POLICY INTEGRATION FOR COMPLEX POLICY PROBLEMS:

WHAT, WHY AND HOW

by

Helen Briassoulis, professor

Department of Geography, University of the Aegean
University Hill, Mytilini, Lesvos 81100, Greece
e-mail address: [email protected]

Abstract

Contemporary policy problems concern complex, cross-cutting socio-environmental

issues that, combined with changing environmental, socio-economic and political
conditions and the quest for sustainable development, place new demands on policy
making. The principal problem is that the policy market does not provide a
satisfactory supply of arrangements to meet this demand. Policies are often found to
overlap or be in conflict and the policy system is unduly complicated, producing
inefficient or even ineffective solutions and generating new problems. Policy
integration (PI) comes as a potentially suitable answer to address this institutional
misfit and facilitate the transition to sustainable development. The paper negotiates
selected conceptual, theoretical, methodological and practical issues concerning
policy integration with a special focus on complex policy problems, occasionally
using the combat against desertification in Mediterranean Europe as an illustrative
example. Although the discussion refers primarily to European Union policies, it is
relevant to other spatial/organizational levels as well.

The first section of the paper discusses the distinguishing characteristics of

contemporary policy problems, drawing on complexity and institutionalist thinking,
sketches briefly the changing context of policy making and the quest for sustainable
development, and offers a brief account of desertification and the associated policy
needs. The second section, drawing on the pertinent literature, attempts a
comprehensive conceptual exploration of the notion of policy integration, proposes a
particular conceptualization of PI and argues for the need for PI to handle complex
policy problems especially at higher spatial/organizational levels such as the EU. The
third section suggests a methodological framework to analyze PI holistically as well
as to serve as a basis for the development of policy integration schemes. The last
section offers preliminary ideas for the design of PI schemes and suggests future
research directions.

ACKNOWLEDGMENTS

Research reported in this paper was undertaken in the context of the EC-DG XII funded
research project MEDACTION (Policies for land use to combat desertification; Contract No.:
ENVK2-CT-2000-00085). A more extensive treatment of the issues tackled in this paper is
offered in Chapters 1 and 2 of the book: “Policy Integration for Complex Environmental
Problems; The Example of Mediterranean Desertification”, H. Briassoulis (editor), Ashgate
Publishing (forthcoming, May 2005).

Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
1
POLICY INTEGRATION FOR COMPLEX POLICY PROBLEMS:
WHAT, WHY AND HOW

1. INTRODUCTION – THE QUEST FOR POLICY INTEGRATION

Contemporary policy problems concern complex, cross-cutting, socio-environmental

issues that, combined with changing environmental, socio-economic and political
conditions and the quest for sustainable development, place new demands on policy
making. More often than not policies are single-purpose concerning directly or
indirectly particular facets of these problems only. Thus, a single sectoral policy
cannot address the problem as a whole. Moreover, policies are often found to be little
coordinated, to overlap or even to be in conflict. The policy system is unduly
complicated, producing inefficient or even ineffective solutions, giving rise to new
problems and waste of resources. After several decades of policy-making experience,
it became evident that sectoralized, uni-dimensional, uni-disciplinary and
uncoordinated policies do not serve well the cause of sustainable development. The
principal problem is that the policy market does not provide a satisfactory supply of
arrangements to meet the demand for policy making generated by complex problems.
Policy integration (PI) comes as a potentially suitable answer to fill this institutional
void and facilitate the transition to sustainable development. The paper examines the
‘what’, ‘why’ and ‘how’ of policy integration with a special focus on complex policy
problems. The combat against desertification in Mediterranean Europe is used
occasionally as an illustrative example. Although the discussion refers primarily to
European Union policies, it is relevant to other spatial/organizational levels as well.

The rest of this section sketches briefly the changing context of policy making and the
quest for sustainable development that generates the need for PI, presents the
distinguishing characteristics of contemporary socio-environmental problems,
drawing on complexity and institutionalist thinking, and offers a brief account of
desertification and of the associated policy needs. The second section explores the
‘what’ and ‘why’ of PI; it provides a conceptual examination of the notion of PI,
proposes a particular conceptualization of PI and argues for its suitability in handling
complex policy problems especially at higher spatial/organizational levels such as the
EU. The third section addresses the ‘how’ of PI; it suggests a methodological
framework to analyze PI comprehensively and to serve as a basis for the development
of policy integration schemes. The last section offers preliminary ideas for the design
of PI schemes and suggests future research directions.

1.1. The changing context of policy making and the quest for sustainable
development

The complexity of contemporary socio-environmental problems, the quest for

sustainable development, globalization, the Europeanization of policies, the growth of
multi-level governance, the requirements of planning for the needs of client groups
and affected regions generate demand for more encompassing, multi-level, spatially
and temporally integrated policy approaches. As the next section further details,
contemporary policy problems are complex, resulting from the intricate interplay of
biophysical and human driving forces over time, cutting across ecological, social,
economic, administrative and political boundaries and transgressing the functional
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
2
specialization of most current political and administrative systems (Peters, 1998;
Robert et al., 2001; Shannon, 2002). Their systemic nature renders causal
relationships indeterminate and direct, single-purpose policy interventions ineffective.

In the quest for sustainable development, integration is a central concept demanding

the consideration of all, and not only the environmental, aspects of policy problems
and of the web of relationships among environmental media, actors and policies, to
secure the long-term, coherent functioning of human-environment systems.

Socio-economic globalization has dissolved and blurred political and institutional

boundaries, has created or intensified multiple and multi-level interdependencies
among people, environmental media, social and economic issues, and policies and has
changed the spatio-temporal scale at which problems can be effectively addressed.
National and subnational governments are rarely able to tackle alone and provide
effective responses to these problems within a reasonable time frame (OECD, 1996b;
Robert et al., 2001; Hajer, 2003). Moreover, international policies and membership in
supra-national organizations press governments to present a coherent policy picture to
the external world (OECD, 1996a; Lewanski, 2002).

In the EU, the process of European integration and the Europeanization of several
policy areas have spurred decentralization and devolution of decision making power
to lower levels of government (Liefferink and Jordan, 2002) thus augmenting the
institutional complexity of contemporary policy problems as well as the territorial and
temporal reach of policy interventions. The state-centric, hierarchical model of
governance is being replaced by multi-level forms of governance (Marks et al., 1996;
Hooghe and Marks, 2001) where decision making authority and influence are shared
across multiple levels. Solutions to problems are devised and implemented in the
context of polycentric, highly complex and interdependent networks of formal and
informal actors, procedures and instruments, increasing accordingly the need for co-
ordination, development of horizontal cooperative structures, and participatory and
joint development and use of public intervention instruments (Hajer, 2003).

Finally, resources utilized in regional and sub-regional planning to address the needs
of specific client groups and lagging regions or to promote regional sustainable
development more generally originate in numerous policies and administering
organizations1. This requires that policies be organized not on the basis of function, as
it has been historically the case, but on the basis of client groups or of particular
regions (Peters, 1998; Hakkinen, 1999; Gibbs et al., 2003; Moss, 2004).

On the supply side, the nature of public policy making and the current performance of
the institutional and administrative apparatuses in the context of multi-level
governance do not appear to serve satisfactorily the demand identified before. For one
thing, policies are ‘moving targets’ (Wittrock and de Leon, 1986). They are not
discrete, disembodied events occurring in isolation from one another. On the contrary,
policy decisions and their impacts are interlocked making it increasingly difficult for
one policy area to function independently of other areas (Greenberg et al., 1977).

1
Such as regulations, economic instruments, and financial incentives.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
3
Public policymaking follows diverse styles. Policy types differ from one area to the
other2 (Richardson, 1982; Vogel, 1986; Knill and Lenschow, 2000; Moss, 2003).
Historically, relatively autonomous policy sectors have dealt with policy issues,
leading to sectoral/functional specialization and vertical organization of
administration at both the EU and the national level (Avery, 2001; Robert et al., 2001;
Hertin and Berkhout, 2003; Zahariadis, 2003). The result is a well-documented,
general lack of coherence, coordination and cooperation among policies, generating
costs and inefficiencies, taxing limited government budgets and detracting from the
achievement of sustainable development (OECD, 1996a; O’ Riordan and Voisey,
1998; Peters, 1998; Persson, 2002; Shannon, 2002).

The implementation of several EU policies has uncovered problems of cooperation at

the level of the EU and of the member states (MS) (Robert et al., 2001). These
include divergent political objectives and interests, lack of collaboration among the
Directorates-General of the European Commission, lack of clear position of the
Commission, different political prospects among the Community and national and
sub-national actors, high degree of policy and administrative sectoralization and
centralization in the MS, weak administrative coordination and low degree of
consultation of sub-national authorities. Policy implementation has revealed also that
asymmetric integration is ineffective; i.e. one policy incorporates concerns of another
but this is not always met by similar moves in the other policy. In sum, PI is needed to
hold the policy system together and to manage the numerous policy interconnections
so that policy supply meets demand successfully and effectively.

1.2. Complex socio-environmental problems and their policy implications

The diffusion of Complexity theory in the Natural, the Social and the Policy Sciences
since the 1980s has enriched greatly the study of contemporary socio-environmental
problems. A voluminous literature now documents the complexity of natural, social
systems and, more importantly, of human-environment systems, shedding light on
suitable policy making approaches to address them (Holling, 1986; Dryzek, 1987;
Waldrop, 1992; Berkes and Folke, 1998; Byrne, 1998; Marion, 1999; Levin, 1999;
Science, 1999; True et al., 1999; Zahariadis, 1999; Gunderson and Holling, 2002).
The dominant Newtonian-Cartesian, linear, equilibrium-centered view of nature and
society, that dissociates the environment from people, policies and politics, has not
been found to fit the evidence. Instead, complexity-informed, non-reductionist, co-
evolutionary, interdisciplinary, historical, and comparative systems approaches,
adopting integrative modes of inquiry and methodological pluralism, are more
suitable for studying these problems. The distinguishing characteristics of complex
systems that help comprehend and negotiate the complexity of socio-environmental
problems and its policy implications are briefly presented below.

“Complex Adaptive Systems” (CAS) is the term used to refer comprehensively to

systems that are constantly adapting to their environment while maintaining their
structure and coherence (Batty and Torrens, 2001; Holland, 1995; Janssen et al., 2000;
Finnigan, 2003; Gunderson and Holling, 2002; Holling, 2001).

2
Some being regulatory, others market-oriented, still others voluntary-sector oriented.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
4
CAS are open systems whose boundaries are difficult to delineate precisely. They
comprise components, semi-autonomous, individual, self-interested agents3,
hierarchically nested (embedded) within larger aggregate systems4 (Berkes and Folke,
1998; Gibson et al., 1998; Levin, 1999; Gunderson and Holling, 2002). Agents, such
as, flora and fauna in ecosystems, individuals and organizations in social systems,
have different characteristics. Human agents5, in particular, have different socio-
economic features, viewpoints, preferences, goals, aspirations, emotions, possess
different amounts of resources6 (Detombe, 2001) and may belong to more than one
hierarchies (Manson, 2001) and groups. Thus, CAS possess an inherent variety and
flexibility that allow them to be creative, respond fast to unforeseeable events and to
try multiple options simultaneously to survive7 because functions and control are
decentralized, system components are linked in parallel through numerous, non-linear
feedbacks and their multiple memberships may generate redundancy and
contradictions but increase the strength and resilience of CAS to external and internal
shocks.

The connectivity of CAS, the particular ways in which agents connect and relate to
one another, is critical to their evolution and survival. The properties of CAS are
explained by an understanding of the relationships among their parts rather than by an
understanding of these parts separately (Gallagher and Appenzeller, 1999; Manson,
2001; Limburg et al., 2002). The non-linear nature of these relationships distinguishes
complex from simple, linear systems and defines their internal structure, behaviour
and mode of change (Manson, 2001).

Linear systems evolve smoothly and continuously towards a single equilibrium state.
Following a disturbance, negative feedback mechanisms bring linear systems back to
their initial equilibrium. On the contrary, non-linear systems possess multiple
equilibria, changing and co-evolving with their environment through self-organization
and adaptation to changing external conditions and “shocks”, a process based on
learning8 (Lee, 1993; Berkes and Folke, 1998; Manson, 2001; Wilson, 2002).
Following a large, external event or minor, random, and sometimes overtly
insignificant, changes, the transition to another state may be abrupt and discontinuous
as CAS spontaneously re-arrange, reorganize, redistribute and restructure their
components in different patterns, behaviour and structure to better interact with their
environment. This happens because mutually reinforcing, positive feedback
mechanisms amplify changes bringing the system to a new equilibrium state (Manson,
2001; Holling 1986, 2001)9.

Consequently, CAS exhibit Sensitive Dependence on Initial Conditions (SDIC)

(Glasner and Weiss, 1993); i.e. historical events influence importantly selection

3
Seeking to maximize some measure of goodness, or fitness, by evolving over time (Dooley, 1997).
4
Their characteristic hierarchical organization differs from top-down, serial, command-and-control
authoritative structures (Levin, 1999; Limburg et al., 2002).
5
Such as individuals, households, public and private formal and informal organizations, and resources
6
Money, know-how, power, etc.
7
For example, polycentric governance systems have proven less vulnerable to unexpected
contingencies than rigidly organized, centralized systems (Ostrom, 1990, 1998).
8
Through continuous exchanges of information with their environment (feedbacks), the system’s
agents are able to anticipate the results of their actions as well as to adapt to changing conditions.
9
The popularly known ‘butterfly effect’ (Gleick, 1994) signifies major changes precipitated by
amplification of an original minor disturbance.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
5
among the multiple equilibria of a system (Holling 1978; Manson, 2001). Increasing
returns in economics (Arthur, 1989) explain several economic and spatial
phenomena10 and the path dependence of socio-economic and spatial development
(Anderson et al., 1988; Arthur et al., 1997), highlighting the contingent nature of their
evolution and the influence of the institutional settings within which they are
embedded (Liebowitz and Margolis, 1995; Henderson, 2001; Berkhout, 2002). Some
systems may be ‘locked in’ a particular state where change is irreversible (Arthur,
1989; Liebowitz and Margolis, 1995)11.

The evolution of CAS is characterized by the co-existence and complementarity of

long periods of relative stasis (phases), where change is gradual (slow processes) and
the system functions smoothly, punctuated by bursts of evolutionary change (phase
shifts), discontinuities and rapid transformations (fast processes) triggered by either
minor or major disturbances (Berkes and Folke, 1998; Wollin, 1999; Holling, 2001;
Gunderson and Holling, 2002). CAS are in a state of tenuous equilibrium, at the ‘edge
of chaos’ (Langton 1990), tending to collapse into a rapidly changing state of dynamic
evolution12.

A much-celebrated feature of CAS is emergence. Instead of being planned and pre-

determined, order and control emerge from the bottom up as local interactions, based
on simple rules, among individual agents produce over time regular patterns that
feedback on the system, informing agents’ behaviour (Dooley, 1997; Levin, 1999).
Emergent qualities are system-wide characteristics, not features of individual
components. They are a function of synergism, not superposition, among system’s
parts at a particular level, rendering CAS unpredictable and difficult to control (Baas
and Emmeche, 1997; Manson 2001).

Contemporary socio-environmental problems are biophysically, socio-economically

and institutionally complex. They involve numerous, diverse and multifarious actors
and resources interacting, through various formal and informal institutions, over and
across different spatial and temporal levels in both linear and non-linear fashion. The
great number of interactions and differentiation among actors and resources reduce
the chances of understanding one another’s functions and increase the incidence of
problems (Zahariadis, 2003). Problem definitions and the associated goals are
contingent and contextual, a function of the initial system conditions (of the actors
that defined the problem and the state of the system). These evolve in the course of
problem-solving as actors, resources and their relationships change. Hence, these
problems are often ill- or multi-defined, definitions and goals may be conflicting, and
causation difficult to establish. It is uncertain when, why, and where the problems
started, as well as where they lead and when they will end. Consequently, these
problems are unpredictable and hard to analyze and handle with ready-made solutions
(DeTombe, 2001). Problem solving is a continuous and fluid process. Seldom are

10
Such as technical standards (the "QWERTY" standard typewriter and computer keyboard), urban
sprawl and urban concentrations, clustering of economic activities, and so on.
11
For example, overgrazing or climate change can push (‘lock’) vegetated systems into a new stability
domain (desertification) that is reinforced by feedback loops that maintain high temperatures and low
water and nutrients (Limburg et al., 2002).
12
Their property of self-organized criticality means that they reach an equilibrium state that is not
stable but which is the most productive and creative, leading to new possibilities.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
6
these problems ‘solved’; at best, they are ‘resolved’ (Patton and Sawicki, 1986) for
their ‘owners’ and over a finite time period.

The implications of socio-environmental problem complexity for policy making and

management are crucial (Dryzek, 1987; Wiman, 1991; Wollin, 1999; Zahariadis,
2003). Resource-related public policies are collective choice institutions mediating
“the relationship between a social group and the life-support ecosystem on which it
depends” (Berkes and Folke 1998, 9). The policy market supplies multifarious and
largely uncoordinated policies. Most policy systems, at all levels, are complicated
rather than complex, comprising policies built on the linear-world assumption13 that
ignore system complexity and variously reduce it. Policies have generally
unpredictable, uncertain, contextual and contingent impacts at various
spatial/organizational levels, frequently, giving rise to undesirable ‘surprises’ in both
the short and the long run or proving to be perverse14. At times, policies displace,
rather than ameliorate, problems across time, space and medium (Brown, 2000).
Moreover, vertical administrative organization and compartmentalization and
institutional fragmentation do not permit policy systems to cope effectively with
policy externalities15. Hence, they are not fit and able to come to terms with the
complexity of contemporary socio-environmental problems.

Responsive and effective policy making necessitates an understanding of the inherent

uncertainty of policy problems and that “there is no single, universally accepted way
of formulating the linkage between social systems and natural systems” (Berkes and
Folke 1998, 9). Solutions to problems cannot be imposed; rather, they emerge from
the interactions among system components following the rules set. Policy integration,
capitalizing on the flexibility of complex human-environment systems (Zahariadis,
2003), appears to be another avenue worth-exploring to manage problem complexity
and provide for better institutional fit, ‘adding value’ to policies, while economizing
on resources (Sanderson, 2000) as it will be discussed later.

1.3. Desertification and the associated policy needs

Desertification is a contentious issue (Thomas 1997) and a ‘wicked’ policy problem.

The 1994 United Nations Convention to Combat Desertification defines it as “land
degradation in arid, semiarid and subhumid tropics caused by a combination of
climatic factors and human activities” (UNCCD, 1994). Land degradation means
reduction or loss of the biological or economic productivity and complexity of rainfed
cropland, irrigated cropland, or range, pasture, forest and woodlands resulting from
land uses or from a process or combination of processes, including those arising from
human activities and habitation patterns, such as: (a) soil erosion caused by wind
and/or water, (b) deterioration of the physical, chemical, biological and economic
properties of soils and (c) long-term loss of natural vegetation (UNCCD, 1994).

13
Contemporary policy approaches gradually abandon the Newtonian-Cartesian, linear worldview and
espouse the complex, non-linear world model. The adoption of the precautionary principle, of learning-
based and strategic approaches and of the adaptive management paradigm, all signify the influence of
complexity-thinking in policy making (Wiman, 1991; Lee, 1993; Healey, 1997; Brown, 2000; Ascher,
2001; White, 2001).
14
I.e. generating environmental costs but not economic benefits (Berkes and Folke, 1998; Holling et
al,, 1998; Ascher, 2001).
15
I.e. changes in one policy area produced by changes in another area.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
7
Desertification is an emergent phenomenon as it is a higher-level feature of the land
system, the cumulative outcome of numerous individual, local level inappropriate
land management practices induced by the intricate interplay of multi-scale
biophysical and societal forces that are mediated by a considerable number of formal
and informal nature-society institutions. Its occurrence is path-dependent, sensitive to
the initial conditions prevailing in particular spatio-temporal contexts. Once
unsustainable conditions set in an area, positive feedback mechanisms may intensify
degradation leading the land system to other equilibrium states. It is, thus, difficult to
disentangle its multi-scale biophysical and societal causes, predict its consequences
and reversibility and know with certainty whether and when an area will be ‘locked’
in an irreversibly desertified state.

As a result of this complexity, desertification is a socio-culturally defined and

determined construct. Considerable uncertainty surrounds it, leading to controversies
over its definition, relative importance of its anthropogenic causes16, assessment of
land affected or at risk, reversibility and importance of its spatio-temporally variable
impacts (Reynolds and Stafford-Smith, 2002b). The meaning and interpretation of its
determinants and impacts have influenced the collective sense of urgency to combat
it. This explains the preference for particular abatement approaches, the slow progress
towards developing integrated approaches, the relative inaction and the partial,
fragmented, and uncoordinated efforts to address it. Policy making to combat
desertification is a complex enterprise (Briassoulis, 2004a). It requires continuous top
down-bottom up communication and coordination of policies and local level
interventions to alleviate the larger problem.

2. POLICY INTEGRATION: WHAT AND WHY

Although interest in policy integration has a long history in several quarters, the
recent renaissance of the subject is associated mainly with the environmental
repercussions of economic activities that are not properly accounted for (if at all) by
sectoral policies; hence, the proliferation of policy activity and research on
environmental policy integration (EPI). In the European Union, Article 6 of the 1999
Amsterdam Treaty gave EPI its current political significance, constituting the official
statement for the integration of environmental concerns in sectoral policies:
“Environmental protection requirements must be integrated into the definition and
implementation of the Community policies and activities referred to in Article 3, in
particular with a view to promoting sustainable development” (Lenschow, 2002: 14).
Various policy developments aim to provide the necessary procedures to materialize
the injunction of Article 6. The focus on EPI has overshadowed, however, other
concerns. For example, in addition to environmental repercussions, sectoral policies
have spatial, social, cultural and other interdependent repercussions. Moreover,

16
Controversy centers on its causes and consequences. The problem is two-fold: “(a) whereas
desertification is most often attributed to a myriad of human activities, … it may be triggered or
exacerbated by climate variability, … so that the causes are not necessarily solely anthropogenic (at
least at the local land use level), (b) not all such ecological, biogeochemical and hydrological changes
have an immediate or direct economic impact on human activities” (Reynolds and Stafford-Smith,
2002b, 3)

Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
8
neither environmental policies account adequately for their economic and social
impacts nor social policies account for their environmental and economic impacts.

The discourse on policy integration (PI) has met with the common problem of
confusion, loose (or no) definitions, differences in the meaning and multiple
interpretations (and mis-interpretations) of ‘policy’, ‘integration’ and ‘policy
integration’ in various policy contexts that lead to different operational expressions,
proposed designs, and so on. Proper and consistent analysis and design require,
however, clear definitions of ‘integration of what, by whom, where, when, why and
how’. Only then it can be judged how well PI facilitates the resolution of problems
and contributes to sustainable development.

2.1. Conceptual exploration of the notion of Policy Integration

The meaning of EPI and PI depend on how ‘policy’ and ‘integration’ are
conceptualized. According to Merriam-Webster’s Collegiate Dictionary, “integrate”
can mean either “to form, coordinate, or blend into a functioning or unified whole” or
“to unite with something else” or “to incorporate into a larger unit” (cited in
Persson, 2002, 9)17. Figure 118 schematically depicts the fine difference between
“blending into a unified whole” or “uniting with something else” and “incorporating
into a larger unit”.

In discussing PI for sustainable development, Thomas (2003) suggests four

conceptions of the term ‘integration’ – integration as efficiency, as mindfulness, as
institutional coordination and as ‘compatibility-within-a-framework’. Integration as
‘compatibility-within-a-framework’ aims at striking a compatible relationship among
various sectoral goals within an overarching framework, allowing for compromises
among them. One step further, ‘goal integration’ involves the integration of
environmental, economic and social dimensions in ways that secure simultaneous
realization of the respective goals in a single policy, programme or project
intervention (Thomas, 2003). This ideal integrationist conception assumes that “all
apparent irreconcilability can be overcome through attaining an underlying unity of
purpose” (Thomas 2003, 203).

Turning to the meaning of ‘policy’, public policies are defined as purposeful courses
of action, comprising a long series of more-or-less related activities, which
governments pursue to reach goals and objectives related to a problem or matter of
concern and to produce certain results (Friedrich, 1963; Lowi, 1964; Anderson, 1984;
Pressman and Wildavski, 1992). A policy is not a single, discrete, unitary,
disembodied phenomenon, but a series of decisions. It concerns what is actually done
(or not done) as opposed to what is proposed or intended, which is the case of
decisions; policy implementation and enforcement complete the actual policy process.
The main constituent elements of a policy are its object (the characteristics of the
problem considered and the theory about it), interested and/or involved actors, their
goals (reflecting their value systems), the available structures and procedures (for
formulation and implementation), and the instruments used to achieve the goals set.

17
The Oxford Combined Dictionary of Current English & Modern English Usage (1982, 129) leans
towards the latter interpretation (“To integrate is to combine components into a single congruous
whole”).
18
Inspired from Persson (2002).
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
9
Figure 1. The difference between unifying and incorporating parts

PI, then, can be conceptualized as a process either of coordinating and blending

policies into a unified whole, or of incorporating concerns of one policy into another.
The literature contains several definitions of PI. Without an adjective, no priority
among the objectives of PI is assumed; the task is delegated to the political
(democratic) process. ‘Environmental’, or for the same purpose, ‘social’ or
‘economic’, denotes a particular point of view and priority in integrating policies19.
Since the 1980s, most definitions refer to EPI (even when the ‘environmental’ is
omitted). Some define EPI as a process, other emphasize the output of the process, an
integrated policy, while others stress both process and output.

Collier (1994) defines policy integration as aiming at (a) achieving sustainable

development and preventing environmental damage; (b) removing contradictions
between as well as within policies; and (c) realising mutual benefits and the goal of
making policies mutually supportive. She considers that EPI involves different
requirements at different stages of the policy process and suggests that a set of criteria
for determining goal trade-offs is necessary to guide the EPI process.

The OECD (1996a), focusing on the process side, defines EPI as: “Early co-
ordination between sector and environmental objectives, in order to find synergy
between the two or to set priorities for the environment, where necessary.” The
European Environment Agency sees EPI as a process of shifting the focus of
environmental policy away ‘from the environmental problems themselves to their
causes … [and] … from ‘end-of-pipe’ ministries to ‘driving force’ sector ministries’
(EEA, 1998, 283).

For Lafferty and Hovden (2002:15), EPI implies “(a) the incorporation of
environmental objectives into all stages of policy making in non-environmental policy

19
Lafferty and Hovden (2002) offer the integration of economic goals into several policies as an
example.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
10
sectors, with a specific recognition of this goal as a guiding principle for the planning
and execution of a policy, (b) accompanied by an attempt to aggregate presumed
environmental consequences into an overall evaluation of policy, and a commitment
to minimize contradictions between environmental and sectoral policies by giving
principles priority to the former over the latter.”

Other authors (O’ Riordan and Voisey, 1998; Shannon, 2002; Hertin and Berkhout,
2003) emphasize the communicative dimension of EPI; intersectoral PI concerns
certain kinds of cooperative behaviour, forms of institutions, and kinds of
communicative action.

Finally, EEB (2003: 13, 14) offers a more holistic definition: “Environmental Policy
Integration is a long-term process that requires changes in administrative practice and
government culture, institutional adaptation and also specific tools… The integration
of environmental aspects into other policy areas must contribute to policies that
effectively lead us to higher environment protection and greater sustainability.”

To develop operational expressions for EPI and PI and measures to achieve them, it is
necessary to clarify (a) what should be integrated and (b) along which dimensions.
The answers are partly conditioned by the focus on PI as process, output or both, and
by the stage of the policy making process concerned. Two generic approaches to these
intertwined questions are suggested: the vertical/intrasectoral and the
horizontal/intersectoral (Lafferty and Hovden, 2002; Persson, 2002; Hertin and
Berkhout, 2003).

The vertical/ intrasectoral approach, reflecting an incrementalist approach to policy

making, has dominated thinking on PI given the emphasis on EPI. In a rather narrow
sense, the object of PI is to incorporate environmental concerns into a sectoral policy
(that were not included in its original design) through proper procedures to produce an
integrated policy (Lafferty and Hovden, 2002)20. In practice, it is encountered mostly
during agenda (and goal) setting where environmental goals are added to the set of
goals of sectoral policies (e.g. reduce air pollution from transport). Naturally, the
critical question is how one chooses proper means to realize these goals. The literature
leans heavily towards procedural measures and instruments for PI and EPI (Peters,
1998; Persson, 2002; EEB, 2003; Hertin and Berkhout, 2003). Underdal (1980)
suggested the vertical consistency criterion for PI; a policy should be consistent at all
its levels, from policy goals to more detailed guidelines. Although it is not often
explicitly stated, evidently, the result of a vertical/intrasectoral integration process is
an integrated policy (in the sense reflected in its goals).

Vertical PI has a spatial dimension, which is mostly implicit in the literature, with
some exceptions to this author’s knowledge, so far (Buller, 2002; EEB, 2003). The
stages of the policy process span several spatial/organizational levels and different
actors are involved at each stage and level, influencing the ease and success of PI. At
higher levels, PI in terms of goals is easily achieved (Lenschow, 2002; Persson, 2002)
while getting formal and informal policy actors comply with ‘integrating’ procedures
and rules21 at lower levels proves difficult if not infeasible (O’ Riordan and Voisey,

20
Obviously, this may apply to any other kind of concerns such as social equity, gender equality, etc.
21
E.g. implement the EIA requirement.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
11
1998). If linkages among the relevant spatial/organizational levels do not exist or are
not fully operational, the initial PI intentions never materialize on the ground.

The horizontal/intersectoral approach to PI, reflecting a rational comprehensive

approach to policy making, is increasingly recognized as the most appropriate for
effective PI. It concerns relationships among policies with respect to a given issue
(e.g. environmental protection) or to several interlinked issues and may not
necessarily lead to a single integrated policy. Most studies concern the environmental
version of horizontal policy integration (HEPI); i.e. ensuring that the environmental
dimension is integrated in all other policies as a cross-cutting concern. Lafferty and
Hovden (2002) consider that HEPI concerns the extent to which a central authority
has developed a cross-sectoral strategy for EPI. Until now they could not find
evidence of HEPI as it requires negotiation of trade-offs among environmental and
other objectives.

Essential requirements for effective HEPI include horizontal communication,

substantive and procedural cross-sectoral cooperation, collaboration and coordination
and networking between environmental and non-environmental sectors, joint
responsibilities and procedures, sharing of resources and lack of administrative
fragmentation to design solutions to shared (or common) problems (Peters, 1998;
Lenschow, 2002b; Shannon, 2002; EEB, 2003, Hertin and Berkhout, 2003). These
requirements can be extended to the general case of PI to address the demands of
contemporary ecologically, socially, politically, administratively, and legally
crosscutting policy problems.

Horizontal PI has a spatial dimension that is implicit in most of the literature.

Liberatore’s (1997) discussion of the ‘space and time’ dimension requires paying
attention to ‘the interactions between different spaces: the geographical space of the
affected environment, the economic space of the activities that have an impact on the
environment, the institutional space of the relevant authorities and policy instruments,
and the cultural space of values’ (p. 117). She mentions also the ‘organisation’
dimension that concerns the mismatch between the territorial competences of
environmental authorities with the affected environment.

The distinction between a vertical and a horizontal approach to EPI and PI is not
unambiguous or straightforward to make in practice (Persson, 2002). In the
perspective of promoting sustainable development, both approaches should be merged
through appropriate procedures for complete and effective PI extending beyond the
environmental field.

2.2. Proposing and justifying a conceptualization of PI

The preceding discussion suggests that PI is needed to hold the policy system
together, to overcome its tendencies towards disorder, and to manage the numerous
policy interconnections so that policy supply meets policy demand, supporting the
effective resolution of complex problems and the transition to sustainable
development. However, PI should be conceived and analyzed more broadly and
thoroughly along many more dimensions than it is currently the case.

Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
12
It is proposed that PI should be construed as ‘integration of policies’, referring to a
process of sewing together and coordinating various policies, both over (horizontally)
and across (vertically) levels of governance, modifying them appropriately if
necessary, to create an interlocking, hierarchical, loosely-coupled, multi-level, policy
system that functions harmoniously in unity. The output of such an integration
process will be an integrated policy system aiming to achieve multiple
complementarities and synergies among policies. Although a perfectly integrated
policy system may be a utopian ideal, the more policies ‘talk to one another’ and the
right hand know what the left does, the more satisfactory will be the response of
policymaking to the demands of contemporary problems.

The proposed interpretation of PI is more appropriate, compared to the narrow and

still vague notion of EPI, for several reasons. First, given the inevitably multifarious
and departmentalized nature of policy making, integration of the multitudinous
policies better supports the transition to sustainable development. The environment-
biased assumption of EPI that the introduction of environmental considerations in
sectoral policies will contribute to the achievement of sustainable development is not
warranted because it prioritizes the environmental dimension of sustainable
development while the broader conceptualization of sustainable development asks for
a balance among its economic, environmental and social dimensions. Moreover, this
assumption ignores the indubitable linkages that exist between the environmental and
the social as well as between the social and the economic facets of policy problems.

Second, the resolution of contemporary, multidimensional and cross-cutting socio-

environmental problems is rarely possible through single-purpose, uni-dimensional,
and uncoordinated policies or of a super policy that integrates all relevant problem
dimensions. The complexity of these problems introduces important non-linearities in
the operation of both the human-environment and the policy system, as it was
discussed before. Integration of policies may eliminate redundancies (actors,
procedures, instruments), constrain conflicts22, reduce the number of system elements
and their interactions, and turn complicated and chaotic or rigid and inflexible policy
systems into ones possessing organized and, thus, manageable complexity. This can
be done by designing proper hierarchically nested, governance systems, introducing
positive feedback mechanisms to reinforce beneficial and attenuate undesirable policy
impacts, and managing path dependency. Because PI is based on communication and
cooperation, it has the potential to advance policy learning for adaptation to
unforeseen contingencies. Creative combination of policies can help exploit
opportunities and foster novelty in designing policy interventions.

Third, even narrower goals, such as environmental protection, cannot be achieved

simply by (vertically) incorporating environmental concerns into extant policies to
create environmentally integrated policies. The way of introducing environmental
concerns in sectoral policies still remains unspecified reflecting either a vague
conceptualization of ‘policy’ or an instrumental identification of a policy with its
goals and instruments. In any event, environmental concerns relate to more than one
environmental policy. Therefore, a meaningful introduction of, say water quality
considerations, in sectoral policies would necessitate the coordination of a sectoral
policy with water policy. Moreover, policy instruments are not delivered by

22
And, thus, reduce costs and waste of resources.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
13
themselves; they are administered by specific administrative bodies, involve particular
policy actors and relate to other instruments23. It follows that essentially EPI
inevitably implies integration among policies in terms of their objects, goals, actors,
structures/procedures and instruments.

3. HOW TO ANALYZE POLICY INTEGRATION:

A METHODOLOGICAL FRAMEWORK
The analysis of PI is approached from an institutionalist, actor-centered perspective,
which is more fit and responsive to the policy needs of complex problems (Healey,
1997; Opschoor and van der Straaten, 1993; Ostrom, 1990; Pressman and Wildavsky,
1992; Weale, 1996). This perspective places emphasis on the values, goals, theories
about the problem, resources, information-processing capabilities, and multiple
memberships of actors, their stakes in particular action situations, and the diverse
ways through which they pursue their interests within a ‘shared power world’ (Healey
1997, Ostrom 1999). Although they operate within opportunity spaces structured by
higher-level forces and power relations, actors, as reflective beings, are not passively
shaped by but they actively shape their social situation, too, developing relational
bonds of various strengths and reach. Policy problems and associated policies are not
things that just happen and policy procedures and instruments are not applied by
themselves. Actors are central from the definition of the policy problem up to the
implementation of particular policy measures. In the following, the object of PI, the
dimensions along which it can be studied, and criteria for assessing PI are discussed.

3.1. The object of policy integration

In analyzing the integration among two or more policies, the first question is “what
should be integrated”, i.e. the object of policy integration. PI can be approached from
a horizontal direction, on the same spatial/organizational level, and from a vertical
direction, across levels. The present discussion focuses on the horizontal direction
although a complete analysis should consider both directions.

The object of horizontal, intersectoral integration is fuzzy24, multidimensional,

multifaceted, and complex. Here the ‘object of policy integration’ summarizes the
multi-part content of integration among policies that varies with the stage of the
policy process (the ‘when’) and influences the ‘how’, the particular procedures and
means prescribed to promote PI. Operationally, it concerns simple and cross
relationships among the objects, goals, actors, procedures and instruments of two or
more policies (Figure 2). The main components of the object of PI are discussed
below. These concern relationships among policies on the same spatial/organizational
level, which are, however, mediated by policy and nonpolicy factors from higher and
lower levels.

23
E.g. environmental regulations cannot be properly enforced in the absence of financial instruments.
24
Starting from the definition of what constitutes a policy, the conceptualization of the policy process
and ending up with asking whether what is finally implemented on the ground bears any resemblance
to the policy as formulated and legislated (Sabatier 1999).
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
14
Figure 2. The object of policy integration

Relationships among policy objects. Two policies are integrated, or have chances of
being integrated, if they have common scope, treat common or complementary facets
(environmental, spatial, economic, social, institutional) of a problem situation in
congruent or unified manner25; or, equivalently, accommodate or respect variously
one another’s concerns about the social, economic, environmental, cultural and other
features of the issues studied26; i.e. frame environmental and other issues positively.

This implies that the policies draw on common or compatible and non-conflicting
theories and epistemological frameworks, reflecting common perceptions of the
problem and common outlooks of the actors involved, or adopt comprehensive,
interdisciplinary problem definitions and theories and define and operationalize
concepts similarly. Moreover, they most likely possess the same or compatible spatial
and temporal systems of reference and consider cross-scale integration (of global and
local issues).

The congruence of the theoretical and conceptual framings of two or more policies is
perhaps the necessary precondition and the sine qua non for their substantive, and not
only instrumental, sustainable integration.

Relationships among policy actors. Two or more policies are, or have chances of
being, integrated if they share common actors either by design27 or for reasons
unrelated to intentions to facilitate PI. Satisfactory PI among policies can be expected
if the relationships among actors are cooperative, collaborative, non-conflicting, and
non-adversarial in general, and if actors have shared values, common visions,
common goals and abide by the same rules even when these are not within their
organizational mandate (Shannon, 2002). This implies that the corresponding policy
25
Without, however, overlapping and duplicating one another.
26
I.e. rural development, regional development, etc.
27
Such as the integration correspondents, placed in each DG of the European Commission to liaise
with DG-Environment and ensure that environmental concerns are given proper account (Lenschow
2002), interministerial committees, special task forces, etc.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
15
networks are somehow coordinated or intersect either spontaneously or by design. As
discussed below, formal rules of interaction among policy actors may be provided
even when PI is not a policy goal. Besides formal, institutionalised relationships,
however, several important relationships among formal and informal actors may be
informal, providing for inconspicuous routes towards PI even when this is not
required formally (Christiansen and Piattoni 2004).

Several relationships among actors, contributing or detracting from PI, emerge during
implementation where many more actors are activated, the stakes are clearer, and
winners and losers are identifiable (Pressman and Wildavsky 1992). This means that
even if PI is institutionalised at higher policy levels, it may break down during
implementation if formal and informal actors do not agree with the integration idea
and the related implementation arrangements.

When the objects of two or more policies exhibit commonalities it is likely that the
policies have common actors, with common interests and outlooks, established
tradition and lines of communication and collaboration, and a genuine interest in
some form PI (not necessarily generally approved and supported). If, however, actors
are closely tied to ‘favorite’ policies, PI efforts may founder (Shannon, 2002).

Relationships among policy goals. Congruent, compatible, consistent, common or

complementary goals among two or more policies are favourable, necessary but not
sufficient, pre-conditions for their integration. Because policy actors define both the
policy object and the associated goals, if policies share common actors whose
relationships are cooperative their goals may be congruent or common. When the
goals and objectives of one policy include its impacts on objects of other policies28, or
if one policy is considered as a tool for the achievement of the goals of another, the
two may exhibit some degree of integration. Congruence on abstract goal statements
is easier to achieve rather than on concrete measures; sustainable development is a
telling example (Lenschow 2002, Persson 2002). Multidimensional policy goals
provide more points of contact, thus, enhancing the chances of policy integration.

Relationships among policy structures and procedures. Horizontal linkages should

exist among the organizational and administrative apparatuses of individual policies,
such as common, congruent, non-conflicting, cooperative and coordinated structures
and procedures, for properly formulating and carrying out joint, cooperative and
integrated solutions to common problems. Horizontal structures are sometimes
already in place at certain spatial/organizational levels indicating need for and interest
in coordinating specific policies, for practical purposes, at least29; i.e. making sure that
observing the prescriptions of one policy does not counteract the requirements of
some other(s). The same is true for agencies serving particular client groups that have
to exploit and combine diverse policy resources to deliver their services (Peters 1998).

Administrative procedures and other organizational arrangements and requirements

for communication, joint decision making, collaboration and conflict resolution
within and between state and non-state actors, both during policy formulation and
28
E.g. the regional economic impacts of water policy; the environmental impacts of regional policy, the
impacts of water policies on rural development, the impacts of rural policy on water resources, etc.
29
Regional and local planning authorities, formulating and implementing development plans, are
horizontal structures that necessarily carry out some form of PI in the planning process.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
16
during implementation, are necessary to promote PI. Some policies may already
mandate cooperation and collaboration among agencies on common or related issues,
which indicates interest in promoting integration with other policies30. Effective PI
requires also vertical linkages among structures and procedures. It does not suffice to
integrate policies horizontally at higher spatial/organizational levels, say the European
Union, if this scheme cannot be properly transmitted to national and sub-national
levels through proper structures and procedures. The principle of subsidiarity may
pose difficulties because needs felt at higher levels may not be equally appreciated at
lower levels where other priorities may prevail. The reverse is also true; the lack of
integration felt at lower levels, during implementation, may not be perceived easily
and directly at higher levels when communication is perplexed and ‘noisy’.

Relationships among policy instruments. The relationships among policy instruments

necessary for successful PI take various forms depending on their type. Three cases
are discussed (a) relationships among instruments of the same type, (b) relationships
among instruments of different types and (c) use of integrative instruments.

Compatible, non-conflicting, coordinated and/or complementary and mutually

reinforcing policy instruments of the same type – e.g. legal, institutional, or financial
– provided by different policies suggests some form of integration. The respective
policy objects and goals may be compatible and non-conflicting too. Conflicts among
instruments testify to lack of policy coordination (Robert et al. 2001). Although policy
instrument compatibility is tested during implementation, several conflicts can be
avoided if the design of a policy’s instruments takes into account the instruments of
other policies. This may not be always possible, however. For example, most EU
funding for regional development purposes is provided on the basis of the Community
Support Frameworks prepared by each country. It is not, therefore, straightforward to
know a priori whether this funding will be congruent, coordinated with or
complementary to funding deriving from the CAP or the Enterprise Policy to promote
sustainable regional development. This example points to the importance of decision
and policymaking procedures for effective coordination of policy instruments.

Coordinated, non-conflicting and complementary policy instruments of different types

usually promote PI. Non-coordination entails costs and inefficient use of resources,
leading at times to inaction. Financial support provided through the Structural Funds
or the Cohesion Fund has frequently contradicted the legal provisions of
environmental policies, indicating problematic integration between regional and
environmental policy (Lenschow 2002b). However, because the successful
implementation of a particular policy instrument mix is context-dependent, providing
satisfactorily coordinated policy instruments by design at higher levels appears to be
very demanding, if not infeasible, as actual integration materializes at lower levels.
Again, the importance of appropriate procedures to guide the combination of
instruments becomes evident.

Use of integrative instruments is not necessarily an indication of PI. For example, the
requirement for EIA or SEA in transport or in regional policy complies more with the
notion of vertical, intrasectoral (EPI) rather than with the notion of horizontal,

30
The EU Water Framework Directive requires participation and cooperation of representatives from
all state agencies responsible for the various uses of water in drafting River Basin Management Plans.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
17
intersectoral PI. Similarly, economic instruments are used with the purpose of
incorporating economic efficiency considerations in environmental policies and
creating markets for goods and services rather than integrating environmental with
economic policies. At best, integrative instruments assist the integration between
policies indirectly by (a) incorporating concerns of other policies in the object of a
policy, (b) inducing the harmonization of policies on theoretical grounds31, (c)
promoting the development and use of integrated assessment methods and data sets.
Integrative instruments may promote PI only if their use is reciprocal; i.e. not
asymmetric32. Moreover, reliance on integrative instruments only is not a long-term
solution to PI.

In sum, prescribed policy instruments and the rules of their use (implementation
procedures) reflect the definition of the policy object and the associated goals of the
actors participating in policy formulation. If the latter are dissociated, then the
relationships among policy instruments will suffer. The provision of integrative
instruments may be a technical solution that does not remove the fundamental
requirement of encouraging, facilitating and promoting proper relationships among
policy objects, goals and actors primarily. Moreover, during implementation, users
combine instruments to achieve their ends (i.e. solve the problems they perceive) in
context-dependent ways that can be neither anticipated nor specified precisely and
unambiguously during policy formulation. It can be argued that PI emerges during
implementation and as such it cannot be completely prescribed a priori.

Lastly, the variability of the object of PI with the stage of the policy process (agenda
setting, policy formulation and policy implementation) is briefly considered.
Integration among policy objects and goals is more relevant and essential during
agenda setting and policy formulation while integration among policy instruments is
critical at the implementation stage. Integration among policy actors is relevant
throughout the process. Nevertheless, as all constituents of a policy change in the
course of its development, a constant occupation with integration among policy
objects, goals, actors, procedures and instruments and an unending two-way
communication between formulation and implementation to facilitate as
comprehensive PI as possible and feasible are necessary.

3.2. The dimensions of policy integration

The object of PI has several facets that should be analyzed not only in functional and
procedural terms, as it is mostly the case at present, but more deeply, in substantive
terms. Four broad interrelated and interdependent clusters of dimensions of PI –
substantive, analytical, procedural, and practical – are discussed in the following to
expand on the content of PI and to inform its operationalization (Table 1).

Substantive dimensions of policy integration. The substantive cluster encompasses the

thematic, conceptual, and value dimensions that have to do with the constitution of
the policy objects to be integrated. The thematic dimension negotiates the essential
relationships among policy objects that usually concern selected characteristics of an
31
E.g. environmental policies adopt the neoclassical economics paradigm and the goal of efficient
resource use.
32
For example, environmental policies may adopt economic instruments but economic policies may
ignore the economic value of environmental goods and services used in production and consumption.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
18
issue – environmental, social, economic, cultural, etc. The term ‘thematic’ may be
equivalent to the term ‘sectoral’, if the latter refers to policy sectors. The thematic
integration requirement is embedded partly in the vertical EPI concept (VEPI)
(Lafferty and Hovden, 2002). However, VEPI has been framed in narrow (only
environmental), vague and procedural rather than broad, concrete and substantive
terms thus far, leaving untouched the issue of what relationships should be satisfied
exactly.

Table 1. Proposed clusters of dimensions of policy integration

Cluster Components
Substantive Thematic
Conceptual
Value
Analytical Spatial
Temporal
Methodological
Procedural Structural
Procedural
Practical Practical

The need for thematic integration draws directly from the sustainable development
requirement that relationships among sectoral policies account for the relationships of
the characteristics of the issue considered. Hence, thematic integration is not limited
to environmental concerns; it extends to social and economic concerns as well as to
relationships between economic sectors; e.g., agriculture and transport, etc.

Multidimensional policy objects and comprehensive policy problem definitions

facilitate the thematic integration among policies, especially if these draw on common
theories. Integrated, interdisciplinary theories should indicate relationships among the
characteristics of an issue to be addressed by the respective policies, thus constituting
the substantive basis of their integration.

Related to the thematic, but acting on deeper levels, are the conceptual and value
dimensions of policy integration. Different policies frequently define differently
similar terms and concepts; such as ‘rural’, ‘social impacts’, ‘spatial impact’, ‘region’,
‘integrated’, ‘forest’, ‘pasture’, and so on. Frequently, also, the terms and
nomenclature used are not defined or they are defined loosely, opening thus the way
for a multiplicity of interpretations (and mis-interpretations). The conceptual
integration problem is rooted in socio-culturally determined differences in the value
systems of actors in the same or different policy contexts. Essential PI presupposes a
common frame of mind, common interest in and sense of collective responsibility both
for the causes and the solution of a policy problem among all those involved; or, at
least, a positive predisposition towards cooperation for the ‘common good’. From the
point of view of its analysis, common value systems among the actors involved are
manifestations or indications of some form of PI or of its likelihood.

The combat against desertification is an excellent example of a complex policy

problem demanding the coordination of several policies, e.g. regional and rural
development, water, environmental, social, economic, transport, at various spatial
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
19
levels. However, the prospects for the substantive integration of these policies are dim
because at present because a basis for thematic, conceptual and value integration does
not seem to exist. An integrated conception of the phenomenon is missing as it has
been approached and analyzed from different, narrow disciplinary camps and
traditions, with natural sciences dominating social sciences or integrated approaches.
Not surprisingly, policy prescriptions are similarly coloured and biased, with most of
them concerning its environmental dimensions despite the recent emphasis on its
socio-economic and cultural dimensions. The connections between all dimensions are
still thin and not well-elaborated, especially on a spatial basis, exposing the lack of a
common substantive basis for PI.

The substantive dimensions of PI are directly related to its analytical, procedural and
practical dimensions as it is discussed next.

Analytical dimensions of policy integration. The analysis of the object of PI entails

spatial, temporal and methodological considerations whose operational expressions
depend on the substantive framing of PI. In a horizontal sense, the spatial dimension
of PI concerns the congruence among the spatial systems of reference (spatial
classification schemes, spatial units of reference and criteria for the delineation of
spatial areas) of different policies on the same level. Environmental and socio-
economic policies do not always share congruent spatial references. Administrative
units are ultimately used that are not, however, suitable for addressing cross-cutting
socioeconomic problems.

The spatial dimension concerns also cross-level relationships within and between state
and non-state actors involved in policy making. Lack of communication, cooperation
and coordination among those who define the policy problem and formulate the
relevant policy at higher levels, those charged with implementation and the ultimate
policy recipients (e.g. land owners, farmers, etc.) contribute to policy failures.

Similarly, the temporal dimension of PI concerns the congruence among the implicit
or explicit temporal systems of reference (temporal units, time intervals, time
horizons, and timing of actions) that policies adopt. Lack of temporal integration often
leads to ineffective and wasteful policy interventions.

The methodological dimensions of PI concern the relationships of the methods and

techniques for policy analysis associated with different policies. Different and
incompatible methods, owing to epistemological and theoretical differences among
policies, lead to conflicting or, more generally, incompatible, definitions and proposed
solutions of the policy problem. Multidimensional, comprehensive and integrated
policy problem definitions and theories dictate the use of integrated, multi-
dimensional methods and techniques at all stages of policy making.

Procedural dimensions of policy integration. These refer to the structural and

procedural relationships among policies that constitute the means through which PI
materializes and have been discussed in the previous section. Two points are
important in studying the procedural dimensions of PI; first, their relationships with
the substantive and analytical dimensions and, second, the internal consistency and
effectiveness of the procedural arrangements provided. The fact that structures and
procedures exist does not automatically imply their suitability for all cases of PI, their
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
20
adoption and implementation, and their effectiveness with respect to the goals of PI.
Ideally, substantive and analytical considerations should suggest what procedural
arrangements are needed to realize PI goals and objectives (cf. Lenschow, 2002).
However, the context-specificity of policy problems renders the achievement of this
ideal linkage between substance and process difficult in most cases. Even if a perfect
PI scheme is designed at higher spatial/organizational levels, the substantive and
institutional misfits that arise during implementation are difficult to presage at the
formulation stage. It follows that documenting PI in procedural terms does not
suffice; evidence of its effectiveness is needed, in other words, whether the policy
system functions coherently, coordination occurs, and produces benefits while it
minimizes costs. Of course, the possibility of spurious PI should be ruled out; i.e. that
the smooth functioning of the policy system does not owe to integration arrangements
but to external, non-policy factors.

Practical dimensions of policy integration. These concern the plethora of practical

issues related to availability, compatibility, consistency and congruence of data and
information needed to analyze properly the object of PI. They draw directly from the
associated substantive and analytical considerations. Without spatially, temporally
and conceptually integrated data and information systems integrated analysis of policy
problems is not feasible and, consequently, limited essential analytical support is
provided for the design of PI schemes (Briassoulis, 2001).

Table 2 suggests a correspondence between the object and the dimensions of PI. A
complete analysis requires examination of all the components of the object of PI
along the appropriate dimensions because all are interrelated.

Table 2. Relationships between object and dimensions of policy integration

Policy Policy Policy Policy Policy

Object Goals Actors Structures & Instruments
Procedures
Substantive Thematic X X
Conceptual X X
Value X X
Analytical Spatial X X X X
Temporal X X X X
Methodological X X
Procedural Structural X X X X
Procedural X X X
Practical Practical X X X X

3.3. Criteria for assessing policy integration

Based on the proposed conceptualization of the object and dimensions of PI and

drawing on the pertinent literature33, this section suggests criteria for assessing
whether integration among two or more policies already exists or for proposing how it
can be achieved or improved. The criteria are organized according to the components

33
Although the literature concerns EPI mostly, the criteria can be generalized to PI because of their
predominantly procedural orientation.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
21
of the object of PI and reflect the respective dimensions. General, cross-cutting
criteria are included also, considered as enabling conditions for the realization of PI.

The degree of PI achieved is difficult to assess in general. Ideally, if all criteria are
satisfied, then policies are perfectly integrated. The more criteria are satisfied the
higher the achievement of PI. Depending on the particular criteria being satisfied, it
can be gauged whether PI is substantive, analytical, procedural or practical. Lastly,
those criteria that are not satisfied may suggest what should be done to promote
integration between the policies examined.

General criteria
• Political commitment and leadership for PI in general34
• Need for compliance with international and EU commitments35
• Existence of long term SD strategy (or a relevant Report or Forum)
• The environmental, social, economic agendas of different sectors form a
consistent overall strategy (perhaps guided by a SD strategy)
• Favorable policy tradition and administrative culture (open, participatory,
horizontal)
• Shared core belief systems and communication across policy sectors
• Absence of intra-governmental power relations and of vertical alliances
hindering EPI/PI and horizontal networking
• Flexible general taxation.

Criteria related to policy objects

• Congruent, compatible, consistent and/or complementary policy objects and
related theories
• Multidimensional policy objects and related integrated/interdisciplinary theories
• Common and consistent concepts and terminologies

Criteria related to policy actors

• Common formal actors on and across various spatial/organizational levels
• Common informal actors on and across various spatial levels

Criteria related to policy goals and objectives

• Political commitment/ leadership for PI in the case of the policies analyzed
• Common, shared, congruent, compatible and/or complementary policy goals
and objectives
• Stipulation of quantitative, measurable, indicator-based targets and timetables
for PI (included, for example, in the Sustainable Development Strategy)

Criteria related to policy structures and procedures

• Administrative capacity for PI; it concerns, among others:
- Organization in charge of PI; such as, a central unit entrusted with
supervision, coordination and implementation of the integration process;
or assigning existing institutions a new mandate, responsibility and
accountability for PI

34
I.e., existence of a formal policy framework for PI
35
Agenda 21, EU SDS, Article 6 of the 1999 Amsterdam Treaty, etc.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
22
 - Special unit for PI in the competent organization
- Officials charged with integration tasks
- Administrative reform (restructuring) in favor of PI
- Presence of horizontal administrative structures as opposed to vertical
and departmentalized structures; e.g. inter-ministerial committees and task
forces, issue-specific joint working groups, networking schemes, regular
circulation of staff between sectoral departments
• Formal/institutionalized interaction36 among policy actors and actor networks
• Informal interaction among formal policy actors and actor networks37
• Interaction among state and non-state policy actors38
• Consistent, compatible and coordinated procedures and rules of decision-
making in competent administrative bodies39
• Strengthening existing administrative units with regard to procedural rights
and rules relevant for coordination and joint problem-solving
• Joint decision making and joint responsibilities of the policy sectors
considered
• Provisions for implementing PI requirements (e.g. compliance, enforcement
and accountability mechanisms for PI among competent agencies)

Criteria related to policy instruments

• Institutionalizing PI; existence of a legal framework for PI among the policies
analyzed40
• Common legal and institutional instruments41
• Compatible, consistent and coordinated legal and institutional instruments42
• Use of one policy as an instrument to achieve the goals of another policy43
• Use of integrative instruments; such as, legal, economic, financial, planning
• Market-based integration between the two policies44

36
Such as: (a) communication, consultation, routine early consultation on sector policies and projects,
cooperation, coordination and collaboration in implementation, etc., (b) policy formulation actors
interact formally with policy implementation actors and vice versa.
37
E.g. ad hoc meetings and informal discussions and consultations; environment or other issues are
regular agenda items in high-level meetings
38
For example, goal-related consultation and participation processes among them (from agenda setting
to policy implementation); partnerships between government and business on cross-cutting issues
39
These include the right to set formal agendas and develop PI proposals, participation by departments
or agencies in decision-making of other policy sectors, coordinated authorization procedures,
coordination/integration of sector approval/licensing processes, spatial planning, EIA, regulatory
review procedures, evaluation procedures.
40
The specific form depends on state political philosophy, ranging from regulating PI to using
persuasion and voluntary measures. It includes legal/institutional reforms in favor of PI (filling gaps in
legislation), administrative rights (rights and standing to intervene in other administrative bodies), etc.
41
Such as EIA, SEA and other provisions common in both policies; clear and common, or compatible
and congruent systems of resource rights for all resources (e.g. land, water, labor, etc.) in all policies;
rights to participate in decision making within and between policy areas for all types of actors.
42
For example, (a) use of legal provisions of one policy as an instrument to achieve goals of another –
e.g. Good Agricultural Practice codes used in rural policy to achieve water protection goals of the
EWFD, (b) adequacy of existing legal instruments as regards PI, (c) elimination of inconsistencies,
duplications, conflicts among instruments, (d) use of voluntary measures (e.g. negotiated agreements)
43
For example, biodiversity protection regulations may well serve water- or soil-related goals and vice
versa; financing transportation projects to promote regional development, etc.
44
(a) Use of economic instruments to internalize environmental costs of economic activity (e.g.
resource pricing, charges, fees), (b) economic instruments for behavior change, not for revenue raising.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
23
 • Environmental and/or Social Fiscal Reform
• Use of financial mechanisms/ incentives, such as, subsidies for PI
• Favorable budgetary process (e.g. for ‘greening’ budgets)
• Common or coordinated/compatible sector Action Plans (e.g. forest,
biodiversity, desertification, transport)
• Common, shared research resources
• Common, or compatible and consistent, data and information bases
• Common assessment and evaluation methodologies, and tools (PI indicators)
• Common monitoring programmes and infrastructure
• Use of communication instruments for PI
• Education and training services for civil servants, bureaucrats, etc. on PI issues

4. CONCLUDING REMARKS – DESIGNING POLICY

INTEGRATION SCHEMES
The quest for sustainable development, the complexity of contemporary socio-
environmental policy problems, the poor performance of current policy systems, and
several broader socio-political changes in the EU and internationally have generated
demand for PI. This paper examined conceptual and methodological issues related to
the analysis of PI, construed as integration of policies, in terms of their objects, goals,
actors, structures/procedures and instruments. This closing section offers some
preliminary thoughts on designing policy integration schemes (PIS for brevity) and
summarizes important questions that future research should address to advance the
analysis of PI.

The design of effective PIS faces the following broad questions:

• Is a general, all-purpose and all-encompassing PIS possible and desirable or

case- or issue-specific PIS should be developed?
• Is horizontal integration sufficient to tackle crosscutting issues or is vertical
integration necessary too, or both?
• Is PI at a given level sufficient or is cross-level PI necessary, or even a grand
scheme of full-blown integration on and across levels?

The ideal situation might be to design a grand scheme of horizontally and vertically
integrated policies, on the same and across levels, which can accommodate all
possible cases of crosscutting issues. However, the complex and elusive nature of
both socio-environmental problems and policies renders this option utopian and
infeasible and necessitates more flexible approaches that can be adapted to the
particularities of each case. Before proceeding to suggest one such approach, two
points are worth mentioning.

First, a PIS designed to address a given issue may provide for arrangements that also
address other issues. Of course, a PIS suitable for one problem may be unsuitable for
another. In any event, not all possible cases of crosscutting issues need to (or, can) be
considered but only a few strategic ones. This second point is supported by the
statement: “Clearly, everything is connected. But because everything is connected, it
is beyond our capacity to manipulate variables comprehensively. Because everything
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
24
is interconnected, the whole environmental problem is beyond our capacity to control
in one unified policy. We have to find … tactically defensible or strategically
defensible points of intervention” (Lindblom, 1973, 11-34). Therefore, the task is to
find which strategic policies should be integrated at each level to enable the
integration of policies on the same and on other levels.

The proposed approach builds on the adaptive management paradigm (Holling, 1978)
that has been developed to integrate uncertainty into decision-making for complex
systems. Adaptive management, based on learning-by-doing and experimentation, can
be viewed as an approach to managing risks associated with uncertainty. Resource
policies are considered as hypotheses and management as experiments from which
managers learn from their successes and from their failures. It stresses the importance
of two-way feedback between management and the state of the resource in shaping
policy, followed by further systematic experimentation to shape subsequent policy. Its
flexible, iterative, co-evolutionary and science-based character allows for institutional
learning; i.e. changing resource management institutions to fit the nature of the
system being managed (Berkes and Folke 1998).

The evolving, dynamic character of policy objects and of policy implementation and
the resulting uncertainty of policy outcomes justify the adoption of this paradigm for
designing PIS. Transferring its main ideas to the present case, the basic tenet of the
proposed approach is that any PIS is a hypothesis to be scientifically tested on the
ground and revised, through participatory approaches, by incorporating systematically
collected information obtained from PIS implementation. The policies to be integrated
at a given level should relate to critical, strategic factors associated with important
crosscutting issues and the sustainability of human-environment systems on and
across spatial levels45. Given the interconnectedness of policies, PI may start from the
most instrumental and pivotal policy, orchestrating all others around it (assuming that
the associated policy actors are willing to cooperate!), some of which inevitably will
originate in higher or lower levels. The main steps of the approach include:

• Design a PIS following a systematic and participatory approach along the lines
of analysis suggested in this study.
• Design a monitoring and evaluation scheme to gather data to address key
uncertainties.
• Implement the PIS
• Systematically monitor all aspects of implementation and record problems
such as overlaps, conflicts, inconsistencies, etc.
• Revise the PIS (policy objects, hypotheses, linkages) to fit better the particular
situations to which it applies based on feedback from implementation
• Implement the revised PIS; repeat the monitoring and revision cycle.

The adoption, implementation and success of this approach require that certain
conditions be satisfied, which usually depend on the spatial/organizational level
concerned and the scope of the PIS. The most critical requirement is interest in,
commitment to PI and a ‘commons’ mentality46 to satisfy the accountability criterion;

45
For example, economic policies and environmental, mainly water resources, policies.
46
I.e. a concern for the wise management of the wide range of natural, manmade and human Common
Pool Resources (CPRs) (Ostrom, 1990)
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
25
namely, ‘who will be responsible for the PIS?’ and ‘who will be charged with
coordinating the overall PI effort?’ (Peters 1998). It should be noted that context-
specific political expediencies, rejecting the idea of social experimentation that the
adaptive management paradigm somehow implies, may preclude the adoption of an
adaptive management approach to PI, irrespective of its plausibility and suitability.

Despite the above reservations, the proposed approach could serve as a conceptual
guide for designing and testing alternative PIS in diverse environmental and socio-
economic contexts and problem situations47. The PIS may be limited to simply
incorporating environmental, social or economic concerns in sectoral policies48 or
may move to more complete PI, ideally starting from the integration of policy objects
and moving to the integration of policy instruments, ensuring the consistency of the
overall process. As this order may be difficult to follow in reality, the process may
start from whichever component of the PI object is handy, convenient, and easy to
manipulate while trying to build reasonable linkages with the other components.

The study of PI is still in its early stages as one can glimpse from the literature. Future
theoretical and empirical research, used in combination, faces a long list of open
questions waiting to be explored. These include the study of:

• the complexity of contemporary policy problems, for a variety of issue areas such
as desertification, biodiversity protection, rural development, tourism
development, etc., and the contribution of PI to their management
• the appropriate scale for PI and the scope of the task as well as the choice between
intra-policy (vertical) versus inter-policy (horizontal) integration and the influence
of non-policy factors on the feasibility of particular types of PI
• the substantive integration of EU and national policies and its linkages to
procedural integration with emphasis on the actor networks involved in the
various EU policies at the EU and the national/subnational levels through suites of
empirical studies covering a variety of country and regional situations
• the analytical and practical dimensions of PI especially as they relate to the
substantive and procedural ones; integrated methodologies used in common by all
policies to provide compatible analyses of policy problems should be explored
• the integration of policy instruments with reference to particular issue areas so as
to effect or improve the integration of policy objects, goals and actors and the
consistency of the overall process as well as to provide guidelines for the
development of policy instrument mixes promoting substantive PI, considering
emerging new forms of governance, new kinds of market-based, and voluntary
instruments and assessment tools (such as sustainability appraisal)
• the special case of integrative instruments should be re-examined in the light of
the more essential need to promote the substantive in addition to the procedural PI
• the appropriateness of the adaptive management approach advocated here at
various scales and PI of various scope.

47
A first application of these ideas in the case of combating desertification, see Briassoulis (2004b).
48
An extension of the EPI idea.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
26
REFERENCES
1. Anderson, J. E. (1984) Public Policy-Making. 3rd edition. New York: Holt, Reinhart and
Winston.
2. Anderson, P., Arrow K.J., and Pines D. (eds) (1988), The Economy as an Evolving Complex
System I, Addison-Wesley, Reading, Mass.
3. Arthur, B. (1989), ‘Competing technologies, increasing returns and lock-in by historical events’,
Economic Journal, 99, pp. 116-31.
4. Arthur, W.B., Durlauf, S.N., and Lane, D.A. (eds) (1997), The Economy as an Evolving
Complex System II, Addison-Wesley, Reading, Mass.
5. Ascher, W. (2001), ‘Coping with complexity and organizational interests in natural resources
management’, Ecosystems, 4, pp. 742-57.
6. Avery, G. (2001), ‘Policies for an Enlarged Union. Report of Governance Group 6’, White
Paper on European Governance Area no. 6, Defining the framework for the policies needed by
the Union in a longer-term perspective of 10-15 years taking account of enlargement, European
Commission, Brussels.
7. Baas, N.A. and Emmeche, C. (1997), ‘On emergence and explanation’, Intellectica, 25, pp. 67-
83.
8. Batty, M. and Torrens, P.M. (2001), ‘Modeling Complexity: The Limits to Prediction’, CASA
Paper 36, Centre for Advanced Spatial Analysis, University College London
(http//www.casa.ucl.ac.uk/paper36.pdf).
9. Berkes, F. and C. Folke, eds. (1998) Linking Social and Ecological Systems: Management
Practices and Social Mechanisms for Building Resilience. Cambridge, Ma. Cambridge
University Press.
10. Berkhout, F. (2002) ‘Technological regimes, path dependency and the environment’, Global
Environmental Change, 12, pp. 1-4.
11. Brown, M. L. (2000), ‘Scientific uncertainty and learning in European Union environmental
policy making’, Policy Studies Journal, 28(3), pp. 576-96.
12. Byrne, D. (1998), Complexity Theory and the Social Sciences: An Introduction, Routledge,
London.
13. Briassoulis, H. (2001) Policy-oriented integrated analysis of land use change: An analysis of
data needs. Environmental Management, 26(2): 1-11.
14. Briassoulis, H. (2004a), ‘The institutional complexity of environmental policy and planning
problems: The example of Mediterranean Desertification’, Journal of Environmental Planning
and Management, 47, No. 1, pp. 115-35.
15. Briassoulis, H. (2004b) Design of a Desertification Policy Support Framework (DPSF),
Deliverable 36, Module 4, MEDACTION. European Commission, DG-XII, Contract No.
ENVK2-CT-2000-00085 (www.icis.nl/medaction).
16. Bromley, D.W. (1991) Environment and Economy. Cambridge, Ma: Blackwell Publishers.
17. Christiansen, T. and S. Piattoni, eds. (2004) Informal Governance in the European Union.
Cheltenham, Edward Elgar.
18. Collier, U. (1994), Energy and environment in the European Union. Avebury, Aldershot.
19. Detombe, D.J. (2001), ‘Methodology for handling complex societal problems’, European
Journal of Operational Research, 128, pp. 227-30.
20. Dooley, K. (1997), ‘A complex adaptive systems model of organization change’, Nonlinear
Dynamics, Psychology and Life Sciences, 1(1), pp. 69-97.
21. Dryzek, J.S. (1987), Rational Ecology, Environment and Political Economy, Blackwell Publishing,
Oxford.
22. EEA (1998), Europe's environment. A second assessment, European Environment Agency,
Copenhagen.
23. European Commission (1998) Partnership for integration: A strategy for integrating
environment in EU policies. COM (98)0333 final.
24. European Commission (1999) Report on Environment and Integration Indicators to Helsinki
Summit. Commission Working Document, SEC (1999) 1942 final, Brussels.
25. European Commission (2000) Bringing our needs and responsibilities together-Integrating
environmental issues with economic policy. Communication from the Commission to the
Council and the European Parliament. COM(2000) 576 final. Brussels.
26. European Commission (2001) Policies for an Enlarged Union. Report of Governance Group 6.
June 2001, Brussels.
Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
27
27. EC (2004) Third Report on Economic and Social Cohesion. COM(2004) 107. Luxembourg,
Office for Official Publications of the European Communities
28. EEB (2003), Environmental Policy Integration (EPI): Theory and Practice in the UNECE
Region, Background Paper of the European ECO Forum for the Round Table on Environmental
Policy Integration at the Fifth “Environment for Europe” Ministerial Conference, Kyiv, May 21-
23, 2003, European Environmental Bureau.
29. Finnigan, J. (2003), ‘Earth System Science in the Early Anthropocene’, Global Change
Newsletter, Issue 55, pp. 8-11.
30. Friedrich, C.J. (1963), Man and his Government, McGraw-Hill, New York.
31. Gallagher, R. and Appenzeller, T. (1999), ‘Beyond reductionism’, Science, 284, p. 79.
32. Gibbs, D., Jonas, A. and While, A. (2003), ‘Regional Sustainable Development as a Challenge
for Sectoral Policy Integration’, Paper presented at Workshop II of the EU Thematic Network
Project REGIONET, Regional Sustainable Development - Strategies for Effective Multi-level
Governance, Lillehammer, Norway, 29-31 January 2003.
33. Glasner, E and Weiss, H. (1993), ‘Sensitive dependence on initial conditions’, Nonlinearity, 6,
pp.1067-75.
34. Greenberg, G.D., J.A., Miller, L.B. Mohr, and B.C. Vladeck (1977) Developing public policy
theory: Perspectives from empirical research. The American Political Science Review, 7: 1532-
1543.
35. Gunderson, L.H. and Holling, C.S. (eds) (2002), Panarchy; