ArticlePDF Available
Template version 4
System for Environmental and Agricultural Modelling;
Linking European Science and Society
INTEGRATED PROJECT EU FP 6 (contract no. 010036)
Global Change and Ecosystems
Project duration: January 2005 - December 2008
Sustainable Development Indicator Frameworks
and Inititiatives
PD 2.2.1
Ghislain Geniaux, Stéphane Bellon, Christian Deverre,
Blaise Powell
Submission date: 26 Dec 2005
No. 010036
Deliverable number:
21 March 2006
Page 2 of 146
SEAMLESS integrated project aims at developing an integrated framework that allows ex-
ante assessment of agricultural and environmental policies and technological innovations.
The framework will have multi-scale capabilities ranging from field and farm to the EU25
and globe; it will be generic, modular and open and using state-of-the art software. The
project is carried out by a consortium of 30 partners, led by Wageningen University (NL).
Authors of this report and contact details
Name: Ghislain Geniaux Partner acronym: INRA
Address: INRA Ecodéveloppement
Site Agroparc, Domaine St Paul
84914 Avignon Cedex 9
Name: Blaise Powell Partner acronym: INRA
Address: INRA Ecodéveloppement
Site Agroparc, Domaine St Paul
84914 Avignon Cedex 9
Name: Stéphane Bellon Partner acronym: INRA
Address: INRA Ecodéveloppement
Site Agroparc, Domaine St Paul
84914 Avignon Cedex 9
Name: Christian Deverre Partner acronym: INRA
Address: INRA Ecodéveloppement
Site Agroparc, Domaine St Paul
84914 Avignon Cedex 9
“This publication has been funded under the SEAMLESS integrated project, EU 6th
Framework Programme for Research, Technological Development and Demonstration,
Priority Global Change and Ecosystems (European Commission, DG Research,
contract no. 010036-2). Its content does not represent the official position of the European
Commission and is entirely under the responsibility of the authors.”
"The information in this document is provided as is and no guarantee or warranty is given
that the information is fit for any particular purpose. The user thereof uses the information at
its sole risk and liability."
No. 010036
Deliverable number:
21 March 2006
Page 4 of 146
Table of contents
Table of contents 4
General information 7
Executive summary 7
1 Introduction 10
2 Sustainability principles 11
2.1 Brief historical review of basic and complementary principles for sustainability and
sustainable development 11
2.2 Stabilized and thematic pillars 13
3 Indicators 17
3.1 Definitions of indicators and SDI frameworks 17
3.2 Functions of indicators 19
3.3 Reference values: target, threshold and goals. 19
3.4 Indicators classes 20
3.4.1 Composite SDI 20
3.4.2 One unit aggregated SDI 22
3.4.3 Lists of indicators and dashboards 23
4 Main Frameworks 24
4.1 SDI using list of indicators 24
4.1.1 The PSR, DPSIR, DSR approaches 24
4.1.2 Capital-based approach 25
4.1.3 SEEA 26
4.1.4 System-based approach 27
4.1.5 EUS (Environmental Utilization Space) and Ecospace 33
4.1.6 Land quality, land use and land value analysis 33
4.1.7 Material Flow Analysis (MFA), Substance Flow Analysis (SFA) and Life Cycle
Assessment (LCA) 34
4.2 Sub-Composite and aggregated SDI approaches 35
4.2.1 Monetarised indicators: GSR, GPI and ISEW 36
4.2.2 A physical indicator aggregated in area unit: the Ecological Footprint (EF) 37
4.2.3 Two composite indicators: IUCN and 2005ESI initiatives 39
4.2.4 Comparing sustainability indicators initiatives 42
4.3 Transversal ideas used in SDI framework 44
4.3.1 Carrying Capacity and related concepts (Eco-efficiency, Limit to growth and Steady State
Economy) 44
4.3.2 Ecosystem health 45
4.3.3 Biodiversity issues 51
5 SDI Frameworks specific to the agricultural sector 53
5.1 Agricultural, Agri-environmental and Ecological indicators to assess the sustainability of
agro-ecosystems 53
5.1.1 Sustainability of agricultural systems 53
5.1.2 Impact indicators 53
5.1.3 Agri-environmental indicators (AEIs) 54
No. 010036
Deliverable number:
21 March 2006
Page 5 of 146
5.1.4 Ecological indicators 55
5.2 Representations of the agri-environment relationship 55
5.2.1 Views of the environment: analytical, constructivist or problem oriented? 55
5.2.2 Resources and functions 56
5.3 Indicator-based methods for agri-environmental assessment 57
5.3.1 General approach for environmental and sustainability assessment 57
5.3.2 Indicator-based methods for agri-environmental assessment 58
5.3.3 Thresholds, representations and layouts 60
5.4 The Sustainable Rural Livelihood (SRL) approach 61
5.5 Topic discussion and directions for further work 63
5.5.1 Scale issues 63
6 A SDI Framework For SEAMLESS. 65
6.1 Implication of a sectoral and regional approach 66
6.2 A systemic framework 68
6.3 A 4 + 1 scales Framework 71
6.4 A composite Framework 75
6.5 Implementing the framework and policy relevance 77
References 81
Glossary 93
Appendices 97
1 Some methodological aspects of composite indicators 97
2 Main lists of sustainable development indicators 100
2.1 Global initiatives 100
2.1.1 UN 100
2.1.2 European Union 105
2.1.3 OECD 113
2.1.4 WB Sustainability Matrix 1995 119
2.1.5 World Resource Institute of SDI (1996) 122
2.1.6 Balaton Group Indicators and Information Systems for SD (1996) 123
2.2 Agricultural related SDI 124
2.2.1 lists of Agri-environmental and/or rural indicators 124
2.2.2 SRL-FSELM (Rigby and al. 2000) 127
2.2.3 IRENA 130
2.2.4 ELISA 131
2.2.5 PAIS 132
2.2.6 MAFF (UK 2000) 136
2.2.7 Potential Biological, Chemical, and Physical Indicators of Soil Quality 137
3 MFA, SFA and LCA 139
3.1 Physical economy approaches by Daniels and Moore (2002) 139
3.2 IFF MFA-BIF model for Austria 142
4 Criteria for indicator selection 143
5 Major works on agricultural system proprieties 145
No. 010036
Deliverable number:
21 March 2006
Page 6 of 146
No. 010036
Deliverable number:
21 March 2006
Page 7 of 146
General information
Task(s) and Activity code(s): T.2.2; A2.2.2
Input from (Task and Activity codes): A2.2.2
Output to (Task and Activity codes): T2.4; T2.5; A2.2.2; A2.2.3; A2.2.4
Related milestones: M2.2.1; M2.1 2
Executive summary
SEAMLESS aims at designing a tool enabling to assess and compare policy options of the
Common Agricultural Policy (CAP), in a Sustainable Development perspective. This
encompasses agricultural and environmental policies and tecdhnological innovations. Within
the project, WP2 is in charge of elaborating indicators for this purpose. Which indicators are
chosen, and how they are organized, determines the scope of SEAMLESS and its definition
of the agricultural system, contributing to SD in Europe.
In this paper we rely on a critical literature review to analyse what the organisation of
indicators for sustainable development into a framework can bring to WP2 objectives. First
the framework translates the vision of SD carried by the promoter of the assessment, what is
studied, how and for what purpose. So principles and issues used in the SD paradigm are
presented. Then, the framework organizes indicators into a meaningful presentation. Different
frameworks allow to take into account the balance between dimensions of SD, between
issues, between pressure or response indicators… Different frameworks integrate a variable
number of desired properties for indicators, thus ensuring such properties are accounted for.
The properties which are not integrated in the framework are generally mentioned in
indicative criteria lists, leaving their consideration to the individual indicator choice.
We then present available major sustainable development indicator frameworks (SDIF). Their
advantages and limits are discussed, distinguishing indicator lists, aggregated and composite
indicators. We notably point how certain frameworks are explicitly or implicitly adapted to
account for : 1) certain interfaces such as economy/environment or society/environment,… 2)
weak or strong sustainability, 3) inter or intra-generational equity, 4) comparability in space
and time or adaptability to local context and participative methods, 5) dependence relations
involving the studied system. Approaches dealing more precisely with agro-ecosystems are
then presented. From the literature reviewed here, we have notably retained that the systemic
approach permits a complete and balanced SDIF.
But SEAMLESS presents a number of challenges that are not dealt with, or solved in a
satisfactory manner in the literature reviewed. The first challenge is dealing with different
scales, and with the agricultural sector. The nature of the agricultural system and of its
relation to the rest of society is specific at some scales. This could condition the link between
scales, which still has to be modelled. Another point is dealing separately with different
agricultural sub-sectors or policies related to CAP. This involves making SD related ex post
assessments as well as ex ante comparison between competing policy options.
The methodological aspects of elaborating composite indicators are also reviewed. How they
could be adapted to these challenges is discussed in the final chapter, where methodological
choices are made to propose an SDI framework for SEAMLESS. The proposed framework is
characterised by:
No. 010036
Deliverable number:
21 March 2006
Page 8 of 146
- at each scale (farm, community, region, nation, Europe), an adapted systemic
framework allowing to flexibly choose indicators representing systemic properties,
according to local context and specific purpose;
- the articulation of two systems: the agricultural system encompassing the
agricultural land, the people working or living on this land and their families, in
relation with the total system in a sustainable development perspective. At each scale
(level) specific relations between society and the agricultural system are thus
emphasized. The farm scale is not territorially defined and the agricultural system
merges with the total system, so the more general sustainable development
perspective at this level will depend on the link with territorial scales. Possible Data
accessibility problems at community level may lead us to integrate rural aspects
inherent to this scale into the regional scale;
- differentiating ex post and ex ante analysis, for various policies;
- enabling to asses these policies separately and completely, because all indicators
derived within the framework can be relative to the particular policy of interest;
- an aggregation scheme for a candidate list of sub-indicators, along with a
complementary list of indicators and contextual variables to be used for post-model
- The necessity for SEAMLESS to formalise correspondances between properties
relative to the systemic logic, and the indicators chosen to represent them - which
can be originally identified through different logics (issues for example).
Thorough ex ante policy analysis in the proposed framework requires that the user be
provided candidate indicators adapted to concerned policies. It is thus expected from WP1
and WP7, to provide a manageable list of policy fields and policy options useful to final
users, in order to engage in elaborating appropriate lists of indicators within WP2.
No. 010036
Deliverable number:
21 March 2006
Page 10 of 146
1 Introduction
This paper presents and assesses the panorama of main Sustainable Development Indicators
(SDI) frameworks that exist in the literature. Its objective is to facilitate the selection and
development of a framework adapted to WP2 in SEAMLESS. It underlines the great diversity
of available frameworks, within national and international initiatives. Indeed, previous PDs
produced within SEAMLESS give little account of such a diversity and possible
methodological choices, particularly concerning the range of different principles and criteria
for organising these frameworks.
The second chapter of the document proposes a detailed review of stabilised and emergent
principles stemming from the idea of sustainability, as it gradually crystallized on the
political and scientific scene. A brief chronological presentation of this emergence and
stabilisation is given. It aims to illustrate the continuous enlargement of principles underlying
SD and the accompanying conditions for its implementation, being the main cause of the
continuous evolution and revision of developed frameworks. This part ends by a simple
reminder of principles our future framework will have to stand up to.
The third chapter aims to define what an indicator and a framework is, when developed in the
context of sustainable development. We define what we expect from an indicator and propose
a simple typology of indicators which is used to discuss the main differences and
methodological issues, related to aggregated indicators, composite indicators and lists of
indicators. This part is focused on general technical and methodological aspect of the
construction of a SDIF.
The fourth chapter discusses in detail eight (8) existing SDI Frameworks. Theoretical and
methodological presuppositions of these respective frameworks are analysed, and the SD
vision they convey is illustrated. The frameworks are described in four sections by
identifying their main organising principles. In a first section, we present frameworks leading
to indicator lists. They differ from those used for aggregated or composite indicators,
introduced in a second section. In a third section some agricultural and rural specific
frameworks are proposed. The fourth section presents and discusses transversal concepts of
SDIF that did not lead to a formal and applied SDI framework but contribute to our objective.
The fifth chapter explores the variety of indicators used in the agricultural sector when
agriculture is related with environmental and sustainability issues. In the first of the four
sections, we present categories of indicators to assess the sustainability of agro-ecosystems
including: agricultural, agri-environmental, and ecological indicators that can be used in
SEAMLESS. Secondly, we show how the choice of indicators depends on the underlying
vision of the environment attached with disciplines or their integration. The third section
details initiatives corresponding to the previous categories of indicators identified and ways
to represent sets of indicators. The last section includes a brief discussion on temporal and
spatial scaling issues.
The sixth and last chapter proposes a framework adapted to the WP2 objectives and
constraints. A systemic organization of SDI, adapted from Bossel (1999) is presented. The
framework is characterised by:
- at each scale (farm, community, region, nation, Europe), an adapted systemic
framework allowing to flexibly choose indicators representing systemic properties,
according to local context and specific purpose;
- the articulation of two systems: the agricultural system encompassing the
agricultural land, the people working or living on this land and their families, in
relation with the total system in a sustainable development perspective. At each scale
No. 010036
Deliverable number:
21 March 2006
Page 11 of 146
(level) specific relations between society and the agricultural system are thus
emphasized. The farm scale is not territorially defined and the agricultural system
merges with the total system, so the more general sustainable development
perspective at this level will depend on the link with territorial scales. Possible Data
accessibility problems at community level may lead us to integrate rural aspects
inherent to this scale into the regional scale;
- differentiating ex post and ex ante analysis, for various policies;
- enabling to asses these policies separately and completely, because all indicators
derived within the framework can be relative to the particular policy of interest;
- an aggregation scheme for a candidate list of sub-indicators, along with a
complementary list of indicators and contextual variables to be used for post-model
- The necessity for SEAMLESS to formalise correspondances between properties
relative to the systemic logic, and the indicators chosen to represent them - which
can be originally identified through different logics (issues for example).
The material collected for this PD particularly all the lists of indicators proposed in all the
SDI frameworks reviewed - and an analytical grid of criteria for the selection of indicators
constructed by us to compare frameworks are presented in Appendix .
2 Sustainability principles
2.1 Brief historical review of basic and complementary principles
for sustainability and sustainable development
Available Sustainable Development Indicator Frameworks (SDIF) are built, explicitly or
implicitly, on a set of underlying sustainable development principles. These principles have
evolved since the emergence and stabilisation of the concept of sustainable development
itself in the 1987 Bruntland commission report ("Our Common Future", WCED, 1987). As
these principles play a central role for SDIF, it is important to identify these principles and
where they come from. Their evolution is tightly connected to the large international
initiatives concerning reflection about and promotion of SD.
Previous benchmarks dealing with the notion of sustainability can be found in agronomic
and agro-ecosystems related disciplines. According to Conway (1983) sustainability is the
ability of a system to maintain productivity and is considered as a specific property of agro-
ecosystems. This is consistent with the forestry concept of sustainability ("nachhaltigkeit")
developed during the XVIIIth century, evolving through the notion of sustainable yield, and
also applied to agricultural crops (Plucknett and Smith, 1986). Besides this scientifically
based "agro-ecological production" path, Becker (1997) also identifies another emerging
"normative" path for the concept of SD. This second path shifts from the notion of "wise use"
in the RAMSAR convention to the notion of "sound strategies" for the United Nations
Conference on Environment in Stockholm (1972); both terms being even more normatively
connoted than sustainable development itself). These two paths come together in an economic
definition of the concept of SD at the WCED in 1987. However, the exact term of SD seems
to have appeared already in a text from IUCN in 1980 (World conservation Strategy)
No. 010036
Deliverable number:
21 March 2006
Page 12 of 146
The World Commission on Environment and Development (WCED) of 1987 is consequently
one of the first locations of emergence and crystallization of SD principles. The four
following principles are acknowledged:
Inter-generational equity
Intra-generational equity
Environmental protection integral to economic development
Public participation
The three first principles constitute the heart of the notion of SD and have been integrated in
the majority of SDIF (UN, 1996). This has not been the case for public participation which
only later initiated a methodological reflection on the means to manage it.
Recognition of biodiversity protection as a necessary condition to SD enforcement appeared
during the Rio conference in 1992. It is also in this conference that the principle of local
knowledge preservation, entered as an important dimension of sustainable development.
In the middle of the nineties, and in the line of the principles affirmed in Rio, a number of
complementary principles were recognised and were used in different SDI initiatives. Work
done for the Rosenthal Workshop (UN, 1996) and continued by Murcott (1997) for the
conference on SDI (AAAS annual Conference, Seattle, Washington) enables to identify the
rationale of these complementary principles.
The necessity of accounting for concerned population's quality of life, and of having recourse
to the notion of environment's carrying capacity, had explicitly been mentioned as founding
principles as early as 1980 during the Geneva conference (IUCN, UNEP, WWF, Caring for
the Earth: a strategy to sustainable living) and have become unavoidable notions in numerous
frameworks. Measuring the satisfaction of basic needs and the Provision of social self-
determination and cultural diversity had gained official recognition during IICCD (Ottawa
1986), and were partially re-entered by the Rio principles.
Some principles have for their part found official recognition and a political existence within
national SD initiatives and only later were acknowledged by international initiatives. This is
the case for Shared Responsibility (among all levels of government and internationally)
which is part of Canada's Green Plan of 1990. It is in Australia's 1992 Green Plan that ideas
relative to the precautionary principle found a beginning of political recognition at national
level within the SD issue (Murcott, 1997). The principle that could be called international
perspective and aims at linking domestic sustainability to external sustainability is also
affirmed in Australia's Green Plan of 1992 (Murcott, 1997).
More generally, dealing with the fact that sustainability of a dynamic system is related to that
of other (possibly embedded) systems, is the heart of system-based approaches to SD.
Building on work where the method was applied to specific dimensions such as ecosystems
or economy, this approach of SD issues was increasingly recognized within different
initiatives, without being met with sufficient willingness to operationalise it.
A principle, which today is widely used as a basic principle of several SDIF, aims at
promoting integrated life-cycle management and closed material cycles in the chain of raw
materials. Having its roots in the 1970's debates that followed the Club de Rome, it found
explicit political definition in Netherlands's 1990 national environmental policy plan.
No. 010036
Deliverable number:
21 March 2006
Page 13 of 146
2.2 Stabilized and thematic pillars
The extension of the notion of progress to the environmental and social spheres is present in
the founding principles of sustainable development. We can however notice that the elements
to integrate in these spheres become more and more numerous as the principles of SD is
developed and refined. The same can be said about the links between these spheres.
Nonetheless, although the systemic dimension was increasingly recognised in the nineties, we
note that several initiatives satisfy themselves with basing their analysis on a separation of
three relatively dissociated pillars of SD: economy, environment, society. In each of the three,
a large choice of indicators aims at satisfying the above mentioned principles.
Before describing the sub themes of the three dimensions, their content and logic, we discuss
some major implications of these main principles on the objective of developing indicators:
extension of the notion of progress to environment and social domains, inter and intra-
generational equity, systems based approach.
Economic and environmental valuations
Research and empirical efforts to assess SD focussed on the difficulty to assess –piece by
piece- the contribution to SD of the variation of environmental state and social context. The
relatively strong convention in the economical sphere around Gross National Product (GNP)
probably did not facilitate the efforts to valuate of other dimensions of SD. Evaluating the
contribution of the state of the environment in terms of SD indices calls for mobilising a
theory about the "value of environment". Becker (1997) and Geniaux (2001), among others,
showed that an anthropocentric approach remains a more coherent choice than competing
philosophies (theological argument whereby only the creator can sustain his creation;
pathocentric approach; biocentric individualism and biocentric holism that imply that non-
humans are moral subjects carrying an intrinsic value (see Hampicke, 1994). The choice of
anthropocentrism is clearly defined in the first paragraph of the Rio declaration: "human
beings are the centre of concern for sustainable development". However, there is no globally
satisfying method for delivering information on the intensity of human preferences about the
value of environmental components. An important part of the debate, beyond the
effectiveness of these methods, is over the question of knowing just to what limit
substitutability should be the pivot in nature's evaluation (weak versus strong sustainability
approach, compensatory versus non-compensatory aggregation).
Social progress : principles and indicators
There exists great diversity of theoretical foundations allowing different indicator choices to
assess social progress and enhancement of individual and collective well being. Among these
we can clearly identify three principles that has structured social indicator production in the
SDI initiatives: intragenerational equity, satisfaction of basic needs and quality of life. The
first principle deals with indicators of wealth distribution within a society, usually using the
Gini, Herfindahl or Atkinson indices. The second one can be seen as a prerequisite to
development itself, and indirectly participating to reducing pressure on natural resources. The
third one which was expressed and acknowledged at a later stage than the others, is meant to
remind us that sustainable development is part of a global progress perspective and does not
only aim at sustaining an organising a production system. Moving from social acceptability to
quality of life marks the transition from the status of prerequisite condition to the one of full
component of sustainable development.
No. 010036
Deliverable number:
21 March 2006
Page 14 of 146
These two last principles have been very structuring on the social dimension of SD and on
selected indicators in the different initiatives. Especially where basic needs are concerned,
with quality of life generally coming down to a simple listing of socio economic variables
leading to a rather inexplicit vision of what quality of life is, this in spite of strong demands
for developing composite indicators in this category. One should indeed note that this
evolution was imposed by the base (experience assessment, holist approach development).
This prudent recourse to relatively autonomous socio-economic variables to account for
quality of life can moreover be explained by the heterogeneity of cultural and legal contexts
in which these variables evolve, making any measure fragile. Indeed the kind of all-enclosing
concepts of "social capital" and "human capital" puts us in the larger context where the social
and institutional components contributing to a better quality of life at both levels (individual,
society) are included.
One of the founding principles of SD concerns intergenerational equity. Two philosophical
principles support the interest for future generations. The first is Kant's categorical
imperative, or ontological principle. The second is Rawls's theory of justice, which is more
precise about the modalities of its operationalisation. Although they seem to fit as political
principles, Rawls's theory ignores society's dynamic, so the problem of extrapolation from
today's individual preferences arises. However, using the definition of justice suggested by
Rawls (1971), we can define a particular criterion to maximise the utility of the least well
endowed generation so as to respect an intergenerational equity constraint (Costes and al.,
2003). Interestingly, this debate is essentially relevant for taking into account future
generations when evaluating "benefits" or "well being" associated to different development
paths or projects. Its formalisation, taking the form of actualisation rates, renders it limited to
ACB approaches (monetary normalisation), and to issues dealing with sharing natural
resource in time. In frameworks using indicator lists and where the appreciation of the
contribution of an indicator to SD is left to the final user, it remains difficult to explicitly
formalise taking intergenerational equity into account, other than case per case or by
relatively loose rules. Moreover, few really normative approaches can be found accordingly
in SDI initiatives, and intergenerational equity is in the end simply affirmed by accepting the
that certain resources are important for future generations and should be taken into
consideration by sustainability evaluation.
Pillars and stakeholders
The notion of "pillar" is frequently used do define the essential dimensions or themes of
sustainable development. The large initiatives mainly refer to three pillars to account for
sustainability: economy, environment and social concerns ("triple bottom line", with varying
relative importance). In some initiatives, such as the capitals approach by the World Bank,
the social pillar is divided into human capital and social capital. In the same way, initiatives
that consider institutional issues sometimes include them in the social dimension whereas
others make them a specific pillar. These pillars are then broken down into different themes,
and this structure which precedes the choice of indicators is an essential stage in framework
Framework evolution is best illustrated by the evolution of the components of the
environmental pillar. One must indeed note that for lots of actors, the initiative to develop SD
indicator was first initiated in relation to environment indicators. The way of structuring the
themes of the environmental pillar, was therefore importantly affected by "methodological
inheritance" form environmental impact evaluation, along natural sciences classical partitions
of environment into "compartments". We thus start with a poorly structured list of
environmental indicators by the OECD in the early nineties, and a systemisation of a
biogeophysical definition of environment in main compartments: atmosphere, water, soil
No. 010036
Deliverable number:
21 March 2006
Page 15 of 146
(land), biota. The main processes involved are in terms of pollution flows and resources stock
and ratio of use (see OECD 1991 list in appendices page 113).
