Content uploaded by Jury Gualandris
Author content
All content in this area was uploaded by Jury Gualandris on Jun 09, 2021
Content may be subject to copyright.
1
The association between supply chain structure and transparency:
a large-scale empirical study
This paper has been accepted for publication in the forthcoming regular issue of the
Journal of Operations Management
Article DOI: 10.1002/joom.1150
Jury Gualandris*
Ivey Business School
Western University
jgualandris@ivey.ca
Annachiara Longoni
ESADE Business School
University Ramon Llull
annachiara.longoni@esade.edu
Davide Luzzini
EADA Business School
dluzzini@eada.edu
Mark Pagell
Smurfit Business School
University College Dublin
mark.pagell@ucd.ie
* Corresponding author
Authors’ note/Acknowledgments: We are very grateful for the insightful guidance provided
by Departmental Editor Brian Jacobs throughout the revision process, and for the feedback
given by three anonymous reviewers. We thank Estefania Ruiz and Marco Suma for their
research assistance, and Robert D. Klassen, Tom Lyon and Wren Montgomery for feedback
on earlier versions. This paper has benefited from input received at Erasmus University.
Feedback received at INFORMS Phoenix, AoM Boston and ARCS Milan was also helpful.
2
The association between supply chain structure and transparency:
a large-scale empirical study
ABSTRACT
An emerging body of work acknowledges the challenges focal firms face in gathering material
information about their extended supply chains and begins to point to the role of supply chain
structure in influencing supply chain transparency. Still, large-scale empirical evidence on this
complex association remains elusive, especially at the supply chain level of analysis. We begin
to bridge this empirical gap by examining whether supply chain structure systematically
associates to supply chain transparency in the context of the collective public environmental,
social, and governance (ESG) disclosures made by a focal firm’s customers, suppliers and sub-
suppliers. To shed light on this underexplored empirical phenomenon we gather Bloomberg
SPLC data and Bloomberg ESG data about 4803 firms and 20,504 contractual ties organized
in 187 extended supply chains. We find that supply chain density positively associates with
supply chain transparency, whereas supply chain clustering holds a negative association. We
also find that supply chain geographical heterogeneity positively associates with supply chain
transparency. Our results significantly expand the literature on supply chain transparency and
are relevant to supply chain professionals because they emphasize the central role of supply
chain structure in enabling or constraining supply chain transparency.
Keywords: Transparency, Sustainability, Supply Chain Heterogeneity, Supply Chain
Interconnectedness
3
1 INTRODUCTION
Focal firms are struggling to increase the transparency of their extended supply chains, in spite
of increasing pressure from investors, non-governmental organizations and regulators (Sodhi
and Tang, 2019). Research recognizes that supply chains are not designed to be transparent
(Parmigiani et al., 2011; Bateman and Bonanni, 2019) and has started to point to the role of
supply chain structure in influencing the extent to which supply chain members disclose
information about their internal practices and performance.
Kim and Davis (2016) used secondary data from the U.S. Securities and Exchange
Commission to investigate the determinants of focal firms’ ability to certify their supply chains
as ‘conflict minerals free’, and found that a smaller number of supply chain members and a
narrower geographical scope significantly increased the chances of success. Qualitative studies
indicate that geographical proximity and denser ties between supply chain members create
normative dynamics that shape how these firms collectively approach the improvement of their
extended operations and report their efforts (Dahlman and Roehrich, 2019; Fontana and
Egels‑Zandén, 2019). In summary, emerging evidence suggests that supply chain transparency
may strongly associate with some structural characteristics of the supply chain.
However, while research on supply chain transparency has made good progress, no large-
scale study provides robust empirical evidence concerning the association between supply
chain structure and supply chain transparency. This is primarily due to three reasons. First,
collecting accurate data about tens of thousands of contractual ties to map the structure of
extended supply chains remains a difficult, time consuming task (Shao et al., 2018). Second,
focal firms’ extended supply chains partially overlap as customers and suppliers belong to
multiple chains (Yan et al., 2015), which introduces methodological challenges in terms of data
aggregation and analysis. Finally, supply chain level phenomena are difficult to deductively
theorize because many organizational theories such as information processing and transaction
4
cost economics were not developed to explain the structuring and functioning of overlapping
networks of organizations that cut across industrial and geographical boundaries.
Our large-scale, descriptive study begins to overcome these challenges to answer one
important question: What is the association between supply chain structure and supply chain
transparency? To answer this question, we gather secondary data about 4803 firms and 20,504
contractual ties organized in 187 extended supply chains. We examine supply chain
transparency in terms of collective ESG disclosure, broadly defined as the aggregate level of
material environmental, social and governance (ESG) information that a focal firm’s customers,
suppliers and sub-suppliers (the extended supply chain) make available to the public. Then, in
line with emerging evidence and the broader supply chain literature (Choi and Krause, 2006)
1
,
we examine supply chain structure in terms of interconnectedness, which refers to the density
and clustering of ties connecting a focal firm’s extended supply chain members, and
heterogeneity, which refers to the degrees of geographical and industrial differentiation among
these members.
From a scholarly standpoint, our article provides two important empirical contributions.
First, instead of focusing on the transparency of focal firms (Marshall et al., 2016) or the
transparency of individual suppliers (Jira and Toffel, 2013; Villena and Dhanorkar, 2020), our
large-scale descriptive study empirically examines transparency at the extended supply chain
level of analysis. Second, we build upon and expand studies that have adopted qualitative
methods (Dahlman and Roehrich, 2019; Fontana and Egels‑Zandén, 2019) or coarse measures
of supply chain structure and transparency (Kim and Davis, 2016) to offer detailed insights into
the complex association between supply chain structure and supply chain transparency as a
collective outcome.
1
Choi and Krause (2006) identify three structural dimensions, namely numerosity, interconnectedness and heterogeneity.
This study focuses on the last two whereas the first i.e., numerosity is included as a control variable. We justify the selection
of independent variables in the hypotheses development section.
5
Our descriptive study is also relevant for practice because it enables focal firms to better
tailor strategies to facilitate supply chain transparency in light of their existing supply chain
structure. Supply chain structure tends to change only minimally in the short term (Osadchiy et
al., 2016) and is often not designed for transparency. Hence focal firms may pay little attention
to the structure of their extended supply chains, with negative implications for their ability to
achieve important collective outcomes. Our findings should motivate supply chain
professionals to invest more time and effort in understanding supply chain structure. We also
propose possible strategies that focal firms could consider to increase the transparency of their
supply chain.
2 LITERATURE REVIEW AND HYPOTHESES
Transparency is a combination of visibility, meaning that the focal firm possesses material
information about upstream and downstream operations, and the public disclosure of this
information (Sodhi and Tang, 2019; Chen et al., 2018; Swift et al., 2019). The scope of supply
chain transparency has expanded beyond the focal firm and its direct suppliers and customers
to include multiple supply chain tiers (Bateman and Bonanni, 2019). Similarly, the content of
transparency now incorporates a broad range of data such as: orders, forecasts and plans
(Akkermans et al., 2004); supply chain costs and customs duties (Sinha, 2000); and ESG
information (Sodhi and Tang, 2019), which reports the environmental impact of products and
services (New, 2010; Tapscott, 2003), carbon emissions (Villena and Dhanorkar, 2020), and
factories’ compliance with environmental and social regulations (Eccles and Klimenko, 2019).
Current studies acknowledge that an individual focal firm will struggle to provide supply
chain transparency alone (Sodhi and Tang, 2019), but do not empirically tackle the challenges
of transparency across the extended supply chain. Both research and practice indicate that focal
firms are largely failing to disclose ESG data for their supply chain because it is costly,
complicated, and time consuming (Chen et al., 2018; Sodhi and Tang, 2019; Kalkanci and
6
Plambeck, 2020), while controlling all tiers is practically impossible (Villena and Gioia, 2018).
We contribute to the literature by studying transparency in the collective ESG disclosure to the
public by the focal firm’s extended supply chain.
Our focus on ESG information as an empirical context is motivated by the mounting
interest of internal and external stakeholders in the environmental and social impacts of supply
chains (New, 2010; Tapscott, 2003) and the broad generalizability of ESG information to the
operational information that constitutes a source of competitive advantage for the focal firm.
The ability to access ESG information along the supply chain can improve operational
efficiency and prevent supply chain disruptions (Busse et al., 2017; Swift et al., 2019) as well
as enhance the supply chain’s reputation (Marshall et al., 2016). However, the availability of
ESG information might give regulators and activists reasons to question certain practices, have
negative repercussions for less transparent supply chain members (Villena and Dhanorkar,
2020), and allow knowledge spill over to competitors (Dubbink et al., 2008).
The majority of previous studies focused only on focal firms’ role in ESG disclosure, but
a few studies explore the role of suppliers (Jira and Toffel, 2013; Villena and Dhanorkar, 2020;
Kalkanci and Plambeck, 2020). For instance, Dahlmann and Roehrich (2019) study the role of
suppliers in sharing climate change information with the focal firm and other supply chain
members. They identify transactional and relational mechanisms that support focal firms’
efforts to stimulate suppliers to share information. Similarly, Jira and Toffel (2013) find that
suppliers are more likely to disclose information when their customers show a commitment to
using the data to make future decisions and or when more of their customers request such data.
Villena and Dhanorkar (2020) build on these studies to show the relevance of managerial
incentives and institutional pressures for the public disclosure of carbon-related information by
individual suppliers. However, to understand the transparency of the supply chain it is necessary
7
to both move from a focal firm to an extended supply chain perspective and to explore collective
disclosure, as opposed to the disclosure of individual suppliers.
The central tenet of this study is that supply chain structural dimensions might be
associated to supply chain transparency. We frame supply chain transparency as the result of
collective ESG disclosure by the members of a focal firm’s extended supply chain and focus
on the influence of supply chain structure in developing this collective outcome. The strategy
literature indicates that a focal firm’s ability to certify its supply chain as ‘conflict minerals
free’ is associated with the number of suppliers and the geographical scope of the supply chain
(Kim and Davis, 2016). Other studies in the business ethics literature indicate that geographical
proximity and dense ties between supply chain members shape how they collectively develop
a more transparent and sustainable supply chain (Fontana and Egels‑Zandén, 2019). These
studies either use coarse measures or are based on a few qualitative observations, but they offer
relevant guidance to begin to identify structural aspects and intervening mechanisms that can
help shed light on the link between supply chain structure and collective ESG disclosure.
We posit that a quantitative study focusing on structural measures will be useful to
provide a better understanding of the structure-collective outcome association. To date, the only
quantitative studies relating supply chain structure and supply chain transparency are in the
context of environmental disclosure and takes the focal firm’s perspective, without
investigating supply chain collective outcomes (Kim and Davis, 2016; Bellamy et al., 2020).
Research on supply chain transparency has made good progress, however no large-scale study
has investigated it from a collective perspective that considers the interrelations between
customers, suppliers and sub-suppliers that comprise a focal firm’s extended supply chain.
Therefore, our study contributes to the literature by empirically exploring the systematic
association between an extended supply chain’s structure and transparency.
8
ESG disclosure is an ideal context to investigate this association as it inherently assumes
a collective perspective: Each supply chain member possesses a non-substitutable piece of ESG
information that—once shared—can reduce information asymmetries across the supply chain,
and more broadly with stakeholders. Therefore, the ESG disclosure of a supply chain member
is valuable to the focal firm, other members and external stakeholders. Another reason to focus
on collective ESG disclosure is the possibility to observe and measure it, unlike other forms of
information sharing or collective outcomes (such as forecasts and inventory levels) that the
supply chain members may enact, but which do not have a public disclosure element.
2.1 Supply chain structural dimensions
The general logic behind our predictions is that certain supply chain structures enable collective
outcomes, like supply chain transparency, as they favor common actions among a focal firm’s
extended supply chain members. Our effort bridges complementary perspectives, ranging from
supply chain management, to strategy and business ethics, that establish a connection between
supply chain structure and collective outcomes. These perspectives propose a variety of
mechanisms that intervene in the supply chain structure-collective outcome association. For
example, the strategy and business ethics literatures investigate the presence of common goals
and interests, common norms and practices, and common know-how and capabilities as the
intervening mechanisms in the supply chain structure-collective outcome association (e.g.,
Vurro et al., 2009; Roberts, 2003). The supply chain management literature focuses mainly on
relational mechanisms pointing to collective outcomes in the presence of supply chain
structures that enable common norms, information sharing and collaboration (e.g., Dalhmann
and Roenhrich, 2019; Carnovale and Yeniyurt, 2015). We do not measure these intervening
mechanisms, both because they vary widely across previous work and are difficult to isolate in
large scale studies based on secondary data. However, we will refer to these mechanisms to
9
support our hypothesis development and acknowledge the associated empirical challenges as a
limitation of our study.
A supply chain can be measured and understood from two levels of analysis; that of a
single node (typically the focal firm) or the supply chain (Kim et al., 2011). Rather than
focusing on the attributes of single nodes (such as brokerage or centrality), our interest is in the
association between the structure of the supply chain and supply chain transparency, in the
guise of collective ESG disclosure. Our focus on supply chain level attributes builds on other
studies (e.g., Kim et al., 2011) that have shifted from focusing on individual nodes’ structural
properties to better understanding the overall pattern of ties and the composition of the focal
firm’s extended supply chain.
The supply chain management literature has investigated the effects of supply chain
structural dimensions on a variety of focal firm outcomes, including: frequency of disruptions
(e.g., Bode and Wagner, 2015; Brandon-Jones et al., 2015), new product creativity (Gao et al.,
2015), innovation performance (Sharma et al., 2019a; Bellamy et al., 2014), international
business performance (Sharma et al., 2019b) and financial performance (Lu and Shang, 2017).
In this literature the structural dimensions were operationalized as either the supply base,
including direct suppliers (e.g., Bellamy et al., 2014), the ego-network, including customers
and direct suppliers (e.g., Park et al., 2018; Lu and Shang, 2017), or the supply network,
including direct suppliers and sub-suppliers (e.g., Sharma et al., 2019; Kim et al., 2011).
However, limited attention has been paid to the extended supply chain of customers, suppliers
and sub-suppliers. Moreover, to the best of our knowledge, only Carnovale and Yeniyurt (2015)
relate these structural dimensions to supply chain collective outcomes; in their case innovations
generated in the joint ventures between the focal firm and the members of its ego-network.
Qualitative studies in the business ethics literature have explored the association between
the structure of extended ties between supply chain members and collective environmental and
10
social outcomes. Some of these studies show that highly interconnected supply chain members
develop a collective commitment to the adoption of social and environmental practices
(Fontana and Egels‑Zandén, 2019; Vurro et al., 2009). Other qualitative studies show that firms
within manufacturing clusters develop collective sustainable standards thanks to higher local
density. However, such collective standards rarely diffuse to the extended supply chain due to
cluster isolation (Battaglia, 2010; Puppim, 2014). Similarly, limited collective behavior is
found in the presence of geographical heterogeneity due to physical separation and different
institutional contexts between interconnected supply chain members (Knorringa and Nadvi,
2014; Fontana and Egels‑Zandén, 2019).
Based on these findings, we believe that to understand collective ESG disclosure in the
extended supply chain, it is relevant to assess the supply chain’s structure including all ties with
the focal firm’s customers (Lu and Shang, 2017; Villena and Dhanorkar, 2020) as well as with
suppliers and sub-suppliers (Roberts, 2003; Fontana and Egels‑Zandén, 2019). Therefore, we
next discuss how supply chain structure relates to collective ESG disclosure, considering
structural dimensions that encompass the extended supply chain.
Choi and Krause (2006) group these diverse dimensions into: numerosity, representing
the number and concentration of supply chain members; interconnectedness, referring to the
pattern of inter-organizational ties; and, heterogeneity, referring to the degree of differentiation
in attributes like home country and industry. Table 1 summarizes the most relevant studies in
the supply chain management literature through this tripartition.
