motivation for
certification may be external. As King et al. (2006) point out,
collapsing these two decisions in the analysis can be misleading. 2. Conceptual framework and
The paper provides a number of advances on previous principal hypotheses
literature. Firstly, the database covers a broad range of
facilities, with observations from facilities with more than 50 As noted above, the introduction of an EMS involves defining a set
employees in all manufacturing sectors in seven OECD of formal environmental policies, goals, strategies and adminis-
countries (France, Norway, Hungary, Germany, Canada, Uni- trative procedures for improving environmental performance.
ted States, Japan). This allows for much greater variation in However, one should keep in mind that implementing an EMS
the data, particularly the policy variables. Since regulatory involves the adoption of a set of practices rather than achieve-
ment of a certain level of environmental performance. The
emphasis is on the process rather than on actual performance.
1
For a review of the literature on signaling see Riley (2001). Facilities may decide to implement an EMS for the benefits
2
See Terlaak and King (2006) for a discussion related to ISO it brings in terms of improved environmental management,
9000. and ultimately environmental performance. The quality and
3
According to the European Commission, ‘An EMS is a problem even presence of environmental management systems may
identification and problem solving tool that provides organiza-
not be observable to those outside the facility. However,
tions with a method to systematically manage their environ-
certification of an EMS is public, and as such facilities may
mental activities, products and services and helps to achieve their
environmental obligations and performance goals.’ [Link] choose to certify their EMS in order to communicate the
[Link]/comm/environment/emas/about/enviro_en.htm accessed existence of a recognised environmental management sys-
4/01/2006. tem. In the remainder of this section, we discuss the
 EC O L O G IC A L E C O N O M IC S 6 8 ( 2 0 09 ) 71 9 – 7 30 721

motivations for both the introduction of an EMS and the programmes also had a beneficial effect, but it arose indirectly
subsequent (or simultaneous) decision to have it certified. through their effect on financial performance. Anton et al.
(2004) used instrumental variables estimation techniques to
2.1. EMS implementation as a means to examine the role of a variety of environmental management
improve environmental performance practices on reported environmental performance. Interest-
ingly, they find that such measures had a positive role, and
Why would environmental management systems and tools that their importance was particularly significant for those
result in improved environmental performance? Firstly, firms that were originally the worst environmental perfor-
environmental management systems and tools can reduce mers. Similarly, in a study of American manufacturing
the probability of unintentional non-compliance with regula- facilities over the period 1991–2003, Toffel (2005) found that
tions. Secondly, environmental management systems and facilities which had certified ISO 14001 had improved their
tools may help managers identify and implement the most environmental performance relative to non-adopters with
cost-effective means of meeting their environmental objec- respect to facility-wide toxic emissions and in terms of health
tives, allowing for improved performance, whether it is hazards associated with toxic emissions. For instance, facility-
induced by regulation or not. Environmental management wide toxic emissions (as reported in the US Toxic Releases
systems and tools are thought to achieve these objectives by Inventory) from ISO-certified facilities declined by 39% relative
generating information about both regulatory requirements to the trend among those facilities which had not adopted ISO
and internal environmental practices; by helping to resolve 14001.
internal agency control issues which may result in adverse There is, however, some countervailing evidence, at least
environmental impacts; and by institutionalizing environ- with respect to ISO certification. In a panel data study of 37
mental performance as a corporate objective (Potoski and pulp and paper plants in Quebec over the period 1997–2003
Prakash, 2005a; Grolleau et al., 2007). Barla (2007) does not find significant evidence of reductions in
If the motivation for the introduction of EMS's is primarily pollution following ISO 14001 certification. While BOD emis-
related to a perception that they generate environmental sions decline slightly, there is no significant change in TSS
benefits or ensure compliance, one would expect to find that emissions or in the quantity of process water discharged.
the public policy factors are determinant in motivating their Indeed, there is some evidence that uncertified plants
introduction. Factors such as the frequency of inspections and experience greater emission reductions than certified plants.
the stringency of the policy framework would play a determi- Similarly, in their study of almost 8,000 American manufac-
nant role in their introduction. Moreover, the benefits turing facilities,King et al. (2006) find evidence that EMS
associated with their introduction may differ by facility or adoption results in improved environmental performance, as
firm. For instance, larger facilities may have greater informa- measured changes in TRI releases over the period 1995–2001.
tion requirements and may be more subject to agency control However, they find no evidence that certification of an EMS to
problems than smaller facilities, increasing the benefits of the ISO 14001 has an effect on changes in TRI releases over the
EMS adoption. period. In the context of the present paper this result is
However, there is limited evidence on the role of EMS's in particularly striking.
encouraging regulatory compliance or improved environmental
performance. In one of the few studies to use compliance as a H1 (PERFORMANCE IMPROVEMENT). Facilities which are
dependent variable, Dasgupta et al. (2000) undertook an analysis subject to more stringent environmental policies (i.e., regula-
of Mexican manufacturing facilities in which they evaluated the tory pressure) are more likely to implement an EMS in order to
importance of the presence of various environmental manage- improve their environmental performance. They may not
ment and policy indicators on their self-reported degree of have it certified by a third party if this does not yield any
compliance relative to the regulatory standard. They found that further improvements in performance.
the number of steps completed in ISO 14000 EMS and a dummy
variable for whether or not environmental training had been 2.2. EMS as a signal to regulators
given to “non-environmental” workers both had a positive effect
on compliance. Similarly, in an analysis of 3,700 U.S. manufac- In addition to bringing about improved environmental per-
turing facilities Potoski and Prakash (2005a) find that ISO 14001 formance, the certification (rather than just the adoption) of
“improves facilities” compliance with government regulations' an EMS may also be used as a signal for regulatory authorities.
with respect to air quality. Findings by Potoski and Prakash (2005a) suggest that environ-
There is rather more evidence on the effect of the adoption mental regulators tend to equate certified EMS with a
of different environmental management tools on environ- commitment to good environmental performance. With
mental performance. For instance, in a study of Italian scarce resources, inspectors will not be able to visit all
facilities by Siniscalco et al. (2000), it was found that between facilities in every period, and will have to “sort” facilities
the years 1994–1997 reported SOx and NOx pollution rates fell according to available information such as previous environ-
more for firms with environmental management tools (com- mental performance. The presence of good environmental
pensation schemes, audits and award schemes) than those management may also be perceived as a useful indicator, and
without such schemes. In multivariate analysis they found as such the presence of a certified EMS may be used in the
that the quality of the information provided in environmental targeting of facilities for inspection. The logic is as set out
reports had a negative (i.e. beneficial) and significant effect on above — a belief that if appropriate management systems are
pollution rates. Audits, compensation schemes and award in place, superior environmental performance will follow.
722 EC O LO GIC A L E CO N O M ICS 6 8 ( 2 00 9 ) 7 1 9 – 7 30

However, if the signal is not “credible” this can be a risky in which the facility operates and the type of competition in
strategy for the regulator to adopt, an issue explored in which the facility engages.5 In addition, the benefits of EMS
Johnstone et al. (2007b). At least initially, the benefits from adoption as a signaling device may be greater for smaller
signaling will, in fact, be greatest for the worst performers. facilities than for larger ones due to the relatively high fixed
Indeed, as noted above King et al. (2006) finds that weaker costs associated with their implementation, increasing the
environmental performers are more likely to have a certified credibility of the signal provided (Sinclair-Desgagne and
ISO 14001. Gozlan, 2003).
Nonetheless, regulators have started introducing formal Information asymmetries are likely to be more salient
policies to encourage facilities to adopt certified EMS (see when the two parties are physically and culturally distant
Coglianese and Nash, 2001 for a discussion.). The policies (Grolleau et al., 2007; King et al., 2006). This is thought to be
increase the benefits associated with adopting and certifying the case because contacts are likely to be less frequent and
an EMS, such as by providing preferences in public procure- cultural differences make communication more difficult. In
ment or by easing the administrative requirements of certified addition, it is more costly to closely monitor suppliers that are
facilities.4 Even if the policy incentive is not explicit there may further away. As Nakamura et al. (2001) have noted “foreign
be a perception amongst managers (perhaps based on customers may have less chance to monitor the performance
experience) that certification will reduce inspection frequency of a firm or have knowledge about its actions and intentions.”
or relax regulatory stringency. Clearly, this type of policy Partly due to this information deficit, it has been argued that
incentive may encourage the use of certification of an EMS as a customers may trust domestic firms more than foreign firms,
signaling device. The database exploited for this study thus increasing the potential benefits of introducing and
includes information on whether procurement preferences certifying and EMS for those who trade on international
were provided, permitting was expedited, and inspections markets. As such, the incentives to certify an EMS will increase
were thought to be less frequent. with the geographical scope of the market in which the facility
The benefits of signaling to regulatory authorities may be operates.
particularly valuable for “foreign” firms, since regulators may The value of certifying an EMS as a signaling device is likely
be less familiar with their practices. In addition, the size of a to be dependent upon the type of competition in which a
facility may result in increased visibility to regulatory autho- facility engages. Under perfect competition the primary
rities, as well as increased applicability of regulatory require- information transmitted (product price) is transparent and
ments. This would further increase the benefits of EMS's as readily accessible. Competition is on the basis of homoge-
signaling devices (Ruiz-Tagle, 2006). neous commodities, and as such there is no need to transmit
information about the management practices of the supplier.
H2 (REGULATORY SIGNALING). Facilities which gain regula-
This is not true in the case of monopolistic competition, in
tory advantages from introducing an EMS (e.g. preferential
which facilities compete on the basis of product quality or
procurement or reduced inspections frequency for certified
brand image. Where product differentiation or firm branding
facilities) are more likely to have it certified. By “signaling”
is important non-price factors can be key factors in the
their good practices to regulators certification has a value
market. This is not an area which has not been previously
which extends beyond the effect that it has on improved
explored in the literature. However, it has been argued that the
environmental performance.
benefits of EMS certification are greater if the facility is in long-
term vertically-integrated relations with buyers, since the
2.3. EMS certification as a signal to buyers
potential to exploit information asymmetries may be greater
(Zenger and Hesterly, 1997). Both Nakamura et al. (2001) and
In the management literature it is commonly argued that one
Bansal and Hunter (2003) find support for this hypothesis.
of the primary motivations for the certification of a manage-
Certification can also serve as an important information
ment system (such as ISO 9000) is to signal to others in the
tool within the firm. For instance, in multi-facility firms,
market of the high-quality management system in place. As
information asymmetries might arise between individual
King et al. (2006) and Potoski and Prakash (2005a) have pointed
facilities and the firm headquarters. Firms concerned for
out, this is likely to be true of ISO 14001 and other certificates
their brand reputation will closely monitor their facilities'
of environmental management as well. In the absence of
commitment to environmental action. For the reasons high-
complete information on all potential suppliers, potential
lighted above, the costs associated with this monitoring will
buyers of a facility's outputs may choose to purchase from
be higher when the facility and headquarters are located in
facilities that are certified, interpreting the certificate as a sign
different countries. As such, facilities with corporate head-
of good environmental management.
quarters located in a foreign country may be encouraged to
However, the value of this information will vary depending
certify an EMS (King et al., 2006).
upon a number of factors. We argue that the main factors
influencing the quality of information between a facility and H3 (MARKET AND FIRM SIGNALING). Facilities which gain
its potential buyers are the geographical scope of the market market advantages from introducing an EMS (e.g., firms

4 5
See Delmas (2002) for a discussion of the case of Germany in For a particularly interesting study in this regard see Christ-
this regard. mann and Taylor (2006).
 EC O L O G IC A L E C O N O M IC S 6 8 ( 2 0 09 ) 71 9 – 7 30 723

competing on the basis of brand image) are more likely to have Table 1 – Response rate by country
it certified since it helps them distinguish themselves from Response rate
other market participants.
Canada 25.0%
France 9.3%
2.4. Cost factors and other considerations
Germany 18.0%
Hungary 30.5%
Adopting an EMS can be very costly in terms of both start-up and Japan 31.5%
operating costs. Moreover, the costs associated with certifying Norway 34.7%
an EMS (rather than adopting one) may also be subject to United States 12.1%
significant fixed transaction costs. While Delmas (2002) found Total 24.7%

that for 75% of her sample of 152 manufacturing firms had

certification costs of less than $100,000, for large firms the costs
due to the high fixed costs associated with EMS adoption and
can be much higher. Potoski and Prakash (2005b) estimate that
certification, this may also discourage adoption and certification
medium-size firms incur costs of more than $250,000. According
amongst smaller facilities.7 As such, differences across facility
to one EPA official transaction costs for a large firm can be as
size are examined for each of the four main hypotheses.
great as USD 1million (Potoski and Prakash, 2005b).
As such a facility's financial resources are expected to
matter a great deal in the decision to adopt and certify an EMS
(Potoski and Prakash, 2005a). Given that this is likely to 3. The data
financed out of retained earnings, the profitability of a facility
is thought to be important influences on the decision to First, the data was collected by means of a postal survey
introduce an EMS, and Henriques and Sadorsky (2007) find undertaken in seven OECD countries (Canada, France, Ger-
support for the role that profitability has on the likelihood of many, Hungary, Japan, Norway and the United States) at the
introducing more comprehensive environmental manage- facility level (see [Link]/env/cpe/firms for a discussion
ment. In some cases governments seek to defray some of of sample procedure). Annex A provides information on the
the costs of adopting and certifying an EMS (e.g. providing survey protocol. The data covers facilities in all manufacturing
financial support or technical assistance, training and support sectors and not only those in the more polluting sectors. The
for third-party certifiers, etc.) (Delmas, 2002). diversity in countries and sectors sampled implies a greater
In addition, since many of the more general “corporate variation across policy frameworks, technological opportu-
governance” requirements associated with having a stock nities, and other factors that will allow for the generation of
market listing are likely to be strongly complementary with more reliable estimates of different potential determinants of
transparent and accountable environmental management, environmental innovation.
there are likely to be economies of scope between the two.6 Respondents were CEOs and environmental managers.
Henriques and Sadorsky (2007) find support for the role that Response rates range from approximately 9% to 35%, with a
profitability and stock market listings have on the likelihood weighted mean of almost 25% (see Table 1). For a postal survey
of introducing an EMS. There may also be economies of scope this is satisfactory, particularly since previous industrial
between the presence of a more general “quality management surveys undertaken in the environmental sphere in many of
system” and environmental management systems (see King the countries included in the survey have tended to have very
and Lenox, 2001; Delmas, 2003). low response rates. For instance, in a review of 183 studies
based on business surveys published in academic journals
H4 (COST FACTORS). More profitable facilities which already
Paxson (1992) reports an average response rate of 21%.
have advanced management practices in places are more
While surveys undertaken as part of official data collection
likely to implement and certify an EMS.
exercises may have higher response rates, in many such cases
There are good reasons to expect that facility size affects all there are legal obligations to respond. Other studies also focus on
the different potential motivations for the adoption and large firms (e.g. Standard and Poor 500) or firms with other
certification of an EMS. Firstly, since smaller facilities are not attributes (i.e. listed on the stock exchange) which are likely to
usually under as much pressure with respect to their environ- have higher response rates. Indeed, given the population
mental impacts, the benefits of adopting an EMS in terms of sampled, the response rate was higher than had been anticipated.
improved environmental performance will be less than for a Table 2 provides data on the number of respondent
larger facility (H1). Moreover, since they face less regulatory facilities by industrial sector for the seven countries. While
scrutiny, the benefits in terms of “signaling” to public authorities the sectoral data is available at the ISIC two digit level
will also be less (H2). Conversely, credence goods (such as a (24 sectors), the data is presented in somewhat aggregated
certificate) may be more valuable for smaller facilities whose form below. Annex B provides a comparison of the population
characteristics are less well-known to outsiders (H3). And finally, of facilities at the two-digit level with our sample for five
countries. In the case of Norway, on the basis of chi-square
test the sample is not significantly different from the
6
For instance, in France the ‘Nouvelles Régulations Economi-
ques’ (article 116) mandates that firms listed on the stock
7
exchange provide environmental reports, a significant compo- Both Dasgupta et al. (2000) and Johnstone et al. (2007a,b) find
nent of implementing an EMS. evidence for the importance of economies of scale.
724 EC O LO GIC A L E CO N O M ICS 6 8 ( 2 00 9 ) 7 1 9 – 7 30