As said in an OECD report (1998) « Initially many approaches to describing the environment
were limited to information describing environmental quality and quality change, in terms of
pollutant load or some other biochemical or biophysical indicators. However, it became
apparent that while this might be directly linked to some specific change in the environment
such as the loss of habitat or species, this sectoral approach did not necessarily support the
decision maker in better management of the environmentally damaging activity. »
Shifts in environmental stakes and extension of SD initiatives
The increasing use of the PSR approach in the early nineties was therefore one of the ways
enabling a double shift (Zuinen, 2004):
from state indicators to a broader perspective accounting for phenomena at the origin
of the evolution of these states as well as for modes of management and regulation of
from an approach focussing on environmental components to a more "environmental
problematic" oriented approach.
This change in approach, expressed through PSR and its successors (DSR, DPSIR), with the
search of causal links and the recourse to response indicators more directly linked to
regulatory action, is also apparent in other initiatives. In such initiatives, the framework of
stress-response type is not being used, there is a reforming of environmental compartments
around environmental themes close to the problematisation mode, having its basis whether in
environmental regulation or in NGO's (ecologists) instances. It also resulted in the
restructuring of the environmental pillar into human activities. So from the middle of the
nineties we find a reinterpretation of initial blocs (water, soil, air) into environmental issues
(waste, biodiversity...) which are then crossed with sub-systems (forests, agricultural
rangelands...), or a classification of resources that can be interpreted as revealing how
environmental regulation is managed. We take as example the WB 1995 sustainability matrix
(see appendices 2.1.4 page 119) who reinterprets blocs into problems of resource
consumption (Sources), environment's purification services (Sink), habitat (life support), and
sanitary effect (human health impact). Or, the same year with the propositions of the UN
University (see Murcott, 1997), who crosses relevant ecosystems for environmental action
with a list of environmental problems (landscape structure, production of goods and services,
biodiversity, air quality, water quality...).
Other initiatives have integrally defined their classification and themes of SD indicators
through political action modes, such as the European Union with Eurostat where structural
indicators has been widely used. "Following United Nations' experience and
recommendations, the commission conceived a framework relying on themes and sub themes
directly associated to the priorities of the EU policies" (Almunia, 2005). The ten themes, that
are recognised to be developed or amended in the future, are:
1) Economic development
2) Poverty and social exclusion
3) Ageing society
4) Public health
5) Climatic change and energy
6) Production and consumption modes
No. 010036
Deliverable number:
21 March 2006
Page 16 of 146
7) Natural resources management
8) Transports
9) Good governance
10) Global partnership
The present (December 2004) theme/sub-theme approach of UNCSD is a go-between these
two tendencies because inside the decomposition of SD into four pillars, themes are defined
with focus on policy issues within economy-social-institutional dimensions, while the
environmental pillar is divided in mainly biogeophysical themes and policy issue sub themes.
Nonetheless, there has been a massive recourse to the PSR framework. Despite
benefiting efforts that were made to structure indicators through a more problematic and
policy issue vision, such efforts ignored that the underlying framework was incomplete:
giving little information on the choice of dimensions, their relative importance, and not
translating a defined vision of SD. As a result, any new issue was translated into a possible
new indicator. This lead to an inflation in the number of indicators at the end of the nineties,
which ended producing lists that were hardly readable and manageable in terms of global
performance of economies in their SD perspective.
Though willingness to respect a parsimony principle figures in all initiatives, it was explicitly
expressed through the research of key or headline indicators from 2000 on. In some cases, it
lead to shrinking the final indicator list, in others to proposing different lists, keeping a large
list accompanied by headline lists (CEI vs KEI for OECD 2003, see from pages 115 to 119).
Hierarchical and problem oriented indicators
It was the study of themes and classification of key indicators that resulted in this movement
of recentering, and balancing of pillars and themes, to bring out the dominant themes:
Eurostat uses a "political problematic" classification (see appendices page 105), and
proposes a three level classification. The first level consist of 12 key indicators, destined to
the public and high level political deciders, level 2 has 45 indicators, useful for evaluating
political domains essential to SD and also for the public, the third level of 98 indicators
corresponding to intervention domains and destined to a more specialised audience. Sub
themes structuring level 2 gives a good idea of essential SD themes from a political
intervention point of view (see appendices page 105).
For others, the course isn't as clear. OECD stayed for a long time with pillars separating
economy, environmental, social, and possibly institution, in its SDI. The end of the 1998-
2001 mandate, they separates environmental, economic and human capitals under the theme
of resources, and add a list of socio-economic variables under the theme of results.
During the 2001-2004 mandate, specific SD dimensions are essentially added to the
environmental pillar (40 indicators among 69), and two other themes appear, with 5
indicators relative to pensions, and 14 relative to living levels in "developing" countries. This
consists in an approach where pillars of SD only develop themes that are not in other
indicator production initiatives: whether structural, but mainly environmental (CEI OECD
2004, appendices page 115), or sectoral (such as the list of agri-environmental
indicators, OECD 2001, appendices page 124).
For the UNCSD (see appendices page 103), we clearly have the 4 dimensions of SD
in the lists, reducing from 132 SD indicators in the middle of the nineties, to a list of 59 from
2001 on. The social pillar is covered by 6 themes: equity, health, education, habitat, security
and population, totalising 18 indicators within 12 sub-themes. The 5 compartments of
No. 010036
Deliverable number:
21 March 2006
Page 17 of 146
environment are: atmosphere, land management, seas and coasts, freshwater and biological
diversity, 18 indicators within 14 sub-themes. The two themes of the economic pillar are
economic structures and modes of production and consumption, with 14 indicators set in 7
sub-themes. Finally, the institutional pillar adds two themes to the list: institutional frame and
institutional capacity, 6 indicators in 6 sub-themes. We can note that in the majority of sub-
themes a unique indicator was systematically chosen, contrary to the sub-themes of poverty,
energy consumption and waste management differing with 3 or 4 indicators.
Balancing pillars
Original preoccupations of the logic behind SDI initiatives strongly condition equilibrium
between pillars, the diversity and level of refinement within themes and sub-themes. We can
distinguish initiatives building on general preoccupations on economic development (OECD,
EU), from those more marked by specific development questions such as poverty (WB
UNDP), alimentation and agriculture (FAO), and others clearly centred on environmental
preoccupations (UNEP, FoE), with natural resources concern for some (WRI), nature
conservation and protected areas for others (IUCN, WWF). Redifining Progress (RP) seems
to position itself in the beginning on a more balanced status for the three pillars of SD.
Indicators and thematic contents of each pillar in major SDI initiatives carried by these
organisations are presented in appendices 2.1 and 2.2.
3 Indicators
3.1 Definitions of indicators and SDI frameworks
Indicators: definition, utilisation and interpretation
Indicators are quantitative tools that synthesise or simplify relevant data relative to the state
or evolution of certain phenomena. They are tools for communication, evaluation and
decision making that can take quantitative as well as qualitative form depending on the
purpose of the indicators (Gallopin, 1997). The sustainable development indicators we
consider here emerges form a particular class of indicators of progress, whose aim is to take
into consideration sustainable development by integrating environmental, economic, social
and human dimensions. Their technical and scientific content, which we will consider in this
chapter, should not make us forget that indicators rest on conventions and that their
legitimacy therefore builds on social conventions on what progress is and on how to evaluate
it (Gadrey and Jany-Catrice, 2005). Such conventions are pre-requisite for the recognition
and durable use of indicators or indices by various actors.
Relations between different variables used in indicator development are often represented
with the help of the “information pyramid” (see Figure 1 below). This figure shows the
different levels of aggregation and synthesis of information. On the first level, raw data
consist of phenomena measurements (variables) in time and space for different populations.
Then, indicators synthesise or simplify relevant data relative to the state or evolution of some
phenomena. Some indicators are the result of aggregation, with or without weighting, of very
diverse data and therefore carry a synthetic message. This is the case for instance with the
Human Development Indicator calculated by the United Nations program for development,
that aggregates 3 indicators (life expectancy at birth; educational attainment; GDP per capita).
These aggregated indicators are also called indices (or index). When they aggregate
No. 010036
Deliverable number:
21 March 2006
Page 18 of 146
indicators having different measuring units, they transform these indicators into indices
(mapped indicators) to make measuring units disappear.
Indicators enable to represent and analyse a specific phenomena. Their signification is in
principle larger than that of the variables composing them; and they permit to build a model
that represents reality, but that is only a simplified image among others of this reality. The
choice of indicators, as well as their ex post interpretation, are founded in at least partly
subjective judgement, whereas their construction is a scientific and technical work. So
indicator selection is partly a political activity.
Figure 1 : Pyramid of information
Frameworks: linking SD principles with indicators, whether partially or completely
A SDI framework aims at translating a vision of sustainable development to an organising
frame for indicator production. It organises the presentation of information relative to the
different dimensions of sustainable development and their links, based on a set of principles
forming the chosen SD vision. Instead of having a “one-problem, one-indicator” approach,
SDI framework has to bring the economic, social and environmental aspects of society
together, emphasising the links between them” (Olsson and al., 2004). A SDIF thus has a
double role, more or less balanced depending on the SDI initiative: to rationally answer the
sustainability paradigm, and concretely organise the production of indices on the observed
phenomena. When a real organizing frame is missing in the framework, one often finds in the
process to design indicators a list of criteria for their selection. Such a process is more or less
explicit and documented, and its role is usually more comprehensive than prescriptive.
Examples of formalised processes are given in the next section. Conversely, when the
framework is mainly devoted to organising indices, such as PSR and its derivatives (see
section 4.1.1), the framework can be endorsed by projects having very different visions of
sustainability. While mainly considering the outputs of the framework, end users
No. 010036
Deliverable number:
21 March 2006
Page 19 of 146
(policymakers, politicians…) can overlook how such outputs depend on the conceptual model
sustaining indicators development. This entails consequences on framework selection. A
complete framework like (Bossel, 1999) system based approach (see section 4.1.4) remains
hardly operational due to its high demand on integration. However, the "completeness" of the
framework is an important aspect insofar as what it does not take account for will end up (or
not) in an indicative list of criteria for indicator choice. An incomplete framework leaves in
the end to much latitude into the indicator selection process, without ensuring that all
dimensions of SD and certain links between them will really be accounted for.
3.2 Functions of indicators
A SDI is a quantitative tool enabling analysis of phenomena that concern sustainability of
development, and their evolution. It thus has as a function to synthesise the information
permitting to assess SD performance, with regard to states of various "dimensions" (man,
society, economy, nature) and the evolution of states, taking place over different territorial
entities. An SDI can also have the function of being an alert system by means of tendency
prolongation, informed prospective or simulation.
Indicators can as well have a less territorial and more sectorial (or corporate) vocation, to
evaluate the sustainability of sectorial policy, of a community's mode or project of
management, of a business.
Finally, a related function to the two above concerns indicators link with action, and reactions
to indices. Indicators, should point out, when possible, the associated actions and priorities
within public policies (norms and rules, inciting policies,...) that could be carried out to
modify the behaviour of the concerned categories of actor. Indicators can also have the
function to facilitate auto-evaluation of sustainability of a production process or a practice for
3.3 Reference values: target, threshold and goals.
In an SD perspective an indicator without pre-specified value or without context has little
meaning (Rigby and al., 2001). Thresholds, targets and benchmarks or reference levels are
necessary to assess the contribution to SD of an indicator change. These references values can
be expressed as a negative, a zero or a positive value in the indicator unit, and when the
indicator passes this reference value, it reveals an unsustainable path. They are crucial in
sustainability assessment of agro-ecological systems “given the propensity of ecological
system to ‘flip’ from one state to another”(Moxey, 1998).
However, for numerous phenomena, it has been shown that establishing thresholds only with
biological and physical criteria can turn out to be impossible or unsound (Pannell and Glenn,
2000). Today scientists’ definition of standard and reference levels is becoming less
influential and the negotiated dimension of standards and reference levels in most of human
activities is now widely accepted (Olsson, 2003). Indicators, with fix quantified goals, exceed
the frame of scientific expertise and deal with political trade-offs between means and ends
concerning problems related to sustainable development.
It may be difficult to produce absolute reference levels, coming from the scientific study of a
phenomena or from negotiation processes. This can make approaches based on relative
reference levels attractive. For instance, trend indicators are a way to relate the evolution of
No. 010036
Deliverable number:
21 March 2006
Page 20 of 146
phenomena. More generally, based on relative or absolute levels, it is possible to rank
territorial entities, from which relative targets can be defined as the distance to the best
performance, the median or worst performance, …
Methodological aspects of these reference values are discussed in more detail for the
agricultural problematic in chapter 5.
3.4 Indicators classes
There are different indicator typologies, but the one which has been most influential on
different framework distinguishes list of indices from aggregated indicators and composite
indicators. There is great variety within each category. In what follows, we mainly
distinguish lists of indicators, who measure an indicator for each element of sustainable
development with an appropriate unit, from aggregated indicators who combine in one indice
(or sub-indice) different variables (or sub-indice) related to different dimensions of SD.
Within aggregated indicators, one usually distinguishes composite indicators that aggregate
different sub-indicators measured in distinct units, from simple aggregated indicators that
involve measuring different phenomena with the same unit. In the first case, aggregation uses
pondering so methodological choices are involved, whereas in the second case an evaluation
theorization must be put forward (economic, biological, physical...).
In addition, although we will not develop this distinction in detail, one should clearly separate
indicators based on objective data involving little value judgement on a lived situation, and
data collected through survey involving opinions and feelings. If we take well-being
indicators for example, an indicator of the first type can be rebuilt from these objective data,
or it can be directly measured by an opinion survey.
3.4.1 Composite SDI
Composite indicators are based on sub-indicators that have no common meaningful unit of
measurement and there is no obvious way of weighting these sub-indicators.
They normalise the judgement set upon the set of sub-indicators from which they are built
and, by the reduction they operate, they facilitate global comparisons between territorial
entities and communication to the public, as for example the Ecological Footprint index or
the Human Development Index. However, this reduction implies information loss, which
should be minimised. Applications to the thematic of sustainable development are also
numerous (Saisana and Tarantola, 2002) and one can rely on an important literature. A
specific development of this literature will be used in task 2.6, and we will here limit
ourselves to only developing the most determining aspects for WP2 tasks involved in
thematic indicators.
A formal presentation of this type of index ca be expressed as follows:
Composite indicator I = f ( M )
Dimensions M = (m1 ,…, m k) where mk =h (Y )
Sub-indicators Y = (y1 ,…, y l) where yi =h (X )
No. 010036
Deliverable number:
21 March 2006
Page 21 of 146
Variables X= (x1 ,…, x n) where xi =g (Z)
Data Z = (z1 ,…, z p)
Contrary to indicator lists where one usually considers that the users will be able to articulate
a set of indicators to evaluate their contribution to SD, aggregation into a single index- or a
limited number, one for each pillar for instance imposed on composite indicators calls for
an exhaustive explicitation of the potential contribution of an indicator to SD or to each SD
pillar. Theoretically this means that whatever the value of each sub-indicator and whatever
the value of other sub-indicators, the aggregation mode enables to attest the real “marginal
contribution”. To give a simple example, if the aggregated indicator is the simple sum of sub-
indicators, this implies that whatever the level of each sub-indicator the contribution of a
supplementary unit of each sub-indicator is constant, or positive or negative, and that
indicators are independent between themselves in their contribution to sustainable
development (no cross effects). This condition is naturally inconceivable given the
complexity of the problem, but it remains an ideal objective. Without fixing such a
constraining frame the following should be ensured:
- make variation ranges of sub-indicators comparable,
- that a consensus emerges, for each retained indicator –in literature or among
SEAMLESS participants, on a univocal, that is positive or negative, contribution to
SD or SD pillar. If such a consensus is not reached, the sub-indicator can be used if
threshold values changing its contribution to SD are identifiable.
This last constraint is by the way a simple mean to avoid the explosion of the number of
indicators in SDI initiatives.
We will analyse here a certain number of methodological aspects because this type of
indicator regroups the range of difficulties other indicators face. This analysis will rely
essentially on works lead within 2005ESI which is a reference where quality in the
methodological treatment of each step of composite indicator construction is concerned.
One of the first methodological determinants is knowing whether the ultimate objective is to
enable ranking of different countries or regions or obtain an index. Clearly SEAMLESS
should enable to compare different policy options in a determined region, which implies that
the final output cannot be only such a ranking and invalidates certain methodological options.
Seven steps are generally distinguished:
- Choice of variables: this step aims at determining the variables that will enter the
construction: deliberate choice, relying on a list of criteria involving availability of
data, quality, comparability, scientific pertinence, and/or relying on statistical
multivariate analysis (correlation analysis, principal component analysis, Factor
analysis, Cronbach’s alpha, Cluster analysis). It also aims at choosing the unit in
which a variable is expressed.
- Imputation mode: missing data may introduce distortions in the information vehiculed
by the indicator. The choice of imputing data or not and of the eventual imputation
method determines, in part, the quality of the indicator. At the heart of imputation
methods intervenes the identification of the statistical properties of the missing data.
They can for instance correspond to a MAR process (the distribution of missing data
doesn't depend on the results associated to missing data), or MCAR (particular MAR
case where the probability that a value is missing is random). Relying on hypothesis
on these proprieties, different statistical methods enable to impute values to missing
values from the analysis of the set of data (example of techniques: listwise deletion,
No. 010036
Deliverable number:
21 March 2006
Page 22 of 146
Last value carried forward, Hot desk closest match, average closest match,
Expectation-Maximum algorithm, multiple Imputation (see Conway, 1993b;
Fairclough, 1998; Little and Rubin, 2002).
- Transformation and normalization: one can apply a mathematical transformation to
data in order to gives it desirable statistical proprieties (homoscedasticity, normality,
reduction of outliers, non-colinearity). Salzman (2003) identifies several possible
techniques for standardization: no standardization (justifiable essentially with ratios),
normalization by Z-score or gaussian normalization, LST(rescalling), ratio percentage
and distance to target, satisfaction level stated by experts.
- aggregation : aggregation can than be additive, multiplicative or put to a power and
also involves fixing a method to determine the weights assigned to each sub-indicator.
Some discrete rules (lexicographic, weak link, number of sub-indicators below
thresholds, …) can also be used to assess sustainability index in a more strong
sustainability perspective.
- Evaluation: sensibility analysis and uncertainty analysis. Using sensitivity analysis,
we can study how variations in the composite index derive from differen t sources
of variation in the assumptions. Sensitivity analysis also demonstrates how each
indicator depends upon the information that composes it.
- ex post statistical inference. In this stage where ex post means that the composite
indicator has been calculated, statistical inference tools can be used to identify the
relation between the composite (or some transformation of it) and a set of contextual
variables not included for different reasons in the composite. 2005ESI (Esty and al.,
2005) for example, also identify what are the included variables that infer mostly on
the ranking results.
- Choice of the visualization method.
See Appendices 1 page 97 for more details on methodological aspects.
3.4.2 One unit aggregated SDI
The design of aggregated SDIs entails measuring, in a single physical or monetary final unit,
the different phenomena considered. Aggregated indicators therefore call for a theoretical and
methodological framing (natural sciences, physical sciences, sociology or economy) to
operate the transition to a common measuring unit. On the other hand, by requesting a same
unit, it simplifies the considerations discussed above.
One can distinguish two types of aggregated indicators: those passing through a
monetarisation of non-market goods and services, examples of this type of indicators is GPI
(Genuine Progress Indicator), and other indicators relying on the aggregation in physical
terms of resource use, such as Ecological Footprint (EF). In both cases the question as to
what phenomena can be apprehended through this type of normalization arises. For instance,
the EF aims at measuring in a single unit expressed in a geographical surface, the biologically
productive area required to produce the natural resources that an entity (community, region,
city, individual) consumes, expressed in global square meters. It is based on a number of
assumptions about sustainable technology, to transform environmental degradation and
resources consumption into surface (see section 4.2 for detail).
Despite the numerous desirable properties of this index, that we will discuss later, the
sustainable development dimensions it can account for is limited since it only concerns
environmental degradation and resource consumption.
When using monetarisation, three approaches are possible:
No. 010036
Deliverable number:
21 March 2006
Page 23 of 146
to measure market or near-market values and thus underestimate the total value of
non-market goods and services,
to measure total values that are established by an expert panel,
to use valuation methods for non-market goods.
Only the third approach calls for an explanation. These methods rely on identification of
consent to pay (or receive) by individuals to benefit (or cease benefiting) from a non-market
good or service, generally environmental, even though there exist a few examples in the
social field. Such valuation methods are whether based (i) on behavioural intentions in
hypothetical situations provoking trade-offs between non-market goods and money
Contingent Valuation Method (or non-market good-Conjoint Analysis)-, or (ii) on analysis of
the market of complementary goods or substitutes – Cost Travel Method, hedonic Method,
Mitigation Costs Method-, like for instance money spent to get to a natural site. Even though
their use is justifiable from a theoretical point of view, their results are weak depending on
the nature of measured goods and services. Some examples can be found in the following
individual familiarity with goods and services proposed to evaluation (Heberlein,
perception of financing modus and public action modalities (Geniaux, 2001) and/or
quality of implementation of methods (ability to reduce methodological bias, see
Mitchell and Carlson 1989 and context effects in Brown and Slovic, 1988).
In addition to doubts about validity measures dedicated to given and contextualized
components of environment, the harshness of these methods (technical and financial) in large
scale initiatives (regional or supra) imposes important recourse to transfer of values between
components judged comparable or between measures taken in different places and contexts,
which make results that much more opaque and unreliable.
The most known indices such as ISEW (Index of Sustainable Economic Welfare), GPI
(Genuine Progress Indicator), GSR (Genuine Savings Rate), use the two first approaches
essentially privileging evaluation of the market values of the considered environmental and
social components (Lawn, 2003).
3.4.3 Lists of indicators and dashboards
An indicator list is a set of indicators measured in units appropriate to the considered
phenomena. These lists are generally ordered in themes and sub themes, so as to enhance
legibility of the performance of a nation or a region in the different dimensions of sustainable
development. Indicator lists remain the most common used in SDI initiatives), even though
they catch less attention from media and public, as compared with EF.
The principal advantages explaining the preference for these types of lists in SDI initiatives
the ability to produce information for those who make the decisions Decision makers
receive information directly centred on their domain of action and potentially on
evaluation of their past actions,
they comply with multi-disciplinary and not necessarily inter-disciplinary
approaches, possibly skipping the work of conceiving a real framework,
No. 010036
Deliverable number:
21 March 2006
Page 24 of 146
lastly, they enable the implementation of substitutability hypothesis between certain
components of sustainable development. And substitutability constitutes for certain
frameworks a condition for addressing SD issues.
One of the principal shortcomings of these lists is the inflation of the number of indicators
over time. The reason for this inflation is that this type of approach implicitly aims at
completeness without a clear definition of essential and universal properties of SD. Inflation
is a result of the fact that it is always possible to identify new components of SD which is not
represented in the existing list. This problem naturally grows in large scale initiatives with
great diversity of environmental, social, institutional and political contexts. Lists also make
global performance for a country or a region complex to appreciate, as far as SD is
System-based approaches on the other hand define a set of properties essential to the desired
viability of the considered system (country, farm, cow...), and for other vitally related
systems, so it is an organising means that avoids repetition and is limited by the number of
included systems. However systematic study of interrelations can be a heavy burden.
System-based approaches indeed allow to avoid one of the recurrent flaws of aggregation,
being that the loss of an essential element of SD can be compensated by another without any
clear expression .
4 Main Frameworks
The main frameworks are described in four sections by identifying their organising
principles. In a first section, we present frameworks leading to indicator lists. They differ
from those used for aggregated or composite indicators, introduced in a second section. In a
third section some agricultural and rural specific frameworks are proposed. The fourth section
presents and discusses transversal concepts of SD that did not lead to a formal and applied
SDI framework but contribute to our objective in SEAMLESS.
4.1 SDI using list of indicators
4.1.1 The PSR, DPSIR, DSR approaches
The "P-S-R" or " Pressure-State-Response" framework was elaborated in the 80's to organize
environmental analysis into causal chains: centred on the state of environment, pressure is
seen as exerted by human activity through pollution flows and resource consumption, while
Response encloses societal measures taken in reaction to the state or change of state of the
environment. (see OECD lists page 113 to 119)
This framework was later refined into the "D-P-S-I-R" or "Driving forces-Pressure-State-
Impact-Response" model. Here, an attempt is made to distinguish the cause of the pressure on
environment, human activity mainly through consumption and production (Energy
consumption,...), from the pressure itself (CO2 emissions for instance). A difference is made
between the state of the environment, of a particular stock, and the impact it has on other
stocks within environment or other dimensions (pollution's impact on human health for
instance). Note that identifying indicators to these categories is not always straightforward.
Nutrient balances, for example, might be a pressure indicator, but could similarly also be
identified as an impact indicator.
No. 010036
Deliverable number:
21 March 2006
Page 25 of 146
An advantage of this family of frameworks is that causal chains are easy to conceive, and are
particularly adapted to some straightforward interactions between economy and environment.
Stocks and flows can be treated correctly within the chain. Another central feature of this
approach is that, economic activity is identified as the main driving force. So responsibility is
clearly put on consumption and production modes, as suggested by principles 7 and 8 of the
Rio Declaration. This serves policy oriented thinking, and gives a place for the institutional
However, the limitations of PSR were only partly, and for some not at all, answered by its
derived successors. These frameworks organise causes and effects of environmental state or
changes, this implicitly emphasise the environmental dimension of sustainable development.
If these initiatives were to put economy in the centre, influencing factors (general economic
trends, sociological situation...) would not so simply fit into this type of causal chain.
Furthermore, while the introduction of impacts can show some consequences of
environmental damage on individuals and society, social condition (poverty) often influences
this damage but through economic activity, thus the link is too indirect to be clearly taken
into consideration by these frameworks. In general, the framework is adapted to situations
where environment is at stake. In these situations types of pressures are identified: specific
(e.g.: use of N fertilisers in nitrate pollution…); straightforward (no multiple causes creating
different equilibriums among non-linear interactions); impacts are not important indirect
causes for further degradation; sectors of activity where efficient measures can be taken are
pre-determined (e.g.: agriculture, industry…)...
In our discussion, the fact that this simplifying model misses some complexity and some
phenomena is an important part of its general shortcoming. Basically, apart from implicit
emphasis on environment, the PSR model gives no hint to what it is important to look at,
what balance should be found between dimensions of sustainable development, or even
between environmental issues, and why certain links are important or not. As discussed in
preceding sections, it is not an indicator framework translating a vision of sustainable
development, except for communicating the minimum common agreement that there is a
problem with man's use of environment. This explains why PSR-type chains are used in
initiatives lacking such a vision and having very different normative goals.
Finally, because causal chains fit well into a sectoral breakdown of issues, simply by splitting
pressures, PSR tends to advantage sectoral policies. Although this may be positive side of the
approach, this can miss the SD objective as a whole, requiring policies across sectors
concerning consumption, production and trade.
4.1.2 Capital-based approach
In a SD perspective, it is essential to have an equally balanced approach between the 3
dimensions of SD: social dimension, economic dimension and environmental dimension.
Only such an approach enable to systematically emphasise interdependence relations between
them, without under-estimating either one. “To make SD more concrete, several writers have
transformed these pillars into different types of capitals to be able to more easily illustrate
the linkages and trade-offs between them (Bossel (Bossel, 1999), 1999). A frequent
classification is four types of capital, namely, manufactured capital, natural capital, human
capital and social capital (Daly, 1990; European Commission, 2002) (Olsson and al.,
This classification plays an important role in equilibrating dimension, but leaves much
flexibility in defining the contents of each capital. In this approach it is essential to precise the
dominant interpretation of the notion of human, natural, social and economic capital.
No. 010036
Deliverable number:
21 March 2006
Page 26 of 146
Following Schuller (2000), Human capital concerns essentially economic behaviour of
individuals, especially the accumulation of competence and knowledge serving their
productivity and income. It can thus integrate certain indicator on necessary means and
conditions for this accumulation (health, well-being).
Social capital involves in particular networks, interpersonal relations, as well as shared social
norms and values they go by. It can be defined at community level, or other entities.
Woolcock and Narayan (2000) differentiate bonding social capital (relatives and close
friends) and bridging social capital (acquaintances, other social groups). When considered as
a stock of social resources, it is also distinguished from capacity, i.e. the ability to draw on
capital for valued purposes. However, social capital is not always attached to a positive view:
mutually supportive interpersonal relationships within a network or community may be
accompanied by prejudice or hostility towards outsiders; resources can also be "objects of
struggle" between individuals and families. Finally, the possible distinction between social
and institutional capital is often blurred in practice. Social capital would have a privileged
position in discussions on sustainability, namely facilitating the creation or use of other types
of capital.
4.1.3 SEEA
The System of integrated Environmental and Economic Accounting (SEEA) is based on an
enlarged national account. It origin can be found in the work of UNEP and the WB, which
was formalised in 1993 in the "Integrated Environmental an Economic Accounting”
handbook”. It develops a satellite system of the System of National Account (SNA) with a
specific focus on the links between economy and environment. It is presented by its
developer to be more flexible because of the use of different statistics modules using both
physical and monetary units. The main differences with other lists of indicators is that it:
- focuses on an elaboration of “all” environment-related flows and stocks,
- provides a linkage of physical and monetary accounts,
- constructs modules with systematic monetary evaluation of man-made and
natural capital, and environmental costs using the link between physical and
economic assets, and provide an estimation of the environmental protection
expenditures (Alfieri, 2000),
- proposes an environmental adjusted national wealth and some other aggregated
indicators in its lists.
The latter indicators are calculated by subtracting depletion and damage cost from the
economic indicators. The most currently used indicators are the EVA (Environmentaly-
adjusted value added), EDP (net domestic product), and ECF (capital formation). A detailed
document on implementation of this framework can be found in (UN and al., 2003).
The main limits of this framework are the same as for other frameworks using monetary
evaluation of environmental benefits (see section 3.4.2). Another problem is that some
attempts of damage valuation have been systemically excluded because they were seen as too
controversial. Moreover, these types of approaches has a complicated organisation of satellite
systems and deals exclusively with physical supports of production, of consumption, and of
environment (accounting approach of resources). This hampers interpretation in terms of
well-being or any measure of quality of life linked to socio-economic variables non
expressible in resource units. This framework also remains ambiguous concerning the
coupling between the monetary and physical units, since it is based on a strong
substitutability principle to define resource depletion in physical terms, but uses monetary
No. 010036
Deliverable number:
21 March 2006
Page 27 of 146
evaluation of these losses in other modules, thus mobilising a weak substitutability principle.
Finally, this framework does not provide directly SDI, but may be used as a database or as
intermediate indicators.
The main advantage of this types of approaches is that they provide a tool that enables to
highlight (partly at least) interrelations between human activities and resource depletion, by
identifying these links.
4.1.4 System-based approach Systems approach and modelling
A systems approach to sustainable development is of interest because through its construction
it enables to explicit what features are considered essential to SD, and it provides an
organization for an indicator set following these features.
Complex dynamic systems and their interactions have been studied at length, in many areas
of science, and theories accompanying a body of organisational principles have emerged
independently of their particular setting (Ewert and al., 2005). A review of its application to
ecosystems is presented in Becker (1997). Bossel (1999) exposes a systems framework and
method for deriving sustainable development indicators, which will be commented here.
A system is an entity composed of elements structured in some way (a person, a community,
a car...), with a boundary defining exchanges between interior and the system's
environment, i.e. its exterior.
The first step in deriving an indicator set for sustainable development is:
Defining the scope and purpose of the study. This involves determining what is the
system to be studied, which we will call the total system, its boundary (spatial area,
nature of the concerned system). What is exterior to this system may be accounted for
with less importance than the system itself, so this subjective choice is important.
Then a modelling of the total system (ex: human society in its environment in
Europe) has to be determined. First, relevant subsystems should be identified, that are
parts of the total system having essential meaning relatively to its functioning (ex:
economy, environment, society).
Among them a particular subsystem can carry the prioritary objectives of the study
(ex: European society); the other subsystems, and the total system seen as embedding
this particular subsystem are then essentially supports for it. Note that identifying if
such a subsystem is pertinent, or if on the contrary all subsystems have the same
status and should be sustainable for their own sake (hence the total system also)
reflects particular visions of SD. For instance, the natural environment can be viewed
as supporting the human society, or its development can be pursued also
independently of its utility to humans. Anthropocentrism was identified in Becker
(1997) as the only widely accepted basis for SD, with the argument that it enables to
consider environmental preservation for its own anthropocentered reasons. Bossel
(1999) comes to an analogue conclusion, and suggests that the indicator set should,
whatever the ideological background, cover the widest horizon of attention, though
some indicators may not reflect actual priorities or dominant values.
Finally, subsystems and their relative importance is also determined by choosing
what essential relations between them are accounted for.
No. 010036
Deliverable number:
21 March 2006
Page 28 of 146
The choice of essential supporting systems and the model for the total system are also
subjective. This choice can emphasise or neglect:
certain relations (advantaged or not by the chosen subdivision),
certain types of relations (decision making, sociological, environmental, economic...),
adapted or not to the scale of the subsystem. Matching a system to its environment: properties and basic orientors
Once the total system, its essential subsystems and their relations are determined, the second
step consist in choosing the properties these systems should bear to be sustainable. Bossel
(1999) argues that human society cannot be static, so its viability has to encompass
development, which is seen as equivalent to sustainable development. Following the
orientation theory (Bossel, 1999, 2002), the properties of its environment shapes how a
dynamic system is structured and how it behaves. Indeed, the system orients its structure,
function and behaviour to adapt and take advantage of exterior and interior conditions. The
theory identifies six universal properties (supposed to be non-redundant and cover all
situations) of a system's environment. So six corresponding orientors, meaning corresponding
categories of concern for the system to deal with its environment, are derived. A
psychological needs orientor, being a concern for sentient beings (system-determined, as
opposed to environment-determined) is necessary for human related systems. These seven
orientors, presented below, will be represented by relevant indicators representing the
system's answer to the concern in a particular situation. They are supposed to cover all
eventual viability concerns for all systems:
Existence. The system must be compatible with and able to exist in the normal
environmental state. The information, energy, and material inputs needed to sustain
the system must be available.
Effectiveness. The system should, on balance over the long term, be effective (not
necessarily efficient) in its efforts to secure required scarce resources (information,
matter, energy) and to exert influence on its environment when necessary.
Freedom of action. The system must have the ability to cope in various ways with
the challenges posed by environmental variety.
Security The system must be able to protect itself from the detrimental effects of
environmental variability, i.e., variable, fluctuating, and unpredictable conditions
outside the normal environmental state.
Adaptability The system should be able to learn, adapt, and self-organize to
generate more appropriate responses to the challenges posed by environmental
Coexistence. The system must be able to modify its behaviour to respond
appropriately to the behaviour of the other systems in its environment.
Psychological needs. These constitute an additional orientor for sentient beings.
So the choice of properties can:
resume to these universal basic orientors for each system, those of them that are
relevant (some simple subsystems do not deal with their entire environment),
be chosen following the interest of the indicator set producer, using or not the
universal orientors presented by Bossel (1999).
Note that the main interest of this systems approach is its attempt to universality, ensuring
that no vital area of sustainability of a system is neglected. Furthermore, the use of these
No. 010036
Deliverable number:
21 March 2006
Page 29 of 146
orientors is unique, not that there doesn't exist other sets of orientors, but because each
orientor is necessary in that it covers concerns ignored by the other six orientors . So
choosing the option of using the orientor set takes subjectivity out of this step (apart from
accepting the pretended universality). Using basic orientors to guide indicator selection
The third step is identifying indicators. A main principle in systems analysis is that the total
system is viable only if all subsystems are. This leads to deriving representative indicators of
the basic orientors of (or the properties assigned to) all systems (total system, other essential
subsystems), and indicators of the participation of each subsystem to the basic orientors of the
total system. If one indicator is chosen for each basic orientor, this leads to a maximum of 14
times the number of subsystems, plus seven for the total system which is limited, but still
Even though indicator choice is subjective, all areas of sustainability are covered. If
indicators are relevant, the sustainability assessment should be relatively independent of their
particular choice.
The indicator set is designed for the purpose of determining if the total system is sustainable(
we can take advantage of it for other purposes, as we will see below): all orientors will have
to satisfy a determined minimum satisfaction (corresponding to an indicator threshold) to
ensure sustainability. If one orientor is bad off, the whole system is unsustainable. This
implies that trade-offs between minimum satisfaction of different orientors is impossible.
And only when basic satisfaction for any defecting orientor is ensured, should one be
preoccupied with ameliorating other performance. However, this systems approach ensures
"all" important questions are addressed concerning the viability of the system model
(meaning the subsystems chosen). So it is also a way of being eventually complete in
deriving an indicator set, of determining what relations are considered important. The
particular choice of indicators can than be chosen to address only the orientors considered
relevant, and with diverse purposes in mind such as policy relevance, early warning,
aggregation or policy comparison... The following points should be kept in mind:
Indicators should account for the links between subsystems: for this a systematic
review of the participation of a subsystem to the basic orientors of another can be
conducted. The chosen indicators should highlight these links.
If one indicator per orientor is chosen, it can be because: it is a good representative of
the corresponding properties, it represents the weakest point, it is an aggregation of
representative indicators (and/or other criteria, depending on the purpose of the study,
sustainability assessment or other).
Indicator's scope should be clearly presented, only its participation to the considered
orientor should be assessed, even when it can be related to another orientor. Assessing viability and performance
This leads to the fourth step: assessment of basic orientor satisfaction, that is quantifying the
indicator's participation to sustainable development.
This can be done thoroughly, for instance by determining assessment functions mapping
indicator performance onto an orientor satisfaction scale, with at least one threshold
corresponding to basic satisfaction. It can also and often has to be done less precisely, the
sustainability diagnostic depending only on evaluation of basic satisfaction of all orientors of
all systems.
No. 010036
Deliverable number:
21 March 2006
Page 30 of 146
The choice of these assessment functions, and the breaking points for indicator values, the
point where the contribution of the indicator to orientor satisfaction changes (in value, in
rate..) are subjective. However, the notion of what is satisfactory should be related to
properties such as adaptativity, which then are considered independently, so assessment
functions should be free of trade-off considerations.
To summarise, the systems approach has the following characteristics:
it is systematic, so a complete view of SD is searched for, and the areas where
indicators, thresholds, breaking points or data are missing are explicited,
it is a construction, so steps where subjective choice is made can be broken down and
made apparent,
it supposes a model of interacting subsystems is derived, so their functioning is
researched, and interactions are emphasized,
within this construction which reflects the chosen vision of SD, a straightforward
organization for the indicator set is given,
assessment of SD is simple: all indicators have to be above a determined threshold, if
not the whole system is unsustainable, no trade-offs are possible between the system's
basic orientors. On the other hand, for comparing different policy options for
instance, indicators have to be chosen that highlight their advantages and
shortcomings, the thresholds being of less central importance.
the indicators are relative to general functions of a system, so, a priori, policy
relevance (understood as interpretable for decision making), even though the
subsystem choice can be related, has to be kept in mind and ensured when choosing
the particular indicator.
Sustainability is assessed through general properties, indicating areas and concerns to
be considered. The indicators that will represent them can be chosen to account for a
particular context, and Bossel 1999 provides guidance on participative methods to
seize such a context. Why the system approach can help us design a framework
It seems interesting to discuss how a system based approach allows to treat certain elements
that the WP2 SEAMLESS Framework has to deal with : scales, sectors, rurality and policy
First, the definition of the studied system influences how these elements intervene. For
instance, we can assume that sustainable development is our objective, and that we want to
build an indicator set for Europe. Sustainable development of Europe (our system) can be set
as the primary goal, with consequences on its world environment assessed in a more or less
organized manner since it is exterior to the considered system, taking for granted a number of
trade, economic, and other world wide functions (that do not depend only on Europe). This
will lead to a set of indicators different than that derived, for Europe, from a world
sustainable development point of view. In the latter, the world SD is the goal and the world
our system, and how Europe finds its place within this goal will shape a European indicator
No. 010036
Deliverable number:
21 March 2006
Page 31 of 146
System and scale
This is true whatever the option is from a world point of view. The first option is to consider
a framework for world SD (or whatever other purpose at world scale), and then a framework
for Europe's SD within paths compatible with the first one, making an SD assessment at
different scales (the link between scales to be determined). The second is to consider a
framework for world SD, and make Europe a subsystem of the world, then having a systemic
relation between both scales: the participation of Europe to the orientors of world
Whatever the purpose, SD assessment or other, whether the smaller scale should be treated as
a subsystem of the larger is a question. Note that in Bossel (1999) the scale is specific and the
notion of subsystem is used at whatever scale for domains or dimensions of SD; this is more
directly similar to considering a sector of these dimensions as a subsystem in a sectoral
The same thing is of course relevant when considering the European level related to lower
levels (in scale), such as country, region, etc. What territorial entities should be used as a
studied system in SEAMLESS, with a complete individual framework for its set of
Moreover, if sub-sectors and/or smaller scale regions are to be fully assessed, new indicators
may be needed, possibly they will not just be an adaptation of the types of indicators figuring
in the higher level sets (such as looking at a particular type of event in a region where it is
traditional, as revealing importance of cultural activity; versus assessing rurality functions,
that can be relevant only at a community level, and not at higher levels).
So it seems pertinent that a framework be designed that can thoroughly analyse the
sustainability questions specifically at each scale, or at some important scales.
System and Sectors
The definition of the system also has consequences on the sectoral approach. Indeed, at any
scale, sustainability of agriculture (our system is then the agricultural system) versus place of
agriculture within the sustainability of its imbedding community (our system is then a village,
a region, a country..) are hopefully not incompatible. But different interest will hardly give a
same indicator set, and the systems framework construction enables to translate this
If the starting point of SEAMLESS is sustainable development of agriculture in Europe,
meaning the studied system is the agricultural sector, the same discussion and choice has to
be made concerning sub-sectors relevant to the different issues (sugar beat issue, different
agricultural systems...). Should sub-sectors be assessed completely and individually, deriving
as many thorough sets as necessary; are they subsystems to the agricultural system?
System and rurality
The concept of rurality is put forward in SEAMLESS. It can be defined relatively to a spatial
area imbedding agriculture, having characteristics of land use, or relatively to a community
living in such an environment, and having particular sociological functioning.
The meaning of rurality has to be defined in SEAMLESS. On the one hand, if rurality is
defined as a specific area, for instance land cover discriminating rural from semi-urban from
urban area, pertinent as for as agricultural land is valuated, is this definition compatible with
available data for all EU countries ? On the other hand, if rurality is defined as functions of a
No. 010036
Deliverable number:
21 March 2006
Page 32 of 146
community in rural area, exceeding land management, what is the relevant community and do
we have data for such community?
As will be presented below, we consider the heart of rurality to naturally take place at the
level of community (village scale in France…). At this scale the concrete relations between
people living in the same community take place, with particular tensions (in particular
between farmers and other inhabitants).
In the framework proposed in the last chapter we distinguish, at all territorial scales, and
within the total system (the region and its inhabitants at regional scale), an agricultural
system: it encompasses the agricultural land, the people working or living on this land and
their families. Both systems and their respective logic, agricultural system maintenance and
sustainable development, are articulated. At farm scale, the total system is the agricultural
system (not territorially defined), so issues related to the more general sustainable
development perspective are only indirectly modelled considered, through the eye of
agricultural maintenance, or in the link (to be modelled) of this scale with the larger territorial
scales. This is why the community level, defined as the first available elective entity, is
important. Hence at community level, the total system is the community, and within it the
agricultural system is assessed, but also the sector's relation to the whole community through
its participation to the sustainable development properties of the community. Not only is this
the scale of direct "physical" interaction, but the contribution of the agricultural system to
sustainable development is mainly dedicated to such interaction. At higher territorial scales,
the contribution to sustainable development will also have to consider and emphasize other
questions such as rural-urban relations, regional balance between activities... Because this
scale may be skipped in SEAMLESS for lack of data, we suggest how to possibly account for
some of it at regional scale.
This community level is also the smallest entity where all functions of human society can be
found, following Bossel (1999).
A difference can also be made between functions (environmental management) and activities
(farming) that are specific to communities in rural area, and the functions and activities that
exist everywhere but are influenced by the fact of taking place in a rural setting (water use,
public services, land use planning...). These issues are within SEAMLESS scope as they are
influenced directly by CAP policies as well as all environmental policies. The indirect
consequences are also important because rurality deals with the closely meshed agricultural
practice and social construction (modes of social reproduction, local political power of
agricultural sector, dependent on the type of farming system and farmer’s origin ...).
System and policy relevance
If policy relevance is the fact that an indicator gives information on a phenomenon that can
be influenced through known political initiatives, it is a desirable feature for indicators, but
should pertinence and completeness of the systems approach be sacrificed to it? This type of
policy relevance is a criteria that is often evoked as desirable if not necessary when choosing
individual indicators. But it can also reduce the scope of the indicator assessment,
concentrating on areas and issues we have responses for, instead of completeness criteria. The
P-S-R type framework has a natural setting for policy relevant indicators with response
indicators, whereas this criteria usually has to be checked individually in choosing particular
indicators. In the case of the (complete) system-based approach this can be done when
determining representative indicators for the system orientors.
Policy relevance can also mean that globally the indicator set should assess if policies have a
positive action on sustainability concerns or contrary. This can be attempted by integrating in
the set a number of indicators relative to policies, still fulfilling the above criteria for
No. 010036
Deliverable number:
21 March 2006
Page 33 of 146
individual indicator choice. But, this seems unsuitable to compare different options of any
particular policy. Since comparing the merits of competing policy options within the CAP is a
desired feature in SEAMLESS, we propose that it be integrated in the framework by ensuring
that every indicator representing properties is chosen for its importance in comparing aspects
of the concerned policies. This is possible because a framework based on universal orientors
(properties) is thought to be adaptable for any system, at whatever scale and in whatever
particular purpose. So instead of having a set of indicators in which different issues and
policies are represented by a number of respective indicators, it is possible consider each
policy separately: all indicators representing the properties in the framework are chosen to
assess this particular policy (to compare options of the policy for instance).
4.1.5 EUS (Environmental Utilization Space) and Ecospace
The origin of this framework goes back to works of Horst Siebert in 1982, and application to
SD in (Opschoor, 1992; Opschoor, 1987). Ecospace uses a spatial equity principle in the
usage of resource and their degradation. It claims that all individuals have the same right to
use an equal amount of natural resources and to pollute the global commons. According to
Hans Opschoor, ecospace is a metaphor to capture the notion of limits and the need for
redistribution of access to resources (OECD 1998, Rosenthal workshop). Unlike direct
application of the carrying capacity concept, measures of EUS or Ecospace have to account
for human demand and its evolution. This can be for example, in the definition of "functional
unit" which measures the size of a resource, modified according to the (competing) demands
made on it and the quality required accordingly.
FoE Europe uses this framework with a “physical” focus and proposes some reference levels
for the North industrialised nations to allow the developing south nations to reach acceptable
development standards. Indicators in construction deals with the environmental performance
of input resources with a reference reduction factor of 4 for the greenhouse gases (based on
calculations that greenhouse gas emissions will double over the next fifty years) and
reduction factor of 10 for the consumption of natural resources. This is based on a per capita
consumption of natural resources that is about five times higher in OECD countries than in
developing countries, which mean that a sustainable level of material turnover is only
attainable if OECD countries reduce their resource consumption by a factor of ten (Spapens
and Buitenkamp, 2001).
4.1.6 Land quality, land use and land value analysis
The central resource for agriculture is the soil, this resource is the focus of certain
frameworks. Some of them are based on land quality indicators. As mentioned in Bindraban
and al. (1999) (“Land quality refers to the condition and capacity of land, including its soil,
weather, and biological properties, for purposes of production, conservation, and
environmental management (Pieri and al., 1995) Maintenance of the agricultural production
capacity of land resources is a fundamental element in the discussion on sustainable land use.
Changes in land quality should be monitored to provide early warning of adverse trends and
to identify problem areas. Monitoring land quality and promotion of land management
practices that ensure production and sustainable use of land resources require development of
quantitative Land Quality Indicators (LQI) (Pieri and al., 1995).”
Some Frameworks hence articulate their sustainability evaluation system around indicators
directly related to this resource: some are based on land use, classified in broad categories,
No. 010036
Deliverable number:
21 March 2006
Page 34 of 146
others are indicators of land quality, and the more complex frameworks deal with agricultural
practices or their impacts (Darwin and al., 1996). Just as Sustainable Rural Livelihood (SRL,
see page 61) framework, some of these frameworks enable to deal with rural/urban relation,
particularly in concurrent usage phenomena. They are perhaps the frameworks where the use
of models (agricultural, agro-ecological) in SDIF is the easiest. This should be considered in
relation with WP3.
If we assume that agricultural land use is preferable to industrial or urban use from an
environmental perspective, agricultural surface in activity can be a first level indicator for
SD. This is particularly relevant when agricultural areas are coupled with indicators
classifying agricultural practices by their diverse respect of environment, as implemented in
some vulnerable zones (Bellon and al., 2000). These two types of indicators are pressure
indicators. Conversely, indicators of LQI type, are state indicators susceptible to participate in
ranking agricultural practices relatively to sustainable development.
An interesting framework than can be used with land cover data is based on Ecosystem
Service Product (ESP) concept (Costanza and al., 1997; Sutton and Costanza, 2002). In this
approach, seventeen land cover classes representing the major biomes of the world where
defined using GIS technology (IGBP data with 1 Km2 resolution) and linked to the value, in
terms of ESP, as estimated in Costanza and al. (1997). It seem suitable to deal with the
spatial organization of ecosystems at large scale using a limited class of ecosystem, but some
development would be necessary to take into account diversity of agricultural lands. The ESP
value can be improved by farm level analysis of environmental impact in order to describe
better the links between agricultural practices and functional services of ecosystem connected
to agricultural land.
4.1.7 Material Flow Analysis (MFA), Substance Flow Analysis (SFA) and Life
Cycle Assessment (LCA)
Numerous SDI Frameworks are based on physical accounts of energy or material flows. They
can be included in quantified tools of physical economy approaches and are based on the
concept of physical economy of humankind (Ayres, 1998). As Daniels and Moore (2002)
point out, “Physical economy approaches exist as evidence of the inadequacy or
incompleteness of monetary measures of the parameters of the relationship between the
human economy and its habitat.” Since several decades, the importance of identifying and
tracing physical flows of materials and energy, and not only monetary flows, is well known
(Ayres and Ayres, 1970, Georgescu-Roegen, 1971, Leontief, 1970). Moreover, the growing
concern of environment degradation, and the large reconnaissance of sustainability principles
has since the 1990’s lead to a revival of this kind of quantified techniques.
Daniels and Moore (2002) identify nine techniques – that are briefly described in appendices
3.1, Box 3, page 139 – separating the ones operating on the entire economy or major
economic activity fields on a specified geographical unit and those focusing on specific
goods, services or process regardless of their location. The first group of techniques includes
“total material requirement and output” (TMRO) analyses, the “bulk internal flow” national
MFA (or the IFF model), substance flow analyses (SFAs), physical input-output tables
(PIOT), ecological footprint analysis (EFA), and environmental utilization of space (EUS)
models. The second group of techniques deals with specific goods, services, or processes
includes lifecycle assessment (LCA), materials intensity per unit service (MIPS), and the
sustainable process index (SPI). One can find a metabolic view point in each of the
techniques, in some part comparable to the Ecosystem Health approach. However the vision
No. 010036
Deliverable number:
21 March 2006
Page 35 of 146
is here more static and closely linked to the acceptance of the principle “material balance”
governed by the first law of thermodynamics (Kneese and al., 1975). This vision makes it
important to identify and control physical limits, notably through measure of physical
indicators on which long-term sustainability assessment will be possible. These approaches
also constitute a useful framework for the comprehension of the linkage between economic
demand and activity, and environmental flows.
While extremely coherent with the “material” condition of sustainable development, the
MFA framework, in which other techniques can be reformulated (TMRO, SFA, PIOT), is
disconnected from the others pillars of SD and only deals with the economic/environment
interface. All of these frameworks have to be supported by complementary approaches to
assess economic performance and social well-being. In most of the previously described
frameworks using the capital or SD pillars based approaches, the disconnection between
pillars allows to support such a complementary framework where physical indicators in
economic/environment interface are estimated in a unified MFA framework, and economic
and social comes from less unified and more participatory procedures. Moreover, ecosystem
services and their spatial distribution, which may not be correlated with global substance
flows, may not be sufficiently taken into account within these types of approaches.
The main initiatives using MFA seem to focus on chemical substance and material trade,
except proposition from Eurostat (Eurostat, 2001) and WRI that analyse a large spectre of
material and commodities (Wernick and Irwin, 2005). Daniels (2002) argues that “One major
potential of the MFA-BIF information system as sustainability indicator is evident in the
research into the human appropriation of the net primary production of plants, MFA-BIF
studies are ideally placed to provide detailed biomass flow data for augmenting sustainability
research of this genre”. (see appendices 3.2 for an illustration of the Biomass material flows
in the IFF MFA-BIF model).
For the agricultural sector and sub-sectors, major MFA initiatives on entire economy flow
analysis frameworks integrate agricultural sector mainly through land use (classification of
use). These land-based frameworks seem useful as a basis for applying scenarios with
quantified evolution of land use in order to evaluate environmental pressure, but they have to
be complemented to capture services and value related to the ecosystem quality and their
spatial structure. For Agricultural LCA literature, detailed analysis of substance and process
analysis (see works of Bentrup for example) exists but one of the most important
methodological challenges is related to the mode of spatial aggregation of individual analyses
of production process or agricultural system. In LCA, local and regional impacts are often
assumed to be equal to the sum of the impacts of each farm, using a simple system of
classification (Payraudeau and van der Werf, 2005) with a hypothesis of uniform practices
between farmer (Dalsgaard and al., 2003). Here again the focus on impact scheme does not
allow to access broad sustainability. Halberg and al. (2005) provide an overview of related
techniques on green accounting at farm level.
4.2 Sub-Composite and aggregated SDI approaches
This kind of indices will be treated more in depth within Tasks 2.5 and 2.6 and it will be
presented in a future PD. Development of composite indicators is one of the options retained
from the early start of SEAMLESS, and the main question is how we will develop such
composite indicators. In this section we present the main principles and constraints involved
in the use of framework that are able to handle composite and aggregated indicators,
illustrating the discussion with some existing frameworks and their comparison.
No. 010036
Deliverable number:
21 March 2006
Page 36 of 146
We will distinguish 1) monetary and 2) physical simple aggregated indicator which involve
the transformation of sub-indicators into a single unit from 3) composite indicator which are
aggregated indicator measured in different units. Composite indicators are different in the
sense that they involve rescaling and/or normalisation in order to make its sub-indicators
4.2.1 Monetarised indicators: GSR, GPI and ISEW
Indicators and frameworks such as Index of Sustainable Economic Welfare (ISEW), Genuine
Progress Index (GPI), Genuine Saving Rate (GSR) and indicators within a green GDP
approach, aim to monetarise elements of social and environmental capital which are not
accounted for in GDP with an endeavour to introduce inter and intra-generational
considerations (Lawn, 2003; Neumayer, 2000; Hamilton, 2000; Hamilton and al., 2003).
This type of indicator was designed in response to certain limits of GDP measurements such
- The lack of accounting for income distribution,
- The lack of accounting for non market activities,
- Incorrect accounting for defensive cost,
- No accounting for variation in the value of natural capital.
The idea of the "Genuine Saving Rate" or GSR (World Bank) is that it is important to
consider that there exist other forms of capital (human capital and natural capital) and thus
derive an improved concept of saving. Saving is interpreted in economy as an income transfer
from today to a future period. If the saving is negative, that is if expenses exceed incomes, it
will result in debt. The price of these income transfers from future toward present is the
interest rate. This interpretation of saving applied to saving of natural capital makes it a
measure of sustainability. If the value of genuine saving is negative, then society builds a debt
to future generation. Indeed, it is incorrect, in a sustainable development perspective, to
consider that genuine saving of a country grows when its natural resources are depleted or
degrading to the point where investments in productive capital cannot compensate the losses.
The same is true if knowledge and capacities of the population (human capital) diminishes.
GSR however suppose that human capital does not depreciate, and finally takes into account
the capacities to save natural resources and the social costs of cleaning up of pollution.
ISEW and GPI are indicators conceptually and methodologically are very close to the GSR.
They have been developed, with some variations, in different countries. These indices
combine, with different weights, economic factors (also non market activities related to well
being), social and environmental factors.
These indicators are based on personal expenses to which are added or deduced gains and
losses relative to different consumption. They adopt a monetary approach by evaluating
defensive costs and non-defensive costs. ISEW and GPI make adjustments in order to
account for the impact of intra generational inequalities. We can formulate the main
categories of ISEW as follows:
- Personal expenditure
- Crime and family breakdown
- Household and volunteer work
- Income distribution
- Resource depletion
- Pollution
No. 010036
Deliverable number:
21 March 2006
Page 37 of 146
- Long-term environmental damage
- Changes in leisure time
- Defensive expenditures
- Lifespan of consumer durables and public infrastructure
- Dependence on foreign assets
These frameworks can only imperfectly respond to SEAMLESS project purposes. The
analysis that will be conducted within task 2.6 will essentially aim at identifying the different
economy/environment interfaces which could respectively (i) fit a monetary equivalent frame
and (ii) measure an important part of non-market values (to be determined from existing
meta-analysis in literature on transferability of value).
Such frameworks which entail a monetary unit for a set of used sub indicators or variables
considered only integrate part of the dimensions of SD. They refer to weak sustainability and
usually call for complementary approaches. Although other dimensions are mentioned in
initiatives using this type of framework, they are rarely explicit and operational. The main
reason is that if other dimensions or non-monetary elements were integrated, they would not
fit with the statute of aggregated indicators. Indeed some non-market components of SD are
not well adapted to the monetarisation exercise or are for the moment beyond the scope of
existing evaluation methods. Such a situation does not fit with the SEAMLESS perspective,
at least for 3 among the 4 thematic indicators applied to the agricultural sector (environment,
social, institutional). Nevertheless form an theoretical and methodological perspective this
approach will be given particular attention within task 2.2.3 concerning the application of the
economic pillar at interregional and national scale. A similar conclusion can be applied to the
notion of Ecological Footprint detailed below.
