Scholarly article on topic 'Analyzing networks in industrial ecology – a review of Social-Material Network Analyses'

Analyzing networks in industrial ecology – a review of Social-Material Network Analyses Academic research paper on "Sociology"

CC BY
0
0
Share paper
Academic journal
Journal of Cleaner Production
OECD Field of science
Keywords
{"Network analysis" / "Inter-organizational relations" / "Industrial ecology" / "Industrial symbiosis" / Embeddedness / Proximity}

Abstract of research paper on Sociology, author of scientific article — Frank Schiller, Alexandra S. Penn, Lauren Basson

Abstract This review concerns the methodological challenges that industrial ecology faces in integrating natural and social sciences. Network analysis can be seen as the most promising method to mediate between industrial ecology's overall systems approach and the complex structures found in society. It is a well established method across scientific disciplines, including the social sciences. It has been successfully applied in industrial ecology, in which localized phenomena of industrial symbiosis have been a key focus, and where metrics from both the social and natural sciences are used to understand socio-metabolic structures. In this paper we classify such studies as Social-Material Network Analyses and we discuss the body of work, drawing on network analyses from various disciplines. A challenge is the hierarchical nature of industrial networks and how it can be addressed socially. We discuss the opportunities and limitations of metric-driven network analysis and offer a review of methodological options for Social-Material Network Analyses.

Academic research paper on topic "Analyzing networks in industrial ecology – a review of Social-Material Network Analyses"

Accepted Manuscript

Analyzing networks in industrial ecology - a review of Social-Material Network Analyses

Frank Schiller, Alexandra S. Penn, Lauren Basson

PII: S0959-6526(14)00250-9

DOI: 10.1016/j.jclepro.2014.03.029

Reference: JCLP 4136

To appear in: Journal of Cleaner Production

Received Date: 24 August 2012

Revised Date: 4 March 2014

Accepted Date: 10 March 2014

Please cite this article as: Schiller F, Penn AS, Basson L, Analyzing networks in industrial ecology -a review of Social-Material Network Analyses, Journal of Cleaner Production (2014), doi: 10.1016/ j.jclepro.2014.03.029.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Analyzing networks in industrial ecology - a review of Social-Material Network Analyses

1 12 3

Authors: Frank Schiller , Alexandra S. Penn ' , Lauren Basson

1 University of Surrey, Evolution and Resilience of Industrial Ecosystems

2 University of Surrey, Center for Environmental Strategy

3 The Green House, South Africa

Corresponding author

Dr. Frank Schiller University of Surrey

Department of Sociology, Evolution and Resilience of Industrial Ecosystems - ERIE

37BC02

Guildford,

Surrey GU2 7HX

Abstract

This review concerns the methodological challenges that industrial ecology faces in integrating natural and social sciences. Network analysis can be seen as the most promising method to mediate between industrial ecology's overall systems approach and the complex structures found in society. It is a well established method across scientific disciplines, including the social sciences. It has been successfully applied in industrial ecology, in which localized phenomena of industrial symbiosis have been a key focus, and where metrics from both the social and natural sciences are used to understand socio-metabolic structures. In this paper we classify such studies as Social-Material Network Analyses and we discuss the body of work, drawing on network analyses from various disciplines. A challenge is the hierarchical nature of industrial networks and how it can be addressed socially. We discuss the opportunities and limitations of metric-driven network analysis and offer a review of methodological options for Social-Material Network Analyses.

Keywords: network analysis; inter-organizational relations; industrial ecology; industrial symbiosis; embeddedness; proximity.

1. Introduction - industrial ecosystems in social context

The field of industrial ecology can be considered to be "the study of the flows of materials and energy in industrial and consumer activities, of the effect of these flows on the environment, and of the influence of economic, political, regulatory and social factors on the flow, use and transformation of resources" (White 1994, p.v.) As is recognized in the field (e.g. Fischer-Kowalski and Weisz 1999), this requires an understanding of both the physical and the social realms and, significantly, their complex interrelations. The study of society's metabolism within nature has not only seen a variety of different material flow analysis methods emerging, but also different methods covering the social aspects of industrial metabolism (see Binder 2007a).

This is not surprising in the face of plural motivations for analyzing flows as well as the different social contexts in which these flows manifest. What makes industrial ecology challenging is that the universal laws of the natural sciences meet the particulars of the social world, which create and maintain the social metabolism. Industrial ecology's analytical studies have tended to cut right through the complexity of the social reality by focusing on the unsustainable levels of industrialized societies' physical metabolism (cf. Rockstrom et al. 2009). Its research designs are commonly understood as responses to 'wicked' problems, which are difficult if not impossible to solve because of incomplete, sometimes contradictory, and ever evolving new requirements (Rittel and Webber 1973).

In the past industrial ecology's methodological discussions have often taken place elsewhere (Fischer-Kowalski and Weisz, 1999; Hoffman, 2003; Kronenberg 2006). Several authors have referred to ecological economics as industrial ecology's twin discipline suggesting that it would offer clues for industrial ecology's own methodology (e.g. Kronenberg 2006; van den Bergh and Janssen 2004). As distinct from the epistemological debates in ecological economics, however, industrial ecologists have so far addressed the wicked problems pragmatically, e.g. as boundary setting problem of metabolic systems. Setting the boundary for analysis does not necessarily occur without reflection e.g. it is acknowledged that "it is the investigator, not the system, that in most cases creates the (necessarily contingent) boundaries" (Allenby 2006: 30).

Given the wicked nature of the problems in industrial ecology and their different social contexts, existing methodology to integrate physical accounting and social research methods represents a considerable achievement (Daniels and Moore 2002). Some researchers have addressed specific metabolic phenomena such as industrial symbiosis and have applied different social research methods to study it (e.g. Boons and Baas 1997; Chertow 2007; Boons and Howard-Granville 2009). Yet these studies have been criticized for being only descriptive in nature, neither producing hypotheses nor testing theory (Ehrenfeld 2009). Indeed, from a methodological point of view most of these studies represent case studies, which bring about specific challenges such as generalizing from small samples (see Ragin 2008).

This review focuses on network analysis as a unifying method capable of integrating these different strands of research. We identified relevant articles by searching appropriate databases for the terms 'industrial symbiosis' or 'eco-industrial park' or 'network analysis'.1 This resulted in a total of 191 hits. Analyzing the content of these articles and consulting network analytical research from other fields and disciplines led us to identify several methodological challenges for industrial ecology.

These are a) to integrate the industrial ecology community across the natural and social scientific interface; b) to acknowledge the epistemological differences between these two ontological realms (social and natural); and c) to establish common ground with related disciplines in the social sciences to increase the chances of successful implementation (methodological reflexivity). It could be argued that the last aspect has already taken off with the rise of new, widely recognized research agendas such as sustainable consumption, eco-efficient production systems and others; yet industrial ecology has thus far failed to penetrate the social sciences in the same way as for instance ecological economics has done.

In this review we propose that network analysis indeed represents the most promising method to integrate industrial ecology across disciplinary lines (challenge a); and by drawing on an existing review framework of social network analyses we derive an analytical framework for Social-Material Network Analyses. It acknowledges the differences between the social and natural realms by enabling the intentional (re)design of more eco- and resource efficient industrial networks (challenge b). To this end we selectively draw on network research from geography, sociology, economics, political science and interdisciplinary fields such as transition management (challenge c). We exemplify its potential by addressing the hypothesis of self-organizing industrial symbiosis (Chertow and Ehrenfeld 2012).

The literature implicitly or explicitly claims that network analysis can facilitate integrated analyses of material/energy and emissions flows and the social realm. Addressing challenge a) we discuss the method in Section 2 and argue that it could indeed become a core method for industrial ecology since it assimilates the natural and the social sciences and is well established in both camps. Here we also introduce the analytical framework to signify Social-Material Network Analyses, which touches upon aspect b. Extending the argument, we discuss in how far Social-Material Network Analyses can rely on evolutionary social theory by drawing on related research in geography and economics. Another facet of aspect b relates to the different social dimensions of industrial networks. This has been addressed as the embeddedness of industrial networks in our field and as proximity in geography and elsewhere. Social embeddedness forms a common theme of social network analytical research in industrial ecology with 9 % of all

1 The databases included Google Scholar, Scopus; Science Direct, Web of Knowledge, and Sustainability Science Abstracts and the keywords were constrained to English.

papers reviewed utilizing this framework. We compare both concepts in Section 3. Different metrics have been deployed to analyze the structure of industrial networks, most of them social scientific in origin. We scrutinize these and extend their perspective by looking at ecological metrics that have been utilized in complexity economics (Section 4). In Section 5 we discuss the potential and limitations of network analysis from a methodological point of view (aspect c). From our review we conclude in Section 6 that Social-Material Network Analyses are developing towards more consistent bottom-up studies for industrial ecology and convergence with other disciplines inside and outside this field.

2. Networks of organizations and flows

Network analysis has been a well-established method in the social sciences for several decades (see Scott 2000 for an overview). Social network analysis was developed to study the relational aspects of social structures. The unit of analysis has been a cluster of individuals or organizations and the linkages among them (Easley and Kleinberg 2010; Kadushin 2012). The method has been applied to the study of groups and organizations with the majority of the empirical studies devoted to dyadic relationships. Over the past decades social network analysis has developed from a descriptive approach metaphorically referring to networks, to a highly analytical method which studies networks by, for instance, means of graph and game theory (Jackson 2008). However, social scientists have yet to show that the social construction of a networked world endorses the transition to sustainability because endorsing these meso-level structures may also give rise to particular macro-level feed-backs.