Table 2 – Survey respondents by sector and by country

ISIC Classification Canada France Germany Hungary Japan Norway USA Total

Food beverage and tobacco Sectors 15–16 23 44 77 68 138 33 37 420

Textiles, apparel, leather Sectors 17–19 8 13 40 50 72 10 12 205
Wood products and furniture Sectors 20 and 36 32 12 26 27 32 49 34 212
Paper, publishing and printing Sectors 21–22 22 17 92 21 129 25 24 330
Fuel, chemicals, rubber, plastics Sectors 23–25 40 48 149 54 195 24 126 636
Non-metallic mineral products Sector 26 13 13 34 21 34 14 20 149
Basic and fab'd metals Sectors 27–28 42 53 211 52 286 54 129 827
Machinery and instruments Sectors 29–33 50 47 227 119 439 55 59 996
Motor vehicles and transp. eqpmt Sectors 34–35 23 19 32 22 113 44 37 290
Recycling and other Sectors 37–39 3 2 10 29 29 1 5 79
Total 256 268 898 463 1467 309 483 4144

population of facilities in terms of size classes (50–99 employ- Measuring the “comprehensiveness” of an environmental
ees; 100–249 employees; 250–499 employees; and, N500 management system is not straightforward. Fortunately data
employees). In the case of Germany the distribution of the was collected on the implementation of a variety of specific
sample is statistically different from that of the population by environmental management tools which may or may not be
sector. Facility size data is not available. In the case of Japan, included as part of an environmental management system.
the sectoral distribution of the sample is representative, but Fig. 1 gives an indication of the comprehensiveness of
not the size distribution. For France and Hungary, only firm- environmental management for facilities which do not have
level data is available when using a cut-off of 50 employees. and environmental system at all, those who have implemen-
There are a large number of observations from smaller ted an EMS but have not certified it, and those who have an
facilities for which response rates are usually much lower in ISO-certified an EMS. While there is little question that
such surveys. Indeed, in many previous studies small and certified EMS's are more “comprehensive” than uncertified
medium sized enterprises are not sampled at all, a significant EMS's, for a number environmental management tools
shortcoming as regulators are increasingly seeking to influ- difference is statistically indistinguishable at the 5% level.
ence the behaviour of smaller sources. In the sample, over Conversely, there is a statistically significant difference
2500 facilities can be characterized as small or medium sized between facilities without an EMS and an uncertified EMS for
enterprises (b250 employees). Given that many of these same all environmental management tools.
facilities are part of multi-facility firms, the true representa- Respondents were also requested to assess the importance
tion of SMEs in the database at the level of the firm is (1 = not important; 2 = moderately important; 3 = very impor-
somewhat lower, but still very significant. tant) of different factors when considering whether or not to
introduce an environmental management system. While
3.1. The EMS variables there is some overlap, these can be broadly distinguished
between factors which are related to the role of EMS's as
The environmental management decision is reflected in the
supply-side efficiency-improving measures (cost savings in
binary response to the question: “has your facility actually
terms of resources and waste management, information
implemented an environmental management system?”
about operations), demand-side market factors (firm image
Respondents were also requested to indicate whether or not
and product differentiation), and those which relate to
it is certified, and if so under which scheme. The percentage of
regulatory concerns (ensuring regulatory compliance, improv-
facilities reporting that they did not have an environmental
ing relations with regulators, and preventing or controlling
management systems (EMSNONE) is just over 60%, but it varies
pollution). Table 3 provides the distribution of responses. On
widely across countries, with figures ranging from just over
this basis it would appear that EMS's are first and foremost
70% (Germany and Hungary) to approximately 43% (United
seen as means of providing information which can help
States). Of the 40% of facilities with an EMS, over 1000 facilities
facilities avoid being unintentionally non-compliant. This is
(25% of the sample) reported having some form of certification,
consistent with other work which has indicated that environ-
whether EMAS or ISO 14001 (EMSCERT). Thus, approximately
mental violations are frequently accidental rather than
15% of facilities have an uncertified EMS (EMSNOCRT). As
intentional.9
expected, smaller facilities are less likely to have environ-
And finally, it is interesting to compare the frequency of
mental management systems in place. For the smallest class
inspections with the different classes of EMS implementation
(b100 employees) less than 20% of facilities have an EMS, while (see Fig. 2). Interestingly, facilities with certified EMS's are
for the largest class (N500 employees) the figure is over 60%.8 more likely to report that they had not been inspected in the
previous three years than those with uncertified EMS's.
However, they are also more likely to report that they have
8
been inspected more than 6 times or more often. They are also
While there are differences between EMAS and ISO 14001, in
more likely to report more than 6 inspections than those
this study we only take facilities with ISO 14001. There are
relatively few facilities which have implemented EMAS, but not
ISO 14001 and so the construction of the dependent variable will
9
have insignificant impacts on the results. See Harrington (1989) and Cohen (1998).
 EC O L O G IC A L E C O N O M IC S 6 8 ( 2 0 09 ) 71 9 – 7 30 725

Fig. 1 – Implementation of environmental management tools.

