Accepted Manuscript

Knowledge sharing and scientific cooperation in the design of research-based

policies: The case of the circular economy

Alessandro Marra, Marialisa Mazzocchitti, Alessandro Sarra

PII: S0959-6526(18)31504-X
DOI: 10.1016/j.jclepro.2018.05.164
Reference: JCLP 13021

To appear in: Journal of Cleaner Production

Received Date: 25 September 2017

Revised Date: 23 April 2018
Accepted Date: 20 May 2018

Please cite this article as: Marra A, Mazzocchitti M, Sarra A, Knowledge sharing and scientific
cooperation in the design of research-based policies: The case of the circular economy, Journal of
Cleaner Production (2018), doi: 10.1016/j.jclepro.2018.05.164.

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.
 ACCEPTED MANUSCRIPT

Knowledge sharing and scientific cooperation in the design of

research-based policies: the case of the circular economy

Alessandro Marra1, Marialisa Mazzocchitti2, Alessandro Sarra3

PT
1
“G. d’Annunzio” University of Chieti and Pescara; Dipartimento di Economia; viale Pindaro

RI
42 – 65127 – Pescara, Italy, alessandro.marra@unich.it
2
“G. d’Annunzio” University of Chieti and Pescara; Dipartimento di Economia; viale Pindaro

SC
42 – 65127 – Pescara, Italy, marialisa.mazzocchitti@unich.it
3
[Corresponding Author] “G. d’Annunzio” University of Chieti and Pescara; Dipartimento di
Economia; viale Pindaro 42 – 65127 – Pescara, Italy, alessandro.sarra@unich.it

U
AN
Abstract
M

Building on the idea that the more general and systemic the policy objectives, the
more cooperation and scientific support is needed for sound policy design, this paper
D

explores the potential for knowledge sharing and research-based cooperation in the
policymaking process in the field of circular economy. A sample of 1,244 research
TE

publications on circular economy from the Scopus database was collected to assess:
i) how the research on CE is distributed among the main disciplines of scientific
research, and how much attention each of them pays to the subject; ii) what potential
they exhibit for interdisciplinary communication and knowledge sharing; and iii) the
EP

potential for developing a transdisciplinary cooperation under a common conceptual

policy framework. The analysis shows that policies for circular economy at the
micro/meso level (firm or inter-firm initiatives) can rely on a much more adequate
C

knowledge base and potential scientific support than those at the macro level
(circular economy as a model of systemic sustainable development). To improve the
AC

results on the macro side, a number of disciplines (including social sciences,

economics and some behavioural sciences) should be solicited to devote more
space to the study of the subject and to develop the basis for mutual understanding
with other traditionally more involved research areas.

Keywords: circular economy; knowledge sharing; research-based policy; multi-inter-

trans-disciplinarity; cluster analysis; network analysis

1
 ACCEPTED MANUSCRIPT

1. Introduction

Since its introduction (Pearce and Turner, 1990), the concept of circular
economy (CE) has attracted great attention, because of the innovative way it looks at
the relationship between industry and the environment, with a view to improving the
long-term sustainability of production and consumption processes. CE has recently
been described as the promotion of “an economic system that replaces the ‘end-of-
life’ concept with reducing, alternatively reusing, recycling and recovering materials in

PT
production, distribution and consumption processes” (Kirchherr et al., 2017: 229). A
number of definitions of this kind have been proposed by the many researchers who
have studied the problem, sometimes with (slightly) different meanings, but having in
common the idea of the centrality of cyclical closed loops in the organization of the

RI
economic system (Murray et al., 2017). Many advantages can be obtained in this
way: a reduction of wasted energy, a more efficient use of natural resources, and, in
general, better control of the environmental impacts of production and consumption.

SC
They were described in the literature. Ayres (1989) assimilates the industrial
economy to the biosphere, and points out that industrial processes should be
designed like natural cycles to increase reliance on regenerative processes and to

U
increase efficiency, both in production and in the use of by-products. McDonough
and Braungart (2002) argue that natural resources (once extracted from nature)
AN
should be used over and over again within circular industrial processes. Stahel
(1997) claims that the ‘river’ structure of the industrial system needs to be converted
into a ‘lake’ structure in order to decouple economic success from resource
throughput.
M

A wide range of topics have found room within such a broad theoretical
framework, from product-related issues (ownership over lifetime, for example) to
reorganization of the whole economic process, with study of the profound and
D

interrelated changes in the industrial practices, business models, and patterns of

consumption required by a transition to CE (Ferdousi and Qiang, 2016).
TE

Acknowledging such variety, the literature has shown that CE can be approached on
three different levels—micro, meso, and macro—and that the implementation of CE
in the economic system and the relative policies can be studied on the same three
EP

levels (inter alia, Balanay and Halog, 2016; Banaité, 2016; Geng et al., 2009;
Ghisellini et al., 2016; Yuan et al., 2006; Zhu and Huang, 2005).
CE implementation at the micro level concerns relatively standard sustainable
development initiatives regarding the single business or product—i.e. practices
C

regarding eco-design and cleaner production (Franco, 2017). CE implementation at

the meso level concerns inter-firm initiatives—i.e. practices regarding the creation
AC

and maintenance of symbiotic relationships between firms and eco-industrial

networks, where firms are able to utilize industrial by-products such as heat energy,
wastewater and manufacturing waste (Geng and Doberstein, 2008). CE
implementation at the macro level concerns cities, provinces, regions, nations; it is
based on initiatives that involve redesigning the industry, the infrastructure, the
cultural framework, and the social system, from local to global (Saidani et al., 2017).
According to Sauvé et al. (2016), there is a clear distinction between macro and
other levels: “at the macro level—and in contrast to the micro and meso levels—
production and consumption become integrated” (Sauvé et al. 2016: 44). While the
implementation of CE at micro and meso levels takes place through initiatives

2
 ACCEPTED MANUSCRIPT

involving single or multiple firms or production processes, at the macro level it

requires initiatives involving socio-economic systems on different scales (national,
regional or even urban: for example, in the case of so-called ‘circular cities’). This
implies that, moving from micro/meso to the macro level, the complexity increases
(Su et al., 2013), and, as a consequence, the design of policy interventions becomes
more challenging, and their effectiveness more difficult to predict. Also in China—one
of the first countries, together with Japan and Germany, to consider CE as a part of a
wider policy for socioeconomic transformation and development, and to promote its

PT
implementation within a national programme—policy interventions supporting the
adoption of circularity principles at the macro level are reported to be less dynamic
and effective than those at the micro/meso levels by Lehmann et al. (2014) in a work

RI
where three country case studies and their different approaches to implementing CE
are compared, and by Su et al. (2013) in their review on practices for the
implementation of CE adopted in China.

SC
In the European Union, the situation does not greatly differ. Policymakers and
operators (industrial associations, large corporations, and financial institutions) seem
convinced that CE should be the leading conceptual benchmark for systemic
development strategies aimed at facilitating the decoupling of economic growth from

U
resource availability and other environmental constraints. The Communication from
the Commission to the European Parliament, the Council, the European Economic
AN
and Social Committee and the Committee of the Regions entitled ‘Closing the loop -
An EU action plan for the Circular Economy’ (European Commission, 2015a)
corroborates this view, also supported by (Franklin-Johnson et al., 2016). But despite
M

the many acknowledgments and declarations of principles, policies for the CE at the
macro level (also systemic policies for CE, hereinafter) have not found successful
implementation other than in the parts relative to more conventional issues, such as
D

material recovery and waste recycling (European Commission, 2016a). As pointed

out by McDowall et al. (2017), scale and place issues should have received greater
attention, i.e. through the individuation and funding of experimental zones at an
TE

urban, provincial or regional scale; moreover, a general approach could also have
taken into account other aspects, such as the promotion of resource-saving
behaviours, and of an environmentally friendly society through a process of
EP

‘ecological civilisation’ based on a common vision of environmental ethics (holistic vs

individualistic). In the same line Bačová et al. (2016: 5) argue, for example, that more
incisive education and awareness campaigns promoting sharing economy as well as
encouraging reuse and repair should have been part of the European approach.
C

They also observe that together with technological issues, the transition towards a
AC

CE calls for a strong dose of social innovation, and especially requires innovative
solutions for political and economic governance (new forms of cooperation among
administration, citizens and other stakeholders) also at the local level (Bačová et al.,
2016: 4). In such a framework, a number of legal issues, ranging from public law and
administrative law (involved in the design of these renewed relations) to private law
(rules of product liability as well as extended producer responsibility) become a
relevant part of the problem.
In our view, such a state of affairs partially depends on the intrinsic features of
the knowledge base on which policies for CE should be designed. As set forth by the
European Better Regulation Guidelines (European Commission, 2017) and the Better
Regulation Toolbox (European Commission, 2015b), rational policymaking requires a

3
 ACCEPTED MANUSCRIPT

conscious design of the measures to be adopted, and a careful evaluation of their

potential impacts. When it is the successful design of macro level policies for CE that
is at stake, such evaluation needs to be based on a thorough and complete
comprehension of the complex and (often) de-structured relationships between
natural, social, and economic systems. But, as will be shown in the following, the
knowledge base on CE seems not to be sufficiently varied, and poorly suited to
easing cross-disciplinary sharing. For this reason, major difficulties can emerge in
promoting effective cooperation among researchers/experts from many different

PT
fields, even if it is needed to support the quality of the policymaking process.
The idea that the formulation of sound and effective policy strategies for CE
calls for the synergic contribution of several fields of knowledge is not new in the

RI
literature. Ghisellini et al. (2016) claim that good prospects for policies aimed at
gradually improving the current production and consumption models depend on the
interdisciplinary framework underpinning CE. Heshmati (2016) argues that the

SC
implementation of CE requires greater intensification of the interdisciplinarity in
research. Lieder and Rashid (2016) emphasise that effective implementation of CE
can be ensured only by multidisciplinary approaches, involving business
perspectives, technological developments, and policies. Murray et al. (2017) discuss

U
the interdisciplinary perspectives inherent in the concept of CE that apply to the
implementation of sustainable businesses. Sauvé et al. (2016) affirm that the
AN
complexity and novelty of the CE model raises a number of practical challenges that
ask for solutions from experts in different disciplines, including natural sciences,
engineering, economics, and management.
M

As noted above, when the level of the policies increases from micro to macro,
the set of knowledge required for effective policymaking consistently widens. For
example, environmental studies, process engineering, management, and business
D

studies could provide the bulk of the competences mainly required within a micro
level policy approach, as Stahel (2016) points out when he discusses tipping points
in diffusion of circular business models, and also as Sauvé et al. (2016) claim, but if a
TE

macro level policy approach is in question, and CE is promoted as a model of

systemic development, a much broader range of disciplines could be useful. They
range from chemistry and industrial engineering to environmental and growth
EP

economics, from technical and economic regulation to industrial architecture, from

territorial and urban planning to biology, ecology, and sociology, as Bruce et al.
(2004) say in reference to the knowledge base necessary for the development of
effective solutions for complex problems, and Kanbur (2002) argues in a commentary
C

paper where he discusses critical issues related to the design of development

AC

policies. In particular, given the involvement of economic systems rather than

productive systems alone, the role of economics and social sciences as instruments
to evaluate the systemic implication of the measures adopted becomes particularly
relevant. For example, behavioural sciences can be useful to suggest the best
strategies to encourage changes in the consumption models (Thaler, 2016), and to
foster their transition from linear to circular.
As it is evident, not all such knowledge can be held by each participant in the
policymaking process. Researchers, operators, stakeholders, politicians and other
technicians, instead, should be able to share and combine their competences
throughout the process, from the early consultation stages to that of impact analysis.
Assuming that research on CE is the main source of knowledge for participants in the

4
 ACCEPTED MANUSCRIPT

policy process, one way to assess the potential of this kind of knowledge sharing is to
look at some characteristics of the scientific literature on CE to understand if it is wide
enough, and how easily an expert from a field can interact (understand and be
understood) and effectively cooperate with others in the complex evaluations implied
by the policymaking process.
With this framework in mind, three basic questions will be addressed in the
following: i) how the research on CE is distributed among the main disciplines of
scientific research, and how much attention each of them pays to the subject; ii) what

PT
potential they exhibit for interdisciplinary communication and knowledge sharing; and
iii) which disciplines are the best candidates to develop the kind of knowledge
sharing and strict research-based cooperation that is needed to sustain better

RI
policymaking processes. The answer to these questions helps to delimit the
knowledge base on which policymaking might rely, and to understand why policies
for CE seem to perform better when the emphasis is on the micro rather than on the

SC
macro level.
The analysis is performed on a sample of 1,244 research publications on CE
from the Scopus database. The three questions are answered using an index of
dispersion across different disciplines; an original and quantitative measure of the

U
potential for cross-disciplinary communication and a k-means clustering; and a
network analysis.
AN
The paper is organized as follows. Section 2 proposes a brief review of the
literature on the need for knowledge sharing and scientific cooperation across
disciplines in the study of CE. Section 3 presents the dimensions for investigating
M

knowledge sharing and research-based cooperation in policymaking. Section 4

describes the dataset and provides a methodological overview to address each
dimension. Section 5 illustrates and discusses the results. Section 6 concludes.
D

2. The macro level policies for circular economy and the need for knowledge
sharing and research-based cooperation: a review
TE

As noted above, the more general and systemic the purpose of the policies, the
broader is the spectrum of knowledge required for their design, and the stricter
EP

should be the cooperation among the actors in the policymaking process. This is
because the wider the scope of the policies, the more spheres of human activity can
be interested by the interventions, and the more complex and multi-faceted the
problems involved.
C

It is not a coincidence that one of the general principles stated by the European
Better Regulation Guidelines (European Commission, 2017) and Better Regulation
AC

Toolbox (European Commission, 2015b) is that policies and instruments must be

‘comprehensive’. It means that policies and tools should result from a decision-
making process in which all the relevant economic, social and environmental issues
and all the relevant impacts underlying alternative solutions are considered. This is a
central point. For example, the policies for ‘circular cities’ as a part of the CE policy
platform require the involvement of researchers, experts and other technicians in a
wide, diversified and coordinated spectrum that goes from the organization and
regulation of public services to the setting of environmental and quality standards for
economic activities, from public finance to urban planning, from social assessment to
integration and education policies, as it emerges from a study focussed on the

5
 ACCEPTED MANUSCRIPT

sharing economy and the role it can have in accelerating sustainable consumption
and production patterns in cities around the globe (Cohen and Muñoz, 2016), and a
study identifying common policy strategies to accomplish practical implementation of
sustainable solutions throughout the city (Prendeville et al., 2016).
Another better regulation principle concerns participation, according to which all
key stakeholders should be involved, including professionals with different
experiences and scholars with diverse backgrounds. In such a scenario, enhanced
reciprocal comprehension can reduce misalignments along the process. Interaction

PT
among people becomes easier and more effective if they are able to comprehend
each other and share their respective points of view and bodies of specialised
knowledge, and this can improve the quality of the policy solutions adopted (Marzano

RI
et al., 2006).
Not all stages of the policymaking process benefit from knowledge sharing and
scientific cooperation in the same way. Taking into consideration the European Union

SC
policy cycle (European Commission, 2017, p.5), the stages that more than others ask
for the availability of broad-spectrum scientific knowledge are those concerning
‘design and preparation’, in which policy proposals are formulated, and those
including ‘evaluation and revision’, in which the potential impacts of different policy

U
choices are assessed.
The need for knowledge sharing and research-based cooperation for effective
AN
policymaking is widely recognized in the literature. Mutz et al. (2014) claim that the
complexity of world problems asks for cross-disciplinary cooperation at both the
research and policy levels. Pohl (2008) stresses the utility of a co-production of
M

knowledge and argues that cross-disciplinary cooperation and collaboration in

knowledge production is crucial to ensure an adequate transfer of information among
people involved internally in the policymaking (and towards other external
D

stakeholders), and a comprehensive assessment of the social, ecological and

economic pros and cons of different scenarios. Owens and Driffill (2008) argue that
through effective cooperation among researchers, policy actors and stakeholders, it
TE

is possible to identify suitable policy levers and improve our ability to achieve
environmental and economic goals simultaneously. This is the case for several
energy/environmental policies. Due to their complexity, they cannot be designed by
EP

referring to technological, social, political or economic factors in isolation, but need a

multi-dimensional and integrated approach to develop viable solutions, as Howarth
and Monasterolo (2017) and Klenk and Meehan (2015) claim when they discuss
decision-making processes around climate change. Similarly, Ghisellini et al. (2016)
C

argue that only the development of an interdisciplinary knowledge base can offer an
AC

appropriate framework for policy actions in the field of CE to boost, for example, the
adoption by firms of cleaner production schemes, increase consumers’ awareness
about end-life products, encourage the use of renewable technologies and materials,
and thus minimize the overall employment of resources.
Policymakers, too, have perceived the need to improve the knowledge base on
which policies are designed. This gap is particularly suffered in the case of macro
level policies for CE, which have been defined as less vibrant than the micro/meso
ones (Heshmati, 2016). They ask for a multifaceted knowledge base, which is still
under construction (Murray et al., 2017). Some authors ask for more attention to this
specific aspect (Pomponi and Moncaster, 2017). Unfortunately, it seems that such a
demand has not been adequately met by researchers, despite the availability of

6
 ACCEPTED MANUSCRIPT

considerable financial support, for example, within the Horizon 2020 Programme
(European Commission, 2016b) and some best practices (see, for example, the
DYNAMIX and POLFREE projects included in the Seventh Framework Programme,
exploring the potential of resource efficiency in Europe).

3. Investigating the potential for knowledge sharing and cooperation in the

policymaking process: hints from the literature on the cross-disciplinarity
of research

PT
As explained in the introduction, the quality of policymaking is strongly
influenced by the specialised knowledge owned by the actors in the process (think for
example at the wide range of technical knowledge – economics, finance, chemistry,

RI
physics – involved in the impact assessment of the EU Emission Trading System),
and by the intensity of the cooperation they can put in place. Passing from micro to
macro level policies for CE, a wider body of knowledge has to be involved and a

SC
stronger integration of different competences is required. The question arises,
therefore, of whether the knowledge base is appropriate to support the kind of
cooperation required by macro level policymaking. The answer is strongly related to

U
some characteristics of the scientific literature on CE. The concepts of
multidisciplinarity, interdisciplinarity and transdisciplinarity of research offer useful
AN
hints on how to explore them.

3.1. The concept of multidisciplinarity

M

Several authors defines multidisciplinarity as a situation in which more than one

discipline works separately on the same issue, considering it from different
perspectives (inter alia, Morillo et al., 2003; Mugabushaka et al., 2016; Van Den
D

Besselaar and Heimeriks, 2001). In such a case, there is usually a mere juxtaposition
of disciplines (Jacobs, 1989), each working in parallel (Aboelela et al., 2007), or
TE

sometimes sequentially (Rosenfield, 1992), and independently of each other (OECD,

2010), each according to its own discipline-specific perspective – this last point is
also supported by Barry and Born (2013) and Stokols et al. (2008). Researchers from
different disciplines also operate separately (Alvargonzález, 2011), maintaining their
EP

distinctiveness (Choi and Pak, 2006). Klein (1990) describes this style of research as
additive and not integrative, a definition that seems to be consistent with those who
argue that the outcome of multidisciplinary research is the sum of individual parts
C

(Wilson and Pirrie, 2000). Klein (1990) cites the example of a centre for Asian studies
that houses specialists from Oriental history, economics and sociology.
AC

A measure of the level of multidisciplinarity of a research topic could be the first

step towards an assessment of the potential for knowledge sharing and research-
based cooperation in the policymaking process. According to the given definition, the
level of multidisciplinarity in CE should strictly depend on: a) the number of
disciplines dealing with CE; b) the number of research outputs for each discipline;
and c) how equitable the distribution of research outputs is across the different
disciplines. A high level of multidisciplinarity of the scientific knowledge base
underlying the CE brings many benefits for the policymaking process. It means that
the research topic is well known within a relevant number of research areas;
therefore, there is a higher probability that the scientific base includes the knowledge
required to find solutions to more complex and emerging policy problems. Moreover,

7
 ACCEPTED MANUSCRIPT

it is easier for different disciplines to help answer relevant questions raised in the
design or in the evaluation stages, or for a common interest to be solicited in specific
issues, according to the needs of the policy cycle.

3.2. The concept of interdisciplinarity

A number of authors claim that research is interdisciplinary if it involves actual

interaction between two or more disciplines (see for example: Atkinson, 2012; Moran,

PT
2002; Rosenfield, 1992; Stokols et al., 2008; Tress et al., 2005). Interdisciplinarity
has been defined by the US National Academy of Sciences as a mode of research
involving teams or individuals that integrate information, data, techniques, tools,
perspectives, concepts and/or theories from two or more disciplines or bodies of

RI
specialised knowledge to advance fundamental understanding or to solve problems,
the solutions to which are beyond the scope of a single discipline or area of research
practice (National Academy of Sciences, 2005). The distinctive feature of

SC
interdisciplinarity is the integrative nature of the cognitive content of research
(Huutoniemi et al., 2010); it leads to outcomes that are greater than the sum of its
individual disciplinary parts, as supported by Wilson and Pirrie (2000) and

U
demonstrated by Larivière et al. (2015) on the success and failure of interdisciplinary
papers.
AN
Interdisciplinarity needs some degree of cooperation across disciplines (Moran,
2002), and such cooperation, in turn, needs researchers from different disciplines to
be able to integrate sector-specific knowledge effectively, to discuss research topics
with full reciprocal understanding, and to exchange ideas. Its fundamental
M

prerequisite is efficient communication between experts from different research fields

– this is emphasised in a number of studies (Bracken and Oughton, 2006; Hebert et
al., 2009; Marzano et al., 2006) – based on a common and easily understandable
D

language, or even meta-language (Hübenthal, 1994).

A high potential for interdisciplinary cooperation among the actors in the
TE

policymaking process can therefore be deduced by extensive use of a common

language in the research outputs on CE. Knowledge sharing and research-based
cooperation among the actors in policymaking are easier if they derive their
EP

specialised knowledge from research areas that use a common language in the
dissemination of their findings. When this is the case, communication among experts
from different disciplines is facilitated, and cooperation is expected to be much
smoother and more efficient.
C

3.3. The concept of transdisciplinarity

AC

The third dimension of the potential for cooperation is strictly linked to the
attitude of different research areas towards developing a transdisciplinary approach
to research. Transdisciplinarity refers to the opportunity to systematise
interdisciplinary research according to a unifying vision (Klein, 1990). According to
Gibbons et al. (1994), transdisciplinarity arises when researchers from different
disciplines are encouraged to surpass their separate conceptual, theoretical and
methodological orientations to develop an approach based on common conceptual
and epistemological frameworks. This last is a recurrent element in the definitions of
‘transdisciplinarity’ formulated by the various authors, albeit under slightly different
denominations: Rosenfield (1992) uses ‘comprehensive theoretical framework’,

8
 ACCEPTED MANUSCRIPT

Stokols (2006) defines it as a ‘shared conceptual framework’, and Klein (2008, 2003)
uses ‘common axiom that transcends separate disciplinary perspectives’.
These common frameworks can be grounded in theory or in policy (Klein,
1990). In the latter case, they should mainly be defined by policymakers and shared
with the researchers, operators, experts, stakeholders and other technicians called to
support the process of policy design. A common conceptual policy framework will
ensure a shared comprehensive vision of the policy goals and a common paradigm
for the assessment and discussion of the implications of policy measures, as

PT
required by the guidelines for better regulation. It should be based on a strong
recognition of the reciprocal utility of the different disciplines for the policymaking
process, and should consist of a common perception of: 1) the policy objectives and

RI
the general strategies to achieve them; 2) the relevant implications of policy choices
and the methods to assess their impact; and 3) the relevance of other disciplines in
contributing to the knowledge base and the methodological tools for the analysis

SC
required by a rational policymaking cycle. If a relevant number of different disciplines,
coherent with the needs of macro level policies for CE, show a high potential to
develop a cooperation within such a conceptual framework, the effectiveness of the
interaction among the participants in the policymaking process and the ability to

U
address complex challenges within a macro level approach are significantly
improved.
AN
A first hint of the attitude towards such cooperation can be achieved by
performing a network analysis aimed at assessing the intensity of the connections
across academic disciplines involved in the research on CE. Strong connections are
M

a basic prerequisite for the development of a transdisciplinary approach, both to

research and to applied reasoning like that implied by the policymaking process. The
problem is to understand which disciplines are more interlinked, and what level of
D

policymaking—micro, meso or macro–they are suited to support.

4. Data and methods

TE

4.1. Data

The potential for knowledge sharing and research-based cooperation in the

EP

policymaking process is assessed through the analysis of a sample of research

outputs published on CE. In particular, the dataset is composed of bibliometric data
on all the research outputs indexed in Scopus that contain the term ‘circular
C

economy’ in the title, abstract or keywords. Over the period January 2001 to
December 2016, the dataset includes a total of 1,281 research outputs. Having
AC

removed retracted articles and duplications, 1,244 research outputs remain.

Figure 1 illustrates that the number of CE-related publications has grown
substantially since 2004. Their annual number doubled between 2005 and 2007,
increased almost four times between 2007 and 2014, and peaked at 345 in 2016.
The main types of publication driving the growth in the sample are journal articles
and conference proceedings. While from 2004 to 2009 and after 2014 the number of
journal articles was greater than conference proceedings, from 2009 to 2014 the
reverse was true.

[Here Figure 1]

9
 ACCEPTED MANUSCRIPT

Figure 1. Number of CE-related outputs per year.

In the period 2001–2016, the researchers who contributed to the scientific

production on CE were affiliated to institutions located in 49 countries: 27 European
countries, 13 Asian countries, seven American countries, Morocco and Australia.
Only a small portion of researchers active in the CE field collaborated with colleagues
affiliated to institutions in other countries. As a result, an international collaboration
network emerges where nodes are countries (in which the institution of affiliation is

PT
located) and links between nodes are given by the collaboration between authors
from two different countries. The network shows 39 interconnected nodes (see
Figure A1 in the Appendix A). The relationship between China and the United States
is the strongest (nine co-authorships), followed by those between China and Japan

RI
(six), China and Canada, China and Italy, and Germany and the Netherlands (five).
The remaining ten countries, where authors carried out their research not in
collaboration with colleagues from other countries, are Argentina, Brazil, India, Israel,

SC
Latvia, the Philippines, Slovakia, Switzerland, Thailand and Turkey.
With regard to the research areas involved in the CE field, the Scopus database
also offers an easy way to attribute the research outputs to the research areas to

U
which they belong. Document sources (journals, book series and conference
proceedings) are classified under four broad subject clusters (level 1), which are
AN
further divided into 27 major subject areas (level 2), and 313 minor subject areas
(level 3). The four subject clusters coincide with the four main scientific areas
(physical, social, health and life), while the 27 major subject areas (SAs) correspond
to the (academic) disciplines (Table 1); finally, the 313 minor subject areas are sub-
M

disciplines or thematic topics. Each research output automatically inherits the subject
code (in some cases there is more than one) assigned to the source in which it is
published.
D
TE
C EP
AC

10
 ACCEPTED MANUSCRIPT

Table 1. Scopus major SAs (grouped by subject cluster).

Physical Sciences Social Sciences Life Sciences Health Sciences

Agricultural and
Chemical Arts and Humanities
Dentistry (Den) Biological Sciences
Engineering (ChEn) (A&H)
(ABS)
Biochemistry,

PT
Business
Health Professions Genetics, and
Chemistry (Chem) Management and
(HePr) Molecular Biology
Account. (BM&A)
(BGMB)

RI
Computer Science Decision Sciences Immunology and
Medicine (Med)
(CoSc) (DeSc) Microbiology (I&M)

SC
Economy
Earth Sciences
Econometrics and Nursing (Nur) Neuroscience (Neu)
(EaSc)
Finance (EE&F)

U Pharmacology,
AN
Engineering (Eng) Psychology (Psy) Veterinary (Vet) Toxicology, and
Pharmaceutics (Phar)
M

Social Sciences
Energy (Ener)
(SoSc)
D

Environmental
Science (EnSc)
TE

Material Science
(MaSc)
EP

Mathematics (Math)
C
AC

Physics and
Astronomy (P&A)

th
Note: 26 subject areas are reported in the table above. The 27 SA is multidisciplinary, which by
definition cannot be considered a subject area per se and thus has been excluded from the analysis.

11
 ACCEPTED MANUSCRIPT

4.2. The level of multidisciplinarity of the literature on CE

In order to assess the level of multidisciplinarity, the multidisciplinarity index (MI)

proposed by Heinze and Bauer (2007) was adapted to explore the dispersion of
research outputs across SAs:



 = 1 − 1 −
2 ∑ 
  + 1 ∑  (1)

PT
where ∑   denotes the number of SAs actually involved in the study of the

RI
subject under investigation out of the total possible SAs, and w is the cumulative
percentage of the outputs per SA over the total number of research outputs in the
sample. The MI sums up the information on the number of SAs involved in the study

SC
of CE with the dispersion of research outputs across the disciplines.
Since this formula includes the Gini index, it can also be expressed as follows:

 =
1 −  ∑  
U (2)
AN
The value of the index equals the number of the SAs involved in the study of
CE when the research outputs are equally distributed across SAs; it equals 0 if all the
M

research outputs are concentrated in a single SA or discipline.

4.3. The potential for interdisciplinarity

D

The second relevant dimension to evaluate the potential for knowledge sharing
TE

and research-based cooperation in the policymaking process for CE is associated

with the extent to which research outputs from different SAs make use of a common
language. The use of common technical terms is a prerequisite for interdisciplinary
cooperation, since it makes communication easier among researchers, experts,
EP

stakeholders and other subjects participating in the policymaking process with

specialisation in different SAs, and favours a synergic approach to the policy
problems implied by CE. Furthermore, a common language increases the
C

comprehensibility of research results for experts from other disciplines, thus making
their diffusion across a wider scientific audience easier and enhancing the potential
AC

for cooperation.
To capture these aspects, an original measure of potential interdisciplinarity is
proposed. It is based on the idea of representing quantitatively, rather than
qualitatively, the measure by which research products use a language shared with
disciplines (SAs) different from those from which they originate.
To that end, a keyword analysis in four stages was performed. In step 1, the
keywords of all research publications relating to each SA were grouped together.
This process was repeated for every SA. The number of ‘baskets’ of keywords
equals the number of major SAs to which are attributed research outputs relating to
CE. In step 2, the weight of each keyword within each single basket was calculated
(that is, if keyword xi is mentioned three times in basket Zj out of a total of 100 entries

12
 ACCEPTED MANUSCRIPT

of keywords, the weight of xi in basket Zj is 3%). In step 3, the values of each

keyword within each basket were reattached to the research outputs in which they
were used, so that publications were depicted not in terms of qualitative keywords,
but in terms of the quantitative relevance of each keyword within each SA. In step 4,
the final matrix was constructed with a number of rows equal to the number of
research outputs considered and a number of columns equal to the number of major
SAs to which are attributed research outputs relating to CE. The generic element aij
of the matrix represents the total relevance of the keywords employed per research

PT
output in each single basket or SA. Values along each row of the matrix can thus be
interpreted as indices of the degree of affiliation of each research output to different
SAs. From another point of view, such values express the degree to which research

RI
outputs associated with a specific SA might be understood by researchers from other
SAs. This is a good measure of the interdisciplinary comprehensibility of each
research output and of the extent to which they could contribute to knowledge

SC
sharing and to fertilizing research in other SAs.
A k-means analysis (one of the simplest unsupervised learning algorithms for
resolving the well-known clustering problem) is then proposed to classify the sample
through k clusters, with k (the number of clusters) fixed a priori, according to the level

U
of the potential for interdisciplinary cooperation and knowledge use. The main goal of
such an analysis is to determine the share of the research outputs with a relevant
AN
degree of interdisciplinary comprehensibility of the total number of research outputs.
The higher this share, the stronger the potential for knowledge sharing and research-
based cooperation in policy-related issues.
M

4.4. The potential for transdisciplinarity

The proposed measure of interdisciplinarity works at the research output level

D

and not at the discipline level: we can realize how many and what share of the total
number of publications are actually characterized by high interdisciplinary
TE

comprehensibility, and from this derive the potential for knowledge sharing and
research-based cooperation in the policymaking process. It is not possible instead to
derive from such a measure any indication about which specific SAs effectively show
EP

the strongest potential for transdisciplinary cooperation, and what kind of policy
approach to CE, micro/meso or macro, can be best supported.
This question can be approached through a network analysis linking SAs on the
basis of the co-occurrence of keywords used for both. The hypothesis is that
C

keywords synthesize characterizing concepts, methodologies and approaches. When

they are consistently shared by different SAs (when the edges of the network through
AC

which they are connected are stronger), these SAs could develop the reciprocal
recognition and intense knowledge sharing which is the basis for transdisciplinary
cooperation under a common conceptual policy framework. Obviously, the more SAs
are strictly linked, the wider, more inclusive and systemic the common conceptual
policy framework can be under which they can cooperate.

5. Results and discussion

5.1. The level of multidisciplinarity of the literature on CE

13
 ACCEPTED MANUSCRIPT

Almost all disciplines (22 out of the 26 SAs) have been directly involved in the
study of CE (with research outputs directly citing CE in the title, abstract or keywords)
over the entire period under consideration (see Table 1 for the abbreviations of the
SAs): BGMB, I&M and Phar in the subject cluster of life sciences; A&H, BM&A,
DeSc, EE&F and SoSc in the subject cluster of social sciences; all SAs composing
the subject cluster of physical sciences; HePr, Med and Vet in the subject cluster of
‘health sciences’. Other disciplines could have contributed in an indirect way, for
example by providing theoretical and empirical background for the research outputs

PT
attributed to the directly involved SAs. Therefore, the disciplinary involvement in the
study of CE or in related aspects seems to be very wide.
However, data show that there is a non-homogeneous dispersion of research

RI
outputs across SAs. According to the MI index, which in this case can assume values
from 0 to 22, multidisciplinarity in the CE research field is low at 8.8.

SC
[Here Figure 2]

Figure 2. Number of CE-related outputs per single SA.

U
Note: The colours indicate the subject clusters: dark grey=physical sciences; light grey=social
sciences; white=life sciences; black=health sciences. See Table 1 for the abbreviations of the SAs.
AN
The study of CE (Figure 2) is heavily concentrated in physical sciences, and in
particular in Eng (536), EnSc (432) and Ener (208), which represent the top three
M

SAs in the list of disciplines by number of research outputs. BM&A (199) and SoSc
(164), respectively the fourth and the fifth in ranking per number of research outputs,
are the most productive disciplines within the social sciences area, which is the most
D

productive scientific subject cluster behind physical sciences. Note that DeSc (58)
encompasses some sub-disciplines closely linked to BM&A, such as management
TE

science and operational research, and information systems and management. By

summing up the research outputs published by researchers in DeSc and outputs
from researchers involved with BM&A, the involvement of business-related
disciplines is significant. The same cannot be said for economic disciplines that study
EP

the structure and the development of socioeconomic systems: EE&F (83) is ranked
eighth.
Disciplines belonging to life sciences and health sciences offer only a negligible
C

contribution: excluding ABS (within the life sciences area), which is ranked tenth in
the list per number of research outputs (with 76 publications), all other disciplines
AC

belonging to these subject clusters are ranked very low, with positions below
fourteenth.
Finally, it has to be noted that two of the four SAs not yet involved in the CE
research field are Psy and Neu. Until 2016, no research output belonging to such
SAs has been indexed in Scopus as directly addressing CE-related issues. They are
two important sections of the behavioural sciences, the utility of which in the study of
the evolution of social systems is more and more recognized in recent times. As said
in the introduction, an effective implementation of CE requires profound changes not
only to industrial practice but also to patterns of consumption; psychological
researches could help to understand the determinants of consumer behaviour
models and how they are related to the shift from linear to circular.

14
 ACCEPTED MANUSCRIPT

The evolution of the level of multidisciplinarity in the literature on CE across

time has been also studied. The period under observation has been divided into
three temporal windows (2001–2006, 2007–2011 and 2012–2016) and the MI has
been calculated for each window. Results show that the number of SAs has
increased, even though the dispersion of research outputs has gradually decreased
over time (see Figure A2 in the Appendix A). In fact, the MI index assumes a value of
7.54 (with the maximum possible value equal to 13, given by the number of involved
SAs) for the first temporal window (2001–2006); a value of 8.35 (with the maximum

PT
possible value equal to 19) for the second window (2007–2011); and a value of 8.62
(with the maximum possible value equal to 21) for the third window (2012–2016).
This trend is mainly due to a rapid growth in the research outputs relative to technical

RI
disciplines useful for both micro/meso and macro approaches (Eng, EnSc, Ener), and
BM&A, a typical discipline of micro interest.

5.2. The potential for interdisciplinarity

SC
As seen above, the mere dispersion of research outputs across different SAs
gives only a preliminary clue with regard to the potential for knowledge sharing and

U
cooperation in research. Eng, EnSc, Ener, BM&A, SoSc and CoSc represent the bulk
of scientific publication on CE and seem to be the best candidates to cooperate and
AN
share knowledge. Additional information can be obtained by looking at the level of
interdisciplinary comprehensibility of research outputs (see section 4.3).
The reduction of the number of SAs to the most relevant ones (Eng, EnSc,
Ener, BM&A, SoSc, CoSc, ChEn, EE&F, MaSc, ABS, EaSc, DeSc and Math), while
M

still accounting for more than 97% of the total number of research outputs, eases the
reading of the results of the analysis. The k-means algorithm partitions the dataset
into eight clusters of research outputs, from A to H. Table 2 shows the value of
D

centroids for each cluster per single relevant SA.

TE

Table 2. Centroids per single cluster of research outputs.

Eng EnSc BM&A Ener CoSc SoSc ABS DeSc EaSc EE&F ChEn MaSc Math
EP

G 0.44 0.55 0.50 0.54 0.39 0.57 0.64 0.46 0.35 0.58 0.44 0.28 0.29
E 0.29 0.35 0.32 0.34 0.26 0.35 0.38 0.29 0.24 0.38 0.27 0.19 0.25
B 0.18 0.22 0.20 0.21 0.17 0.21 0.23 0.19 0.17 0.24 0.19 0.14 0.17
C

A 0.12 0.14 0.14 0.14 0.12 0.14 0.15 0.14 0.11 0.15 0.12 0.09 0.11
AC

D 0.08 0.09 0.08 0.09 0.07 0.08 0.09 0.08 0.07 0.09 0.08 0.06 0.07
H 0.05 0.05 0.05 0.04 0.05 0.04 0.05 0.05 0.05 0.05 0.03 0.03 0.05
C 0.03 0.03 0.03 0.02 0.03 0.03 0.03 0.02 0.04 0.02 0.07 0.08 0.03
F 0.02 0.01 0.02 0.01 0.02 0.01 0.01 0.01 0.02 0.01 0.02 0.01 0.01

Note: See Table 1 for abbreviations of the SAs.

Clusters G and E include those publications with the highest degree of

interdisciplinary comprehensibility. Despite outputs in cluster G relating to a lower
number of SAs than in cluster E, they show a similar pattern when only the most
relevant SAs are considered (values of centroids are relatively high in ABS, EE&F,

15
 ACCEPTED MANUSCRIPT

SoSc, EnSc, Ener, BM&A and Eng). The research outputs grouped in these clusters
share a language that can be easily understood by researchers from disciplines that
are different and distant from each other (the subject clusters of physical, social and
life sciences are involved). Unfortunately, they are a large minority of the total (their
shares of the total research are 2.3% and 4.4% respectively).
Cluster B, the third by degree of interdisciplinary comprehensibility, accounts for
8% of all research outputs and presents a similar pattern to that of clusters G and E.
Cluster A accounts for 10.9% of all research outputs, while clusters D (18.4%), H

PT
(19.7%), C (8.8%) and F (27.5%) show gradually lower levels of interdisciplinary
comprehensibility. In particular, cluster C includes research outputs with a low degree
of affiliation to most relevant SAs, except ChEn and MaSc, SAs that are far less

RI
central in all other clusters, mainly because they are characterized by very technical
language and knowledge use that is difficult to understand outside these specific
disciplines.

SC
In conclusion, the share of the research outputs with a significant degree of
interdisciplinary comprehensibility is very low (6.7% or 14.7%, depending on the
inclusion of the third cluster). However, it seems not to be constrained to certain SAs,
but rather to be distributed across all relevant SAs. The potential for knowledge

U
sharing across SAs therefore remains very low. Moreover, no significant sign of a
trend towards greater integration emerges from the average date of publication (year
AN
and month) of research outputs per single cluster (2012.2 for cluster G versus 2011.7
for F, respectively the most and the least interdisciplinary).

5.3. The potential for transdisciplinarity

M

To investigate the potential for transdisciplinary cooperation, the open-source

software R (R Core Team, 2017) was used to build a network with 20 nodes (SAs)—
D

all disciplines involved in the CE research field except Vet and I&M, which have no
connections—and 190 edges (Figure 3). The thickness of the edges is proportional to
TE

the number of times two SAs share a keyword. As said in section 4.4, the stronger
the edges which link two SAs, the stronger is their potential for transdisciplinary
cooperation.
EP

[Here Figure 3]

Figure 3. Output of the network analysis across SAs (restricted view).

C

Note: See Table 1 for abbreviations of the SAs.

AC

The most connected nodes were found to be those corresponding to technical

disciplines useful for both micro/meso and macro approaches. In fact, among the
strongest links are those between Eng and EnSc (391 keywords and terms in
common), Ener and EnSc (328), and Ener and Eng (319). BM&A is strongly
connected with Eng (336), EnSc (318) and Ener (290). The connection would be
even stronger were DeSc considered a part of managerial disciplines. To sum up,
SAs more interested at the micro/meso level of CE (industrial processes and
systems) show good potential for mutual recognition and for developing
transdisciplinary cooperation in policy design.