Numerosity, unlike interconnectedness and heterogeneity, has been linked directly to
visibility and transparency. Most previous research on numerosity and transparency, considered
horizontal complexity, measured by the concentration of direct suppliers in a focal firm’s
supply base, and vertical complexity, measured by the average concentration of sub-suppliers
in the supply chain. This research generally concludes that numerosity negatively affects supply
11
chain visibility and transparency by hampering the focal firm’s ability to observe suppliers and
sub-suppliers (Kim and Davis, 2016). This negative association is reinforced in several studies
that relate numerosity to different supply chain-related outcomes, ranging from the focal firm’s
financial performance (Lu and Shang, 2017), to risk management (Bode and Wagner, 2015),
and transparency (Kim and Davis, 2016). Because numerosity is well understood in the context
of transparency, we control for it, but do not explicitly elaborate on its potential association
with collective ESG disclosure. Instead, our focus is the other two structural dimensions of
interconnectedness and heterogeneity.
Supply chain interconnectedness and collective ESG disclosure
Supply chain interconnectedness’ association with transparency is unknown, but
interconnectedness’ association to other supply chain outcomes has been previously explored.
We expect supply chain interconnectedness to be crucial for collective ESG disclosure, as
interconnected supply chains are usually positively related to collective outcomes. Supply chain
interconnectedness has been linked to achieving collective outcomes (other than transparency)
via mechanisms such as common norms (Vurro et al., 2009; Roberts, 2003) as well as
information sharing and collaboration between supply chain members (Gould, 1993; Ostrom
and Walker, 2003; Bellamy et al., 2014).
To investigate possible nuances in the association between supply chain
interconnectedness and collective ESG disclosure, we consider two sub-dimensions of
interconnectedness: supply chain density, which captures cohesiveness in terms of the
distributed ties between supply chain members (Kim et al., 2011), and supply chain clustering,
that is, the extent to which suppliers form loosely coupled sub-groups that are nested within the
extended supply chain (Pathak et al., 2014).
The most frequent conceptualization of interconnectedness is supply chain density.
Previous studies generally propose a positive relationship between supply chain density and
12
outcomes at the level of the focal firm (Bellamy et al., 2014) and supply chain (Carnovale and
Yeniyurt, 2015), while also indicating that high levels of supply chain density could negatively
affect such outcomes (Gualandris et al., 2015; Sharma et al., 2019a). Several disparate
observations suggest that dense connections between firms favor information sharing and create
a normative lock-in whereby good practices and outcomes are ensured through a concern for
reputation and reciprocation (Burt, 2002; Coleman, 1990). For example, dense ties have been
positively associated with the achievement of collective outcomes in communities of practice
and multi-party alliance networks (Fonti et al., 2017; Rivera et al., 2010; Wenger, 2000).
Similarly, supply chain density has been shown to be positively associated with collective
innovation (Carnovale and Yeniyurt, 2015) or risk mitigation in the supply chain (Bode and
Wagner, 2015). Studies focusing on these associations suggest that the mechanism linking
supply chain density to collective outcomes is that dense ties enable the creation and
enforcement of common norms as well as communication and information sharing across
supply chain members.
In the sustainability context, supply chain density is suggested to be a vehicle for
improving environmental and social outcomes in supply chains (Vurro et al., 2009; Roberts,
2003). These studies identify common sustainability norms, trust and information sharing
among supply chain members as the mechanisms that link structure to collective sustainability.
For example, Fontana and Egels-Zanden (2019) found that direct and indirect information
exchanges between suppliers enable the implementation of common labor practices and
improved working conditions. Similarly, Dalhmann and Roenhrich (2019) propose that supply
chain density enables trust which then favors information sharing regarding climate change
between supply chain members.
Overall, multiple streams of literature suggest that supply chain density positively
associates to collective outcomes via enabling mechanisms such as information sharing and
13
common norms and practices. We expect the same to be observed in relation to collective ESG
disclosure. Hence, our first hypothesis is:
H1. Supply chain density is positively associated with collective ESG disclosure.
More recent studies recognize the importance of moving beyond supply chain density,
which is the most basic measure of interconnectedness, to capturing other structural patterns
linking supply chain members. We focus on supply chain clustering, which describes the degree
to which members of the extended supply chain belong to loosely connected sub-groups (Pathak
et al., 2014). Recent research that explored the importance of supply chain complexity (Lu and
Shang, 2017) and supply chain fragmentation (Kim et al., 2015) for focal firm financial
performance and resilience suggests that clustering could influence collective outcomes.
However, this supposition is unexplored.
Supply chain clustering is a measure of the entire network’s structure and how sub-groups
of supply chain members engage with each other. Supply chain clustering is seen as
complementary to density because interconnectedness might not be homogeneous across the
supply chain and even in dense but clustered networks, collective outcomes may be hampered
due to limited interactions between clusters and more distant, disconnected members (Burt,
2002).
While not as well studied as supply chain density, the research that exists on supply chain
clustering suggests a negative association between clustering and collective outcomes.
Specifically, firms within a cluster can generate collective outcomes for the cluster. This occurs
because the members of a cluster are embedded in a local network of tight relationships which
creates collective outcomes for the cluster through mechanisms such as common norms and
information sharing (Schmitz and Nadvi, 1999; Nadvi, 2008). However, the presence of clusters
limits the achievement of collective outcomes at the extended supply chain level because
14
structural fragmentation inhibits information sharing and norms propagation between clusters
(Bimber et al., 2005; Fu and Shumate, 2016).
Supply chain clustering may hinder collective sustainability efforts and outcomes in a
supply chain. Qualitative studies in the business ethics literature show that in clustered supply
chains the collective sustainability outcomes tend to be poor (Battaglia et al., 2010; Knorringa
and Nadvi, 2014; Puppim et al., 2014). This is because while some clusters may achieve high
levels of sustainability, the lack of common norms and information sharing between clusters
acts as key barrier to the development of common sustainability practices throughout the
extended supply chain. Similarly, in a study about greenhouse gas emissions, Dooley et al.
(2018) find that in the presence of clusters, local outcome improvements will prevail rather than
collective outcomes. This is due to the fact that supply chain clustering favors autonomy for the
individual cluster and reduces the need for engagement and collaboration between clusters.
These findings suggest that supply chain clustering limits the sharing of information, and
common norms and practices across the supply chain, which then inhibits collective outcomes
such as ESG disclosure. Hence, our second hypothesis is:
H2. Supply chain clustering is negatively associated with collective ESG disclosure.
Supply chain heterogeneity and collective ESG disclosure
Several streams of literature suggest that in addition to the patterns of ties, the
heterogeneity of nodes in an extended supply chain might also relate to collective outcomes
(Heckathorn, 1993; Poteete and Ostrom, 2004). Previous literature on supply chain structure
investigates two forms of heterogeneity: geographical, which refers to supply chain members’
dispersion across different countries (Bode and Wagner, 2015; Kim and Davis, 2016); and
industrial, which relates to the spread of supply chain members across diverse industries
(Brandon-Jones et al., 2015; Gao et al., 2015). We explore both dimensions of supply chain
heterogeneity to capture its possible nuanced association to collective ESG disclosure.
15
The literature provides conflicting predictions on the link between supply chain
heterogeneity and collective outcomes. In general, different forms of heterogeneity have been
negatively associated to collective outcomes. Authors propose that mechanisms such as
different goals and interests, different know-how and capabilities (Poteete and Ostrom, 2004)
and limited information sharing and collaboration (Bode and Wagner, 2015; Sharma et al.,
2019b) explain the negative association between heterogeneous supply chain structures and
collective outcomes. In geographically dispersed supply chains different rules, regulations,
sociopolitical issues, infrastructural and cultural contexts between supply chain members from
diverse countries inhibit collective outcomes such as risk mitigation in the supply chain (Gereffi
et al., 2005; Bode and Wagner, 2015). Similarly, the literature suggests that supply chain
industrial heterogeneity hinders collective outcomes, due to the different goals, interests and
know-how of supply chain members operating in diverse industrial contexts (Brandon-Jones et
al., 2015; Gao et al., 2015). Because of these differences, previous studies found that common
norms and practices struggle to develop in the presence of supply chain geographical and
industrial heterogeneity. Instead, supply chain members tend to collaborate with and adopt
similar practices to their peers who are in the same industry and or country (Brandon-Jones et
al., 2015).
In the specific realm of environmental and social collective outcomes, several studies
suggest that geographical heterogeneity is negatively related with collective efforts to develop
sustainability along a supply chain due to different goals and interests (Knorringa and Nadvi,
2014) and different know-how across supply chain members (De Neve, 2014; Niforou, 2015).
These studies indicate that sustainability standards developed in western countries or in specific
geographic areas do not diffuse in geographically dispersed supply chains due to differences in
the way the supply chain members understand sustainability and integrate it in their operations.
Similarly, supply chain industrial homogeneity facilitates the establishment of collective
16
outcomes, such as the widespread adoption of common sustainability standards for an industry,
thanks to common goals and interests of members (Campbell et al., 2007). However, standards
developed for one industry rarely diffuse to other industries. Van den Brink and Van der Woerd
(2004) follow this logic to attribute different levels of public disclosure among industrially
heterogenous firms to different goals and interests as well as diverse ESG disclosure standards.
Overall, these arguments suggest that supply chain heterogeneity will be less conducive
to collective ESG disclosure because of different goals and interests and the poor transferability
of practices that were developed to fit specific local institutions and operational settings. Thus,
our third hypothesis is:
H3. Supply chain geographical and industrial heterogeneity are negatively associated with
collective ESG disclosure.
While much of the extant research suggests a negative association between supply chain
heterogeneity and collective outcomes, there are studies that suggest the same association could
be positive. For example, some studies identify different know-how and capabilities between
supply chain members as enabling collective outcomes, such as innovation (Gao et al., 2015),
new market formation (Lee et al., 2018) and executing other complex collective tasks (Rodan
and Galunic, 2004; Sammarra and Biggiero, 2008). In these studies, supply chain geographical
heterogeneity stimulates information sharing as a form of reciprocity to strangers and distant
supply chain members who are not viewed as direct competitors (Constant et al., 1996; Marques
et al., 2019). Similarly, supply chain industrial heterogeneity generates collective outcomes by
favoring collaboration, as supply chain members provide complementary know-how to the
innovation process (Enkel and Gassmann, 2010; Leven et al., 2014). Therefore, it is possible
that association between supply chain heterogeneity and collective ESG disclosure is positive,
not negative. Thus, the competing formulation of our third hypothesis is:
H3-competing. Supply chain geographical and industrial heterogeneity are positively
associated with collective ESG disclosure.
17
Table 1. Summary of supply chain management studies on supply chain structure
Author-Year
Supply chain structural dimensions
Observational unit
Dependent Variable
Numerosity
Interconnectedness
Heterogeneity
Bellamy et al. (2020)
-
Supply chain density
-
Supply Base (tier 1)
Focal firm environmental
disclosure
Sharma et al. (2019a)
Horizontal and vertical
complexity
PageRank centrality
Spatial complexity
Supply network
(tier 1 and 2)
Focal firm innovation
performance
Sharma et al. (2019b)
-
Supply chain density;
betweenness centrality; average
clustering coefficient; PageRank
centrality; average path length
-
Supply network
(tier 1 and 2)
Focal firm international
business performance
Park et al. (2018)
-
Focal firm accessibility;
interconnectedness
-
Alliance network
(tier 1)
Focal firm accessibility; Ego-
network interconnectedness
Basole et al. (2017)
-
Structural prominence; supply
chain density
-
Alliance network
(tier 1 and 2)
Focal firm economic
performance
Serpa and Krishnan
(2018)
Customer base concentration
-
Customer-Supplier
geographical distance;
Customer-Supplier Operational
similarity
Supply chain dyads
Supplier productivity
Lu and Shang (2017)
Horizontal and vertical
complexity
Eliminative and cooperative
complexity
Spatial complexity
Ego-network
(tier 1)
Focal firm financial
performance
Kim and Davis
(2016)
Number and concentration of
suppliers
-
-
Supply base
(tier 1)
Focal firm conflict mineral
declaration
Bode and Wagner
(2015)
Horizontal and vertical
complexity
-
Spatial complexity
Focal firm
Focal firm frequency of
disruptions
Gao et al. (2015)
-
Relational strength; supply chain
density
Technological diversity
Focal firm
Focal firm new product
creativity
Dong et al. (2015)
-
Relational embeddedness;
centrality
-
Focal firm
Focal firm opportunism
Carnovale and
Yeniyurt (2015)
-
Betweenness centrality; supply
chain density; brokerage;
weakness
-
Alliance network
(tier 1)
Ego-network innovation
Brandon-Jones et al.
(2015)
Scale complexity
-
Differentiation and
delivery complexity;
supply chain geographical
dispersion
Focal firm
Focal firm frequency of
disruptions
Bellamy et al. (2014)
-
Accessibility; interconnectedness
-
Supply base (tier 1)
Focal firm innovation
performance
Kim et al. (2011)
Number of suppliers
Supply chain density; supply
chain centralization
-
Supply network
(tier 1 and 2)
-
This study
Control variables
Supply chain density; Supply
chain clustering
Supply chain geographical and
industrial heterogeneity
Extended supply chain
(customers, tier 1 and 2)
Collective ESG disclosure
18
3 METHODOLOGY
3.1 Data collection and cleaning process
We test our hypotheses using secondary data from the Bloomberg SPLC database and the
Bloomberg ESG database. Bloomberg SPLC gathers data about contractual ties from the U.S.
Securities and Exchange Commission, news articles, trade publications, firm websites, and
private communications.
2
We collected SPLC data in May 2016 but supply chain ties were
reported within the 12-month window before our data collection, allowing us to capture supply
chain structure in 2015. We conducted a second data collection in December 2016 and a third
in May 2017 to assess the stability of supply chain structure over time. Similar to prior research
(Osadchiy et al., 2016), we find that structure changes only minimally in the short term.
The Bloomberg ESG database has been extensively used in operations management and
accounting research (e.g., Bellamy et al., 2020). It gathers data about firms’ internal ESG
practices and performance through direct communications, including meetings, phone
interviews and surveys with firms, but also from corporate sustainability reports, regulatory
filings, web-pages and news articles. The time stamp of our ESG data is December 31, 2016,
but firms publicly reported their 2016 ESG information in 2017.
The flowchart in Figure 1 summarizes our data collection and cleaning process. To build
measures at the extended supply chain level, we started our data collection process by
identifying the focal firms of interest. The 2015 Forbes 2000 list constituted an optimal
sampling frame as it offered a comprehensive annual ranking of the world’s largest focal firms
that, because of their size and success, were equally visible to external stakeholder groups. In
particular, we selected focal firms that operated in 10 four-digit Global Industry Classification
Standard (GICS) groups, all under the general category of manufacturing. Constraining our
2
To cross-validate our customer/supplier lists from Bloomberg SPLC we used Compustat’s segment database and the Thomson
Reuters value chains database. A comparison between data from Bloomberg, Compustat, and Thomson Reuters suggests that
the former is indeed more comprehensive than the latter two sources. This comparison is available from the authors upon
request.
19
analysis to manufacturing industries diminished the potential for unobserved heterogeneity
among cases, for example, related to less supply chain focus (Swift et al., 2019), customer-
supplier duality (Sampson and Froehle, 2006) and other unobserved features characterizing a
service setting.
Figure 1. Data collection and cleaning process
The major challenge with Bloomberg SPLC data is that there is no provision to collect
the data in bulk, rendering large-scale empirical analysis difficult. Similar to recent studies
(Sharma et al., 2019a), we tried to overcome this challenge by randomly sampling a limited
number of focal firms, using multiple Bloomberg terminals and hiring a team of research
assistants. A power analysis for the F-test with a medium effect size (0.15) for ten controls and
four predictors and a desired significance level of 0.01 suggested that a target sample size of
150 cases (supply chains) would guarantee strong statistical power for our study (>90%). Since
we assumed only 55-60% of the sampled focal firms could ultimately present satisfactory
Step 1
Identify focal firms in the 2015 Forbes2000 list.
Step 2
Apply a proportionate stratified random sampling with industry-country strata.
Step 3
(a) Identify customers, suppliers and sub-suppliers from the Bloomberg SPLC database;
(b) Search and match redundant tickers (e.g., ABB Ltd appears as ‘ABB SS Equity’ on the
Stockholm Stock Exchange and as ‘ABBN VX Equity’ on Swiss Stock Exchange).
(c) Gather ESG data and demographics from the Bloomberg ESG database and the
Bloomberg Financials database;
Step 4
(a) Remove 35 focal firms that, based on refined descriptors like ‘NACE’ and ‘CIE_DES’ in
the Bloomberg Financials database, can be classified as retailers or distributors.
(b) Remove supply chain members that are not involved in COGS ties, so to approximate
‘physical’, stable supply chains (Wang et al., 2018);
Output:
- 280 focal firms;
- 5852 supply chain members;
- 27973 contractual ties.
Output:
Representative list of 280 focal firms.
Output:
List of 617 focal manufacturing firms.
Output:
- 245 focal firms;
- 4803 supply chain members;
- 20504 contractual ties.
Step 5
(a) Isolate 245 extended supply chains from a large network by following the two-step
procedure illustrated in Supplement A;
(b) Remove outliers that could distort empirical results:
(a) 55 supply chains each containing less than 5 direct suppliers;
(b) 1 supply chain that contained only unquantified contractual ties i.e., ties with
unknown transaction amount.
(c) 2 supply chains with standardized values of dependent variables < - 4 or > 4.
Output:
- Final sample of 187 extended
supply chains.
20
supply chain data (cf. Bellamy et al., 2014, p. 362), we decided to collect data for a larger
sample of 280 focal firms. We followed a proportionate stratified random sampling approach
that achieves high levels of generalizability by dividing the original sampling frame of 617
focal manufacturing firms into strata (industry-country pairs) and selecting a proportionate
random selection of focal firms from each stratum (Rice, 2006, p. 247).
For each sampled focal firm, we collected customer/supplier lists and other tie-level data,
including a percentage of the supplier’s revenue from each customer, a percentage of the
customer’s spend on each supplier, and whether a tie involved the supply of goods (COGS),
capital expenditures, research and development, or administrative services (see Bloomberg
2011, 2013 for details). Measuring our dependent and independent variables involves
customer/supplier information not only for the focal firm, but also for its suppliers. Hence, we
repeated the data collection process for each of the direct suppliers of the sampled focal firms.
Once we compiled the full list of supply chain members, we had a dataset of 4,803 firms and
20,504 contractual ties.
We then collected ESG disclosure scores for all 4,803 firms from the Bloomberg ESG
database, which identifies firms based on the same tickers. The Bloomberg ESG disclosure
score ranges from 0 for firms that do not disclose any ESG information to the public to 100 for
those that disclose every indicator examined by Bloomberg.
Lastly, we applied an innovative two-step procedure to isolate extended supply chains
like the one presented in Figure 2 from our large network of intertwined supply chains. Details
about the mechanics of this mapping procedure are provided in Supplement A (available
online).
21
Figure 2. Simplified illustration of the extended supply chain of ABB Ltd
Table 2 presents the demographics of our final sample. The one-way analysis of variance
(ANOVA) suggests that data collection and cleaning did not lead to bias as the excluded cases
(supply chains) do not differ significantly from our final sample in terms of supply chain
transparency (p-values > 0.15). In other words, our 187 supply chains can be considered
representative of the 617 manufacturing supply chains found in the Forbes 2000 list.
In our final sample, the average extended supply chain consists of 16 customers (min 0,
max 55), 22 direct suppliers (min 5, max 150), 251 sub-suppliers (min 13, max 898) and 1,522
contractual ties (min 27, max 7,942). Our descriptive analysis in Supplement B (available
online) indicates that ESG disclosure scores vary significantly across supply chain tiers, with
sub-suppliers being the least transparent. ESG disclosure scores vary significantly even within
supply chain tiers, with firms managing a richer portfolio of buying and selling ties generally
disclosing more ESG information to the public relative to their peers.
Bombardier
(Canada)
Siemens AG
(Germany) AVNET
(US)
Baoshan
Iron & Steel
(China)
SAP SE
(Germany)
Skanska AB
(Sweden)
ABB
(Switzerland)
TATA Steel
(India)
Infineon Tech
(Germany)
Advanced
Semiconductor Eng.
(Taiwan)
IBM
(US)
Synaptics Inc
(US)
Kaga Electronics
(Japan)
Intel Corp
(US)
…16 more
LG
Chemical
(South Korea)
DOW
(US)
SSE PLC
(Britain)
RIO
Tinto
(Britain)
…13 more …18 more …8 more
…16 more
…19 more
…292 more
SKF AB
(Sweden)
…18 more
Airbus SE
(France)
…15 more
Legend:
Examples of excluded firms;
Examples of excluded contractual ties;
Supply chain member;
Selling contractual tie;
Reciprocated contractual tie involving buying and selling contracts.
22
Table 2. Sample demographics
3.2 Dependent variable
The Bloomberg ESG disclosure score measures a firm’s transparency in regards to a
number of ESG practices and outcomes such as employee training costs, employee turnover,
injury rates, number of environmental spills, greenhouse gas emissions, and recycling. The
Bloomberg ESG disclosure score is tailored to different industry groups so that each firm is
only evaluated in terms of the data that are material to its industry (Bloomberg ESG, 2018).
Thanks to this robust industry standardization and to the diversity of its data sources (both firm-
produced material and direct communications), the Bloomberg disclosure score alleviates
methodological concerns due to a lack of comprehensiveness, selectivity, and greenwashing,
and therefore captures the amount of material ESG information a firm discloses to the public
(Bellamy et al., 2020).
Population:
Focal firms Focal firms Supply chain members
Capital goods - 2010 68 37 674
Food, Beverage and Tobacco - 3020 86 31 252
Technology Hardware and Equipment - 4520 66 25 504
Automobiles and components - 2510 30 24 280
Materials - 1510 196 22 620
Pharma & Biotech - 3520 44 19 209
Consumer Durables & Apparel - 2520 20 13 175
Health Care Equipment - 3510 46 9 131
Household and Personal Products - 3030 39 6 37
Seminconductors - 4530 22 1 254
Others (67 CIGS Industry groups) 0 0 1667
Total 617 187 4803
United States 185 66 1091
Japan 99 46 914
Taiwan 21 11 413
United Kingdom 19 9 123
China/Hong Kong 64 8 762
France 19 8 85
Germany 25 8 94
South Korea 15 8 492
Switzerland 16 6 43
Sweden 12 4 51
Brazil 11 2 36
Finland 5 2 22
Mexico 7 2 31
Others (64 countries) 119 7 646
Total 617 187 4803
Final Sample:
23
We measure supply chain transparency as collective ESG disclosure (Disclosure_ESG),
a variable that captures the aggregate level of material ESG information that a focal firm’s
supply chain members make available to the public. To compute this variable, we calculate the
percentage of supply chain members i of each supply chain j that have a Bloomberg ESG
disclosure score greater than or equal to 26; the median of the Bloomberg ESG disclosure in
our large network after removing those supply chain members that do not disclose any ESG
information. This operationalization is adapted from studies that examined the diffusion of
TQM practices within networks of public hospitals (Young et al., 2001).
For robustness, we considered two alterative operationalizations of collective ESG
disclosure; Bloomberg’s environmental disclosure score (Disclosure_Env) and a signal-to-
noise variable (Disclosure_Signal), calculated as the ratio between average ESG disclosure
among supply chain members i of supply chain j (signal) and the ESG disclosure standard
deviation in the same set (noise). In our research, it is reasonable to combine these arithmetic
values in a signal-to-noise ratio because high supply chain transparency is observed when the
average disclosure per supply chain member i is high and disclosure standard deviation among
them is low. ESG disclosure scores from different supply chain members are non-substitutable
as each refers to their internal practices and performance. For this reason, a given extended
supply chain j can be considered highly transparent only when most supply chain members
converge around a large average ESG disclosure score. Arithmetic values of average and
standard deviation were used in prior operations management research that examined collective
innovation performance across ego-networks (Carnovale and Yeniyurt, 2015).
3.3 Independent variables
Our independent variables are computed by considering the extended supply chain members N
operating in each focal firm’s supply chain j. First, supply chain density (SCDensity) captures
interconnectedness as cohesion (Kim et al., 2011), which is calculated as the number of ties
24
among supply chain members E expressed as a function of the number of pairs among these
supply chain members (N(N−1)) (Kolaczyk and Csárdi, 2014, p. 55):
SCDensity = E/(N(N−1))
E counts the number of edges between supply chain members in the set N of supply chain
j. The adopted density score treats each extended supply chain j as a binary directed network
where reciprocated contractual ties rightfully count as two separate edges. SCDensity ranges
from 0 to 1, where 0 indicates that none of a focal firm’s supply chain members share ties with
each other, and 1 indicates that every supply chain member is linked to every other supply chain
member.
Second, supply chain clustering (SCClustering) captures the extent to which supply chain
members are divided into loosely coupled communities (Schilling and Phelps, 2007; Pathak et
al., 2014). To measure this variable, we adopt a recent generalization of the clustering score for
binary directed networks (Fagiolo, 2007; Kolaczyk and Csárdi, 2014, p. 82):
SCClustering=
, with
where N is the number of supply chain members i in a given supply chain j and is the
clustering score of each member i, calculated as the ratio between all directed closed triads
actually formed by i and the number of all possible closed triads (triples) that i could possibly
form in its supply chain j. A closed triad represents an instance in which the supply chain
member i is connected to two other members that are also connected to one another whereas a
triple represents an instance in which the supply chain member i connects any two other supply
chain members that need not be connected. As illustrated in Figure 3, in undirected networks
the three members can form only one sort of closed triad, whereas in directed networks, there
are eight distinct closed triad types, including, for example, feedback loops (FB) and feed-
forward loops (FF) (Ahnert and Fink, 2008). SCClustering considers all these potential closed-
25
triads and, ultimately, can range from 0 to 1, with larger values indicating higher supply chain
clustering.
Figure 3. Closed triads in binary directed networks
Third, supply chain geographical heterogeneity (SCGeographicalH) captures the extent
of geographical differentiation between supply chain members in the set N of extended supply
chain j (Choi and Krause, 2006) and is measured by categorical heterogeneity as follows
(Borgatti and Li, 2009):
SCGeographicalH = 1 −
where is the proportion of supply chain members i in supply chain j that fall within country
k. Our measure will vary from 0 if all supply chain members are from the same country to
(1−1/k) if they are distributed evenly across k countries. Note that supply chain members in our
sample are located in 77 countries (see Table 2).
Finally, supply chain industrial heterogeneity (SCIndustrialH) captures the extent to
which supply chain members in the set N of extended supply chain j are evenly distributed
across a large number of industries (Choi and Krause, 2006). Similar to supply chain
geographical heterogeneity, industrial heterogeneity is also measured by categorical
heterogeneity (i.e., 1 −
) where k identifies diverse GICS industry groups.
i
(Feed-Back triads)
(Feed-Forward triads)
ii
ii
ii
i i
i
Legend:
Supply chain member i
in supply chain j.
Other supply chain member
e.g., customer, focal firm
supplier or sub-supplier.
Directed tie.
Undirected tie.
One type of closed triad
in undirected networks.
Eight possible types of closed triads
in directed networks.
26
3.4 Control variables
We consider two sets of control variables to minimize the probability of obtaining spurious
estimates in our focal effects.
The first set of five controls addresses supply chain structural dimensions beyond supply
chain interconnectedness and heterogeneity that could be related to supply chain transparency.
Two controls are measured at the focal firm level of analysis whereas three controls capture
supply chain-level effects.
First, the literature indicates that dynamic industries are associated with fewer visible
sustainability issues and attract less disclosure activism (Reverte, 2009). Accordingly, we
control for clock speed (IndustryClock) by considering a focal firm’s industry group and then
dividing our sample into dynamic and static supply chains based on Fine’s (1998)
classification.
Second, focal firms from Asia may present higher levels of collectivism in their supply
chains, which is expected to correlate with both denser supply chain structures and higher levels
of supply chain transparency. Hence, we control for collectivism by considering the home
country of the focal firm (America = Low; Europe = Medium; Asia = High) (Oyserman, 2011).
Because increments in collectivism may not be equally sized across continents, we include
dummies for America and Europe in our regression models and hold out Asia as the reference
category. In our sample, 37.43% of focal firms are from America, 21.93% from Europe and
40.64% from Asia.
Third, because the literature shows that larger firms tend to be more transparent about
internal practices and performance (Darnall et al., 2010), we control for supply chain
organizational size (SCOrgSize), which is calculated as the average number of employees per
supply chain member i in the set N of each extended supply chain j.
Finally, the literature suggests that collective outcomes are more difficult to achieve in
27
large communities (Ostrom, 1990, 2003). Thus, we control for supply chain horizontal
complexity (SCHorizontal) and supply chain vertical complexity (SCVertical). The former
subtracts a focal firm’s percentage spend on its five top suppliers from 1 (Kim and Davis,
2016). The larger the percentage of spend on its five main suppliers, the smaller the horizontal
complexity of supply chain j. For vertical complexity, we average the values of horizontal
complexity calculated for direct suppliers in supply chain j (Lu and Shang, 2017).
A second set of four controls addresses multiple sources of pressure that could affect both
the structuring and the transparency of the supply chain. Three controls are measured at the
focal firm level of analysis whereas one control captures important supply chain-level effects.
First, the literature suggests that central actors often drive the construction of common
practices and the achievement of collective outcomes through establishing specific group
structures and governance mechanisms (Gould, 1993). Therefore, we control for the Bloomberg
ESG disclosure score of the focal firm (FFDisclosure).
Second, as the closeness to final markets might shape incentives and practices (Bellamy
et al., 2020), we control for End-Market Distance, a categorical variable based on the industry
of the focal firm (1 = Consumer goods; 2 = Technology; 3 = Industrials; 4 = Materials). In our
sample, 40.64% of focal firms operate in the Consumer Goods sector, 27.81% in the
Technology sector, 19.79% in the Industrials sector and the remaining 11.76% operate in the
Materials sector.
Third, pressure from external stakeholders may drive disclosure efforts, and such efforts
are likely to be higher in industries that produce salient externalities on society and the natural
environment (Darnall et al., 2010). Thus, we control for industry-level reputational risk
(IndustryRepRisk) through the index provided by RepRisk (2016), the only provider that
systematically analyzes adverse business conduct data reported by media, activists, and
consumer associations (Kölbel et al., 2017). Higher scores of this industry index indicate more
28
frequent and severe incidents involving focal firms and its supply chain members. We employ
the 2016 reputation risk index score of the industry group of the focal firm.
Finally, the literature demonstrates that firms are more predisposed to disclose ESG
information in countries with higher regulatory quality (Jira and Toffel, 2013). The variable
supply chain regulatory pressure (SCRegPressure) controls for the average quality of national
regulations per supply chain member i in supply chain j. This variable is based on the average
of six indicators – control of corruption, political stability, government effectiveness, rule of
law, regulatory quality, voice and accountability – as provided by the World Bank (2016).
Table 3 summarizes the operationalization of our variables and Table 4 presents
descriptive statistics and the correlation matrix.
29
Table 3. Variables summary
`Variable Variable long name So urce Ca lculatio n References
(1a) Disclosure_ES G Collective ESG disclosure Bloomb erg ESG
% of supply chain m embers i of sup ply chain j that have an ESG disclosure score = or > than 26.
(1b) Disclosure_Env Collective Environmental disclosure Bloomb erg ESG
% of supply chain m embers i of supply ch ain j that have an En vironmental disclosure score = or > than 26.
(1c) Disclosure_Sig nal Signal to Noise ratio Bloomb erg ESG
Ratio between the average (signal) and the standard deviation (noise) of ESG d isclosure scores of sup ply chain members i
within supply chain j
(2) SCDensity Supply chain density Bloomberg SPLC
# of supply chain ties E expressed as a fu nction of the # o f pairs of supp ly chain p artners (N(N-1)). Kim et al. (2011)
(3) SCClustering Supply chain clusterin g Bloomb erg SPLC
Averag e clustering score p er supply chain member i of a given sup ply chain j . While accounting f or contractual
directionality, the clustering score considers the n umber o f closed triad s formed by supp ly chain m ember i ex pressed as a
function of the possible n umber o f closed triad s (triplets) it cou ld form.
Adapted from Fagiolo (2007) as per
Kolaczyk and Csárdi (2014, p. 82)
(4) SCGeog raph icalH Sup ply chain geograp hical heterogeneity Bloomb erg SPLC
Categorical heterogeneity score where each catego ry k represents a different country. Borgatti and Xun Li (2009)
(5) SCInd ustrialH Sup ply chain industrial h eterogeneity Bloomb erg SPLC
Categorical heterogeneity score where each catego ry k represents a different indu stry g roup. Borgatti and Xun Li (2009)
(6) SCHorizontal Supply chain ho rizontal com plexity Bloomberg SPLC 1 - (% of a focal firm ’s cost of go ods sold that is spent on its top five supp liers).
(7) SCVertical Supply chain ver tical com plexity Bloomb erg SPLC
Averag e horizontal complexity per supply chain m ember i in the sub-set Ha of a given supply chain j.
(8) SCOrgSize Sup ply chain organizational size Bloomb erg Financials
Averag e number of employ ees per supp ly chain m ember i of a given supp ly chain j.Adapted from Darnall et al. (2010).
(9) IndustryClock Industry clockspeed Bloomb erg Financials Dummy indicating if the industry of the focal firm is d ynamic or static. Fine, (1998).
(10) Collectivism America Collectivism for American supply chains Bloomb erg Financials Dum my indicating if the country of the focal firm is lo cated in America. Adapted from Oyserman (2011)
(11) Collectivism EU Collectivism for EU sup ply chains Bloomb erg Financials Dummy indicating if the cou ntry of the focal firm is located in Euro pe. Adapted from Oyserman (2011)
(12) FFDisclosure Focal firm disclo sure Bloomber g ESG Bloom berg disclosure score of the focal firm . Lai et al., (2016).
(13) End-Market Distan ce Supply ch ain end-m arket distance Bloomb erg Financials
Categorical variable b ased on the ind ustry of the fo cal firm (1 : Consumer Good s; 2: Tech; 3 : Ind ustrials; 4: Materials). Bellamy et al. (2020)
(14) IndustryRepRisk Industry repu tational risk RepRisk Reputational risk index o f the industry of the focal firm. Adapted from Dai et al., (2018)
(15) SCRegP ressure Supply chain regulatory pressure World Bank
Averag e national r egulatory quality per sup ply chain mem ber i in a given supply chain j.De Villiers and Marques, (2016)
a While each supp ly chain j contains a set N of supply chain m embers i, the sub-set G co nsiders upstream members g on ly i.e., suppliers and sub-suppliers (G<N). Then , the sub-set H inclu des direct sup pliers only (H<G<N).
Adapted from Kim and Davis (2016)
and from Lu and Shang (2017).
Novel - not empirically tested. Adapted
from Young et al. (2001) and from
Carnovale and Yeniyurt (2015)
30
Table 4. Summary statistics and correlation matrix (N = 187)
Variable (1a) (1b) (1c) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15)
(1a) Disclosure_ESG 1.000
(1b) Disclosure_Env 0.916 1.000
Sign ificance level 0.0 00
(1c) Disclosure_Signal 0.720 0.725 1.000
Sign ificance level 0.0 00 0 .000
(2) SCDensity 0.284 0.351 0.253 1.000
Sign ificance level 0.0 00 0 .000 0.00 1
(3) SCClustering -0.103 -0.023 -0.173 0.524 1.000
Sign ificance level 0.1 62 0 .757 0.01 8 0.000
(4) SCGeographicalH 0.263 0.220 0.327 -0.381 -0.344 1.000
Sign ificance level 0.0 00 0 .003 0.00 0 0.000 0 .000
(5) SCIndustrialH -0.126 -0.232 -0.242 -0.492 -0.223 0.049 1.000
Sign ificance level 0.0 85 0 .001 0.00 1 0.000 0 .002 0.50 9
(6) SCHorizontal 0.014 -0.007 -0.059 0.283 0.164 -0.293 0.167 1.000
Sign ificance level 0.8 52 0 .922 0.42 0 0.000 0 .025 0.00 0 0.022
(7) SCVertical -0.024 -0.106 -0.065 0.079 -0.096 -0.229 0.285 0.347 1.000
Sign ificance level 0.7 46 0 .151 0.38 1 0.283 0 .192 0.00 2 0.000 0 .000
(8) SCOrgSize 0.460 0.410 0.178 0.424 0.065 -0.105 -0.076 0.231 0.164 1.000
Sign ificance level 0.0 00 0 .000 0.01 5 0.000 0 .379 0.15 3 0.304 0 .002 0.02 5
(9) IndustryClock -0.048 -0.001 -0.030 0.126 0.006 0.152 -0.287 -0.102 -0.215 0.076 1.000
Sign ificance level 0.5 12 0 .991 0.68 1 0.087 0 .939 0.03 8 0.000 0 .163 0.00 3 0.304
(10) Collectivism Am erica -0.171 -0.254 -0.427 0.023 0.093 -0.353 0.069 0.015 0.028 0.188 -0.010 1.000
Sign ificance level 0.0 20 0 .001 0.00 0 0.760 0 .206 0.00 0 0.349 0 .837 0.70 0 0.010 0 .888
(11) Collectivism EU -0.021 -0.033 -0.080 -0.087 0.037 0.108 0.199 0.177 0.289 0.171 -0.110 -0.410 1.000
Sign ificance level 0.7 74 0 .657 0.27 8 0.234 0 .618 0.14 0 0.006 0 .016 0.00 0 0.019 0 .134 0.00 0
(12) FFDisclosure 0.039 0.023 0.069 -0.319 -0.107 0.188 0.191 -0.034 -0.040 -0.076 -0.268 -0.223 0.354 1.000
Sign ificance level 0.5 94 0 .752 0.34 6 0.000 0 .147 0.01 0 0.009 0 .641 0.58 4 0.304 0 .000 0.00 2 0.000
(13) End- Market Distance 0.209 0.206 0.147 -0.085 -0.052 0.153 0.098 0.031 0.011 -0.082 -0.449 -0.041 -0.039 0.103 1.000
Sign ificance level 0.0 04 0 .005 0.04 4 0.247 0 .477 0.03 7 0.183 0 .673 0.88 0 0.268 0 .000 0.57 5 0.600 0 .160
(14) IndustryRepRisk 0.139 0.065 0.116 0.019 -0.143 -0.045 0.365 0.303 0.442 -0.009 -0.206 -0.020 0.068 -0.084 0.506 1.000
Sign ificance level 0.0 57 0 .379 0.11 6 0.796 0 .052 0.54 1 0.000 0 .000 0.00 0 0.907 0 .005 0.78 6 0.356 0 .254 0.00 0
(15) SCRegPressure 0.365 0.313 0.005 -0.017 0.077 -0.156 0.156 0.048 -0.031 0.139 -0.190 0.060 0.066 0.064 0.081 0.033 1.000
Sign ificance level 0.0 00 0 .000 0.94 8 0.815 0 .298 0.03 4 0.033 0 .519 0.67 8 0.057 0 .009 0.41 6 0.369 0 .386 0.26 9 0.651
Mean 33.17 26.04 1.85 0.02 0.15 0.74 0.92 89.10 92.68 37133.48 0.75 0.37 0.22 35.96 2.03 39.03 1.16
SD 4.88 4.47 0.14 0.01 0.04 0.07 0.03 14.04 4.93 9424.55 0.43 0.49 0.41 20.69 1.04 5.67 0.11
Min 14.13 13.04 1.57 0.01 0.07 0.33 0.84 1.65 73.74 11506.82 0.00 0.00 0.00 0.00 1.00 22.00 0.19
Max 51.65 42.98 2.46 0.07 0.32 0.88 0.96 99.54 99.68 79999.76 1.00 1.00 1.00 72.31 4.00 45.00 1.36
31
4 EMPIRICAL ANALYSIS
Section 4.1 discusses how our study addresses diverse sources of endogeneity. Section 4.2
presents our main analysis and results. Section 4.3 examines concerns related to multiple
memberships across supply chains. In addition, the online supplements provide detail on a
number of robustness checks. Supplement C1 considers a different unit of analysis (the
upstream supply chain). Supplement C2 evaluates alternative thresholds for calculating
collective ESG disclosure. Supplement C3 considers an alternative operationalization of supply
chain heterogeneity. Supplement D assesses the extent to which variation in the degree of
supply chain data coverage across supply chains constitutes a source of systematic bias.
Supplement E presents additional analysis concerning the comprehensiveness of the Bloomberg
ESG database.
4.1 Potential endogeneity concerns
Although the cross-sectional nature of our exploratory research does not support causal
inference, the employed methods and the selected empirical context mitigate several potential
sources of endogeneity (Lu et al., 2018). First, reverse causality seems unlikely because changes
in the transparency of multiple supply chain members would hardly trigger significant structural
changes at the supply chain level of analysis in the short term. Yearly variations in supply chain
structure have been reported as minimal (Osadchiy et al., 2016), and we validated this finding
through a second data collection in December 2016 and a third in May 2017. Second,
simultaneity should not be a strong concern, since our independent and dependent variables are
lagged by more than a year i.e., we captured supply chain structure in 2015 whereas ESG
disclosure data were reported in 2017. Third, we construct ten control variables that associate
with both our dependent and independent variables (Table 4), thereby minimizing endogeneity
issues that may arise through omitted variables bias. Fourth, although the selection of supply
chain members within and across extended supply chains may be endogenous, supply chains
32
have not been designed for transparency (Bateman and Bonanni, 2019) but rather for
operational responsiveness or efficiency (Parmigiani et al., 2011). Research also indicates that
sustainable procurement was largely under-developed at the time of our data collection (Golini
and Gualandris, 2018).
4.2 Regression results
We test our hypotheses using ordinary least squares (OLS) regressions with robust clustering
by industry (Abadie et al., 2017). Because the dependent and independent variables are all at
the supply chain-level of analysis, hierarchical linear models would not be appropriate.
After computing our dependent, independent and control variables, we checked for
symmetric distributions that could contribute to non-normal standard errors in our regressions
(Greene and Zhang, 2003). We applied a natural logarithmic transformation to
Disclosure_ESG, Disclosure_Env, Disclosure_Signal, SCDensity, SCClustering and
SCOrgSize to correct for positive skewness. A power transformation was applied to
SCGeographicalH, SCIndustrialH, SCHorizontal, SCVertical and FFDisclosure to correct for
negative skewness. Standardization enhanced comparability across variables that were
originally expressed in different scales and units. The Breusch-Pagan / Cook-Weisberg test
(Breusch and Pagan, 1979) and the White test for heteroskedasticity (White, 1980) do not reject
the null hypothesis of constant variance in our standard errors (p-values > 0.5 and 0.10,
respectively). The variance inflation factor for our regressors ranges between 1.15 and 3.41
(mean VIF 1.90), suggesting that spuriously high multicollinearity is not a concern
(Wooldridge, 2015).
Table 5 presents the results of our OLS regressions for Disclosure_ESG, including
robustness checks using Disclosure_Env and Disclosure_Signal. Models 1a and 1b in Table 5
indicate that the inclusion of supply chain structural dimensions in our regression model
significantly improves fit statistics like R-squared (+12.4%) and RMSE (-0.09).
33
Models 1b, 2 and 3 in Table 5 provide strong, consistent support for H1, indicating that,
in the context of ESG disclosure, supply chain density positively associates with supply chain
transparency. Models 1b, 2 and 3 provide weak but consistent support for H2, which indicates
a negative association between supply chain clustering and supply chain transparency. Then, in
line with H3-competing, Models 1b, 2 and 3 suggest that supply chain geographical
heterogeneity positively associates with supply chain transparency. Finally, our analysis in
Table 5 suggests that supply chain industrial heterogeneity does not significantly associate with
supply chain transparency.
Additional analysis in supplement C1 evaluates collective disclosure for the upstream
portion of a supply chain; only suppliers and sub-suppliers. Supplement C2 evaluates the
sensitivity of our main results to variations in the selected threshold for collective ESG
disclosure (i.e., 25th, 40th, 60th and 75th percentiles). Supplement C3 considers entropy-based
measures of heterogeneity. These robustness checks largely confirm the results in Table 5.
In conclusion, our large-scale descriptive study suggests that supply chain
interconnectedness plays an important but nuanced role, supply chain geographical
heterogeneity positively associates with supply chain transparency and supply chain industrial
heterogeneity is not influential.
Table 5. Regression results
(3) Disclosure_Signal
Coefficient Std Errors P-values Coefficient Std Errors P-values Coefficient Std Errors P -values Coefficient Std Errors P-values
SCDensity 0.474 0.119 0.003 0.497 0.113 0.001 0.483 0.221 0.054
SCClustering -0.200 0.120 0.126 -0.159 0.080 0.075 -0.254 0.114 0.050
SCGeographicalH 0.388 0.075 0.000 0.278 0.126 0.052 0.316 0.087 0.005
SCIndustrialH -0.061 0.045 0.201 -0.080 0.071 0.289 -0.142 0.072 0.076
SCHorizonta l -0.106 0.065 0.136 -0.055 0.045 0.250 -0.062 0.056 0.300 -0.069 0.070 0.350
SCVertical -0.023 0.126 0.858 0.002 0.081 0.983 -0.046 0.070 0.529 -0.025 0.086 0.775
SCOrgSize 0.589 0.070 0.000 0.401 0.079 0.000 0.367 0.080 0.001 0.226 0.142 0.142
Indu stryClock 0.062 0.340 0.859 -0.182 0.257 0.494 -0.054 0.256 0.838 -0.410 0.171 0.038
Collectivism America -0.819 0.149 0.000 -0.385 0.179 0.057 -0.619 0.220 0.018 -0.927 0.210 0.001
Collectivism EU -0.748 0.107 0.000 -0.550 0.125 0.001 -0.553 0.092 0.000 -0.823 0.184 0.001
FFDisclosure 0.083 0.058 0.184 0.105 0.051 0.066 0.097 0.051 0.088 0.142 0.029 0.001
End- Market Distance 0.140 0.074 0.088 0.048 0.068 0.501 0.127 0.106 0.258 -0.079 0.030 0.026
Indu stryRepRisk 0.129 0.082 0.149 0.134 0.049 0.021 0.056 0.081 0.506 0.206 0.071 0.016
SCRegP ressure 0.475 0.129 0.004 0.598 0.055 0.000 0.521 0.041 0.000 0.115 0.135 0.417
Constant 0.101 0.417 0.814 0.331 0.369 0.391 0.170 0.368 0.655 1.063 0.222 0.001
Observations 187 187 187 187
(pseud o) R-squar ed 0.509 0.630 0.599 0.561
(1a) Disclosure_ESG
(1b) Dislcosure_ESG
(2) Disclosure_Env
34
Although not directly related to our research questions, factors beyond supply chain
structure also emerge as important considerations. Above all, Collectivism plays a strong and
consistent role across our models, indicating that supply chains whose focal firms are Asian
achieve higher levels of supply chain transparency than their European and American counter-
parts. In addition, and contrary to what the literature would suggest (Schmidt et al., 2017), the
coefficient of End-Markets Distance is not significant. A closer examination of this result using
ANOVA finds that, in the context of ESG disclosure, supply chain transparency is significantly
lower for focal firms in the Technology sector relative to those operating in the Consumer
Goods sector (+2.62%), Industrial sector (+3.68%) and the Materials sector (+6.53%).
4.3 Concerns related to multiple memberships
In line with Carter et al. (2015), we draw the extended supply chain boundaries from the
perspective of the focal firm and compute our dependent, independent and control variables
accordingly. This approach could potentially bias our estimates as, according to Yan et al.,
(2015), supply chain members normally belong to multiple, diverse extended supply chains and
are influenced by focal firms and supply chain members across these chains. Thus, multiple
memberships could inflate within-supply chain variation relative to between-supply chain
variation for any of the variables considered in our models. In addition, variation in the degrees
of multiple memberships across supply chains may constitute a source of unobserved
heterogeneity that systematically correlates with our dependent and independent variables,
therefore biasing our main results. To explore these potential issues, we run two additional
analyses.
First, in accordance with the literature (Bliese, 2000), we calculate ICC(1) and ICC(2) to
assess if the extended supply chain represents a meaningful level of data aggregation, in spite
of multiple memberships. ICC(1) captures the degree to which ESG disclosure scores from
firms in the same extended supply chain are influenced by and depend on some characteristics
35
of that supply chain. ICC(1) can be calculated from a one-way random effects ANOVA
following Bartko’s (1976) formula, with values equal to or greater than 0 suggesting significant
group-level effects. ICC(2) captures the degree to which ESG disclosure scores for multiple
members of the same supply chain are consistent when expressed as deviations from their mean.
ICC(2) is typically estimated with the use of mean squares from a one-way random effects
ANOVA following McGraw and Wong’s (1996) formula; values greater than 0.6 suggest that
a score from one randomly extracted observation provides a reliable estimate of the group mean.
In our research, an ICC(1) of 0.009 and an ICC(2) of 0.72 suggest that the extended supply
chain represents an important level of data aggregation and analysis, in spite of multiple
memberships.
Second, to address potential issues of unobserved heterogeneity of multiple memberships
across supply chains, we constructed a novel variable called supply chain relevance to measure
the relative economic importance of each extended supply chain j for its supply chain members.
Supply chain relevancej is calculated as the average percentage of revenues that members of
supply chain j receive from each other (Mean: 37.50%; Median: 35.67%; Min: 14.29%; Max:
64.02%). We then split our sample based on the median of this variable and ran two separate
regressions (Table 6); in each sub-sample, minimal variation of supply chain relevance is likely
to reflect minimal variation in the degrees of multiple memberships across cases (supply
chains). Moreover, when supply chain relevance is high, multiple memberships are less likely
to constitute a major influencing factor because supply chain members are more relationally
oriented to each other’s than they are, on average, with firms in other supply chains.
In Table 6, Model 4a for high supply chain relevance remains largely consistent with the
main results in Table 5, despite being lower in statistical power. Then, Model 4b for low supply
chain relevance indicates a weaker association between supply chain structure and
transparency, suggesting that our overall estimates in the main analysis are likely to be
36
downward biased (conservative) due to the presence of cases with low supply chain relevance.
This analysis also indicates that firms belonging to multiple supply chains may choose a
dominant supply chain based on its economic relevance and then adjust practices and outcomes
based on the structural characteristics of this supply chain.
Table 6. Robustness check – Multiple memberships
5 DISCUSSION
Our study addressed an important question: what is the relationship between supply chain
structure and supply chain transparency? The answer is complex, with supply chain density
having a positive association with supply chain transparency, supply chain clustering holding
a negative effect and supply chain geographical heterogeneity having a positive effect. In
addressing this empirical question via a large-scale descriptive study at the supply chain level
of analysis, we extend the current understanding of supply chain transparency as a collective
outcome that systematically associates to supply chain structure in a nuanced fashion. The study
makes two important empirical contributions to the literature on supply chain transparency.
High SC relevance Low SC relevance
Coefficient Std Errors P-values Coefficient Std Errors P-values
SCDensity 0.589 0.115 0.000 0.348 0.210 0.137
SCClustering -0.357 0.096 0.004 -0.152 0.165 0.383
SCGeographicalH 0.349 0.088 0.003 0.323 0.107 0.017
SCIndustrialH 0.009 0.079 0.907 -0.180 0.069 0.031
SCHorizontal 0.016 0.050 0.752 -0.103 0.066 0.160
SCVertical 0.027 0.091 0.776 -0.068 0.104 0.531
SCOrgSize 0.419 0.126 0.008 0.396 0.135 0.018
IndustryClock -0.217 0.274 0.448 -0.079 0.201 0.704
Collectivism America -0.300 0.152 0.077 -0.380 0.396 0.365
Collectivism EU -0.637 0.191 0.008 -0.479 0.252 0.094
FFDisclosure 0.113 0.041 0.020 0.098 0.067 0.182
End-Market Distance -0.040 0.064 0.540 0.192 0.098 0.086
IndustryRepRisk 0.106 0.052 0.070 0.023 0.079 0.782
SCRegPressure 0.702 0.035 0.000 0.454 0.087 0.001
Constant 0.624 0.346 0.101 -0.003 0.452 0.995
Observations 93 94
(pseudo) R-squared 0.806 0.564
(4a) Disclosure_ESG
(4b) Dislcosure_ESG
37
First, we extend studies that have adopted qualitative methods (Dahlman and Roehrich,
2019; Fontana and Egels‑Zandén, 2019) or coarse measures (Kim and Davis, 2016) to offer
important insights into the multi-faceted association between supply chain structure and supply
chain transparency. Particularly, our study provides evidence that, in the context of ESG
disclosure, interconnected supply chains can achieve markedly different levels of supply chain
transparency, depending on the pattern of ties connecting supply chain members.
Research in a variety of empirical settings – from innovation (Carnovale and Yeniyurt,
2015) to corporate social responsibility (Fontana and Engles-Zanden, 2019) and climate change
(Dalhmann and Roenhrich, 2019) – suggests that denser ties between supply chain members
enable information sharing and the development of common norms and practices. Our
observation of high collective ESG disclosure in relation to high supply chain density
empirically validates this supposition and supports hypothesis 1. Moreover, a limited but
expanding body of studies suggests that high supply chain clustering may lead to high levels of
information sharing within a cluster, but limited information sharing and poor norms diffusion
among clusters (Bimber et al., 2005; Fu and Shumate, 2016; Dooley et al. 2018). Our
observation of low collective ESG disclosure in relation to high supply chain clustering
empirically validates this supposition and supports our second hypothesis, especially when
focusing on the transparency of suppliers and sub-suppliers (supplement C1).
We also find that supply chain heterogeneity only partially associates with supply chain
transparency. Research in a variety of empirical settings – from risk management (Bode and
Wagner, 2015) to conflict minerals (Kim and Davis, 2016) – suggests that supply chain
heterogeneity may inhibit the attainment of important collective outcomes due to the limited
transferability of practices that were developed to fit specific local institutions and operational
settings. However, research in other settings – from new product development (Gao et al., 2015)
to knowledge diffusion (Marques et al., 2019) and new market formation (Lee et al., 2018) –
38
suggests that supply chain heterogeneity may create a collaborative climate whereby the
attainment of collective outcomes is enabled by the lack of competition, the need to overcome
information asymmetries and the opportunity to learn and combine complementary know-how.
We find that supply chain geographical heterogeneity is positively associated with supply chain
transparency, supporting H3-competing. Finally, we find that supply chain industrial
heterogeneity does not associate with supply chain transparency.
Our second empirical contribution relates to membership in multiple supply chains.
Instead of focusing on the transparency of focal firms (Marshall et al., 2016) or the transparency
of individual suppliers (Jira and Toffel, 2013; Villena and Dhanorkar, 2020), our large-scale
study examines transparency at a supply chain level of analysis. It is notable that our calculation
of ICC1 and ICC2 in section 4.3 provides evidence that the transparency of individual supply
chain members is influenced by and depends on some characteristics of the extended supply
chain, in spite of multiple memberships. Our analysis in Table 6 also indicates that supply chain
structure is more influential when the extended supply chain accounts for a large portion of its
members’ revenues. In summary, contrary to what the literature implicitly assumes (Min et al.,
2008), our observations suggest that firms belonging to multiple extended supply chains may
choose a dominant supply chain based on its economic relevance and then adjust practices and
outcomes based on some structural characteristics of this supply chain.
We also make a methodological contribution. Prior studies have typically offered a partial
account of a focal firm’s supply chain, as studies focusing on the supply base or the ego-network
usually lack information about sub-suppliers (Carnovale and Yeniyurt, 2015), while studies
focusing on the supply network lack information about customers and their extended ties with
other supply chain members (Sharma et al., 2019a). In this paper, we build upon these studies
to offer a novel two-step procedure to isolate extended supply chains, each consisting of a focal
firm’s customers, suppliers, and sub-suppliers (see Supplement A for details). This procedure
39
has two advantages: first, it enables us to isolate supply chains as overlapping sub-graphs in a
large directed network; second, it allows us to analyze asymmetric supply chains where
boundaries must stretch more upstream than downstream or vice-versa. Future studies can use
this procedure to continue to empirically investigate constructs and theoretical associations at
a supply chain level of analysis (Carter and Washispack, 2018).
5.1 Managerial implications
The managers in our sample all have a fiduciary duty to their shareholders. However, external
stakeholders are becoming increasingly assertive and sophisticated in holding focal firms to
account for the global ESG performance of their extended supply chains. For example, some
1,051 large firms from 49 countries have recently been asked to disclose environmental data
under the Non-Disclosure Campaign organized by CDP, the non-profit global environmental
disclosure platform (Scott, 2020). Simultaneously, a group of 101 international investors
representing over $4.2 trillion in assets under management has joined forces to call on
governments to put in place regulatory measures requiring focal firms to conduct ongoing
human rights due diligence in their supply chains (Gorsen and Sandick, 2020). And on June
2020, the European Parliament formally adopted the EU’s Sustainability Taxonomy Regulation
with the aim to provide a consistent policy framework to reduce “greenwashing”, where
products are marketed as environmentally sustainable without sufficient factual basis for their
claims (Valentine et al., 2020). Our study helps focal firms to understand the important role
that supply chain structure will play in addressing these raising demands.
A survey of 1,128 supply chain professionals during the fall of 2019 indicates that 35%
of the respondents lack clear ESG goals for their supply chains (MIT CTL & CSCMP, 2020).
Research goes on to indicate that these professionals do not understand the structure of their
extended contractual ties (Villena and Gioia, 2018) and the implications for supply chain
transparency and sustainability (Bateman and Bonanni, 2019). Our study tackles this lack of
40
managerial understanding by examining the systematic association between supply chain
structure and transparency in 187 extended supply chains. Our results help supply chain
professionals to better justify investments into mapping and altering their existing supply
chains.
In light of our empirical observations, investments into mapping the extended supply
chain can help supply chain professionals to understand why supply chains are opaque or
underperforming. If a sparse, clustered or overly homogenous structure is identified, then our
study offers important insights on how it could be corrected as well as on how professionals
could try to mitigate its negative implications.
First, in addition to selecting more transparent supply chain members, supply chain
professionals should invest more time and efforts into brokering new supply chain ties. To
begin, focal firms could act as tertius iungens i.e., the third who joins disconnected supply chain
members. The typical tertium iungens strategy discussed in the literature would suggest
relinquishing control over structural holes in exchange for synergy and self-coordination in the
resulting closed triads (Pathak et al., 2014). Based on our empirical observations, we instead
suggest focal firms connect with supply chain members that do not share ties with the same
customers and belong to different manufacturing clusters, to increase supply chain density
without exacerbating supply chain clustering. Supplier forums and conferences could offer
important opportunities to broker ties between these supply chain members (Gualandris and
Barg, 2020).
Second, when selecting new supply chain members, supply chain professionals should
pay more attention to how the addition of these actors could modify the structure of the extended
supply chain. The number and nature of ties that new suppliers or customers could potentially
develop with existing supply chain members, based on their technical and relational
capabilities, should become important criteria for selection.
41
Third, supply chain professionals should be cognizant of the unintended consequences of
localizing their supply chains. Although supply chain geographical homogeneity may generate
positive competitive effects in terms of reduced production lead times, lower costs and
improved quality, it may also hamper supply chain transparency and the ability of the focal firm
to address the ever-growing informational demands of external stakeholders.
Fourth, supply chain professionals and their focal firms should demonstrate a level of
commitment toward transparency that exceeds their own requirements for supply chain
members: The Models in Table 5 suggest that an increase in the transparency of the focal firm
positively associates with an increase in the transparency of supply chain members.
Fifth, based on the Models in Tables 5, we suggest American and European supply chain
professionals to more actively promote a collectivistic culture in their extended supply chains.
Although unobservable in our study, it is reasonable to suggest that focal firms from Asia secure
higher levels of supply chain transparency by building keiretsu, a Japanese practice where
supply chain members take small equity stakes in each other while remaining operationally
independent (e.g., Tiessen, 1997).
Finally, we recommend supply chain professionals publicly disclose the names, locations
and operations of their supply chain members on web-sites and media, enlisting investors, non-
governmental organizations and local regulators to monitor their activities. While disclosing
such information is often perceived as risky or expensive, increased scrutiny will likely help
deter suppliers from undesirable practices which have been empirically linked to reduced
competitiveness (e.g., Foerstl et al., 2010). Supply chain professionals may want to balance this
trade-off by prioritizing the disclosure of those supply chain members that are more passive
and difficult to influence.
42
5.2 Limitations and future developments
This study has a few limitations that represent additional opportunities for future research.
First, our study does not support causal inference. While we test our hypotheses using ordinary
least squares (OLS) regressions with robust clustering by industry, two-stage least squares,
discontinuous regressions and quasi-natural experiments would be preferable (Serpa and
Krishnan, 2018; Dai et al., 2020). Unfortunately, important theoretical and empirical challenges
discussed in the literature currently compromise the ability of such methods to estimate supply
chain-level effects (Lu and Shang, 2017, p. 34). Therefore, as customer/supplier data become
progressively more available and easier to collect, future studies can build panels of quantitative
data to tackle these limitations. For example, quasi-natural experiments could estimate if supply
chains with diverse structures follow diverse patterns of supply chain transparency when
responding to common exogenous shocks like the introduction of new regulations or industry
scandals.
Second, there is room to investigate other dimensions of interconnectedness such as
clustering signature that considers the predominant type of closed triads in a supply chain
(Figure 3; Ahnert and Fink, 2008) and other structural measures that capture the extent to which
supply chain members compete with one another (Gulati and Gargiulo, 1999, p. 1447; Lu and
Shang, 2017). Similarly, there is also room to further unpack the association between supply
chain heterogeneity and supply chain transparency by adopting more granular measures of
institutional, cultural and operational heterogeneity that could better isolate heterogeneity’s
positive and negative mechanisms.
Third, and related, we do not measure what intervening mechanisms explain the
association between supply chain structure and supply chain transparency. As illustrated in
recent research (Bansal et al., 2020), qualitative big data analyzed via unsupervised machine
learning could help to disentangle these hidden mechanisms. For example, our findings on
43
interconnectedness align well with the social theory of network closure proposed by Coleman
(1990) and its application to collective action problems (Ostrom, 1998). Coleman suggests that
diffused norms regulating important collective outcomes emerge and are better enforced when
distributed ties span many structural holes, typical of high density and low clustering. Future
supply chain studies could use qualitative big data to reveal if and how supply chain structure
relates to the emergence of effective norms and how these norms in turn ensure the achievement
of collective outcomes.
Finally, future research should investigate how suppliers’ structural position within and
between supply chains relate to their transparency and sustainability. Our descriptive analysis
in Supplement B finds that transparency varies significantly within supply chain tiers, with
higher disclosure scores achieved by nexus firms that manage a richer portfolio of buying and
selling ties. This empirical observation aligns well with prior qualitative studies (Villena &
Gioia, 2018) but demands more targeted investigation.
6. CONCLUSION
The literature centered on focal firms’ efforts at increasing transparency in their extended
supply chains shows how challenging such a process can be. This study, in the context of ESG
disclosure, empirically demonstrates that supply chain structure has a strong, nuanced
association with supply chain transparency. It contributes to the growing understanding of how
supply chain structure can facilitate or hamper more transparency and sustainability. We hope
that our empirical exploration will spur a variety of novel studies and practical developments.
44
On-line Supplement A – Isolating intertwined supply chains from a larger network.
Our large network in Figure A1 (below) is composed of intertwined extended supply chains,
each capturing a four-tier system. While identifying customers and direct suppliers of a focal
firm is straightforward (the ego-network), identifying sub-suppliers requires more careful
consideration because these supply chain members can easily be confused with other firms that
are indirectly connected to the focal firm like, for example, a supplier’s customer representing
a direct competitor of the focal firm (See Figure 2 for a simplified illustration of these
problematic network members).
Isolating extended supply chains requires a two-step analytical procedure:
• In the first step, we build a directed, binary adjacency matrix reflecting the
network of contractual ties among the 4803 firms in our dataset. From a graph-
theoretic standpoint, all information about the supply chain members of a focal
firm is contained in such a matrix and, more specifically, encapsulated in the sub-
matrixes presented in Figure A2. Each sub-matrix represents a special type of
“Organization Architecture Design Structure Matrix” (Browning, 2015) that
provides a decomposition of the network of direct and indirect ties surrounding a
specific focal firm. A focal firm-specific sub-matrix allows to categorize supply
chain members into different functional roles (customer, supplier, sub-supplier,
potential competitor) based on their structural position relative to the focal firm.
In a generic sub-matrix, ‘A’ represents the focal firm, ‘B’ represents a customer
or a direct supplier and ‘C’ can identify a sub-supplier, depending on the pattern
of ties linking ‘C’ to ‘A’. Sub-matrixes will be asymmetrical because supply
chains represent ‘directed’ networks where contractual ties are oriented and firms
may simultaneously act as customers and suppliers. In each sub-matrix, ‘X’ takes
the value of 1 whenever two firms directly exchange products and services for
production, 0 otherwise.
• In a second step, we isolate a focal firm’s extended supply chain by retaining
supply chain members identified as ‘B’ (i.e., customers or suppliers) plus a
selection of ‘C’ firms for which the combination X3≠0, X5=0 and X6≠0 applies
(i.e., sub-suppliers). For details on how to implement this procedure in R studio,
please refer to Kolaczyk and Csárdi (2014, p. 18).
Unfortunately, the Bloomberg SPLC database does not provide information about partial
directed-buy agreements between focal firms and sub-suppliers. As a consequence, when a
partial directed-buy agreement is formalized into a contract between a focal firm and a sub-
supplier, then X5 in the sub-matrix will be equal to 1 and our procedure will classify this sub-
supplier as a direct supplier.
45
Figure A1. Large network of intertwined supply chains
Figure A2. Sub-matrixes and selection of a focal firm’s legitimate sub-suppliers
a) Generic sub-matrix identifying potential supply chain members:
A B C
A
0 X1 X2
B
X3 0 X4
C
X5 X6 0
b) Sub-matrixes that identify legitimate sub-suppliers of a generic focal firm 'A'
A←B←CA←B↔C A↔B←CA↔B↔C
A B C A B C A B C A B C
A
0 0 0
A
0 0 0
A
0 1 0
A
0 1 0
B
1 0 0
B
1 0 1
B
1 0 0
B
1 0 1
C
0 1 0
C
0 1 0
C
0 1 0
C
0 1 0
c) Sub-matrixes that identify firms to be excluded from the extended supply chain of a generic focal firm 'A':
A→B→C
A→B↔C A→B←C A↔B→C A←B→C
A B C A B C A B C A B C A B C
A
0 1 0
A
0 1 0
A
0 1 0
A
0 1 0
A
0 0 0
B
0 0 1
B
0 0 1
B
0 0 0
B
1 0 1
B
1 0 1
C
0 0 0
C
0 1 0
C
0 1 0
C
0 0 0
C
0 0 0
Legend:
Firm;
Contractual tie.
46
On-line Supplement B – Transparency differences across and within supply chain tiers.
This descriptive analysis examines the variation in the average value of ESG disclosure scores
across and within diverse supply chain tiers of our large network of intertwined extended
supply chains.
As indicated in Table B, we developed four conditions to identify eight mutually
exclusive groups of firms that occupy different structural positions within our large network.
The resulting typology distinguishes between firms based on the supply chain tier and the
breath of contractual relationships. Condition A and condition C aggregate firms in different
supply chain tiers based on their selling and or buying ties with focal firms. Condition B and
condition D distinguish firms that manage a richer portfolio of buying and selling ties, labeled
as nexus firms (Yan et al., 2015), relative to other firms within the same tier.
ANOVA and the Scheffe test indicate that ESG disclosure scores vary substantially
across supply chain tiers with suppliers – especially sub-suppliers – presenting on average the
lowest ESG disclosure scores. The analysis of variance also indicates that ESG disclosure
scores vary significantly within tiers, with ‘nexus’ firms achieving higher disclosure scores
than other firms at the same tier.
Table B. ESG Disclosure differences across and within tiers
Group ID Group Na me Obs. Examples in Figure 2
A: Buy from
focal firms
B: Buy from
other firms
C: Sell to
focal firms
D: Sell to
other firms
Average Std Dev
Anova , scheffe a
1Customers (nexus) Yes Yes No Yes 159 A VNET 24.99 22.25 2; 3; 4; 5; 6; 7; 8
2Customers Yes Yes No No 495 Siemens AG 14.55 18.51 1; 3; 4; 5; 7
3Focal firms
245cABB Ltd 33.59 20.76 1; 2; 4; 5; 6; 7; 8
4Direct suppliers (nexus) Y es Yes Yes Yes 479 S kanska AB 19.56 21.19 1; 2; 3; 5; 7; 8
5Direct suppliers No Yes Yes Yes 691 Infineon Tech. 10.55 16.11 1; 2; 3; 4; 7
6Sub-suppliers (nexus) No Yes No Yes 330 Synaptics Inc. 14.12 19.14 1; 3; 4; 7
7Sub-suppliers No No No Yes 1249 A dvanced Semiconductor Eng. 4.96 11.3 1; 2; 3; 4; 5; 6; 8
8
Potential competitorsbNo Yes No No 1155 Kaga Electronics 12.12 17.33 1; 3; 4; 7
Tot. 4803
a F-Statistics 115.14; chi2(7) 462.73
c As indicated in our fowchart (Figure 1), we collected data for 245 focal firms. Only 187 survived to step 5 (i.e., outliers)
- Sa mpled from Fo rbes2000 -
b Focal firms' potential competitors were identified as suppliers' customers with no different connection with focal firms. Our two-step procedure to isolate extended supply chains ultimately
Conditions
Bloomberg ESG Disclosure Sco re
47
On-line Supplement C1 – Concerns related to our unit of analysis.
While our main analysis in Table 5 considers the full set of supply chain members N in each
extended supply chain j, the sub-set G of direct suppliers and sub-suppliers (G < N) may present
important idiosyncrasies. By considering supply chain members N, our main analysis could
over-estimate supply chain transparency and its association with supply chain structure because
customers tend to publicly disclose more ESG information than upstream members (see
Supplement B for details). These upstream members greatly affect the operational,
environmental and social performance of the supply chain (Blanco et al., 2016) but are
incredibly difficult to influence and motivate (Villena and Dhanorkar, 2020), which ultimately
constrains the amount of material ESG information that can be made available to the public and
used for improvement. To assess this potential concern, we perform a regression analysis that
evaluates how the upstream portion of a supply chain (i.e., sub-set G) is influenced by the
structural characteristics of the complete network (i.e., ties between customers, suppliers and
sub-suppliers in the set N). Results are reported in Table C1.
Table C1. Robustness checks – Supply chain transparency of sub-set G
First, Models 5, 6 and 7 in Table C1 confirm the positive association between supply
chain density and supply chain transparency. Second, compared to our main analysis in Table
5, Models 5, 6 and 7 in Table C1 provide evidence of a stronger negative association between
supply chain clustering and supply chain transparency (Wald test’s p-value = 0.002). This
suggests that supply chain clustering hampers the development of collective outcomes across
upstream members more than it does across other supply chain members (i.e., customers).
Third, compared to the main analysis in Table 5, Models 5, 6 and 7 in Table C1 indicate a
weaker and less consistent association between supply chain geographical heterogeneity and
transparency (Wald test’s p-value = 0.000). Finally, Models 4, 5 and 6 in Table 6 confirm that
supply chain industrial heterogeneity is not associated with transparency.
Overall, this robustness check provides additional support to our main results in Table 5
but also add interesting nuances concerning the association between supply chain structure and
transparency.
(7) Disclosure_Signal
Coefficient Std Errors P-values Coefficient Std Errors P-values Coefficient Std Errors P-values
SCDensity 0.482 0.115 0.002 0.494 0.138 0.005 0.519 0.203 0.028
SCClustering -0.327 0.082 0.003 -0.259 0.042 0.000 -0.369 0.095 0.003
SCGeographicalH 0.188 0.077 0.035 0.136 0.117 0.272 0.196 0.101 0.081
SCIndustrialH -0.041 0.046 0.393 -0.100 0.073 0.199 -0.098 0.060 0.135
SCHorizontal -0.091 0.056 0.135 -0.083 0.057 0.178 -0.073 0.075 0.355
SCVertical -0.055 0.096 0.576 -0.043 0.080 0.598 -0.040 0.079 0.625
SCOrgSize 0.276 0.069 0.003 0.296 0.074 0.003 0.147 0.121 0.252
IndustryClock -0.201 0.265 0.466 0.067 0.269 0.807 -0.236 0.168 0.192
Collectivism America -0.484 0.190 0.029 -0.633 0.192 0.008 -0.889 0.243 0.004
Collectivism EU -0.603 0.156 0.003 -0.563 0.174 0.009 -0.782 0.189 0.002
FFDisclosure 0.079 0.049 0.137 0.075 0.055 0.201 0.138 0.035 0.003
End-Market Distance 0.014 0.050 0.787 0.117 0.099 0.264 -0.036 0.042 0.406
IndustryRepRisk 0.127 0.067 0.088 0.006 0.090 0.946 0.119 0.075 0.143
SCRegPressure 0.556 0.081 0.000 0.452 0.087 0.000 0.209 0.144 0.176
Constant 0.507 0.377 0.208 0.145 0.427 0.741 0.826 0.243 0.007
Observations 187 187 187
(pseudo) R-squared 0.520 0.526 0.478
(5) Dislcosure_ESG
(6) Disclosure_Env
48
On-line Supplement C2 – Robustness check with multiple percentiles.
This robustness check assesses the sensitivity of our main results in Table 5 to variations in the selected threshold for constructing our dependent variable.
It is important to note that, as the selected threshold progressively moves away from the median, the distribution of our dependent variable becomes flatter
(Table C2a) and the ability of this measure to discriminate across different levels of supply chain transparency deteriorates. Overall, however, the analysis
in Table C2b provides additional credence to our main results; the coefficients of our four structural regressors are reasonably consistent across models
8a-8d and 9a-9d. In particular, supply chain density remains the strongest and most consistent predictor, followed by supply chain geographical
heterogeneity and then supply chain clustering. The coefficient of supply chain industrial heterogeneity remains statistically insignificant across all our
models.
Table C2a. Distribution of Collective ESG Disclosure for different percentiles.
Table C2b. Robustness check with multiple percentiles.
Mean Std. Dev. Min Max
25th percentile 44.01 4.43 28.26 60.40
40th percentile 35.36 4.79 21.74 54.55
50th percentile (primary measure) 33.17 4.88 14.13 51.65
60th percentile 26.46 4.23 13.04 40.91
75th percentile 17.86 3.52 7.20 29.58
Coefficient Std Errors P-values Coefficient Std Errors P-values Coefficient Std Errors P-values Coefficient S td Errors P-values Coefficient Std Errors P-values Coefficient
Std Errors
P-values Coefficient Std Errors P -values Coefficient Std Errors P-values
SCDensity 0.423 0.222 0.086 0.547 0.173 0.010 0.618 0.151 0.002 0.529 0.202 0.026 0.519 0.175 0.014 0.682 0.151 0.001 0.445 0.183 0.036 0.473 0.161 0.015
SCClustering -0.147 0.112 0.217 -0.234 0.135 0.114 -0.154 0.091 0.124 -0.003 0.087 0.973 -0.220 0.111 0.076 -0.243 0.090 0.022 -0.163 0.076 0.057 -0.027 0.090 0.774
SCGeographical 0.440 0.143 0.012 0.557 0.098 0.000 0.456 0.131 0.006 0.476 0.172 0.020 0.512 0.141 0.005 0.448 0.147 0.012 0.349 0.212 0.131 0.257 0.171 0.165
SCIndustrial -0.015 0.156 0.925 -0.100 0.107 0.374 -0.074 0.107 0.502 -0.061 0.226 0.795 -0.107 0.112 0.365 -0.087 0.108 0.438 -0.134 0.206 0.531 -0.121 0.182 0.521
SCHorizonta l -0.035 0.037 0.358 -0.037 0.044 0.416 -0.070 0.058 0.251 -0.073 0.051 0.183 -0.043 0.050 0.409 -0.048 0.057 0.417 -0.033 0.055 0.562 -0.053 0.060 0.393
SCVertical 0.056 0.104 0.601 -0.016 0.088 0.861 -0.065 0.077 0.419 -0.121 0.095 0.232 -0.028 0.086 0.756 -0.024 0.077 0.760 -0.062 0.077 0.441 -0.176 0.082 0.057
SCOrgSize 0.442 0.088 0.001 0.372 0.083 0.001 0.356 0.062 0.000 0.349 0.102 0.006 0.340 0.091 0.004 0.251 0.070 0.005 0.383 0.127 0.013 0.267 0.108 0.033
Indu stryClock -0.081 0.120 0.515 -0.134 0.193 0.502 -0.043 0.240 0.861 0.153 0.242 0.541 -0.045 0.218 0.839 0.011 0.223 0.961 0.177 0.194 0.384 0.164 0.231 0.493
Collectivism America -0.487 0.174 0.019 -0.675 0.194 0.006 -0.642 0.171 0.004 -0.496 0.206 0.037 -0.690 0.174 0.003 -0.755 0.175 0.002 -0.682 0.179 0.003 -0.564 0.157 0.005
Collectivism EU -0.590 0.155 0.003 -0.782 0.116 0.000 -0.575 0.108 0.000 -0.286 0.111 0.027 -0.779 0.094 0.000 -0.738 0.101 0.000 -0.616 0.120 0.000 -0.400 0.086 0.001
FFDisclosure 0.123 0.059 0.062 0.107 0.049 0.052
0.088+0.050 0.110 0.102 0.058 0.108 0.101 0.048 0.059 0.099 0.054 0.095 0.111 0.067 0.128 0.089 0.084 0.316
End- Market Distance 0.109 0.038 0.017 0.088 0.054 0.134 0.119 0.087 0.203 0.278 0.132 0.063 0.102 0.093 0.297 0.130 0.103 0.234 0.240 0.151 0.143 0.247 0.147 0.123
Indu stryRepRisk 0.094 0.076 0.242 0.091 0.054 0.125 0.074 0.067 0.294 -0.067 0.111 0.557 0.035 0.047 0.481 0.026 0.065 0.699 -0.098 0.096 0.334 -0.110 0.107 0.327
SCRegPressure 0.508 0.112 0.001 0.429 0.088 0.001 0.498 0.034 0.000 0.282 0.098 0.017 0.616 0.054 0.000 0.545 0.031 0.000 0.480 0.094 0.000 0.310 0.058 0.000
Constant 0.178 0.690 0.504 0.342 0.270 0.234 0.189 0.342 0.645 -0.431 0.444 0.355 0.239 0.351 0.512 0.189 0.342 0.592 -0.226 0.446 0.624 -0.303 0.424 0.492
Observations 187 187 187 187 187 187 187 187
(pseud o) R-squared 0.576 0.598 0.605 0.538 0.621 0.599 0.582 0.468
(9d) Dislcosure_Env
(8b) Disclosure_ESG
(8c) Dislcosure_ESG
(8d) Dislcosure_ESG
(9a) Dislcosure_Env
(9b) Disclosure_Env
(9c) Dislcosure_Env
(8a) Dislcosure_ESG
40th percentile
60th percentile
75th percentile
Collective ESG Disclosure
Collective Environmental Disclosure
25th percentile
40th percentile
60th percentile
75h percentile
25th percentile
49
On-line Supplement C3 – Robustness check with entropy-based measures.
We employ alternative entropy-based measures of heterogeneity which consist of the sum
of pi *log(pi), where p is the fraction of supply chain members i in set N of supply chain
j in a given category k (Dooley et al., 2018). For supply chain geographical heterogeneity
k represents different home countries. For supply chain industrial heterogeneity k
represents different industry groups. In the context of ESG disclosure, the analysis in
Table C3 provides additional support to H3-competing for supply chain geographical
heterogeneity and confirm the absence of a statistically significant association between
supply chain industrial heterogeneity and transparency.
Table C3. Robustness check with entropy-based measures.
(10) Dislcosure_ESG (11) Disclosure_Env (12) Disclosure_Signal
Coefficient Std Errors P-values Coefficient Std Errors P-values Coefficient Std Errors P-values
SCDensity 0.642 0.162 0.003 0.581 0.146 0.003 0.586 0.285 0.067
SCClustering -0.212 0.109 0.072 -0.175 0.076 0.044 -0.271 0.112 0.036
SCGeographical_entropy 0.602 0.111 0.000 0.381 0.166 0.045 0.457 0.097 0.001
SCIndustrial_entropy -0.040 0.113 0.733 -0.079 0.126 0.547 -0.149 0.186 0.442
SCHorizontal -0.043 0.047 0.380 -0.059 0.057 0.322 -0.066 0.075 0.403
SCVertical 0.000 0.083 0.999 -0.053 0.074 0.492 -0.034 0.095 0.729
SCOrgSize 0.334 0.086 0.003 0.331 0.082 0.002 0.179 0.148 0.254
IndustryClock -0.073 0.206 0.732 0.031 0.223 0.891 -0.316 0.152 0.064
Collectivism America -0.617 0.179 0.006 -0.789 0.188 0.002 -1.122 0.199 0.000
Collectivism EU -0.767 0.096 0.000 -0.690 0.097 0.000 -0.991 0.186 0.000
FFDisclosure 0.110 0.049 0.050 0.101 0.057 0.107 0.146 0.033 0.001
End-Market Distance 0.093 0.077 0.258 0.166 0.109 0.160 -0.033 0.036 0.384
IndustryRepRisk 0.064 0.042 0.162 0.008 0.078 0.923 0.144 0.077 0.091
SCRegP ressure 0.570 0.060 0.000 0.493 0.032 0.000 0.081 0.131 0.549
Constant 0.252 0.322 0.453 0.098 0.352 0.787 0.986 0.199 0.001
Observations 187 187 187
(pseudo) R-squared 0.644 0.591 0.552
50
On-line Supplement D – Assessing the impact of supply chain data coverage.
The estimates from the main analysis (Table 5) may be biased if supply chain members
that disclose more ESG information to the public are also those with more complete
supply chain data available in the Bloomberg SPLC database. This intrinsic endogeneity
bias may occur if some firms decide not to publicize ties with other supply chain members
that are less sustainable or less transparent, in order to mitigate a potential reputational
backlash. Motivated by this consideration, we constructed two variables that measure (a)
the total percentage of a customer’s known spend on its direct suppliers and (b) the total
percentage of a supplier’s known revenues from its customers. While customer data seem
bias free, supplier’s revenues correlate with the ESG disclosure score (correlation p-value
= 0.000), suggesting the existence of an intrinsic endogeneity bias.
We then attempted to quantify the extent to which this bias influences the results
from the main analysis (Table 5). We constructed a novel variable called supply chain
data coverage to capture the completeness of customer/supplier data for any given
extended supply chain j. This variable is calculated as the average percentage of “known”
revenues per supply chain member i in a given supply chain j (mean = 21.00%; median
= 21.02%; min = 7.92%; max = 33.8%). Finally, we split our sample based on the median
of this variable and ran two regression analyses (Table D). In each sub-sample minimal
variation across cases limits the influence of the identified intrinsic bias on our estimates.
Overall, many of the hypothesized direct effects remained statistically significant,
despite being lower in statistical power. Of particular note, is that we find consistent
effects in Models 13a and 14a, suggesting that our overall estimates in the main analysis
are likely to be downward biased (conservative) due to the presence of low data coverage
observations.
Table D. Robustness check – Supply chain data coverage
High SC data coverage Low SC data coverage High SC data coverage Low SC data coverage
Coefficient S td Errors P-values Coefficient S td Errors P-values Coefficient Std Errors P-values Coefficient S td Errors P-values
SCDensity 0.640 0.101 0.000 0.402 0.143 0.021 0.641 0.075 0.000 0.403 0.147 0.023
SCClustering -0.465 0.137 0.008 -0.142 0.154 0.378 -0.455 0.148 0.013 -0.255 0.099 0.029
SCGeographicalH 0.374 0.114 0.009 0.346 0.079 0.002 0.305 0.128 0.044 0.108 0.079 0.203
SCIndustrialH -0.027 0.084 0.760 -0.095 0.114 0.424 0.019 0.081 0.816 -0.095 0.110 0.407
SCHorizonta l 0.007 0.063 0.908 -0.129 0.071 0.104 -0.012 0.047 0.811 -0.197 0.101 0.082
SCVertical -0.003 0.132 0.981 -0.046 0.113 0.694 0.021 0.137 0.884 -0.147 0.117 0.242
SCOrgSize 0.539 0.246 0.056 0.418 0.154 0.024 0.341 0.268 0.235 0.326 0.117 0.021
Indu stryClock -0.344 0.363 0.368 0.019 0.260 0.944 -0.257 0.336 0.464 -0.095 0.265 0.728
Collectivism America -0.270 0.074 0.005 -0.421 0.365 0.278 -0.294 0.098 0.015 -0.608 0.329 0.097
Collectivism EU -0.562 0.288 0.083 -0.459 0.337 0.206 -0.676 0.411 0.135 -0.555 0.297 0.094
FFDisclosure 0.130 0.067 0.083 0.081 0.107 0.468 0.120 0.073 0.137 0.066 0.093 0.497
End- Market Distance -0.057 0.048 0.270 0.238 0.066 0.006 0.015 0.049 0.772 0.130 0.072 0.105
Indu stryRepRisk 0.140 0.072 0.084 -0.041 0.078 0.616 0.091 0.080 0.282 -0.033 0.101 0.749
SCRegPressure 0.558 0.117 0.001 0.602 0.100 0.000 0.711 0.142 0.001 0.492 0.084 0.000
Constant 0.798 0.378 0.064 -0.174 0.432 0.696 0.567 0.408 0.199 0.301 0.447 0.518
Observations 94 93 94 93
(pseud o) R-square d 0.685 0.659 0.603 0.539
(13a) Disclosure_ESG
(13b) Dislcosure_ESG
(14a) Disclosure_ESG
(14b) Dislcosure_ESG
Collective ES G Disclosure of all su pply chain members (set N)
Collective ES G Disclosure of up stream memb ers (sub-set G)
51
On-line Supplement E – Assessing the impact of ‘N/As’
The Bloomberg ESG database predominantly covers public firms with medium and large
market capitalization ( US$2billion). In this database, firms that do not disclose anything
will show N/A (i.e., 0) and firms that are not covered by the database will also show N/A
(i.e., 0). About 34% of the firms in our dataset have small market capitalization
(<US$2billion) and a disclosure score of N/A (i.e.,0). Unfortunately, there is no reliable
way to discover what sub-fraction of these firms were active disclosers of ESG
information in 2016. The following analysis, however, suggests that these potentially
problematic firms do not constitute a source of systematic bias in our main analysis.
First, we counted the number of firms i per supply chain j for which two conditions
applied: small market capitalization (<US$2billion) and a disclosure score equal to 0.
Then, we divided this count variable by the total number of firms in any given supply
chain j to calculate their percentage. Figure E shows the distributions of the counting
variable ‘N/As count’ and the percentage variable ‘N/A%’. These two variables are not
significantly correlated to our supply chain structural characteristics (p-value > 0.150)
and the regression analysis in Table E shows that our main results for both set N and sub-
set G are robust to the uneven distribution of these problematic firms across our cases.
Figure E. Distribution of N/As count and N/A%
Table E. Robustness check - controlling for N/A%
(a) N/As count per supply chain j(b) N/A% per supply chain j
Mean: 97.89
Std Dev: 75.29
Mean: 32%
Std Dev 7%
Coefficient Std Errors P-values Coefficient Std Errors P-values
SCDensity 0.537 0.114 0.001 0.547 0.091 0.000
SCClustering -0.225 0.111 0.070 -0.352 0.090 0.003
SCGeographicalH 0.381 0.070 0.000 0.181 0.079 0.045
SCIndustrialH 0.004 0.063 0.956 0.025 0.057 0.665
SCHorizontal -0.024 0.040 0.564 -0.059 0.043 0.200
SCVertical -0.009 0.085 0.916 -0.067 0.099 0.517
SCOrgSize 0.306 0.077 0.003 0.179 0.056 0.009
IndustryClock -0.050 0.296 0.870 -0.064 0.297 0.835
Collectivism America -0.284 0.161 0.109 -0.380 0.170 0.050
Collectivism EU -0.423 0.111 0.003 -0.472 0.122 0.003
FFDisclosure 0.093 0.048 0.084 0.065 0.043 0.157
End-Market Distance 0.098 0.081 0.254 0.066 0.061 0.306
IndustryRepR isk 0.033 0.082 0.701 0.022 0.092 0.814
SCRegP ressure 0.645 0.095 0.000 0.605 0.066 0.000
N/A% -0.328 0.061 0.000 -0.339 0.085 0.003
Constan t 0.064 0.443 0.887 0.232 0.428 0.600
Observations 187 187
(pseudo) R-squared 0.713 0.615
(15a) Disclosure_ESG
(15b) Disclosure_ESG
Set N
Sub-set G
52
REFERENCES
Abadie, A., Athey, S., Imbens, G., Wooldridge, J. (2017). When Should You Adjust Standard
Errors for Clustering? MIT Press. Available at https://economics.mit.edu/files/13927
Ahnert, S. E., & Fink, T. M. (2008). Clustering signatures classify directed networks. Physical
Review E, 78(3), 1-5.
Akkermans, H., Bogerd, P., & van Doremalen, J. 2004. Travail, transparency and trust: A case
study of computer-supported collaborative supply chain planning in high-tech electronics.
European Journal of Operational Research, 153, 445-456.
Bansal, T., Gualandris, J., Kim, N. (2020). Theorizing Supply Chains with Qualitative Big Data
and Topic Modeling. Journal of Supply Chain Management, 56(2), 7-18
Bartko, J. (1976). On various intraclass correlation reliability coefficients. Psycological Bulletin,
83, 762-765.
Basole, R. C., Ghosh, S., and Hora, M. S. (2017). Supply network structure and firm performance:
Evidence from the electronics industry. IEEE Transactions on Engineering Management,
65(1), 141-154.
Bateman, A., and Bonanni, L. (2019). What supply chain transparency really means. Harvard
Business Review.
Battaglia, M., Bianchi, L., Frey, M., & Iraldo, F. (2010). An innovative model to promote CSR
among SMEs operating in industrial clusters: Evidence from an EU project. Corporate social
responsibility and environmental management, 17(3), 133-141.
Bellamy, M. A., Ghosh, S., & Hora, M. (2014). The influence of supply network structure on firm
innovation. Journal of Operations Management, 32(6), 357–373.
Bellamy, M., Dhanorkar, S., and Subramanian, R. (2020). Administrative environmental
innovations, supply network structure, and environmental disclosure. Journal of Operations
Management, 895-932.
Bimber, B., Flanagin, A. and Stohl, C. (2005). Reconceptualizing collective action in the
contemporary media environment. Communication Theory, 15 (4), 365-388.
Blanco, C., Caro, F., & Corbett, C. J. (2016). The state of supply chain carbon footprinting:
analysis of CDP disclosures by US firms. Journal of Cleaner Production, 135, 1189-1197.
Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications
for data aggregation and analysis. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel
theory, research, and methods in organizations: Foundations, extensions, and new
directions (p. 349–381). Jossey-Bass.
Bloomberg (2011). Supply Chain on Bloomberg. Retrieved December 7, 2018, from
https://business.library.emory.edu/documents/faq-handouts/bloomberg-splc.pdf.
Bloomberg (2013). Bloomberg Supply Chain Algorithm: Providing insight into a company
relationships. Retrieved December 7, 2018, from https://kenan-
flagler.instructure.com/files/54372815.
Bloomberg ESG (2018). Environmental, Social & Governance (ESG) product. Retrieved
October 15, 2020, from https://data.bloomberglp.com/professional/sites/10/1148330431.pdf.
Bode, C., & Wagner, S. M. (2015). Structural drivers of upstream supply chain complexity and
the frequency of supply chain disruptions. Journal of Operations Management, 36, 215–228.
Borgatti, S. P., & Li, X. (2009). On social network analysis in a supply chain context. Journal of
Supply Chain Management, 45(2), 5–22.
Bové, A., & Swartz, S. (2016). Starting at the source: Sustainability in supply chains. McKinsey
Company.
Brandon-Jones, E., Squire, B., & Van Rossenberg, Y. G. (2015). The impact of supply base
complexity on disruptions and performance: the moderating effects of slack and visibility.
International Journal of Production Research, 53(22), 6903-6918.
Breusch, T. S., & Pagan, A. R. (1979). A simple test for heteroscedasticity and random coefficient
variation. Econometrica: Journal of the Econometric Society, 1287-1294.
Browning, T. R. (2015). Design structure matrix extensions and innovations: a survey and new
opportunities. IEEE Transactions on Engineering Management, 63(1), 27-52.
Burt, R. S. (2003). The social structure of competition. Networks in the knowledge economy, 13,
57-91.
53
Burt, R. S. (2002). The social capital of structural holes. The new economic sociology:
Developments in an emerging field, 148(90), 122.
Busse, C., Meinlschmidt, J. and Foerstl, K. (2017). Managing information processing needs in
global supply chains: A prerequisite to sustainable supply chain management. Journal of
Supply Chain Management, Vol. 53 No. 1, pp. 87-113.
Campbell, J. L. (2007). Why would corporations behave in socially responsible ways? An
institutional theory of corporate social responsibility. Academy of management Review, 32(3),
946-967.
Carnovale, S., & Yeniyurt, S. (2015). The role of ego network structure in facilitating ego network
innovations. Journal of Supply Chain Management, 51(2), 22–46.
Carter, C. R., Rogers, D. S., & Choi, T. Y. (2015). Toward the theory of the supply chain. Journal
of Supply Chain Management, 51(2), 89–97.
Carter, C.R. and Washispack, S. (2018). Mapping the Path Forward for Sustainable Supply Chain
Management: A Review of Reviews. Journal of Business Logistics, 39(4), 242-247.
Chen, S., Zhang, Q and Zhou, Y. (2018). Impact of supply chain transparency on sustainability
under NGO scrutiny. Production and Operations Management, 28(12), 1-21.
Choi, T. Y., & Krause, D. R. (2006). The supply base and its complexity: Implications for
transaction costs, risks, responsiveness, and innovation. Journal of Operations Management,
24(5), 637–652.
Coleman, J.S., (1990). Foundations of Social Theory. The Belknap Press of Harvard University
Press, Cambridge, Massachusetts.
Constant, D., Sproull, L., & Kiesler, S. (1996). The kindness of strangers: The usefulness of
electronic weak ties for technical advice. Organization Science, 7(2), 119-135.
Dahlmann, F., & Roehrich, J. K. (2019). Sustainable supply chain management and partner
engagement to manage climate change information. Business Strategy and the
Environment, 28(8), 1632-1647.
Dai, R., Liang, H., & Ng, L. (2020). Socially responsible corporate customers. Journal of
Financial Economics, https://doi.org/10.1016/j.jfineco.2020.01.003.
Darnall, N., Henriques, I., & Sadorsky, P. (2010). Adopting proactive environmental strategy:
The influence of stakeholders and firm size. Journal of Management Studies, 47(6), 1072–
1094.
De Neve, G. (2014). Fordism, flexible specialization and CSR: How Indian garment workers
critique neoliberal labour regimes. Ethnography, 15(2), 184-207.
De Villiers, C., & Marques, A. (2016). Corporate social responsibility, country-level
predispositions, and the consequences of choosing a level of disclosure. Accounting and
Business Research, 46(2), 167–195.
Dong, M. C., Liu, Z., Yu, Y., & Zheng, J.-H. (2015). Opportunism in Distribution Networks: The
Role of Network Embeddedness and Dependence. Production and Operations Management,
24(10), 1657–1670.
Dooley, K. J., Pathak, S. D., Kull, T. J., Wu, Z., Johnson, J., & Rabinovich, E. (2019). Process
network modularity, commonality, and greenhouse gas emissions. Journal of Operations
Management, 65(2), 93-113.
Dubbink, W., Graafland, J., & Van Liedekerke, L. (2008). CSR, transparency and the role of
intermediate organisations. Journal of Business Ethics, 82(2), 391–406.
Eccles, R. G., & Klimenko, S. (2019). The investor revolution. Harvard Business Review, 97(3),
106-116.
Enkel, E. and Gassmann, O. (2010). Creative imitation: exploring the case of cross‐industry
innovation. R&D Management, 40(3), 256-270.
Fagiolo, G. (2007). Clustering in complex directed networks. Physical Review E, 76(2), 1-16.
Fine, C. (1998). Clockspeed: Winning Industry Control in the Age of Temporary Advantage. New
York: Perseus Books.
Fontana, E. and Egels-Zandén, N. (2019). Influence of supplier collective behaviour on corporate
social responsibility in the Bangladeshi apparel supply chain. Journal of Business Ethics,
159(4),1047-1064.
54
Fonti, F., Maoret, M., & Whitbred, R. (2017). Free‐riding in multi‐party alliances: The role of
perceived alliance effectiveness and peers' collaboration in a research consortium. Strategic
Management Journal, 38(2), 363-383.
Foerstl, K., Reuter, C., Hartmann, E. and Blome, C., 2010. Managing supplier sustainability risks
in a dynamically changing environment—Sustainable supplier management in the chemical
industry. Journal of Purchasing and Supply Management, 16(2), pp.118-130.
Fu, J. S., & Shumate, M. (2016). Hyperlinks as institutionalized connective public goods for
collective action online. Journal of Computer-Mediated Communication, 21(4), 298-311.
Gao, G., Xie, E., Zhou, K. (2015). How does technological diversity in supplier network drive
buyer innovation? Relational process and contingencies. Journal of Operations Management,
36, 165-177.
Gereffi, G., Humphrey, J., & Sturgeon, T. (2005). The governance of global value chains. Review
of international political economy, 12(1), 78-104.
Golini, R., & Gualandris, J. (2018). An empirical examination of the relationship between
globalization, integration and sustainable innovation within manufacturing
networks. International Journal of Operations & Production Management, 38(3), 874-894.
Gorsen, M.F., and Sandick, P. (2020), Supply Chain Managers Must Prepare for New Mandatory
EU Rules on Human Rights Due Diligence. Alston & Bird. Available at:
https://www.alston.com/en/insights/publications/2020/07/supply-chain-managers-must-
prepare.
Gould, R. V. (1993). Collective action and network structure. American Sociological Review,
182–196.
Greene, W.H. and Zhang, C. (2003). Econometric Analysis, Prentice Hall Upper Saddle River,
NJ.
Gualandris, J., Klassen, R., Vachon, S., Kalchschmidt, M., (2015). Sustainable evaluation and
verification in supply chains: Aligning and leveraging accountability to stakeholders. Journal
of Operations Management, 38, 1-13.
Gualandris, J., Barg, J., 2020. Huawei: Struggling to Develop a More Sustainable Supply
Network. Ontario: Ivey Business School, University of Western Ontario.
Gulati, R., Gargiulo, M., 1999. Where do interorganizational networks come from? American
Journal of Sociology, 104 (5), 1439–1493.
Heckathorn, D. D. (1993) ‘Collective Action and Group Heterogeneity: Voluntary Provision
Versus Selective Incentives’, American Sociological Review, 58(3): 329–50.
Jira, C. (Fern), & Toffel, M. W. (2013). Engaging Supply Chains in Climate Change.
Manufacturing & Service Operations Management, 15(4), 559–577.
Kalkanci, B., & Plambeck, E. L. (2020). Managing Supplier Social and Environmental Impacts
with Voluntary Versus Mandatory Disclosure to Investors. Management Science.
Kim, Y. and Davis, G. F. (2016). Challenges for global supply chain sustainability: Evidence
from conflict minerals reports. Academy of Management Journal, 59(6), 1876–1916.
Kim, Y., Choi, T. Y., Yan, T., & Dooley, K. (2011). Structural investigation of supply networks:
A social network analysis approach. Journal of Operations Management, 29(3), 194–211.
Kim, Y., Chen, Y. S., & Linderman, K. (2015). Supply network disruption and resilience: A
network structural perspective. Journal of operations Management, 33, 43-59.
Knorringa, P., & Nadvi, K. (2016). Rising power clusters and the challenges of local and global
standards. Journal of Business Ethics, 133(1), 55-72.
Kolaczyk, E. D., & Csárdi, G. (2014). Statistical analysis of network data with R, New York:
Springer.
Kölbel, J.F., Busch, T. and Jancso, L.M., (2017). How media coverage of corporate social
irresponsibility increases financial risk. Strategic Management Journal, 38(11), 2266-2284.
Lee, B. H., Struben, J., & Bingham, C. B. (2018). Collective action and market formation: An
integrative framework. Strategic Management Journal, 39(1), 242-266.
Leven, P., Holmström, J. and Mathiassen, L. (2014). Managing research and innovation networks:
Evidence from a government sponsored cross-industry program. Research Policy, 43(1), 156-
168.
55
Lu, G., & Shang, G. (2017). Impact of supply base structural complexity on financial
performance: Roles of visible and not-so-visible characteristics. Journal of Operations
Management, 53, 23–44.
Lu, G., Ding, X. D., Peng, D. X., & Chuang, H. H. C. (2018). Addressing endogeneity in
operations management research: Recent developments, common problems, and directions for
future research. Journal of Operations Management, 64, 53-64.
Marques, L., Tingting, Y., and Matthews, L. (2019). Knowledge diffusion in a global supply
network: a network of practice view. Journal of Supply Chain Management, 56(1), 33-53.
Marsden, P. V. (2002). Egocentric and sociocentric measures of network centrality. Social
Networks, 24(4), 407-422.
Marshall, D., McCarthy, L., McGrath, P., & Harrigan, F. (2016). What’s Your Strategy for Supply
Chain Disclosure? MIT Sloan Management Review, 12, 1-14.
McGraw, K. O., & Wong, S. P. (1996). Forming inferences about some intraclass correlation
coefficients. Psychological methods, 1(1), 30.
Min, S., Kim, S. K., & Chen, H. (2008). Developing social identity and social capital for supply
chain management. Journal of Business Logistics, 29(1), 283-304.
MIT Center for Transportation & Logistics and Council of Supply Chain Management
Professionals (2020). State of Supply Chain Sustainability 2020. Cambridge, MA.
Nadvi K (2008) Global standards, global governance and the organization of global value chains.
Journal of Economic Geography 8(3): 323–343.
New, S. (2010). The transparent supply chain. Harvard Business Review, 88(10), 11.
Niforou, C. (2015). Labour leverage in global value chains: The role of interdependencies and
multi-level dynamics. Journal of Business Ethics, 130(2), 301-311.
Osadchiy, N., Gaur, V., & Seshadri, S. (2016). Systematic risk in supply chain networks.
Management Science, 62(6), 1755–1777.
Ostrom, E. (1990). Governing the commons. New York: Cambridge University Press.
Ostrom, E. (1998). A behavioral approach to the rational choice theory of collective action:
Presidential address, American Political Science Association, 1997. American political
science review, 1-22.
Ostrom, E., & Walker, J. (Eds.). (2003). Trust and reciprocity: Interdisciplinary lessons for
experimental research. Russell Sage Foundation.
Oyserman, D. (2011). Culture as situated cognition: Cultural mindsets, cultural fluency, and
meaning making. European review of social psychology, 22(1), 164-214
Park, H., Bellamy, M. and Basole, R. (2018). Structural anatomy and evolution of supply chain
alliance networks: A multi-method approach. Journal of Operations Management, 63, 79-96.
Parmigiani, A., Klassen, R. D., & Russo, M. V. (2011). Efficiency meets accountability:
Performance implications of supply chain configuration, control, and capabilities. Journal of
operations management, 29(3), 212-223.
Pathak, S. D., Wu, Z., & Johnston, D. (2014). Toward a structural view of co-opetition in supply
networks. Journal of Operations Management, 32(5), 254–267.
Poteete, A. and Ostrom, E. (2004). Heterogeneity, group size and collective action: the role of
institutions in forest management, Development & Change, 35(3), 435-461.
Puppim de Oliveira, J. A., & de Oliveira Cerqueira Fortes, P. J. (2014). Global value chains and
social upgrading of clusters: Lessons from two cases of fair trade in the Brazilian
northeast. Competition & Change, 18(4), 365-381.
Qiu, Y., Shaukat, A., & Tharyan, R. (2016). Environmental and social disclosures: Link with
corporate financial performance. The British Accounting Review, 48(1), 102-116.
RepRisk (2016). RepRisk Scope, Process and Metrics. Available at https://www.reprisk.com/our-
approach#risk-metrics. Accessed 11 Dec 2018.
Reverte, C. (2009). Determinants of corporate social responsibility disclosure ratings by Spanish
listed firms. Journal of Business Ethics, 88(2), 351–366.
Rice, J. A. (2006). Mathematical statistics and data analysis. Cengage Learning
Rivera, M. T., Soderstrom, S. B., & Uzzi, B. (2010). Dynamics of dyads in social networks:
Assortative, relational, and proximity mechanisms. Annual Review of Sociology, 36(1), 91-
115.
56
Ranganathan, R., Ghosh, A., & Rosenkopf, L. (2018). Competition–cooperation interplay during
multifirm technology coordination: The effect of firm heterogeneity on conflict and consensus
in a technology standards organization. Strategic Management Journal, 39(12), 3193-3221.
Roberts, S. (2003). Supply chain specific? Understanding the patchy success of ethical sourcing
initiatives. Journal of business ethics, 44(2-3), 159-170.
Rodan, S., & Galunic, C. (2004). More than network structure: How knowledge heterogeneity
influences managerial performance and innovativeness. Strategic management journal, 25(6),
541-562.
Sammarra, A., & Biggiero, L. (2008). Heterogeneity and specificity of Inter‐Firm knowledge
flows in innovation networks. Journal of management studies, 45(4), 800-829.
Sampson, S. E., & Froehle, C. M. (2006). Foundations and implications of a proposed unified
services theory. Production and Operations Management, 15(2), 329-343.
Schilling, M. and Phelps, C. (2007). Interfirm collaboration networks: The impact of large-scale
network structure on firm innovation. Management Science, 53(7), 1113-1126.
Schmidt, C. G., Foerstl, K., & Schaltenbrand, B. (2017). The supply chain position paradox: green
practices and firm performance. Journal of Supply Chain Management, 53(1), 3-25.
Schmitz, H. & Nadvi, K. 1999. Clustering and industrialization: Introduction. World
Development, 27(9), 1503–14.
Scott, M. (2020). Investors step up Pressure on companies that don’t disclose Environmental
Risks. Forbes.com. Available at:
https://www.forbes.com/sites/mikescott/2020/06/09/investors-step-up-pressure-on-
companies-that-dont-disclose-environmental-risks/#6532695a727f.
Serpa, J. C., & Krishnan, H. (2018). The impact of supply chains on firm-level
productivity. Management Science, 64(2), 511-532.
Shao, B. B., Shi, Z. M., Choi, T. Y., & Chae, S. 2018. A data analytics approach to identifying
hidden critical suppliers in supply networks: Development of nexus supplier index. Decision
Support Systems, 114, 37–48
Sharma, A., Pathak, S., Borah, S. B., & Adhikary, A. (2019a). Is it too complex? The curious case
of supply network complexity and focal firm innovation. Journal of Operations Management.
Sharma, A., Kumar, V.; Yan, J., Borah, S. and Adhikary, A. (2019b). Understanding the structural
characteristics of a firm’s whole buyer–supplier network and its impact on international
business performance. Journal of International Business Studies, 50(3), 365-392.
Sinha, I. 2000. Cost transparency: The net’s real threat to prices and brands. Harvard Business
Review, March–April issues.
Sodhi, M. and Tang, C. (2019). Research Opportunities in Supply Chain Transparency.
Production & Operations Management, 28(12), 2946-2959.
Swift, C., Guide Jr, V.D.R. and Muthulingam, S. (2019). Does supply chain visibility affect
operating performance? Evidence from conflict minerals disclosures. Journal of Operations
Management. 65(2), 406-429.
Tapscott, D. (2003). The naked corporation: How the age of transpar- ency will revolutionize
business. Vol. 128. New York: Free Press.
Tiessen, J. H. (1997). Individualism, collectivism, and entrepreneurship: A framework for
international comparative research. Journal of Business Venturing, 12(5), 367-384.
Valentine, J., Sholem, M., and Smith, C. (2020). EU Parliament Adopts Sustainability Taxonomy
Regulation to Fight Greenwashing. National Law Review, X(174).
Van Den Brink, T. W., & van Der Woerd, F. (2004). Industry specific sustainability benchmarks:
An ECSF pilot bridging corporate sustainability with social responsible investments. Journal
of Business Ethics, 55(2), 187–203.
Villena, V. H., & Dhanorkar, S. (2020). How institutional pressures and managerial incentives
elicit carbon transparency in global supply chains. Journal of Operations Management, 1-38.
Villena, V. H., & Gioia, D. A. (2018). On the riskiness of lower-tier suppliers: Managing
sustainability in supply networks. Journal of Operations Management, 64, 65-87
Vurro, C., Russo, A., & Perrini, F. (2009). Shaping sustainable value chains: Network
determinants of supply chain governance models. Journal of business ethics, 90(4), 607-621.
57
Wenger, E., and Snyder, W. (2000). Communities of practice: the organizational frontier.
Harvard Business Review, 78(1), 139-145.
White, H. 1980. A heteroskedasticity-consistent covariance matrix estimator and a direct test for
heteroskedasticity. Econometrica 48: 817-838.
Wooldridge, J. M. (2015). Introductory econometrics: A modern approach. Nelson Education.
World Bank. (2016). Available at: https://datacatalog.worldbank.org/dataset/worldwide-
governance-indicators. Accessed 12 December 2018.
Yan, T., Choi, T. Y., Kim, Y., & Yang, Y. (2015). A theory of the nexus supplier: A critical
supplier from a network perspective. Journal of Supply Chain Management, 51(1), 52–66.
Young, G. J., Charns, M. P., & Shortell, S. M. (2001). Top manager and network effects on the
adoption of innovative management practices: A study of TQM in a public hospital system.
Strategic Management Journal, 22(10), 935-951.