without an EMS, but given the strong correlation between 45.4% consider the environmental policy regime to be
inspection frequency and facility size, as well as between EMS moderately stringent (POLSTRNG2), requiring some dedi-
implementation and facility size, a simple bivariate compar- cated managerial and technological responses. As noted
ison is not particularly illuminating. above, we also have information on the number of inspec-
tions by public environmental authorities the facility faced
3.2. Explanatory variables during the last three years (INSPFREQ). We use this as an
additional measure of the environmental regulatory frame-
Consistent with the hypotheses set out above respondents work's stringency. The correlation between this variable and
were asked to provide information on factors which are likely STRING is quite low and as such those two variables provide
to reflect supply-side factors, as well as demand-side factors, quite different information.
including those which related to the role of EMS's as a means For the signaling of commercial buyers, it has been argued
to improve performance and as a signal. that the nature of competition faced by the facility can be
With respect to supply-side factors facilities were asked to significant. In order to capture this effect the respondent was
assess their overall business performance over the last three requested to indicate whether it competed on the basis of
years. With this information we create a variable PROFIT that corporate image (COMPIMG) or some other factor (i.e. price
takes the value 1 if a facility has generated positive profit over competition). In addition, it has been argued that the signaling
the past three years and 0 otherwise. We also have informa- of good environmental management is more important the
tion on whether the facility belongs to a firm that is listed on a more “distant” the facility is from market counterparts. As
stockmarket (FRMQUOT). Given the assumption that there are such, facilities were also asked to characterise their market's
economies of scale in the introduction of EMS's, data was
collected on facility size as reflected in the number of people
Table 3 – Motivation for introduction of EMS
employed by the facility on average over the last three years
(FACEMPL). In addition, facilities were requested to indicate Not Moderately Very
whether the regulatory authority provide financial support for important important important
the introduction of environmental management systems To prevent or control 601 844 883
(RAPFSUP). And finally, it is assumed that there will be pollution
economies of scope between the introduction of other To ensure regulatory 544 770 1018
management tools such as Quality Management Systems compliance
To identify 808 852 657
(QMSSYSTM) and EMS's.
environmental
On the demand side, respondents were requested to give
liabilities
their perception of the degree of stringency of the environ- To improve relations 1002 770 539
mental regulatory framework. In our sample only 16.8% of the with regulators
facilities consider the environmental policy regime to be very To differentiate 1166 778 366
stringent (POLSTRNG3), with a great deal of influence on products
decision-making within the facility (ranging from just 3.4% in To improve corporate 610 1029 691
image
Japan to 38.2% in France). 37.8% of the facilities consider that
To save input costs 919 760 628
obligations can be met with relative ease (POLSTRNG1) and
726 EC O LO GIC A L E CO N O M ICS 6 8 ( 2 00 9 ) 7 1 9 – 7 30

The expected signs on the some of the main variables are

presented in Table 4, as well as the expected motivation for
the sign, whether due to cost factors, benefits in terms of
performance improvement, or signaling (market or regula-
tory). Descriptive statistics for the main variables are provided
in Table 5. In the case of certified EMS's, the mean is relative to
those who have an EMS in place.

4. The model and results

In the first instance, the implementation of an EMS is

modelled. Given the binary nature of the dependent variable
(EMS = 0 if no EMS in place; EMS = 1 if EMS in place) a probit
model is estimated. A reduced-form single equation is
Fig. 2 – EMS vs. inspection frequency.
estimated, reflecting both demand and cost factors. In addition
to the variables mentioned above, dummies are included for
scope: local, national, regional and global. With this informa- both countries (with United States as the reference) and sectors
tion we create a variable (INTLMRKT) that takes the value 1 (with food and beverage sector as the reference). These are
when the facility operates on a global market on 0 otherwise. A intended to capture the effects of variables that are not
similar variable is created to indicate whether the firm captured by the other explanatory variables (e.g. differences
headquarters are located in a foreign country (FORGNHQ). in the costs of implementation of an EMS depending upon
Since it is felt that signaling of process-related environmental sectoral characteristics).
practices is important to downstream buyers a variable is Let EMS kji be the variable reflecting the net benefits
included to indicate whether the facility sells primarily associated with having an environmental management sys-
intermediate inputs to other firms (SUPCHAIN). tem in place for facility i in sector j in country k:
In addition, we asked facilities whether regulatory autho-
rities have programmes and policies in place to encourage EMS4kji ¼ bVXjki þ vk þ gj þ ekji ð1Þ
them to use an EMS. Indeed, over 20% of respondents reported
that governments provided explicit incentives for the intro- where β is a vector of coefficients, Xkij the vector facility
duction of environmental management systems and tools. If characteristics and, vk, ηj and εkji the error terms. However, we
do not observe EMSkji ⁎ directly but whether a facility has an
so, facilities provided information on the type of programmes
that are in place. The potential to receive financial support for EMS in place:
the introduction of an EMS has been discussed above, but a

1 if EMS4kji N 0
number of other incentives are also frequently in place some EMSkji ¼ ð2Þ
0 otherwise
of which reflect the value of EMS's in signalling to regulatory
authorities: expediting environmental permits programmes Given that we feel that there may be differences between the
(RAPEPRMB); reduction of the frequency of regulatory inspec- motivation to introduce an EMS and the motivation to have it
tions (RAPINSPB); allowing for access to public procurement certified, and that certification may be considered an indicator of
markets (RAPPREFB). “quality” of EMS we also estimated an ordered probit on the

Table 4 – Motivation for inclusion of variable and expected sign of coefficients

Factor Variable Cost Benefit
name
Performance Signal
improvement
Regulators Market

Perceived policy stringency POLSTRNG +

Frequency of inspections INSPFREQ +
Reduced inspection frequency RAPINSPB +
Expediting of permits RAPEPRMB +
Procurement preferences RAPPREFB +
Provision of financial support RAPFSUPB +
Facility employment FACEMPL + +
Listing on stock exchange FRMQUOT +
Reporting positive profits POSPROFIT +
Compete internationally INTLMRKT +
Hq in a foreign country FORGNHQ + +
Compete on corporate image COMPIMG +
Sell to other firms (not final good) SUPCHAIN +
QMS (i.e. ISO 9000) QMSYSTM +
 EC O L O G IC A L E C O N O M IC S 6 8 ( 2 0 09 ) 71 9 – 7 30 727

Table 5 – Descriptive Statistics

Mean [Link]. Minimum Maximum NumCases

EMSNONE 0.514 0.500 0 1 3664

EMSNOCRT 0.206 0.405 0 1 3664
EMSCERT 0.574 0.495 0 1 1790
POLSTRNG1 a 0.354 0.478 0 1 3832
POLSTRNG3 b 0.162 0.368 0 1 3832
INSPFREQ 3.748 7.044 0 106 3480
RAPPREFB 0.055 0.228 0 1 3832
RAPEPRMB 0.085 0.279 0 1 3832
RAPINSPB 0.072 0.258 0 1 3832
RAPFSUPB 0.084 0.277 0 1 3832
FACEMPL 352.275 880.960 50 28618 3832
FRMQUOT 0.175 0.380 0 1 3799
POSPROFIT 0.574 0.495 0 1 3832
INTLMRKT 0.510 0.500 0 1 3803
FORGNHQ 0.126 0.331 0 1 3808
SUPCHAIN 0.615 0.487 0 1 3765
COMPIMG 2.303 0.584 0 3 3773
QMSSYSTM 0.761 0.426 0 1 3717
a
Environmental policy is reported to be not stringent.
b
Environmental policy is reported to be very stringent.

following three classes of dependent variable: EMSNONE, matter more when uncertified and certified EMS's are
EMSNOCRT, EMSCERT. It is assumed that EMSCERT is “superior” distinguished.
to EMSNOCRT, which is in turn superior to EMSNONE. However, in order to test for the relative importance of cost
Keeping the same notation as above, Let EMSORDkji be the factors and the three demand-side motivations (performance
variable reflecting the environmental management behavior improvement, market signaling, regulatory signaling) for the
of facility i in sector j in country k: introduction of an EMS we report the marginal effects for the
intermediate classification (EMSNOCRT). If the sign is negative
EMSORD4kji ¼ b VXjki þ vk þ gj þ ekji ð3Þ for a variable which is assumed to reflect the returns on
signaling, then such effects are potentially important since a
However, we do not observe EMSORDkji⁎ directly but whether a unit change in the variable has a greater effect on encouraging
facility has an EMS in place and if so, whether or not it is certification than on implementation of an EMS.
certified: For the reasons set out in Section 2, we run the same
8 regressions with the full sample, the sample of facilities with
>
<0 if EMSORD4kji b 0 less than 100 employees, the sample of facilities with 100 to
EMSORDkji ¼ 1 if 0 V EMSORD4kji b l ð4Þ 249 employees and the sample of facilities with more than 250
>
:
2 otherwise

We estimate Eq. (2) as a simple probit. The results are

presented in Table 6. A number of supply-side factors are Table 6 – Results of Probit on EMS
important — facility size, the presence of a quality manage- Coeff. P-value
ment system, positive profits, and a listing on the stock
exchange. In addition, the provision of financial support for INTERCEPT −1.241 0.000
POLSTRNG1 −0.166 0.006
the introduction of an EMS is an important motivating factor.
POLSTRNG3 0.078 0.308
The results for the coefficients which are thought to reflect
INSPFREQ 0.009 0.051
signaling for market partners are more ambiguous. While RAPPREFB 0.041 0.755
supply chain partners and distance (markets and headquar- RAPEPRMB −0.047 0.695
ters) have a positive and significant influence on EMS RAPINSPB 0.425 0.001
adoption, the variable reflecting whether or not the respon- RAPFSUPB 0.430 0.000
dent feels that corporate image is an important factor in their FACEMPL 0.001 0.000
FRMQUOT 0.320 0.000
market strategy is not significant. Signaling regulators is
POSPROFIT 0.190 0.000
important, with the variable reflecting the presence of INTLMRKT 0.255 0.000
incentives related to perceived inspection frequency statisti- FORGNHQ 0.372 0.000
cally significant and positive. SUPCHAIN 0.159 0.006
We estimate Eq. (4) as an ordered probit. The results are COMPIMG 0.053 0.224
presented in Table 7. The only significant difference is that the QMSSYSTM 0.540 0.000
n = 3124
variable reflecting corporate image (COMPIMG) is now statisti-
log-likelihood = −1738.083
cally significant and positive. Market signaling appears to
728 EC O LO GIC A L E CO N O M ICS 6 8 ( 2 00 9 ) 7 1 9 – 7 30

Table 7 – Results of Ordered Probit on EMSNONE, of a quality management system increases the likelihood of
EMSNOCRT, EMSCERT having and certifying an EMS. Market “signaling” does not
Coeff. Std. t-ratio P- M.E. appear to be significant.
Err. value (EMSNOCRT) For medium-sized facilities, other supply-side factors
(facility size, quotation) matter much more than they do for
ONE −1.910 0.147 −12.952 0.000 −0.151
POLSTRNG1 −0.192 0.055 −3.475 0.001 −0.015 smaller facilities. Interestingly, the provision of financial
POLSTRNG3 0.054 0.066 0.821 0.411 0.004 assistance by public authorities for the implementation of an
INSPFREQ 0.013 0.004 3.602 0.000 0.001 EMS is still statistically significant and positive — but the
RAPPREFB −0.075 0.109 −0.690 0.490 −0.006 coefficient is much smaller. In addition, having a corporate HQ
RAPEPRMB 0.003 0.101 0.025 0.980 0.000 overseas and competing in international markets has a
RAPINSPB 0.278 0.101 2.758 0.006 0.022
positive influence on implementing and certifying an EMS.
RAPFSUPB 0.330 0.092 3.602 0.000 0.026
FACEMPL 0.000 0.000 17.035 0.000 0.000
Competing on the basis of corporate image is now a
FRMQUOT 0.315 0.064 4.880 0.000 0.025 statistically significant factor, as was the case for the full
PROFIT 0.173 0.046 3.741 0.000 0.014 sample. Demand-side factors related to policy stringency
INTLMRKT 0.291 0.053 5.488 0.000 0.023 remain insignificant.
FORGNHQ 0.357 0.070 5.128 0.000 0.028 For large facilities the overall fit is better still. For this sub-
CHAIN 0.158 0.051 3.115 0.002 0.013
sample policy stringency and inspection frequency play a
COMPIMG 0.077 0.039 1.977 0.048 0.006
significant and positive role, although for the former variable
QMS 0.623 0.061 10.207 0.000 0.049
n = 3257 respondents who reported moderate and significant policy
Log-likelihood = −2882.18 stringency are not statistically different — i.e. the variable
reflecting whether policy is “not very stringent” is negative
and significant. In this case the role of reduced inspection
frequency is statistically significant and positive. The percep-
employees. This is necessary to allow the effect of variables to tion that the presence of an EMS will reduce inspection
differ across size groups. By splitting the sample we will be frequency increases the likelihood of having one in place and
able to assess whether, for example, SMEs are more likely to certifying it. On the supply side, facility size, profitability and
respond to use a certified EMS as a signal to potential buyers the presence of a quality management system increase the
while larger firms are more likely to respond to policies likelihood of having an EMS and having it certified. The role of
increasing the benefits of having a certified EMS in place. market signaling is supported with variables reflecting inter-
For small facilities, there is a significant change. Regulatory national markets, foreign headquarters and supply chain
signaling is not significant. The only public policy incentive of partners all being positive and significant. However, for this
significance is the provision of financial assistance, with a sub-sample company image is no longer statistically signifi-
much larger coefficient than for the full sample. The presence cant (Table 8).
However, examination of the marginal effects for the
intermediate class dependent variable (EMSNOCRT) indicates
Table 8 – Results of Ordered Probit on EMSNONE, that it is only for the sample with the largest facilities that the
EMSNOCRT, EMSCERT for Sub-Samples effect of a given variable on the decision to certify an EMS is
greater than the effect of having an EMS at all (see Table 9). In
Small facility Medium Large facility
sample (50–99 facility sample sample this case the marginal effects of the variables which reflect
employees) (100–249 (N250 regulatory signalling (RAPINSP) and a number of the variables
employees) employees)