Positivistic design perspectives are oblivious to the character of analytical reductions in empirical social networks. In the social science reducing empirical phenomena follows methodological criteria, which are highly dependent on the context and are therefore generally domain-specific. A network such as the World Wide Web is different from global container shipping or the mafia not because of its structure (which is indeed the same in the first two cases) but because of the normative and epistemic claims we raise when analyzing the world (Habermas 2003). In order to do justice to the complex and normative character of the social world Gluckler (2007) has suggested that the dyadic tie formation should always be the unit of analysis, whereas the network structure should be the object of our epistemic interest.

The nodes in a network can be connected by various types of ties at the same time including amongst others material and energy flows, financial transactions, information, and social interaction. The property of any two or more actors to have several ties at the same time is called multiplexity (Borgatti et al. 2009). Unlike physical or chemical relationships human relationships are generally multi-dimensional (Watts 2004): firms compete in the market place, collaborate in business associations, and communicate with consumers - all concurrently. Despite industrial ecology's primary interest in the functional ties that establish the metabolism of the network, it also needs to consider indirect social influences, e.g. research institutes

spreading knowledge, banks handing out loans or regulators introducing new regulation. Thus, besides functional flows, social relationships between different forms of organizations, which constitute the nodes of the network, are important too. The question is whether these complex socio-metabolic interactions can be integrated into analyses?

FIGURE 1 near here

Figure 1 represents an adaptation and extension of the work of Borgatti et al. (2009) to industrial ecology. At the top level of the schema are nodes and ties. The properties of nodes are listed on the left hand side. Ties are shown to branch into two categories: continuous ties, which operate all the time, and discontinuous ties that are either on or off. One level below properties refers to nodes while ties fall into four different categories: similarities between the subjects of a network, their relationships, their interactions, and their transmissions. Similarities signify contexts that increase the chances for dyadic ties to form. This includes time & space, sectors, business associations, and energy flows, e.g. through the exchange of skills and knowledge. Next, relationships refer to continuous dyadic links such as supplier/purchaser, lender/borrower, regulator/regulatee, management/skill. Interactions include purchases, legal contracts, personal contacts and actor constellations. Transmissions comprise knowledge, payments, technology, material flows including those of recyclates and by-products. Figure 1 illustrates that industrial networks under study will be constituted by multiple layers of physical and social relations; one implication of this is that we should consider the whole industrial network as always consisting of continuous energy flows (supplier/purchaser relationship) while individual layers may have specific characteristics such as symbiosis forming new dyads (by-products under transmissions) that may, for instance, correspond to personal relationships between managers in another dimension. We consider any analysis within the above framework studying these node properties in conjunction with any of these ties a Social-Material Network Analysis. Implicit to Borgatti's schema is the assumption of bounded rationality whereas the communicative openness of the social world is reflected in the particularities of any empirical context (Habermas 2003).

Compared with the generic network schema suggested by Borgatti et al. (2009) the adaptation in figure 1 sees some of the similarities of links and nodes as "pre-social" in the sense that we are born into them. The irreversible, dissipative flows constitute metabolic relationships in time and space (Binswanger 1993), which are produced and reproduced by social relationships extending beyond the immediate realm of necessity including for instance knowledge and legal contracts. Without these, anthropogenic flows would not exist, which is why space & (synchronous) time appear under similarities while material & energy inputs, waste & emission

output, net-addition to stock and organizational form signify the properties of nodes. Industrial ecology studies those social relationships that affect the metabolism. Thus, the relations to the right of figure 1 (relationships, interactions) stand for active social relationships that effect the metabolism while transmissions describe the content that is passed on between dyads. Node properties as well as the presence or absence of any social link are an empirical question. Thus, the schema requires empirical specification and it is not necessarily exhaustive. We will present examples of these relationships as we go along.

Rather than defining ties as pre-social, several disciplines including economics have come to analyze social interactions as evolutionary processes. The mechanisms behind evolutionary explanations are selection, retention (continuity) and variation at firm level (Nelson and Winter 1982). Unlike evolutionary economics where the selection of firms is a consequence of exogenous market competition, evolutionary geography suggests that selection emerges from the formation of ties (Glückler 2007; Boschma and Frenken 2011). This conforms to industrial ecology's epistemic perspective, in particular that of industrial symbiosis (Chertow 2007; Lombardi and Laybourn 2012). These evolutionary mechanisms are also pertinent when industrial ecologists claim an endogenous drive of these networks towards sustainability (Chertow and Ehrenfeld 2012). What is the empirical evidence supporting this hypothesis?

Selection of ties may result from the corporate socially responsible choice of firms in (reverse) supply chain management including industrial symbiosis close to the top of the waste hierarchy (Bansal and McKnight 2009).2 Selection of ties and nodes may also result from competition, accelerated by pecuniary externalities of technological and organizational choice (Rennings 2000) or price risks related to the specific material input portfolio (Busch and Hoffmann 2007). However, selection is also exogenously enforced in industrial ecosystems to internalize ecological effects in markets via the price system. Retention in industrial networks is constituted by the profitability of firms in the market (Jackson and Clift 1998), their location in space (Ter Wal and Boschma 2011), and their capability to (eco)-innovate (Esty and Porter 1998). Variation, by which firms distinguish themselves from others, arises from actual product and process innovation increasing resource-, energy- or eco-efficiency - and ideally all three (Horbach, Rammer, and Rennings 2012). Still, competition, increasing resource and sink constraints and formal institutions implementing these anticipated constraints constantly reduce variation.

Several areas can be identified in industrial ecology developing this research agenda, including industrial symbiosis, (reverse) supply chain management, eco-industrial parks and distribution research amongst others. These fields have introduced new methods and theories from other disciplines where they were considered coherent with the epistemic perspective of industrial ecology. Thus, industrial ecology's evolutionary perspective is consistent for instance with open

2 Reverse supply chains or reverse logistics are supply operations for recycling materials and products.

system theories in organisational studies (Freeman, Audia 2006), the evolutionary perspective of ecological economics (van den Bergh and Gowdy 2000), or transition management (Safarzynska, Frenken, and van den Bergh 2012). In contrast, new economic and evolutionary geography, and also complexity economics have gone relatively unnoticed in the field.

3. Explaining network growth

The approach proposed here is geared towards explaining endogenous network growth as material/energetic relationships emerging from the (dyadic) interaction of firms (nodes) and trade (ties). It allows studying the diffusion and impact of eco-innovations in these networks. Producing companies interact in upward and downward relationships with suppliers and purchasers (White 2004). Some of the resulting dynamics such as competition or price volatility of recyclates are endogenously driven. Other effects are exogenously brought about by hierarchical interactions including for instance the interaction with financial markets or governing institutions. Research on industrial symbiosis has mostly framed this interaction with the help of the embeddedness concept, which relates structural analysis to the wider social context (Boons and Howard-Granville 2009). A related concept which has not yet been considered in industrial ecology, is the proximity approach that has become important in geography (Boschma 2005). This approach focuses on the social enablers of endogenous network growth (see Jones at al. 1997).

The embeddedness and proximity approaches to industrial network analysis both recognize five social dimensions: spatial, cognitive, social, organizational, and institutional (Shaw and Gilly 2000; Boschma 2005; Boons and Howard-Grenville 2009). Unlike the embeddedness approach the proximity approach does not consider culture. The overall convergence between the approaches may cause surprise given the structural complexity of social networks, the usual incompleteness of empirical data and such dissimilar social contexts that they have often given rise to different disciplines. Despite this the dimensions have proven remarkably stable. A methodological difference between the approaches is that Boons and Howard-Grenville (2009) argue that the dimensions cannot be analytically distinguished. By contrast the proximity concept has insisted that the dimensions should be kept analytically distinct and each dimension should be analyzed independently (Boschma 2005).

Most social network analyses in industrial ecology try to demonstrate how social networks facilitate the growth of industrial ecosystems in the face of market structures which are increasing metabolic throughput (Chertow, Ashton, and Espinosa 2008; Ashton 2008; Paquin and Howard-Grenville in press). The research perspective is that of firms adapting to changes in their social environment, and endogenously emerging sustainable industrial ecosystems. To succeed with this agenda we need to specify which social dimensions support endogenous network growth. In order to keep the argument straightforward we will focus on industrial symbiosis.

Spatial dimension of industrial networks

Many conceptual articles and empirical studies in the field have addressed the spatial embeddedness of material and energy flows (see fig.1: similarities). In particular, the anchor tenant approach (Korhonen and Snakin 2001) and the concept of eco-industrial parks (Chertow 1998; Chertow et al. 2008; Shi et al. 2010; Zhang et al. 2010; Tudor et al. 2007) have used the idea of co-location to promote cascading resource use and industrial symbioses between companies and municipalities. Furthermore, regional symbiosis has been observed to arise in numerous locations (Sterr and Ott 2004; Wells and Bristow 2007; Lyons 2008; Jensen et al. 2011). Industrial ecology should perhaps already have come to use spatial metrics - not least because GPS data is readily available. Yet spatial integration of physical flows is not straightforward because it may be material-specific (e.g. Weiss et al. 2007), or challenge simple distance minimization assumptions because the combined effects may represent a lower throughput (e.g. Frohling et al. 2012).

Social-Material Network Analyses might take up this research direction by locating the nodes (firms) in space and by explaining physical (e.g. transport) and social tie formation (e.g. knowledge) as corresponding dependent and enabling processes that may involve path dependencies that are not simply economic. Indeed, at times development might be place-dependent (Martin and Sunley 2006). Clusters can be self-reproducing despite absent or negative localization economies if other dimensions, such as social or institutional, are well-aligned (Boschma and Frenken 2011). Spatial proximity has generally been studied by new economic and evolutionary geography. The predictive power of geography's methods, theories and models can for instance help to identify opportunities for industrial symbiosis by identifying niches of wasted resources (Schiller et al. forthcoming).

The cognitive dimension of industrial networks

While a variety of theories of individual rationality exist3 Social-Material Network Analyses should be concerned with organizations and consider those as the micro-level of analysis

3 In agreement with most industrial ecologists Boons and Howard-Granville (2009) refer to bounded rationality when discussing cognitive embeddedness. Bounded rationality is associated with cognitively restricted agents producing self-organized systems in interaction. Yet, two other forms of rationality deserve equal attention in empirical research: rational choice and communicative rationality. Rational choice theory is rarely applied in industrial symbiosis although it is common in the (reverse) supply chain literature and environmental economics. Many symbiotic relationships could be designed as cooperative games with specific pay-off structures for the individual participants. Developing game theoretical models for different sectors, numbers of actors and cooperative benefits could directly inform industrial symbiosis in practice, if requested, all on the premise of available win-win situations. More recently Ehrenfeld (2009) has proposed to look at communicative rationality (Habermas 2003), which is concerned with the evolvability of society. These rationality theories complement each other, e.g. the strategic rationality of game theory is a boundary case of communicative rationality and likewise game theory has been integrated into complex adaptive system research usually associated with bounded rationality whilst bounded rationality implies communicative rationality.

(Jacobsen and Anderberg 2005; Paquin and Howard-Grenville 2009). Evolutionary economics and transition management theory have suggested that individual cognition co-evolves along with markets (e.g. Nelson and Winter 1982; Foxon 2006). Interdependence between education, skills and capabilities has also been shown to exist for resource management (Bleischwitz 2003). Empirical studies suggest that cognitive frames and innovations are developing interdependently (hence: sectors, business associations under similarities in fig. 1). If these frames include environmental concerns, they appear to be generally broader than just industrial symbiosis at tie-level: they also include eco- and material efficiency at node level (Bleischwitz 2007).

The social dimension of industrial networks

Industrial symbiosis literatures suggest that trust is indispensible for functioning industrial ecosystems (Schwarz and Steininger 1997; Baas 2008; Baas and Boons 2004; Doménech and Davies 2011; Gibbs and Deutz 2005; Mirata 2004; Sterr and Ott 2004; Chertow, Ashton, and Espinosa 2008; Dlouhá 2013). Recent network analyses have therefore carefully looked at individuals to explain and demonstrate the build-up of motivation and trust in establishing symbiotic exchange networks (Ashton, 2008; 2009; Ashton and Bain 2012). However, in a large-scale survey of eco-industrial park implementation in the Netherlands, Boons and Spekkink (2012) found that only the ability to mobilize actors was crucial for delivering industrial symbiosis. According to this study, neither technical knowledge nor specific relationships could explain the uptake of industrial symbiosis. Furthermore, companies in symbiotic networks in Germany and Austria have been found to remain ignorant of their connection to a network (Posch 2010).

This lack of awareness (reflectivity) of firms' functional network role does not tell us about the role of trust in transmission of knowledge through interpersonal bonds. Some individuals might bridge structural holes between different groups of the whole networks (Burt 2004). Empirical innovation research is not conclusive on this matter. Whereas Zaheer and Bell (2005) demonstrate that firms which bridge structural holes tend to generate greater capability, Ahuja (2000) found that direct and indirect ties both matter for innovation. This concerns the structural side and it does not answer whether explicit knowledge would cease to diffuse and whether, if personal trust was absent, tacit knowledge would fade too. From this angle spreading knowledge in industrial networks could be studied in more subtle ways as tacit and explicit knowledge diffusion (Borgatti and Cross 2003). This might also be indicated since the codification of knowledge can influence cooperation in industrial symbiotic relationships, e.g. when data transparency alters relationships between firms (Fichtner et al. 2004). On the other hand, as was pointed out above, even existing symbiotic relationships may not endorse recognition of the whole network. This may suggest a greater role for public discourses and the free provision of codified knowledge, e.g. through internet-based matching tools (Grant et al. 2010) or quality standards for recyclates that reduce uncertainty and create trust.

The organizational dimension of industrial networks

The organizational dimension concerns the dyadic interaction of companies within networked market structures (Halinen et al. 1999). Distinct from existing business research studying relative scarcity, metabolic ties are also subject to absolute scarcity of biophysical resources.4 Much of industrial ecology is particularly concerned with the internalization of external environmental effects under conditions of market competition (see fig.1 relationships). Firms create and adapt to these dynamics or otherwise go bankrupt (van den Bergh and Gowdy 2000).

Resource and eco-efficiency measures, like industrial symbiosis, can be carried out at intra- or inter-firm network level. Win-win situations from symbiotic cooperation may exist between firms (Chertow and Lombardi 2005) that may outweigh intra-firm resource-efficiency measures (Chertow and Miyata 2011), and thus adding a dynamic, dyadic, triadic or even network-wide element to the decision-making of firms (Korhonen and Seager 2008; Boix et al. 2012; Hiete et al. 2012; Wassmer at al. 2012). Three patterns of industrial symbiotic cooperation have been observed (Fichtner et al. 2004): resource recovery networks without common investment, resource recovery networks with common investment, energy cascading networks as a specific form of inter-company cooperation with common investment. The individual context can be altered by innovative business models (Halme et al. 2007; Reiskin et al. 1999). This may require new models of finance (lender/borrower, fig. 1). In contrast to rather well-established efficiency services in the electricity sector (Bertoldi et al. 2006) resource-efficient services face a more difficult market environment despite substantial market potential.5

Because of the greater complexity of material flows compared to energy, services for resource efficiency and industrial symbiosis in particular often struggle to establish themselves. They are driven by ongoing dilemmas of value creation. According to Paquin and Howard-Grenville (in press) facilitators of industrial symbiosis follow an endogenous trajectory when promoting industrial symbiosis. In this account resources included data, information, ideas, potential connections as well as very specific connections to resources or partners. Yet neither these "resources" nor their central network position has made facilitators self-sustaining (Laybourn and Lombardi 2012). Transaction costs might simply be too high (e.g. because of testing, process

4 Daly (1992) has compared the biophysical limits of the macro economy with the plimsoll line of a ship that indicates the maximum load a ship can take. If loaded beyond this mark the ship will sink no matter how well the load is allocated or distributed.

5 According to estimates for the UK the potential cost savings of resource productivity are in the order of £55bn if a one year return period is allowed (DEFRA 2011), for Germany potential savings are estimated to be €48bn annually (Schröter, Lerch, and Jäger 2011). McKinsey & Company (2011) suggest that the world-wide benefits could be as high as $3.7 trillion a year (sic). Some of the associated innovations are in the public domain (http://www.eco-innovation.eu/). By comparison the EU's Roadmap to a Resource Efficient Europe estimates the cost saving potential resulting from industrial symbiosis across the EU only to be €1.4bn a year (with an additional €1.6bn in sales) (COM 2011, p.6).

auditing, traceability requirements) or added-value might be too low on average (since externalities are not internalized by competitors). However, transaction costs can be reduced as shown above. Yet so far the potential of habitats (Jensen et al. 2012) or niches (Kronenberg and Winkler 2009) to create niche markets (Gibbs 2009; Adamides and Mouzakitis 2009; Nill and Kemp 2009)6 has not been seized in predictable ways. Opportunities are rarely exploited beyond waste management (Costa and Ferrao 2010a) and also value-adding eco-innovations have yet to get exploited more systematically. As suggested above, niche markets might be anticipated to emerge alongside path- and place-dependent socio-technological trajectories. As a consequence, promising symbiotic opportunities might be expected to relate at least as much to skills, capabilities and institutions selecting, varying, and retaining these pathways, as to the specific resources (Sartorius 2006).

The institutional dimension of industrial networks

Social network analysis in industrial ecology has often revealed cooperation deficits despite existing coordination mechanisms such as markets. Industrial networks are constantly influenced by regulation, social and market dynamics exerting upward and downward pressures (see fig.1: interactions and relationships) and hence their performance can be evaluated accordingly (Langrock and Bleischwitz 2007). More specifically, symbiotic exchanges may grow to resource recovery networks (Fichtner et al. 2004) with common pool resource (CPR) characteristics. Institutionalized symbiotic common pool resources can mitigate risk of price fluctuation or regulatory uncertainty for firms while increasing their competitiveness.

Salmi et al. (2012) identify three conditions for successful common pool resources management based on industrial symbiosis: first, the CPR network should have clearly defined physical and membership boundaries. Second, the CPR network should apply proportional equivalence between benefits and costs. Third, the CPR network should be organized to enhance participation in collective decision making, to ensure monitoring and fair sanctioning, and to provide local conflict resolution. It will typically exploit a specific eco-industrial habitat. Since these symbiotic networks are private and domain specific, decentralized, horizontal relationships to regulatory agencies will dominate (Atkinson and Coleman 1989). This has received particular attention in Costa's and Ferrao's (2010b) middle-out approach. Symbiotic networks emerge at different scales and comprise different industries (Chen et al. 2012) and different levels of competitiveness. Without cognitive reintegration through global indicators and standards, decentralized spread of industrial symbiosis may lead to policy fragmentation. Even worse, it may reinforce existing path-dependencies of locked-in industries (Shi et al. 2012).

6 It should be noted that the concept of niche markets is a biological metaphor used in transition management studies to signify institutionally protected areas for eco-innovations. It does not directly relate to resource constraints or eco-industrial habitats.

4. Analyzing network structure

In recent years the analysis of industrial networks has been lead by scientists applying social metrics. Ashton and Bain (2012) study the communication between managers by measuring the in/out degree (ties claimed by others about the actor and ties claimed by the actor about others respectively). They also analyze the average degree (number of ties per firm), density (ratio of actual ties to all possible ties), and average constraint (measures how constrained each node is by neighbors). Paquin and Howard-Grenville (2013) have measured the growth of a regional symbiosis network by counting the number of firms and projects in the whole network and derive the network's centralization from firms' individual eigenvector centralities. They opt for eigenvector centrality because it also measures connectedness and accounts for influence and information flows from connections between firms (direct and indirect). Unsurprisingly they find the facilitator occupies the central position. Equally some authors have suggested analyzing waste managing companies as brokers (Ashton 2008; Chertow and Ehrenfeld 2012), which could draw on existing social network research (Burt 2004). The usefulness of such analysis will depend on the maturity of the institutional and organizational dimension since in the presence of waste regulation, waste brokers may simply co-evolve with functioning waste markets (Bleischwitz 2003) and may thus require little further research. Such cumulative causations speak for carrying out network analyses in industrial ecology along all five dimensions like geography has done (e.g. Neffke and Henning 2008). To some degree industrial ecology appears to be moving towards this paradigm. Ashton (2012) studied cognitive and social dimension (together) in a similar framework.

It comes as a surprise that industrial ecologists have applied few ecological metrics to industrial network structures. As an exception Hardy and Graedel (2002) and later Wright et al. (2009) investigated the connectance in industrial networks, which defines the ratio of the number of actual interactions to the number of potential interactions in a community (Ashton and Bain 2012 speak of compactness). Similarly, Templet (2004) and Korhonen and Snakin (2005) have looked at diversity in industrial ecosystems arguing that "diversity can create possibilities for increasing connectedness and cooperation in waste and by-product utilisation within ecosystems and within industrial ecosystems" (ibid., p. 171). Ashton (2009) has studied diversity empirically (applying it to sectors rather than firms).

Different to the network structure and its social dimensions, resilience and robustness address the metabolic network function. Both network properties have only entered the debate more recently (Holling 2001; Folke 2006; Allenby and Fink 2005). Both refer to the ability of a system or network to maintain its essential function and/or structure in response to perturbation, and the terms are often used interchangeably. However, the definition of what constitutes essential function or structure is particular to any given network and its socio-spatial boundary is crucial for any quantitative measurement. Yet they have no universally applicable definition (Brand and Jax. 2007). In general, robustness may be considered as the stability of a certain network

structure (e.g. the presence or absence of particular firms or actors in particular quantities) or property (e.g. connectance, total number of actors or energy generation capacity) to the removal of nodes (bankruptcy) or ties (flows under transmissions, fig. 1). Resilience is often considered to be the ability of a network to maintain (potentially adaptively) its structure and/or function in the face of an external perturbation (e.g. a price spike). However external perturbations may, of course, consist of the direct removal of nodes or ties too, and so this distinction between robustness and resilience is not clear cut.

Leach et al. (2010) make the distinction between robustness and resilience based on the nature of the external perturbation. They consider systems which maintain their structure and/or function in response to short-term shocks to be resilient and those which maintain themselves in the face of long-term stresses to be robust. Others have distinguished further between endogenous and exogenous shocks in this respect (Sornette 2006). In practice resilience and robustness have been defined in industrial ecology by other network or system properties that are assumed to bestow them, e.g. diversity, adaptability, modularity or degree distribution amongst others (Korhonen and Seager 2008). A problematic aspect of many of these measures is the fact that structure and function are not the same, and many structures may map to a given function (cf. van Berkel 2009). So whilst individual firms may have a vested interest in the robustness or resilience of a specific system structure because they are locked-in a particular technology (Koch 2011), whole-system sustainability goals usually pertain to system functions. This suggests that it is necessary to consider the structure of the whole economic system giving rise to the metabolic network function and hierarchical network structures.

The rare use of ecological metrics in industrial ecology is surprising since in recent years economics has seen exciting work here. For example, drawing on ecological research on mutualistic networks Saavedra et al. (2009) analyze organizational networks as bipartite structures of cooperative partner-partner interactions. Unlike the aforementioned metrics, analysis of bipartite structures recognizes hierarchical relationships between entities (nodes). This type of network analysis has found various applications in economics and geography.

Hidalgo and Hausmann (2009) have introduced a similar "economic complexity index", which maps industrial products related to one another by existing knowledge and capabilities, in other word cognitive and organizational proximity. They use a product space map which represents the whole (bi-partite) network of products manufactured in a country or region. Hausmann and Hidalgo (2011) show that some co-occurrences of products are more likely than others. The product portfolio is nested, which enables identification and ranking of opportunities for product development (cf. Bustos et al. 2012). These methods derive an astonishing degree of predictability from analyzing and mapping the "product space" when it comes to the competitiveness of countries and international trade (Hidalgo et al. 2007), which previously built on disaggregated macro-economic sector models.

Neffke and Henning (2008) also assume a bi-partite network structure to study co-location. Unlike the aforementioned product indicators or co-occurrence indicators that study relationships between firms, "revealed relatedness" analyzes the whole economic network/ system in relation to plants. Companies cooperate in industries for developing innovative outputs. These are not emerging randomly but as an effect of proximity, strongly influenced by existing technologies, capabilities and skills. Tie formation between plants is influenced by a number of factors, including average profitability in the industry, intensity of competition, or wage levels. Neffke and Henning's method allows for the control of any of these factors as long as information on them is available at the level of the class aggregate.

These methods allow an astonishing degree of predictability of regional development, countries' competitiveness and of international trade (Hidalgo et al. 2007), which previously built on disaggregated macro-economic sector models. While this work has focused on added-value of the whole economic system analyzing the associated environmental burden of the growing and increasingly complex economic systems with these methods remains a desideratum. The lack of similar research on material and energy flows is surprising in view of the fact that Bustos et al. (2012) even claim the term industrial ecology for their method. Clearly, these methods call for application in industrial ecology, which has seen some relevant work on economic complexity (Wood and Lenzen 2009).

5. Discussion

Network analysis is used across various disciplines and yet the integration of structure and function of networks is theoretically and empirically challenging since the social and metabolic dimensions are interdependent and dynamic. In the study of industrial symbiosis, network analyses have predominantly relied on social scientific metrics to promote sustainability. In as much as these have focused on the (social) formation of symbiotic ties it has arguably turned the danger of inducing a "structural determinism" when analyzing social phenomena (Emirbayer and Goodwin 1994) into a virtue. The preferred framework of these analyses has been the embeddedness concept. The cognitive and social dimensions have been emphasized in network analytical studies of industrial symbiosis while others dimensions such as the institutional have been given little attention. By contrast, we suggest that Social-Material Network Analyses should start from spatial proximity and scrutinize all identified cognitive-socio-organizational-institutional dimensions.

Social-Material Network Analyses should consider individual cognition in as much as it relates to organizations and the development of firms' capabilities. Flimsy pro-environmental frames stabilize when they become institutionalized, for instance, through the position of environmental managers in companies, facilitators, professional networks and others (Boons et al. 2011). Only when these social institutions persevere do firms develop lasting capabilities for eco-innovations. Since firms with strong capabilities to eco-innovate have been found to be

more productive (Rennings and Rammer 2009) an empirical mechanism for variation in eco-industrial networks can be claimed. Likewise, the creation of symbiotic ties constitutes endogenous selection towards an eco-industrial network with a reduced metabolic throughput. The exploration and exploitation of symbiotic opportunities may, however, not be stable over time and firms may in fact ignore these opportunities altogether if they cannot establish the required social and organizational capability in the first place. Eco-innovation research indicates that exogenous selection by institutions can substantially improve the generation of eco-innovations in the market place (Ashford and Hall 2011; Horbach, Rammer, and Rennings 2012).

While industrial symbiosis could learn from fields like innovation research, it might focus on the organizational dimension in doing so since not only does the relevance of each dimension change over time (Paquin and Howard-Grenville 2009; Doménech and Davies 2011) but some dimensions are also more effective than others in bringing about change. Personal trust, for instance, loses importance once recyclate markets are established (Jensen et al. 2011). Standards for recyclates, reliable process control and assured traceability may replace personal trust and assume its social purpose. This involves institutions and governance. Consequently, countries have copied entire institutional arrangements from other countries in order to accelerate their transition processes (Busch et al. 2005).

The dominant view across the social sciences is that social networks eventually become institutionalized and that these institutionalized networks are located somewhere on the continuum between markets and governments (Powell 1990). If we conceive of symbiotic relationships and eco-industrial parks as cooperative games to establish common pool resources these would fit well in the middle of this spectrum (Jones et al. 1997). If, however, we face locked-in industries then exogenous governance involving non-cooperative games between industry and regulators might be inevitable. It should be noted in this respect that structural network analyses have contributed little to network governance or, in other words, institutional proximity (Borzel 1998, Khan 2013).

From a regulatory point of view, networks can be governed by coordinating the strategies of actors promoting mutual adjustment via negotiations and consultations (Kenis and Schneider 1991). These take place in particular actor constellations, see figure 1 (Scharpf 1997). Since most interactions between formal institutions (actors) can be represented as mixed-motive games, understanding actors' cognitive orientation becomes equally important as the constellation or structure in which actors interact if we want to explain retention at the institutional level. The success of internalizing environmental damages via network governance will then crucially depend on factors such as the number of industrial symbiotic common pool resource regimes, their diversity, the coherence of their cognitive and metabolic orientations, and how inward-looking (closed) any network is (Adam and Kriesi 2007). Network governance might then

represent one form of institutional capacity (Boons and Spekkink 2012) that can influence the metabolic network structure.

So far we have omitted the temporal dimension of networks' metabolisms. Despite relevant empirical frameworks for analysis (Baas and Boons 2004; Korhonen and Snakin 2005; Domenech and Davies 2011; Boons et al. 2011) and simulations of network evolution (Baldwin et al. 2004; Cao et al. 2009) there are only a few empirical network studies addressing the temporal dimension (Gibbs and Deutz 2007; Tudor et al. 2007 for industrial parks and Paquin and Howard-Grenville (in press) for a national symbiosis program). The latter studies represent implementation studies. It does not come as a surprise that a recent critique and reformulation of industrial symbiosis has readdressed the temporal dimension (Lombardi and Laybourn 2012, 29). In a growing industrial system, success from symbiosis will not last if the total metabolism continues growing. Even worse, some industrial symbiosis may directly increase total throughput (Shi et al. 2012). This danger is present since industrial symbiosis also represents a form of transmaterialization (Labys 2002) and the associated efficiency gains are potentially liable to cause rebound or even backfire effects (through increased production or consumption or both (Saunders 2000)).

Mattila et al. (2012) show how industrial symbiosis can be evaluated using life-cycle assessment to prevent such negative effects. Park and Behera (2014) make a similar point using eco-efficiency. Social-Material Network Analyses may adapt these perspectives taking in some further thoughts. It should be noted that the original schema underlying figure 1 also reflects methodological concerns over biases in social data. Newman et al. (2002) consider affiliation network data (associated with the column similarities) as more reliable than data from the columns to the right (linked to surveys and case studies). This difference partly applies to material and energy data as well: industrial symbiosis is characterized by case studies and a focus on dyadic relationships. These are functionally marked (as by-products)7 but still ambivalent if network-wide metabolic analyses are missing (van Berkel 2009; Lombardi and Laybourn 2012). Hence figure 1 considers material and energy flows at three points to overcome such ambivalence: at node level as net-addition to stock, at tie level under similarities as energy flows and again at tie level under transmissions as material flows and recyclates/by-products.8 Nodes and dyadic flows are localized, specific and irreversible. They are spatially bounded, concern only particular materials and they are thermodynamically constrained. These last two constraints are well established by industrial symbiosis and

7 Distinguishing research areas such as material and energy efficiency at node level and forward and reverse supply chain analysis (the latter including industrial symbiosis) at tie level is crucial to avoid conflicting orientations in analyzing and governing networks (cf. Seuring 2004; Mattila et al. 2012).

8 We analytically differentiate between energy and material flows in this way to highlight the centrality of energy for production. We do not raise methodological claims about existing material analytical methods.

eco-industrial habitats while the field could learn from geography regarding the first (cf. Schiller et al. forthcoming).

A question that comes up in this respect is in how far eco-industrial paths designed from the dynamic formation of dyadic ties are stable under their endogenous dynamics and against exogenous shocks like the increasingly felt constraints of the biophysical resource base. At its epistemic core, industrial ecology claims that industrial systems are an extension of nature's self-organization. This claim, however, is not straightforward in the social sphere (which is why many industrial ecologists have insisted on an ecological metaphor). Whereas markets have been seen as self-organizing (cf. Witt 1997) it is not well understood how self-organization of network structures and functions relate to society at large. Typically complexity science has modeled emerging, self-organizing phenomena such as scale-free networks or the aforementioned innovation networks along the lines of preferential attachment or in-/decreasing returns but this does not necessarily explain which specific social mechanisms actually reproduce or change society. Likewise the evolutionary hypothesis we have discussed earlier does not provide a satisfying explanation and this leaves a theoretical gap regarding the social structure.

Like the different designation of nodes in the "revealed relatedness" respectively the "product space" approach industrial symbiosis too has discussed whether plants or firms should constitute the nodes of exchanges (cf. van Berkel 2009). The differentiation of research fields speaks for the plant level although the firm level is more relevant overall. The focus on the network function of symbiotic reverse flows has lead to structurally complex studies at the micro-level but few equivalent studies at the macro-level. This might be due to material flow analyses in conjuncture with institutional analyses leading macro-level research. Furthermore, network analytical methods hardly existed for the macro-level until recently. This is changing and as a consequence we may see closer integration of bottom-up and top-down network analyses in the future, e.g. bipartite network analysis enabling to study the international diffusion of eco-innovations.

It is always true that local industrial networks are functionally nested in larger network structures that are socially constituted by intentional and non-intentional structures. We pointed out that while proximity and embeddedness approaches study nearly the same variables only the proximity approach conceptualises cumulative causations as bottom-up processes. We showed that, by providing endogenous explanations, the proximity approach indeed comes closer to the pre-analytical vision of self-organizing industrial systems. The cumulative causations between proximities (Martin and Sunley 2006) are seemingly giving rise to path-dependent self-organization of our socio-environmental metabolism (Chertow and Ehrenfeld 2012). However, industrial self-organization towards sustainability still necessitates several pre-conditions including economics of scale and/or scope, which applies to individual

firms and brokers but would also involve (formal) institutions and governance that internalise externalities.9 Evolutionary economic geography also suggests that self-organisation involves knowledge spill-over to grow regional economies, which seems to hold for resource- and eco-efficient innovations at firm-level.

The analytical framework we introduced seems to favour complex research projects that reconstruct each dimension and show the cumulative causations relevant for the industrial function. While there is a point in covering dimension by dimensions, if only by qualitative analysis (cf. e.g. Nill and Kamp 2009), we do not expect that all research projects will and can cover all dimensions. Rather our scientific endeavor itself will be cumulative when developing bottom-up, implementation-oriented research that strives for universal explanations. Further integration might be achieved by modelling endogenous and exogenous shocks to these emerging networks for instance as fluctuating recyclate prices, delayed investments or increasing energy and resource prices. We hope that the approach offered here will prevent network analyses from becoming distracted by the multitude of possible social factors, processes and structures while enabling comparisons between top-down- and bottom-up-driven implementations of industrial ecology across nations. If applied systematically, Social-Material Network Analyses offer novel opportunities to scrutinize and compare efforts to eco-modernize industrial networks.

6. Conclusion

In reviewing the existing literature on network analysis in industrial ecology this article has come to propose a complexity-derived approach for Social-Material Network Analyses of industrial networks. By focusing on five social dimensions in specifying the social mechanisms that enable dyadic and triadic relationships, the methodology assures multiplexity rather than reductionism in the analysis of industrial networks. It presents a considerable extension of existing Social-Material Network Analyses in the field, which have not considered more than two dimensions. It should bring us closer to the consilience between natural and social sciences while avoiding natural fallacy.

Acknowledgements

The authors gratefully acknowledge the partial support of the UK Engineering and Physical Sciences Research Council for program grant EP/H021450/1 (Evolution and Resilience of Industrial Ecosystems (ERIE)). The views expressed in this article are in the sole responsibility of the authors and may not

9 It is surprising that few studies in industrial symbiosis have focussed on savings in the primary sector although these appear particularly promising from a functional point of view.

reflect the views of funding body. We would also like to thank Nigel Gilbert and the anonymous reviewers for their helpful comments.

References

Adam, Silke, and Hanspeter Kriesi. 2007. The network approach. In Theories of the policy process, 2:129-54. 2nd ed. Boulder, CO: Westview Press.

Adamides, Emmanuel D., and Yannis Mouzakitis. 2009. "Industrial ecosystems as technological niches." Journal of Cleaner Production 17 (2) (January): 172-180.

Ahuja, Gautam. 2000. "Collaboration Networks, Structural Holes, and Innovation: A Longitudinal Study." Administrative Science Quarterly 45 (3) (September): 425-455.

Allenby, Brad, and Jonathan Fink. 2005. "Toward inherently secure and resilient societies." Science 309 (5737) (August 12): 1034-6.

Allenby, Braden R. 2006. "The ontologies of industrial ecology?" Progress in Industrial Ecology 3 (1/2): 28.

Ashford, Nicholas A., and Ralph P. Hall. 2011. "The Importance of Regulation-Induced Innovation for Sustainable Development." Sustainability 3 (1) (January 19): 270-292.

Ashton, Weslynne S. 2008. "Understanding the Organization of Industrial Ecosystems." Journal of Industrial Ecology 12 (1) (February): 34-51.

---. 2009. "The Structure, Function, and Evolution of a Regional Industrial Ecosystem." Journal of

Industrial Ecology 13 (2) (April): 228-246.

Ashton, Weslynne S., and Ariana C. Bain. 2012. "Assessing the 'Short Mental Distance' in Eco-Industrial Networks." Journal of Industrial Ecology 16 (1) (February 23): 70-82.

Atkinson, Michael M., and William D. Coleman. 1989. "Strong States and Weak States: Sectoral Policy Networks in Advanced Capitalist Economies." British Journal of Political Science 19 (01) (January 27): 47-67.

Baas, Leo W. 2008. "Industrial Symbiosis in the Rotterdam Harbour and Industry Complex: Reflections on the Interconnection of the Technosphere with the Social System." Business Strategy and the Environment 17 (5) (July): 330-340.

Baas, Leo W., and Frank Boons. 2004. "An industrial ecology project in practice: exploring the boundaries of decision-making levels in regional industrial systems." Journal of Cleaner Production 12 (8-10) (October): 1073-1085.

Baldwin, James Scott, Keith Ridgway, Belinda Winder, and Robert Murray. 2004. "Modelling industrial ecosystems and the problem of evolution." Progress in Industrial Ecology 1 (1/2/3): 39-60.

Bansal, Pratima, and Brent McKnight. 2009. "Looking Forward, Pushing Back and Peering Sideways: Analyzing the Sustainability of Industrial Symbiosis." Journal of Supply Chain Management 45 (4) (October): 26-37.

Bertoldi, P, S Rezessy, and E Vine. 2006. "Energy Service Companies in European Countries: Current

Status and a Strategy to Foster Their Development." Energy Policy 34 (14) (September): 1818-1832.

Binder, Claudia R. 2007a. "From material flow analysis to material flow management Part I: social

sciences modeling approaches coupled to MFA." Journal of Cleaner Production 15 (17) (November): 1596-1604.

Binswanger, Mathias. 1993. "From Microscopic to Macroscopic Theories: Entropic Aspects of Ecological and Economic Processes." Ecological Economics 8 (3) (December): 209-233.

Bleischwitz, Raimund. 2003. "Cognitive and Institutional Perspectives of Eco-efficiency." Ecological Economics 46 (3) (October): 453-467.

Bleischwitz, Raimund. 2007. Corporate Governance of Sustainability: A Co-Evolutionary View on Resource Management. Ed. Raimund Bleischwitz. Cheltenham: Edward Elgar Publishing.

Boix, Marianne, Ludovic Montastruc, Luc Pibouleau, Catherine Azzaro-Pantel, and Serge Domenech.

2012. "Industrial Water Management by Multiobjective Optimization: From Individual to Collective Solution Through Eco-industrial Parks." Journal of Cleaner Production 22 (1): 85-97.

Boons, Frank, and Jennifer A. Howard-Grenville. 2009. Introducing the social embeddedness of industrial ecology. In The Social Embeddedness of Industrial Ecology, ed. Frank Boons and Jennifer A. Howard-Grenville, 3-27. Cheltenham: Edward Edgar.

Boons, Frank, and Leo W. Baas. 1997. "Types of industrial ecology: The problem of coordination." Journal of Cleaner Production 5 (1-2): 79-86.

Boons, Frank, and Wouter Spekkink. 2012. "Levels of Institutional Capacity and Actor Expectations About Industrial Symbiosis." Journal of Industrial Ecology 16 (1) (February 17): 61-69.

Boons, Frank, Wouter Spekkink, and Yannis Mouzakitis. 2011. "The dynamics of industrial symbiosis: A proposal for a conceptual framework based upon a comprehensive literature review." Journal of Cleaner Production 19 (9-10) (January 13): 905-911.

Boons, Frank. 1998. "Caught in the web: the dual nature of networks and its consequences." Business Strategy and the Environment 7 (4) (September): 204-212.

Borgatti, Stephen P., Ajay Mehra, Daniel J Brass, and Giuseppe Labianca. 2009. "Network Analysis in the Social Sciences." Science 323 (5916) (February 13): 892-5.

Borgatti, Stephen P., and Rob Cross. 2003. "A Relational View of Information Seeking and Learning in Social Networks." Management Science 49 (4) (April): 432-445.

Boschma, Ron A. 2005. "Proximity and Innovation: A Critical Assessment." Regional Studies 39 (1) (February): 61-74.

Boschma, Ron A., and K. Frenken. 2011. "The Emerging Empirics of Evolutionary Economic Geography." Journal of Economic Geography 11 (2) (January 8): 295-307.

Börzel, Tanja A. 1998. "Organizing Babylon - On the Different Conceptions of Policy Networks." Public Administration 76 (2) (January): 253-273.

Brand, F. S., and K. Jax. 2007. "Focusing the meaning(s) of resilience: resilience as a descriptive concept and a boundary object." Ecology and Society 12 (1).

Burt, Ronald S. (2004). Structural Holes and Good Ideas. American Journal of Sociology, 110(2), 349-399.

Busch, Per-Olof, Kerstin Tews, and Helge Jörgens. 2005. "The Global Diffusion of Regulatory Instruments: The Making of a New International Environmental Regime." The ANNALS of the American Academy of Political and Social Science 598 (1) (March 1): 146-167.

Busch, Timo, and Volker H. Hoffmann. 2007. "Emerging Carbon Constraints for Corporate Risk Management." Ecological Economics 62 (3-4) (May): 518-528.

Bustos, Sebastián, Charles Gomez, Ricardo Hausmann, and César A Hidalgo. 2012. "The Dynamics of Nestedness Predicts the Evolution of Industrial Ecosystems." PLoS One 7 (11) (November 19): e49393.

Cao, Kai, Xiao Feng, and Hui Wan. 2009. "Applying agent-based modeling to the evolution of eco-industrial systems." Ecological Economics 68 (11) (September): 2868-2876.

Chen, Xudong, Tsuyoshi Fujita, Satoshi Ohnishi, Minoru Fujii, and Yong Geng. 2012. "The Impact of Scale, Recycling Boundary, and Type of Waste on Symbiosis and Recycling." Journal of Industrial Ecology 16 (1): 129-141.

Chertow, Marian R. 1998. "The Eco-industrial Park Model Reconsidered." Journal of Industrial Ecology 2 (3) (June): 8-10.

---. 2007. "'Uncovering' Industrial Symbiosis." Journal of Industrial Ecology 11 (1) (March): 11-30.

Chertow, Marian R., and D. Rachel Lombardi. 2005. "Quantifying Economic and Environmental Benefits of Co-Located Firms." Environmental Science & Technology 39 (17) (September): 6535-6541.

Chertow, Marian R., Weslynne S. Ashton, and Juan Espinosa. 2008. "Industrial Symbiosis in Puerto Rico: Environmentally Related Agglomeration Economies." Regional Studies 42 (10) (December): 1299-1312.

Chertow, Marian R., and Yuko Miyata. 2011. "Assessing Collective Firm Behavior: Comparing Industrial Symbiosis with Possible Alternatives for Individual Companies in Oahu, HI." Business Strategy and the Environment 20 (4) (May 2): 266-280.

Chertow, Marian R., & Ehrenfeld, John R. (2012). Organizing Self-Organizing Systems. Journal of Industrial Ecology, 16(1), 13-27.COM. (2011). Roadmap to a Resource Efficient Europe. Brussels, Luxembourg: European Commission.

Costa, Ines, and Paulo Cadete Ferrao. 2010a. "Crossroads between resource recovery and industrial symbiosis networks. Evidences from Case Studies." Regional development dialogue 31 (2): 1-18.

Costa, Ines, and Paulo Cadete Ferrao. 2010b. "A Case Study of Industrial Symbiosis Development Using a Middle-out Approach." Journal of Cleaner Production 18 (10-11) (July): 984-992.

Daly, Herman E. 1992. "Allocation, Distribution, and Scale: Towards an Economics That Is Efficient, Just, and Sustainable." Ecological Economics 6: 185-193.

Daniels, Peter L, and Stephen Moore. 2002. "Approaches for Quantifying the Metabolism of Physical Economies - Part 1: Methodological Overview." Journal of Industrial Ecology 5 (4): 69-93.

DEFRA. 2011. Further Benefits of Business Resource Efficiency. London.

Dlouha, J., Barton, A., Huisingh, D., Adomssent, M., Sol, J., Beers, P. J., & Wals, A. E. J. (2013). Social

learning in regional innovation networks: trust, commitment and reframing as emergent properties of interaction. Journal of Cleaner Production, 49, 35-43.

Domenech, Teresa, and Michael Davies. 2011. "The Role of Embeddedness in Industrial Symbiosis Networks: Phases in the Evolution of Industrial Symbiosis Networks." Business Strategy and the Environment 20 (5) (July 5): 281-296.

Easley, David, and Jon Kleinberg. 2010. Networks, Crowds, and Markets: Reasoning About a Highly Connected World. Cambridge: Cambridge University Press.

Ehrenfeld, John R. 2009. "A Critical View on the Social Science Contribution to Industrial Ecology." In The Social Embeddedness of Industrial Ecology, ed. Frank Boons and Jennifer A. Howard-Grenville, 257-272. Cheltenham: Edward Edgar.