4.2.2 A physical indicator aggregated in area unit: the Ecological Footprint
The ecological footprint (EF) puts environmental sustainability in a spatial perspective. It was
first defined as "the area of ecologically productive land (and water) in various classes that
would be required on a continuous basis to provide all the energy/ material resources
consumed, and to absorb all the wastes discharged by a population with prevailing
technology, wherever on earth the land and water is located" (Wackernagel and Rees, 1996).
EF is the main aggregated indicator using a physical equivalent for all variables considered
(see also the Living Planet Index [LPI] published every two-years by WWF). The EF
proposes an indirect measure, expressed in physical terms, of the society-environment
relation, relying on the carrying capacity concept.
The footprint can be compared with nature’s ability to renew the human consumption of
resources. It groups and calculates material and energy requirements of nations or regions for
a limited number of consumption functions, converts these metabolic flows into the
ecologically productive land area required to produce the resources used in these activities,
and compares the required areas to available regional, national, and global ecologically
productive areas. Existing studies have typically been restricted to the ecological resource
output potential of terrestrial areas (Daniels and Moore, 2002).
This index accounts essentially for some environmental aspects through resource
consumption and neglects numerous variations in the values and services produced by
ecosystems or by biodiversity that can be associated with different modes of consumption and
with different modes of territorial managing. It does not account for any social or institutional
aspect, however without pretending to do so. Its main advantages are:
No. 010036
Deliverable number:
21 March 2006
Page 38 of 146
- Its capacity to encompass the whole production and consumption chain through its
consumed energy entry, thus providing a coherent frame to measure impacts induced
by the consumption of a country or a community on the rest of the world.
- Relevance in comparing environmental advantage between competing productive
- It can be used at different scales.
However, EF should be used with caution (van den Bergh and Verbruggen, 1999; Nijkamp
and al., 2004). Three reasons can be mentioned:
a. EF is incomplete.
In dealing with consumer goods1, land equivalents account for:
- the energy mobilised in the fabric of consumer goods (indirectly through CO2
absorbing forest land mobilised by fossil-fuel equivalent of this energy), from basic
materials and agricultural products.
- the energy mobilised in extracting materials and in agriculture.
- the land used for agriculture.
- the land used for growing wood material.
So availability and depletion of non-renewable resources is not addressed, and this distorts
trade-offs. Then, ecosystem services are only present through CO2 absorption related to
energy use: other absorption of pollution is absent. This is why agricultural practices are all
considered sustainable: so intensive fertiliser use can ameliorate the EF because it mobilises
less agricultural land for a given production. This however means that negative effects on
land or water systems involved in absorbing these fertilisers are not accounted for.
b. The EF aggregation scheme is questionable.
All the land mobilised for energy use, agriculture or urbanisation is added on an equal
weighting basis. Improving this approach would mean to start using some kind of
equivalence theories such as monetary evaluation of relative land value, which is what the
principle of land equivalent and physical unit in general is trying to escape from. Then, the
calculation relies on "one land-one use" basis to avoid double counting: this does not relate
the variations in how land plays different roles at the same time, and therefore with the issue
of multifunctionality.
Another restriction concerns the choice of the area units analysed. National boundaries, for
instance, do not have a clear environmental meaning. For instance, comparing (i) a large,
scarcely populated country with a high level of consumption and an ecological surplus with
(ii) a small, densely populated country with, say, an equivalent an equal level of consumption
but a significant ecological deficit2, does not immediately offer relevant information about
which country is on the right track towards sustainability. In addition, at any level apart from
1 Wackernagel and Rees (1996) divide consumption in 5 main categories: food, services,
transportation, consumer goods and housing.
2 Ecological deficit is an indicator of dependence on out-of-boundary ecosystems. It gives an idea of
the extent to which a country, region or city is dependent on extra-territorial productive capacity
through trade or appropriated natural flows.
No. 010036
Deliverable number:
21 March 2006
Page 39 of 146
the global, the EF should be seen as the net input of virtual land from outside the analysed
system. The EF is essentially more a meaningful ecological dependency indicator for a given
area, which is scale dependent. The geographical scale inherent in the EF calculation is,
therefore, the Achille's heel of the methodology (from Nijkamp and al., 2004).
c. The interpretations of Ecological Footprint are not straightforward.
First the shortcomings described above imply that EF may have to be used partially or
completed following particular goals. Moreover the increase in EF caused by transport seems
to plead for self-reliance more than exchanges among entities. This has to be considered
against other advantages, that involve for instance socio-cultural issues exceeding the scope
of the EF. However, such socio-cultural issues may indirectly benefit in monitoring EF
through increased co-operation among relevant entities, since we are not in closed economies
and physical or monetary exchanges occur. An unbalanced EF between cities and rural zones
in a country may reveal societal compromises (when general interest and redistribution is
ensured by government). Conversely if general interest is not ensured, unbalanced EF can
reveal injustice in appropriation of natural resources and services. So the EF can be a policy
relevant tool because it points at balances that have to be addressed (but through cautious
interpretation), allowing to compare certain environmental advantages of competing
productive policies.
The use of a framework adapted from the ecological footprint in an agri-environmental
setting would call for important theoretical and methodological developments. Its present use
can only be justified as a partial contribution within thematic environmental indicators. The
EF indeed represents a powerful prospective tool to account for energy consumption
associated to different options of agricultural policies (consumption induced for production of
inputs, consumption induced in transporting finished goods) in Europe, but also for
consequences to the rest of the world (induced reorganization of agricultural productions and
of natural resource use in other countries). In this sense, it is a tool which is susceptible, from
an environmental point of view, to justify the subsidizing of short distribution networks as
well as the promotion and maintenance of a diversified domestic production.
4.2.3 Two composite indicators: IUCN and 2005ESI initiatives A participatory composite indicator by IUCN (International Union for the
Conservation of Nature and Natural Resources)
Definition: two-subsystems (human and ecosystem) and 2x5 compartments, leading
to indicators assessment
IUCN has developed a comprehensive tool for assessing sustainability of human related
systems, at various scales and with various scopes (IUCN, 2001).
This tool comprises a plan for articulating successive stages of the sustainability assessment,
with guidance on participative involvment when relevant. The first steps of the method are
designed to make explicit the scope and purpose of the particular initiative, its assumptions
and how these purposes may affect the participative elaboration of indicators in the next
steps. IUCN (2001) presents a comprehensive tool, but partial use of the method is possible
depending of the purpose of the use. For instance one purpose can be to evaluate the
evolution of a situation after a previous assessment
No. 010036
Deliverable number:
21 March 2006
Page 40 of 146
In general, the sustainability assessment is not designed to replace project planning or
monitoring, but to structure the material and data needed for informed decisions, because the
scope, spatial area and time span of the assessment are often larger than that assessment
which is necessary for project planning.
This framework divide sustainability into ten compartments, five for the human system
studied and five for the ecosystem forming its environment:
health & population,
wealth (national economy & household economy),
knowledge & culture,
community (peace & order and freedom & governance),
equity (gender equity and household & ethnic equity),
land (quality & diversity),
water (inland waters & sea),
air (global & local atmosphere),
species & populations diversity,
resource use (energy & materials and resource sectors).
These ten compartments are supposed to cover most of the human related sustainability issues
that can be encountered. A composite indicator for each contextually relevant dimension is to
be derived through local participative process, ensuring concerned people design their own
vision of sustainability, provided the number of dimensions between people and ecosystem
are equal. This approach aims to ensures equal treatment of people and ecosystem well being.
The composite indicators representing each dimension are then aggregated on an equal
weighting basis, first into a human well-being index (HWI) using the five human-related
indices and a ecosystem well-being index (EWI) representing the five ecosystem
compartments. This translates IUCN's view that the two subsystems are not substitutable.
Under this condition, the HWI and EWI can then be aggregated or put into a ratio.
The interactions between these two subsystems are considered within the subsystem where
the impacts are felt: human stresses (pollution...) and benefits (conservation) on the
ecosystem are recorded under 'ecosystem' and conversely.
Indicators are judged following the four criteria: measurability, representativity, reliability,
feasibility. Indicators measuring directly the state of an element are considered more reliable
than indicators of pressure on this element or indicators of societal response to perceived
problems derived from the element. State indicators are thus privileged in the process.
Comments on the assumptions supporting IUCN's initiative.
First, equal treatment of people and ecosystem places the framework within an environment-
society interface. The economy is considered explicitly within the people subsystem, but
weighting one tenth in sustainability issues since it is one of the 10 compartments; for the rest
it only has an indirect role through implicit participation in other issues and relations. So this
framework may not be sufficient to relate economic functioning if it is considered essential to
SD. On the other hand, its relative emphasis on social concerns balances many other
frameworks and can help show the importance of the relation between 'social' condition and
environmental degradation or concern.
No. 010036
Deliverable number:
21 March 2006
Page 41 of 146
The separated assessment of human system and ecosystem shows IUCN's non-compensatory
approach relying on strong sustainability (which is uncommon in composite indicator
initiatives), however the usual (non)substitutability pivot between natural capital and
technology is not emphasised because economy has little place.
Second, the manner in which the relations between people and environment are accounted for
is not very clear (as is often the case it is left to indicator fabrication to highlight these links).
For instance, human economic activity produces pollution, which has an impact on
environment; further, degraded environment has an impact on human health. Since relations
are considered within the subsystem where impacts are felt, both phenomena are recorded
separately, whereas in a PSR-type chain the causality relation is put forward from the
economic activity to the health impact. Even though such a causality is a simplification, the
correlation should be presented. Assessing "where the impact is felt" to record phenomena is
implicitly assessing for pressure-state relations, using state indicators to reliably assess a
causing phenomena and an effect phenomena. This may be more reliable, but hardly more
More over, one way of knowing if a response indicator is reliable is to measure it and face it
(with the prudence commanded by uncertain interrelations) to the evolution of the state of the
targeted phenomena. As far as policy relevance is concerned, state indicators give a more
reliable picture of the situation, and hence enable to derive new solutions from new insights.
But response indicators combined with state indicators tell whether policies are implemented
and if they are sound and sufficient. IUCN's preference for state indicators is however
coherent with what the organization expects from a sustainability assessment of policy: to
provide a solid and balanced information and rationale to inform decision making.
An essential feature of this method is the idea that developing a vision of sustainability is
participative and context-specific, so that priorities are derived by those who are concerned
and respect local feasibility: this reflexive procedure and responsibility appropriation is
considered as important as the indicators actually produced.
The relative importance given to societal concerns by the people-ecosystem framework of
IUCN is particularly consistent with the fact that the rest of the assessment is locally involved
and that a thorough participative guide is provided. Indeed, social and quality of life concerns
are often context-specific and the demand for such indicators is expressed in numerous
surveys. A composite indicator dealing with protection and management of
environment: the 2005 Environmental Sustainability Index (2005ESI)
The 2005 Environmental Sustainability Index (2005ESI) lead by Yale University is a
composite indicator dealing with protection and management of environmental resources and
stresses (Esty and al., 2005).
While environmental sustainability refers to the long term maintenance of valued
environmental resources in an evolving human context, the 2005ESI's emphasis is policy-
oriented and focuses on a shorter term period: it provides a gauge of a society's natural
resource endowments and environmental history, pollution stocks and flows, and resource
extraction rates as well as institutional mechanisms and abilities to change future pollution
and resource use trajectories.
It only deals with economic and social issues insofar as they relate to their environmental
sustainability objective, through the following logic. The 2005ESI framework is divided into
five core components (or thematic categories):
No. 010036
Deliverable number:
21 March 2006
Page 42 of 146
- Environmental Systems (are they healthy, deteriorating or ameliorating?)
- Reducing Environmental Stresses
- Reducing Human Vulnerability (food, health, disasters)
- Social and Institutional capacity (institutions and underlying patterns of skills,