Coeff. P-value Coeff. P-value Coeff. P-value

INTERCEPT − 1.652 0.405 − 2.225 0.000 − 1.665 0.000 Table 9 – Marginal effects for sub-samples
POLSTRNG1 − 0.118 0.111 − 0.068 0.459 − 0.209 0.045 Small Medium Large
POLSTRNG3 − 0.117 0.170 0.165 0.162 0.080 0.413
INSPFREQ 0.003 0.011 0.010 0.165 0.011 0.017 INTERCEPT − 0.278 −0.221 0.102
RAPPREFB − 0.068 0.338 0.015 0.937 − 0.112 0.506 POLSTRNG1 − 0.020 −0.007 0.013
RAPEPRMB − 0.164 0.215 − 0.054 0.774 0.077 0.624 POLSTRNG3 − 0.020 0.016 − 0.005
RAPINSPB 0.214 0.230 0.126 0.487 0.420 0.006 INSPFREQ 0.001 0.001 − 0.001
RAPFSUPB 0.928 0.190 0.354 0.040 0.007 0.960 RAPPREFB − 0.011 0.001 0.007
FACEMPL 0.001 0.004 0.002 0.008 0.000 0.000 RAPEPRMB − 0.028 −0.005 − 0.005
FRMQUOT 0.190 0.202 0.367 0.004 0.122 0.158 RAPINSPB 0.036 0.013 − 0.026
POSPROFIT 0.167 0.099 0.143 0.075 0.230 0.003 RAPFSUPB 0.156 0.035 − 0.001
INTLMRKT 0.220 0.110 0.280 0.002 0.182 0.049 FACEMPL 0.000 0.000 0.000
FORGNHQ 0.242 0.219 0.379 0.002 0.327 0.001 FRMQUOT 0.032 0.036 − 0.008
SUPCHAIN 0.169 0.115 0.088 0.300 0.277 0.001 POSPROFIT 0.028 0.014 − 0.014
COMPIMG 0.086 0.084 0.127 0.048 0.010 0.880 INTLMRKT 0.037 0.028 − 0.011
QMSSYSTM 0.430 0.119 0.582 0.000 0.628 0.000 FORGNHQ 0.041 0.038 − 0.020
n = 917 n = 1158 n = 1182 SUPCHAIN 0.028 0.009 − 0.017
log-l'hood = log-l'hood = log-l'hood = COMPIMG 0.014 0.013 − 0.001
−660.71 −1051.03 −1047.20 QMSSYSTM 0.072 0.058 − 0.038
 EC O L O G IC A L E C O N O M IC S 6 8 ( 2 0 09 ) 71 9 – 7 30 729

which reflect market signalling (FRMQUOT, INTLMRKT, Bansal, P., Hunter, T., 2003. Strategic explanations for early
FORGNHQ) are negative for the EMSNOCRT class. adoption of ISO 14001. Journal of Business Ethics 46, 289–299.
Barla, Philippe, 2007. ISO 14001 certification and environmental
performance in Quebec's pulp and paper industry. Journal of
Environmental Economics and Management 53, 291–306.
5. Conclusion Bennear, L.S., 2007. Are management-based regulations
effective? Evidence from state pollution prevention
Motivations for the introduction of EMS's are of increasing programs. Journal of Policy Analysis and Management 26 (2),
importance to policymakers. However, a distinction must be 327–348.
Christmann, P., Taylor, G., 2006. Firm self-regulation through
drawn between factors that relate primarily to the implemen-
international certifiable standards. Journal of International
tation of an EMS, and factors that relate to certification. In
Business Studies 37, 863–878.
many cases these two sets of factors will be the same, but their Coglianese, C., Nash, J., 2001. Policies to promote systematic
relative importance will vary depending upon the importance environmental management. In: Coglianese, C., Nash, J. (Eds.),
of EMS's in bringing about improved environmental perfor- Regulating from the Inside. Resources for the Future Press,
mance and in advertising this to others (signaling). The data Washington D.C.
collected allows us to distinguish between the decision to Cohen, M., 1998. Monitoring and Enforcement of Environmental
Policy. Vanderbilt University, Owen Graduate School of
implement and EMS and have it certified.
Management, Working Paper. August.
The results are telling. On the one hand, the descriptive
Dasgupta, S., Hettige, H., Wheeler, D., 2000. What improves
data indicates that facilities with certified EMS's are more environmental performance? Evidence from Mexican industry.
likely to have a variety of other environmental management Journal of Environmental Economics and Management 39,
tools in place. In addition, policy stringency is marginally 39–66.
more important when the two decisions are distinguished in Delmas, M.A., 2002. The diffusion of environmental management
an ordered probit model rather than a simple probit model standards in Europe and in the United States. Policy Sciences
35, 91–119.
where certified and uncertified EMS's are treated together. On
Delmas, M., 2003. In Search of ISO: An Institutional Perspective on the
the other hand, the variable reflecting whether a company Adoption of International Management Standards. Stanford
competes on the basis of corporate image is only significant Graduate School of Business, Research Paper.
when these two decisions are distinguished. Grolleau, G., Mzoughi, N., Thomas, A., 2007. What drives agro-food
There are, however, differences across different sizes of firms to seek a certified environmental management system?
facility. Cost factors are the most important for smaller European Review of Agricultural Economics 34, 233–255.
Harrington, W., 1989. Enforcement leverage when penalties are
facilities, with both the provision of financial assistance and
restricted. Journal of Public Economics 37, 29–53.
profitability being significant. The latter coefficient is no
Henriques, I., Sadorsky, P., 2007. Environmental management
longer significant for larger facilities. For all samples the systems and practices: an international perspective. In:
presence of a quality management system has a positive Johnstone, N. (Ed.), Environmental Policy and Corporate
effect, reflecting economies of scope between the two Behaviour. Edward Elgar, Cheltenham.
management measures. Signaling regulators only matters Johnstone, N., et al., 2007a. Public environmental policy and
for the largest facilities (above 250 employees), as reflected in corporate behaviour. In: Johnstone, N. (Ed.), Environmental Policy
and Corporate Behaviour. Edward Elgar, Cheltenham, UK.
the positive and significant effect of perceived reduction in
Johnstone, N., et al., 2007b. Many a slip 'twixt the cup and the lip:
inspection frequency. Signaling to supply partners also
direct and indirect policy incentives to improve environmental
matters, but this is also particularly important for facilities performance. In: Johnstone, N. (Ed.), Environmental Policy and
above 100 employees. The likelihood of implementing and Corporate Behaviour. Edward Elgar, Cheltenham, UK.
certifying an EMS is greater if larger facilities are distant from King, A., Lenox, M., Terlaak, A., 2006. The strategic use of
buyers. In addition, if the corporate headquarters is overseas decentralized institutions: exploring certification with the ISO
this also has an important influence for the largest facilities. 14001 management standard. Academy of Management Journal
48 (6), 1091–1106.
On the basis of the results it would seem that “signaling” is a
King, A., Lenox, M., 2001. Lean and green: exploring the spillovers
strong motivation for the adoption and certification of EMS's, at from lean production to environmental performance. Production
least for larger facilities. However, a full assessment of its relative and Operations Management 10, 1–13.
importance is dependent upon reliable environmental and Nakamura, M., Takahashi, T., Vertinsky, I., 2001. Why Japanese
commercial performance data, as well as data on the presence firms choose to certify: a study of managerial responses to
and certification of EMS's, and public policy influences. Unfortu- environmental issues. Journal of Environmental Economics
and Management 42, 23–52.
nately such data is not available for a cross-section of countries.
Paxson, M.C., 1992. Response Rates for 183 Studies. Washington
However, given that governments are providing incentives for the
State University, Working Paper.
introduction of EMS's a fuller understanding of the motivations Potoski, M., Prakash, A., 2005b. Covenants with weak swords: ISO
for, and effects of, their introduction is required. 14001 and facilities' environmental performance. Journal of
Policy Analysis and Mangement 24 (4), 745–769.
Potoski, M., Prakash, A., 2005a. Green clubs and voluntary
REFERENCES governance: ISO 14001 and firms' regulatory compliance.
American Journal of Political Science 49, 235–248.
Riley, J.G., 2001. Silver signals: twenty-five years of screening and
Anton, W., Deltas, G., Khanna, G., 2004. Incentives for environmental signalling. Journal of Economic Literature 39, 432–479.
self-regulation and implications for environmental performance. Ruiz-Tagle, M.T., 2006. Why do Manufacturing Plants Invest in
Journal of Environmental Economics and Management 48, Environmental Management? University of Cambridge,
632–654. Department of Land Economy. WP 20.2006.
730 EC O LO GIC A L E CO N O M ICS 6 8 ( 2 00 9 ) 7 1 9 – 7 30

Sinclair-Desgagne, B., Gozlan, E., 2003. “A theory of environmental Toffel, M.W., 2005. Resolving information asymmetries in
risk disclosure”. Journal of Environmental Economics and markets: the role of certified management programs. UC
Management 45, 377–393. Berkeley, Center for Responsible Business, Working Paper 19.
Siniscalco, D., Borghini, S., Fantini, M., Ranghieri, F., 2000. The Zenger, T., Hesterly, W., 1997. The disaggregation of corporations:
Response of Companies to Information-Based Environmental selective intervention, high-powered incentives, and
Policies. FEEM Working Paper, Milan. molecular units. Organization Science 8, 209–223.
Spence, A.M., 1973. Job market signaling. Quarterly Journal of
Economics 87.
Terlaak, A., King, A.A., 2006. The effect of certification with the
ISO 9000 quality management standard: a signaling
approach. Journal of Economic Behaviour and Organization
60, 579–602.