Emirbayer, Mustafa, and Jeff Goodwin. 1994. "Network Analysis, Culture, and the Problem of Agency." American Journal of Sociology 99 (6) (May): 1411-1454.

Esty, Daniel C, and Michael E Porter. 1998. "Industrial Ecology and Competitiveness." Journal of Industrial Ecology 2 (1) (December): 35-43.

Fichtner, Wolf, Ingela Tietze-Stockinger, and Otto Rentz. 2004. "On Industrial Symbiosis Networks and Their Classification." Progress in Industrial Ecology 1 (1/2/3): 130-142.

Fischer-Kowalski, Marina, and Helga Weisz. 1999. "Society as Hybrid Between Material and Symbolic Realms - Toward a Theoretical Framework of Society-Nature Interaction." Advances in Human Ecology 8: 215-251.

Folke, Carl. 2006. "Resilience: The emergence of a perspective for social-ecological systems analyses." Global Environmental Change 16 (3) (August): 253-267.

Foxon, Timothy J. 2006. "Bounded Rationality and Hierarchical Complexity: Two Paths from Simon to Ecological and Evolutionary Economics." Ecological Complexity 3 (4) (December): 361-368.

Fröhling, Magnus, Frank Schwaderer, Hauke Bartusch, and Frank Schultmann. 2012. "A Material Flow-based Approach to Enhance Resource Efficiency in Production and Recycling Networks." Journal of Industrial Ecology (August 9): no-no.

Gibbs, David, and Pauline Deutz. 2005. "Implementing Industrial Ecology? Planning for Eco-industrial Parks in the USA." Geoforum 36 (4) (July): 452-464.

Gibbs, David, and Pauline Deutz. 2007. "Reflections on implementing industrial ecology through

eco-industrial park development." Journal of Cleaner Production 15 (17) (November): 1683-1695.

Gibbs, David. 2009. Eco-industrial parks and industrial ecology: strategic niche or mainstream

development? In The social embeddedness of industrial ecology, ed. Frank Boons and Jennifer A. Howard-Grenville, 73-102. Cheltenham: Edward Elgar.

Glückler, Johannes. 2007. "Economic Geography and the Evolution of Networks." Journal of Economic Geography 7 (5) (June 13): 619-634.

Grant, Gabriel B., Thomas P. Seager, Guillaume Massard, and Loring Nies. 2010. "Information and

Communication Technology for Industrial Symbiosis." Journal of Industrial Ecology 14 (5) (October 7): 740-753.

Habermas, Jürgen. 2003. Truth and Justification. Cambridge, MA: MIT Press.

Halinen, Aino, Asta Salmi, and Virpi Havila. 1999. "From Dyadic Change to Changing Business Networks: An Analytical Framework." Journal of Management Studies 36 (6) (November): 779-794.

Halme, M, M Anttonen, M Kuisma, N Kontoniemi, and E Heino. 2007. "Business models for material efficiency services: Conceptualization and application." Ecological Economics 63 (1) (June 15): 126-137.

Hardy, Catherine, and Thomas E Graedel. 2002. "Industrial Ecosystems as Food Webs." Journal of Industrial Ecology 6 (1) (December): 29-38.

Hausmann, Ricardo, and César A. Hidalgo. 2011. "The Network Structure of Economic Output." Journal of Economic Growth 16 (4) (October 1): 309-342.

Hidalgo, César A, B Klinger, A-L Barabási, and Ricardo Hausmann. 2007. "The Product Space Conditions the Development of Nations." Science 317 (5837) (July 27): 482-7.

Hidalgo, César a., and Ricardo Hausmann. 2009. The Building Blocks of Economic Complexity. Proceedings of the National Academy of Sciences of the United States of America. Vol. 106.

Hiete, Michael, Jens Ludwig, and Frank Schultmann. 2012. "Intercompany Energy Integration." Journal of Industrial Ecology 16 (5) (October 4): 689-698.

Hoffman, Andrew. J. 2003. "Linking Social Systems Analysis To The Industrial Ecology Framework." Organization & Environment 16 (1) (March): 66-86.

Holling, C. S. 2001. "Understanding the Complexity of Economic, Ecological, and Social Systems." Ecosystems 4 (5) (August): 390-405.

Horbach, Jens, Christian Rammer, and Klaus Rennings. 2012. "Determinants of Eco-innovations by Type of Environmental Impact — The Role of Regulatory Push/pull, Technology Push and Market Pull." Ecological Economics 78 (April): 112-122.

Jackson, Matthew O. 2008. Social and economic networks. Princeton, Oxford: Princeton University Press.

Jacobsen, Noel Brings, and Stefan Anderberg. 2005. Understanding the evolution of industrial symbiotic networks—The case of Kalundborg. In Economics of Industrial Ecology - Materials, Structural Change and Spatial Scales, ed. Jeroen C.J.M. van Den Bergh and Marco A. Janssen, 331-335. Cambridge, MA: MIT Press.

Jensen, Paul D., Lauren Basson, Emma E. Hellawell, Malcolm R. Bailey, and Matthew Leach. 2011.

"Quantifying 'geographic proximity': Experiences from the United Kingdom's National Industrial Symbiosis Programme." Resources, Conservation and Recycling 55 (7) (May): 703-712.

Jensen, Paul D., Lauren Basson, Emma E. Hellawell, and Matthew Leach. 2012. "'Habitat' Suitability Index Mapping for Industrial Symbiosis Planning." Journal of Industrial Ecology 16 (1) (February 15): 38-50.

Jones, C., W. S. Hesterly, and Stephen P. Borgatti. 1997. "A General Theory of Network Governance:

Exchange Conditions and Social Mechanisms." Academy of Management Review 22 (4) (October 1): 911-945.

Kadushin, Charles. 2012. Understanding Social Networks. Oxford etc.: Oxford University Press.

Kenis, Patrick, and Volker Schneider. 1991. Policy networks and policy analysis: scrutinizing a new

analytical toolbox. In Policy networks: Empirical evidence and theoretical considerations, ed. Bernd Marin and Renate Mayntz, 25-59. Frankfurt/Boulder, CO: Campus/Westview Press.

Koch, Jochen. 2011. "Inscribed Strategies: Exploring the Organizational Nature of Strategic Lock-in." Organization Studies 32 (3) (March 24): 337-363.

Korhonen, Jouni, and Juha-Pekka Snäkin. 2001. "An anchor tenant approach to network management: considering regional material and energy flow networks." International Journal of Environmental Technology and Management 1 (4): 444-463.

---. 2005. "Analysing the evolution of industrial ecosystems: concepts and application." Ecological

Economics 52 (2) (January): 169-186.

Korhonen, Jouni, and Thomas P Seager. 2008. "Beyond Eco-Efficiency: a Resilience Perspective." Business Strategy and the Environment 419: 411-419.

Krippner, Greta R., and Anthony S. Alvarez. 2007. "Embeddedness and the Intellectual Projects of Economic Sociology." Annual Review of Sociology 33 (1) (August): 219-240.

Kronenberg, Jakub, and Ralph Winkler. 2009. "Wasted waste: An evolutionary perspective on industrial by-products." Ecological Economics 68 (12) (October 15): 3026-3033.

Kronenberg, Jakub. 2006. "Industrial ecology and ecological economics." Progress in Industrial Ecology 3 (1/2): 95-113.

Labys, W. C. (2002). Transmaterialization. In R. U. Ayres & L. W. Ayres (Eds.), A Handbook of Industrial Ecology (pp. 202-208). Cheltenham: Edward Elgar.

Langrock, Thomas, and Raimund Bleischwitz. 2007. Networks, decision-making of companies and

evaluation: concepts and terminology. In Corporate Governance of Sustainability: A Co-Evolutionary View on Resource Management, ed. Raimund Bleischwitz, 83-96. Cheltenham: Edward Elgar.

Laybourn, Peter, and D. Rachel Lombardi. 2012. "Industrial Symbiosis in European Policy." Journal of Industrial Ecology 16 (1) (February 21): 11-12.

Leach, Melissa, Ian Scoones, and Andy Stirling. 2010. DynamicSustainabilities: Linking Technology, Environment and Social Justice. London: Earthscan.

Lyons, Donald I. 2008. "A Spatial Analysis of Loop Closing Among Recycling, Remanufacturing, and Waste Treatment Firms in Texas." Journal of Industrial Ecology 11 (1) (October 9): 43-54.

Martin, R., and P. Sunley. 2006. "Path Dependence and Regional Economic Evolution." Journal of Economic Geography 6 (4) (April 24): 395-437.

Mattila, Tuomas, Suvi Lehtoranta, Laura Sokka, Matti Melanen, and Ari Nissinen. 2012. "Methodological Aspects of Applying Life Cycle Assessment to Industrial Symbioses." Journal of Industrial Ecology 16 (1) (February 15): 51-60.

Khan, Jamil (2013). What role for network governance in urban low carbon transitions? Journal of Cleaner Production, 50, 1, 133-139.

McKinsey & Company. 2011. Resource Revolution: Meeting the world's energy, materials, food, and water needs.

Mirata, Murat. 2004. "Experiences from Early Stages of a National Industrial Symbiosis Programme in the UK: Determinants and Coordination Challenges." Journal of Cleaner Production 12 (8-10) (October): 967-983.

Nelson, Richard R., and Sidney G. Winter. 1982. An Evolutionary Theory of Economic Change. Belknap Press of Harvard University Press.

Newman, M E J, D J Watts, and S H Strogatz. 2002. "Random Graph Models of Social Networks." Proceedings of the National Academy of Sciences of the United States of America 99 Suppl 1 (February 19): 2566-72.

Nill, Jan, and René Kemp. 2009. "Evolutionary Approaches for Sustainable Innovation Policies: From Niche to Paradigm?" Research Policy 38 (4): 668-680.

Paquin, Raymond L., and Jennifer A. Howard-Grenville. 2009. Facilitating Regional Industrial Symbiosis: Network Growth in the UK's National Industrial Symbiosis Programme. In The Social Embeddedness of Industrial Ecology, ed. Frank Boons and Jennifer A. Howard-Grenville, 103-128. Northampton: Edward Elgar.

---. (2013). "Blind Dates and Arranged Marriages: Longitudinal Processes of Network Orchestration."

Organization Studies, 34(11), 1623-1653.

Park, H.-S., & Behera, S. K. (2014). „Methodological aspects of applying eco-efficiency indicators to industrial symbiosis networks". Journal of Cleaner Production, 64, 478-485.

Posch, Alfred. 2010. "Industrial Recycling Networks as Starting Points for Broader Sustainability-Oriented Cooperation?" Journal of Industrial Ecology 14 (2) (March 30): 242-257.

Powell, Walter W. 1990. "Neither Market Nor Hierarchy: Network Forms of Organization." Ed. Grahame Thompson. Research In Organizational Behavior 12 (1): 295-336.

Ragin, Charles C. 2008. Redesigning Social Inquiry: Fuzzy Sets and Beyond. Chicago, London: University of Chicago Press.

Reiskin, Edward D., Allen L. White, Jill Kauffman Johnson, and Thomas J. Votta. 1999. "Servicizing the Chemical Supply Chain." Journal of Industrial Ecology 3 (2-3) (March): 19-31.

Rennings, Klaus, and Christian Rammer. 2009. "Increasing Energy and Resource Efficiency through

Innovation: An Explorative Analysis Using Innovation Survey Data." Czech Journal of Economics and Finance (Finance a uver) 59 (5): 442-459.

Rennings, Klaus. 2000. "Redefining Innovation — Eco-innovation Research and the Contribution from Ecological Economics." Ecological Economics 32 (2) (February): 319-332.

Rittel, Horst W. J., and Melvin M. Webber. 1973. "Dilemmas in a General Theory of Planning." Policy Sciences 4 (2): 155-169.

Rockström, Johan, Will Steffen, Kevin Noone, Asa Persson, F Stuart Chapin, Eric F Lambin, Timothy M Lenton, et al. 2009. "A safe operating space for humanity." Nature 461 (7263) (September): 472-475.

Saavedra, Serguei, Felix Reed-Tsochas, and Brian Uzzi. 2009. "A Simple Model of Bipartite Cooperation for Ecological and Organizational Networks." Nature 457 (7228) (January 22): 463-6.

Safarzynska, Karolina, Koen Frenken, and Jeroen C.J.M. van den Bergh. 2012. "Evolutionary Theorizing and Modeling of Sustainability Transitions." Research Policy 41 (6) (July): 1011-1024.

Salmi, Olli, Janne Hukkinen, Jyrki Heino, Nani Pajunen, and Maaria Wierink. 2012. "Governing the Interplay Between Industrial Ecosystems and Environmental Regulation." Journal of Industrial Ecology 16 (1) (February 14): 119-128.

Sartorius, Christian. 2006. "Second-order Sustainability—conditions for the Development of Sustainable Innovations in a Dynamic Environment." Ecological Economics 58 (2) (June 15): 268-286.

Saunders, H. D. (2000). "A view from the macro side: rebound, backfire, and Khazzoom-Brookes." Energy Policy 28(6-7): 439-449.

Scharpf, Fritz W. 1997. Games Real Actors Play: Actor-centered Institutionalism in Policy Research. Boulder: Westview Press.

Schiller, Frank, Penn, Alexandra, Basson, Lauren, Druckman, Angela (forthcoming). Exploring Space, Exploiting Opportunities - the Case for Analyzing Space in Industrial Ecology. Accepted for publication in Journal of Industrial Ecology.

Schröter, Marcus, Christian Lerch, and Angela Jäger. 2011. Materialeffizienz in der Produktion:

Einsparpotenziale und Verbreitung von Konzepten zur Materialeinsparung im Verarbeitenden Gewerbe. Karlsruhe.

Schwarz, Erich J., and Karl W. Steininger. 1997. "Implementing Nature's Lesson: The Industrial Recycling Network Enhancing Regional Development." Journal of Cleaner Production 5 (1-2) (January): 47-56.

Scott, John. 2000. Social Network Analysis. Los Angles, London etc.: Sage Publications.

Seuring, Stefan. 2004. "Industrial Ecology, Life Cycles, Supply Chains: Differences and Interrelations." Business Strategy and the Environment 13 (5) (September 18): 306-319.

Shi, Han, Marian Chertow, and Yuyan Song. 2010. "Developing Country Experience with Eco-industrial Parks: a Case Study of the Tianjin Economic-Technological Development Area in China." Journal of Cleaner Production 18 (3) (February): 191-199.

Shi, Han, Jinping Tian, and Lujun Chen. 2012. "China's Quest for Eco-industrial Parks, Part II." Journal of Industrial Ecology 16 (3) (June 17): 290-292.

Sornette, Didier. 2006. "Endogenous Versus Exogenous Origins of Crises." In Extreme Events in Nature and Society, ed. Sergio Albeverio, Volker Jentsch, and Holger Kantz, 95-119. Berlin, Heidelberg, New York: Springer.

Sterr, Thomas, and Thomas Ott. 2004. "The industrial region as a promising unit for eco-industrial development—reflections, practical experience and establishment of innovative instruments to support industrial ecology." Journal of Cleaner Production 12 (8-10) (October): 947-965.

Templet, Paul H. 2004. "Diversity and other emergent properties of industrial economies." Progress in Industrial Ecology 1 (1/2/3): 24-38.

Ter Wal, Anne L.J., and Ron A. Boschma. 2011. "Co-evolution of Firms, Industries and Networks in Space." Regional Studies 45 (7): 919-933.

Tudor, Terry, Emma Adam, and Margaret Bates. 2007. "Drivers and limitations for the successful development and functioning of ElPs (eco-industrial parks): A literature review." Ecological Economics 61 (2-3) (March 1): 199-207.

van Berkel, René. 2009. "Comparability of Industrial Symbioses." Journal of Industrial Ecology 13 (4) (August): 483-486.

van den Bergh, Jeroen C.J.M., and John M. Gowdy. 2000. "Evolutionary theories in environmental and resource economics: Approaches and applications." Environmental and Resource Economics 17 (1): 37-57.

van den Bergh, Jeroen C.J.M., and Marco A. Janssen. 2004. The interface between economics and

industrial ecology: a survey. In Economics of Industrial Ecology: Materials, structural change, and spatial scales, 13-54. MIT Press.

Wassmer, Ulrich, Raymond Paquin, and Sanjay Sharma. 2012. "The Engagement of Firms in

Environmental Collaborations: Existing Contributions and Future Directions." Business & Society (March 28).

Watts, Duncan J. 2004. "The 'New' Science of Networks." Annual Review of Sociology 30 (1) (August): 243-270.

Weiss, Martin, Martin Patel, Hermann Heilmeier, and Stefan Bringezu. 2007. "Applying

Distance-to-target Weighing Methodology to Evaluate the Environmental Performance of Bio-based Energy, Fuels, and Materials." Resources, Conservation and Recycling 50 (3) (May): 260-281.

Wells, Peter, and Gillian Bristow. 2007. "Embedding eco-industrialism into local economies: the search for sustainable business and policy paradigms." Progress in Industrial Ecology 4 (3/4): 205.

White, R. M. 1994. Preface. In The greening of industrial ecosystems. Washington, DC: National Academy Press.

White, H. C. 2004. Markets from Networks: Socioeconomic Models of Production. Princeton, NJ: Princeton University Press.

Witt, Ulrich. 1997. "Self-organization and Economics—what Is New?" Structural Change and Economic Dynamics 8 (4): 489-507.

Wood, Richard, and Manfred Lenzen. 2009. "Aggregate Measures of Complex Economic Structure and Evolution." Journal of Industrial Ecology 13 (2) (April): 264-283.

Wright, Ramsey A., Raymond P. Côté, Jack Duffy, and John Brazner. 2009. "Diversity and Connectance in an Industrial Context." Journal of Industrial Ecology 13 (4) (August): 551-564.

Zaheer, Akbar, and Geoffrey G. Bell. 2005. "Benefiting from Network Position: Firm Capabilities,

Structural Holes, and Performance." Strategic Management Journal 26 (9) (September): 809-825.

Zhang, Ling, Zengwei Yuan, Jun Bi, Bing Zhang, and Beibei Liu. 2010. "Eco-industrial Parks: National Pilot Practices in China." Journal of Cleaner Production 18 (5) (March): 504-509.

Figure 1 Key Tie and Node Properties in Social-Material Network Analyses