16
 ACCEPTED MANUSCRIPT

A very different picture is that of the disciplines which contribution is especially

required for the design of macro level policies for CE. While the connections of SoSc
with EnSc (321) and Eng (312) are strong, much less tight are those between SoSc
and EE&F (72), and the remaining SAs. Moreover, EE&F has only very weak
connections with other SAs and a moderate value of the clustering coefficient (a
measure of how complete the neighbourhood of a node is) within the network, with
values just below the average. Finally, some SAs which would be typically
considered within the behavioural sciences (Psy and Neu) are not even part of the

PT
network, having produced no research output on CE in the considered period.
Due to the loose interconnections of these SAs (SoSC, EE&F, Psy and Neu),
among themselves and with the others, it is unlikely that experts from these fields

RI
could adhere to a widely shared conceptual policy framework. The level of
transdisciplinary cooperation useful to improve the quality of the macro level policies
for CE seems, therefore, difficult to be attained.

SC
To better show the evolution across time of the relationships between different
SAs, a network per year for the entire period has been drawn. Each network has
been drawn adding new relationships established during the year to those
established in the previous years. A static visual summary of the main changes in the

U
network is shown in Figure A3 (in Appendix A), while the evolution of the network can
be watched at the link below (see Appendix B). Figure A3 displays a network with
AN
one node in 2001 because only researchers publishing within the Ener discipline
were involved in the study of CE. In 2004, a network with eight nodes (Eng, EnSc,
Ener, CoSc, SoSc, EE&F, Math and Phar) and 24 edges emerged. From 2005 to
M

2008, every year some SAs have been added to the network: in 2005, ABS, EaSc,
Chem and MaSc, but these last two were isolated nodes; in 2006, BM&A; in 2007,
DeSc; in 2008, BGMB. Med and HePr joined the network in 2010, while Chem and
D

P&A established their first relations with other disciplines studying CE in 2012. A&H
was the last SA to join the network.
What is striking, in this case, is that EE&F emerged early as a node of the
TE

network, but its connections developed and diversified very slowly. Quite different
was the evolution of BM&A, which appeared two years later in the network, but
quickly established and intensified connections with other nodes. In other words, the
EP

earlier interest of economic disciplines in the study of CE was not followed by the
development of a structure of relations comparable to the one built by business-
related disciplines. As said, this circumstance has probably limited the ability of the
experts in economic disciplines to effectively and easily participate in
C

transdisciplinary policymaking processes. Also, this result can help to explain the
AC

lower dynamism of macro level policies for CE compared to micro/meso policies,

often reported in literature.

6. Conclusions

When research-based policies are at stake, a better regulation in the sense

adopted by the European Institutions is strongly influenced by the specialised
knowledge owned by the actors in the process, and by the intensity of the
cooperation they can put in place. Passing from micro to macro level policies for CE,
a wider body of knowledge is necessary and a stronger integration of different
competences is required. In particular, given the involvement of economic systems

17
 ACCEPTED MANUSCRIPT

rather than productive systems alone, the role of economics and social sciences as
instruments to evaluate the systemic implication of the measures adopted becomes
particularly relevant. The same can be said for psychology and other behavioural
sciences with reference to demand/consumption related issues.
The paper builds on the idea that the lower dynamism of macro level CE
policies in comparison to the micro/meso level ones could be partially due to some
intrinsic features of the knowledge base on which policies for CE are designed.
In this paper, an analysis of the potential for knowledge sharing and research-

PT
based cooperation in the policymaking process has been presented in terms of the
level of multidisciplinarity and the preconditions for interdisciplinary and
transdisciplinary research displayed by the scientific literature on CE.

RI
The main findings can be summarised as follows.
First, a wide number of disciplines are involved in the CE research field, but
most of the available scientific knowledge on the subject has been produced by just a

SC
few of them. In particular, less involvement of economics and social sciences has
been registered, and no involvement at all of psychology and other sections of the
behavioural sciences is registered, so that the body of knowledge seems to be best
suited to policymaking at the micro/meso level rather than at the macro level.

U
Moreover, during the entire period taken into consideration, a slight trend towards
greater concentration of scientific production has occurred. In order to provide a
AN
knowledge base suitable to address CE-related issues at the macro level, the body of
scientific knowledge should be enriched, promoting research in the fields more
connected with the socioeconomic systemic implications of CE.
M

Second, the number of research outputs comprehensible to researchers and

experts from different backgrounds represents a very small part of the total research
developed on CE and comes from a reduced number of research areas. The use of a
D

common language, therefore, seems to be very limited. Anyway, among them, both
the disciplines more linked to micro/meso aspects and those linked to the macro
ones are represented, and the research outputs make use of a common language to
TE

a high degree. This is a sign that at least a few aspects of CE can be debated from
an interdisciplinary perspective. This situation does not seem to have significantly
changed over the period of analysis. In order to make the exchange of knowledge
EP

among different disciplines and the dialog among the participants in the policymaking
process easier, a wider effort to promote the use of a common language in the
research on CE should be put in place.
Third, the best candidates among the academic disciplines to develop the kind
C

of transdisciplinary research-based cooperation that is needed to sustain better

AC

policymaking processes seem to be found mainly in the disciplines useful for

micro/meso approaches to CE. The low connectedness of social sciences and
economics, of economics with the rest, and the absence of at least psychology
among behavioural sciences could hinder the development of a common conceptual
policy framework suited to supporting the policymaking cycle in the case of macro
level policies for CE. In order to facilitate the development of a common conceptual
policy framework suited to macro level policymaking, a better integration of social
sciences, economics and behavioural sciences should be encouraged.
In the European Union, these results were usually pursued by promoting
research or cooperation programmes involving a minimum number of different
disciplines. They can give relevant help in relation to points 2 and 3 above. Those

18
 ACCEPTED MANUSCRIPT

instruments, anyway, give the best results when a consolidated research base on CE
is already present in the different disciplines. Unfortunately, in the case of macro
level policy programmes for CE, some relevant disciplines seem not to have a
pertinent research base at present, and this should be built, starting with the
involvement of the university system. For example, doctoral programmes on CE
should be encouraged in the fields of economics, social sciences and behavioural
sciences, as well as a specialised professional formation through Master’s degrees.
Future research efforts should include the use of alternative ways to attribute

PT
research outputs to the SAs in order to make such attribution more adherent to their
specific content. A possibility is that of elaborating on the reference list of each
research output instead of using the SA attributed by the Scopus database to the

RI
document source.
This research can also be enriched gathering information on the main
disciplinary specialisation of the authors of co-authored research outputs. In this

SC
case, more precise information on the real connections among disciplines could be
obtained, building a network where the nodes are the reference disciplines of each
author and the intensity of the links is proportional to the number of times two authors
belonging to different nodes are co-authors in a research output. Moreover, by

U
adding to this the already available information on the affiliation of the authors, it
should be possible to assess if a geographical pattern of specialisation of the
AN
knowledge base emerges.
Finally, a new interesting direction for further research is that of complementing
the analysis proposed here with a similar exercise on actual policy documents. The
M

research should focus on the range of knowledge used, the potential range of
knowledge that could have been used, and on the benefits that might derive from
widening the knowledge base. Nevertheless, some critical issues must be addressed
D

before moving in this direction, as the identification or the development of valid and
reliable methods of extracting keywords from policy documents, together with
rigorous criteria to refer the extracted keywords to subject areas and weight their
TE

relative relevance.
C EP
AC

19
 ACCEPTED MANUSCRIPT

Appendix A

[Here Figure A1]

Figure A1. The international collaboration network in the CE field.

PT
Note: Links between nodes refer to collaborations between researchers affiliated to institutions located
in different countries. Bubble sizes are proportional to the number of research products produced by
the country. The thickness of the lines (edges) between countries represents the intensity of
collaboration (number of co-authored research products between each pair).

RI
SC
[Here Figure A2]

U
Figure A2. Number of CE-related outputs per single SA in 2001–2006, 2007–2011 and 2012–2016.
AN
M

[Here Figure A3]

D

Figure A3. Evolution of relationships between disciplines across time.

TE
EP

Appendix B

An interactive video displaying temporal evolution of the network in Figure 3

(from 2001 to 2016) has been created using the NDTV package in R (Bender-deMoll,
C

2016). In addition to playing and pausing the movie, it is possible to click on nodes
and edges to show their IDs, double-click to highlight neighbours, and zoom into the
AC

network at any point in time.

Acknowledgement

This research did not receive any specific grant from funding agencies in the
public, commercial or not-for-profit sectors.

20
 ACCEPTED MANUSCRIPT

References

Aboelela, S.W., Larson, E., Bakken, S., Carrasquillo, O., Formicola, A., Glied, S.A.,
Haas, J., Gebbie, K.M., 2007. Defining interdisciplinary research: conclusions
from a critical review of the literature. Health Serv. Res. 42, 329–346.
doi:10.1111/j.1475-6773.2006.00621.x

Alvargonzález, D., 2011. Multidisciplinarity, interdisciplinarity, transdisciplinarity, and

PT
the sciences. Int. Stud. Philos. Sci. 25, 387–403.
doi:10.1080/02698595.2011.623366

Atkinson, M., 2012. Key Concepts in Sport and Exercise Research Methods. Sage.

RI
Ayres, R.U., 1989. Industrial metabolism, in: Technology and Environment.
Washington, DC: The National Academies Press. https://doi.org/10.17226/1407.

SC
pp. 23–49.

Bačová, M., Böhm, K., Guitton, M., Herwijnen, M. van, Kállay, T., Koutsomarkou, J.,
Magazzù, I., O’Loughlin, E., Rok, A., 2016. Pathways to a Circular Economy in

U
cities and regions.
AN
Balanay, R., Halog, A., 2016. Charting policy directions for mining’s sustainability
with circular economy. Recycling 1, 219–230. doi:10.3390/recycling1020219
M

Banaité, D., 2016. Towards circular economy: analysis of indicators in the context of
sustaibable development. Soc. Transform. Contemp. Soc. 4, 142–150.

Barry, A., Born, G., 2013. Interdisciplinarity: Reconfigurations of the Social and
D

Natural Sciences. Routledge.

TE

Bender-deMoll, S., 2016. NDTV: Network dynamic temporal visualizations. R

package version.
EP

Bracken, L.J., Oughton, E.A., 2006. ‘What do you mean?’ The importance of
language in developing interdisciplinary research. Trans. Inst. Br. Geogr. 31,
371–382. doi:10.1111/j.1475-5661.2006.00218.x
C

Bruce, A., Lyall, C., Tait, J., Williams, R., 2004. Interdisciplinary integration in Europe:
The case of the Fifth Framework programme. Futures 36, 457–470.
AC

doi:10.1016/j.futures.2003.10.003

Choi, B.C.K., Pak, A.W.P., 2006. Multidisciplinarity, interdisciplinarity and

transdisciplinarity in health research, services, education and policy: 1.
Definitions, objectives, and evidence of effectiveness. Clin. Investig. Med. 29,
351–364.

Cohen, B., Muñoz, P., 2016. Sharing cities and sustainable consumption and
production: towards an integrated framework. J. Clean. Prod. 134, 87–97.
doi:10.1016/j.jclepro.2015.07.133

21
 ACCEPTED MANUSCRIPT

European Commission, 2015a. Communication from the Commission to the

European Parliament, the Council, the European Economic and Social
Committee and the Committee of the Regions. Closing the loop–An EU action
plan for the Circular Economy. COM(2015) 614.

European Commission, 2015b. Better Regulation ‘Toolbox.’ Brussels.

European Commission, 2016a. Circular Economy Package.

PT
European Commission, 2016b. Draft Horizon 2020 Work Programme 2016–2017 in
the area of Leadership in enabling and industrial technologies–Introduction.
European Commission Decision C(2017)2468 of 24 April 2017.

RI
European Commission, 2017. Better Regulation Guidelines. Brussels.

SC
Ferdousi, F., Qiang, D., 2016. Implementing circular economy and its impact on
consumer ecological behavior. J. Innov. Sustain. 7, 3–10.

Franco, M.A., 2017. Circular economy at the micro level: A dynamic view of

U
incumbents’ struggles and challenges in the textile industry. J. Clean. Prod. 168,
833–845. doi:10.1016/j.jclepro.2017.09.056
AN
Franklin-Johnson, E., Figge, F., Canning, L., 2016. Resource duration as a
managerial indicator for circular economy performance. J. Clean. Prod. 133,
M

589–598. doi:10.1016/j.jclepro.2016.05.023

Geng, Y., Doberstein, B., 2008. Developing the circular economy in China:
challenges and opportunities for achieving ‘leapfrog development.’ Int. J.
D

Sustain. Dev. World Ecol. 15, 231–239. doi:10.3843/SusDev.15.3:6

TE

Geng, Y., Zhu, Q., Doberstein, B., Fujita, T., 2009. Implementing China’s circular
economy concept at the regional level: a review of progress in Dalian, China.
Waste Manag. 29, 996–1002. doi:10.1016/j.wasman.2008.06.036
EP

Ghisellini, P., Cialani, C., Ulgiati, S., 2016. A review on circular economy: the
expected transition to a balanced interplay of environmental and economic
systems. J. Clean. Prod. 114, 11–32. doi:10.1016/j.jclepro.2015.09.007
C

Gibbons, M., Limoges, C., Nowotny, H., Schwartzman, S., Scott, P., Trow, M., 1994.
AC

The New Production of Knowledge: Dynamics of Science and Research in

Contemporary Societes. Sage, London.

Hebert, J.R., Brandt, H.M., Armstead, C.A., Adams, S.A., Steck, S.E., 2009.
Interdisciplinary, translational, and community-based participatory research:
finding a common language to improve cancer research. Cancer Epidemiol.
Biomarkers Prev. 18, 1213–1217. doi:10.1158/1055-9965.EPI-08-1166

Heinze, T., Bauer, G., 2007. Characterizing creative scientists in nano-S&T:

productivity, multidisciplinarity, and network brokerage in a longitudinal
perspective. Scientometrics 70, 811–830. doi:10.1007/s11192-007-0313-3

22
 ACCEPTED MANUSCRIPT

Heshmati, A., 2016. A Review of the circular economy and its implementation,
Working Paper Series in Economics and Institutions of Innovation. Royal
Institute of Technology, CESIS - Centre of Excellence for Science and
Innovation Studies.

Howarth, C., Monasterolo, I., 2017. Opportunities for knowledge co-production

across the energy-food-water nexus: making interdisciplinary approaches work
for better climate decision making. Environ. Sci. Policy 75, 103–110.

PT
doi:10.1016/j.envsci.2017.05.019

Hübenthal, U., 1994. Interdisciplinary thought. Issues Integr. Stud. 75, 55–75.

RI
Huutoniemi, K., Klein, J.T., Bruun, H., Hukkinen, J., 2010. Analyzing
interdisciplinarity: typology and indicators. Res. Policy 39, 79–88.
doi:10.1016/j.respol.2009.09.011

SC
Jacobs, H.H., 1989. Interdisciplinary curriculum: design and implematation.
Association for Supervision and Curriculum Development, 1250 N. Pitt Street,

U
Alexandria, VA 22314.

Kanbur, R., 2002. Economics, social science and development. World Dev. 30, 477–
AN
486.

Kirchherr, J., Reike, D., Hekkert, M., 2017. Conceptualizing the circular economy: an
M

analysis of 114 definitions. Resour. Conserv. Recycl. 127, 221–232.

doi:10.1016/j.resconrec.2017.09.005
D

Klein, J.T., 1990. Interdisciplinarity: history, theory, and practice. Detroit, Michigan:
Wayne State University Press.
TE

Klein, J.T., 2003. Unity of knowledge and transdisciplinarity: contexts of definition,

theory and the new discourse of problem solving. Encycl. Life Support Syst. I, 1–
30.
EP

Klein, J.T., 2008. Evaluation of interdisciplinary and transdisciplinary research. A

literature review. Am. J. Prev. Med. 35. doi:10.1016/j.amepre.2008.05.010
C

Klenk, N., Meehan, K., 2015. Climate change and transdisciplinary science:
problematizing the integration imperative. Environ. Sci. Policy 54, 160–167.
AC

doi:10.1016/j.envsci.2015.05.017

Larivière, V., Haustein, S., Börner, K., 2015. Long-distance interdisciplinarity leads to
higher scientific impact. PLoS One 10, 1–15. doi:10.1371/journal.pone.0122565

Lehmann, M., Leeuw, B. de, Fehr, E., Wong, A., 2014. Circular economy. Improving
the management of natural resources. Zürich, Schweizerische Akad. der Tech.
Wissenschaften 19. (www. satw.
ch/publikationen/schriften/kreislaufwirtschaft/a__circulareconomy_with_refere
nces_EN. pdf).

23
 ACCEPTED MANUSCRIPT

Lieder, M., Rashid, A., 2016. Towards circular economy implementation: a

comprehensive review in context of manufacturing industry. J. Clean. Prod. 115,
36–51. doi:10.1016/j.jclepro.2015.12.042

Marzano, M., Carss, D.N., Bell, S., 2006. Working to make interdisciplinarity work:
investing in communication and interpersonal relationships. J. Agric. Econ. 57,
185–197. doi:10.1111/j.1477-9552.2006.00046.x

PT
McDonough, W., Braungart, M., 2002. Cradle to Cradle: Remaking the Way We
Make Things. North point press.

McDowall, W., Geng, Y., Huang, B., Barteková, E., Bleischwitz, R., Türkeli, S., Kemp,

RI
R., Doménech, T., 2017. Circular Economy Policies in China and Europe. J. Ind.
Ecol. 21, 651–661. doi:10.1111/jiec.12597

SC
Moran, J., 2002. Interdisciplinarity (The new critical idiom: New critical idiom).
London, New York: Routledge.

Morillo, F., Bourdons, M., Gomez, I., 2003. Interdisciplinarity in science: a tentative

U
typology of disciples and research areas. J. Am. Soc. Inf. Sci. Technol. 54,
1237–1249. doi:10.1002/asi.10326
AN
Mugabushaka, A.-M., Kyriakou, A., Papazoglou, T., 2016. Bibliometric indicators of
interdisciplinarity: the potential of the Leinster--Cobbold diversity indices to study
M

disciplinary diversity. Scientometrics 107, 593–607. doi:10.1007/s11192-016-

1865-x
D

Murray, A., Skene, K., Haynes, K., 2017. The circular economy: an interdisciplinary
exploration of the concept and application in a global context. J. Bus. Ethics 140,
TE

369–380. doi:10.1007/s10551-015-2693-2

Mutz, R., Bornmann, L., Daniel, H.D., 2014. Cross-disciplinary research: what
configurations of fields of science are found in grant proposals today? Res. Eval.
EP

1–7. doi:10.1093/reseval/rvu023

National Academy of Sciences, 2005. Facilitating Interdisciplinary Research. The

National Academies Press, Washington, D. C. doi:10.1007/s00103-011-1362-6
C

OECD, 2010. Measuring Innovation. A New Perspective. OECD Publishing, Paris.

AC

http://dx.doi.org/10.1787/9789264059474-en

Owens, S., Driffill, L., 2008. How to change attitudes and behaviours in the context of
energy. Energy Policy 36, 4412–4418. doi:10.1016/j.enpol.2008.09.031

Pearce, D.W., Turner, R.K., 1990. Economics of natural resources and the
environment. John Hopkins University Press, Elsevier.

Pohl, C., 2008. From science to policy through transdisciplinary research. Environ.
Sci. Policy 11, 46–53. doi:10.1016/j.envsci.2007.06.001

24
 ACCEPTED MANUSCRIPT

Pomponi, F., Moncaster, A., 2017. Circular economy for the built environment: a
research framework. J. Clean. Prod. 143, 710–718.
doi:10.1016/j.jclepro.2016.12.055

Prendeville, S., Cherim, E., Bocken, N., 2016. Circular cities: mapping six cities in
transition. Environ. Innov. Soc. Transitions. doi:10.1016/j.eist.2017.03.002

R Core Team, 2017. R: a language and environment for statistical computing.

PT
Rosenfield, P.L., 1992. The potential of transdisciplinary research for sustaining and
extending linkages between the health and social sciences. Soc. Sci. Med. 35,
1343–1357. doi:10.1016/0277-9536(92)90038-R

RI
Saidani, M., Yannou, B., Leroy, Y., Cluzel, F., 2017. How to assess product
performance in the circular economy? Proposed requirements for the design of a

SC
circularity measurement framework. Recycling 2, 6.
doi:10.3390/recycling2010006

Sauvé, S., Bernard, S., Sloan, P., 2016. Environmental sciences, sustainable

U
development and circular economy: alternative concepts for trans-disciplinary
research. Environ. Dev. 17, 48–56. doi:10.1016/j.envdev.2015.09.002
AN
Stahel, W.R., 1997. The service economy: ‘wealth without resource consumption’?
Philos. Trans. R. Soc. A-Mathematical Phys. Eng. Sci. 355, 1309–1319.
M

doi:10.1098/rsta.1997.0058

Stahel, W.R., 2016. Circular economy. Nature 531, 435–438. doi:10.1038/531435a

D

Stokols, D., 2006. Toward a science of transdisciplinary action research. Am. J.

Community Psychol. 38, 63–77. doi:10.1007/s10464-006-9060-5
TE

Stokols, D., Hall, K.L., Taylor, B.K., Moser, R.P., 2008. The science of team science.
Overview of the field and introduction to the supplement. Am. J. Prev. Med. 35.
EP

doi:10.1016/j.amepre.2008.05.002

Su, B., Heshmati, A., Geng, Y., Yu, X., 2013. A review of the circular economy in
China: moving from rhetoric to implementation. J. Clean. Prod. 42, 215–227.
C

doi:10.1016/j.jclepro.2012.11.020
AC

Thaler, R.H., 2016. Behavioral economics: past, present and future. Am. Econ. Rev.
106, 1577–1600. doi:10.1257/aer.106.7.1577

Tress, B., Tress, G., Fry, G., 2005. Defining concepts and the process of knowledge
production in integrative research, in: From Landscape Research to Landscape
Planning: Aspects of Integration, Education and Application. pp. 13–26.

Van Den Besselaar, P., Heimeriks, G., 2001. Disciplinary, multidisciplinary,

interdisciplinary: concepts and indicators, in ISSI-2001: International Society for
Scientometrics and Informetrics. International conference (pp. 705-716).

25
 ACCEPTED MANUSCRIPT

Wilson, V., Pirrie, A., 2000. Multidisciplinary teamworking: beyond the barriers? A
review of the issues. Research Report Series. The Scottish Council for Research
in Education, 15 St. John Street, Edinburgh EH8 8JR (6 British pounds). For full
text: http://www. scre. ac. uk/resreport/rr96.

Yuan, Z., Bi, J., Moriguichi, Y., 2006. The circular economy: a new development
strategy in China. J. Ind. Ecol. 10, 4–8. doi:10.1162/108819806775545321

PT
Zhu, D.J., Huang, X.F., 2005. Building up a model for circular economy based on
object, main body, and policy. Naikai Acad. J. 4, 86–93.

RI
U SC
AN
M
D
TE
C EP
AC

26
 ACCEPTED MANUSCRIPT

2001-2006 2007-2011 2012-2016

400

350

PT
300

RI
Number of publications

250

SC
200

U
150

AN
100

M
50

D
TE
0 EP
Subject Area
C
AC
Subject Area 2001-2006
ACCEPTED MANUSCRIPT
2007-2011 2012-2016
Eng 8 139 389
EnSc 15 102 315
Ener 3 46 159
BM&A 4 62 133
SoSc 8 57 99
CoSc 2 78 67
ChEn 6 19 59
EE&F 1 17 65
MaSc 1 12 65

PT
ABS 8 18 50
EaSc 3 32 25
DeSc 0 26 32

RI
Math 1 17 31
Chem 0 4 42

SC
Med 0 2 22
P&A 0 1 20
BGMB 0 5 10
Phar 3 6 6

U
A&H 0 0 8
AN
I&M 0 0 4
HePr 0 1 0
Vet 0 0 1
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

400

350
Number of research outputs

300

PT
250

RI
200

SC
150

100

U
AN
50

M
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
Year

D
Journal Article Book/Book Chapter Conference Paper All

TE
C EP
AC
 ACCEPTED MANUSCRIPT

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
Journal Article 1 0 0 4 11 18 30 21 38 30 21 39 42 62 100 248
Book/Book Chapter 0 0 0 0 1 0 0 1 0 0 2 1 2 2 16 22
Conference Paper 0 0 0 0 3 1 11 14 41 75 90 56 55 84 27 75
All 1 0 0 4 15 19 41 36 79 105 113 96 99 148 143 345

PT
RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

536

432

PT
RI
208 199
164

SC
147
84 83 78 76
60 58 49 46

U
24 21 15 15 8 4 1 1

AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

SA Freq
Eng 536
EnSc 432
Ener 208
BM&A 199
SoSc 164
CoSc 147
ChEn 84
EE&F 83
MaSc 78
ABS 76
EaSc 60
DeSc 58
Math 49
Chem 46
Med 24

PT
P&A 21
BGMB 15
Phar 15
A&H 8
I&M 4
HePr 1
Vet 1

RI
U SC
AN
M
D
TE
EP
C
AC

Chart Title

Other
0%
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT
Knowledge sharing and scientific cooperation in the design of research-
based policies: the case of the circular economy

Highligths

1. A shift to a circular economy (CE) needs effective policy interventions

PT
2. Cooperation across disciplines is required to develop effective policy design

3. The potential for cooperation across disciplines working on CE is investigated

RI
4. The current potential for cooperation is more suited for micro/meso level policies

5. Policies for CE at macro level suffer from an inadequate potential for cooperation

U SC
AN
M
D
TE
C EP
AC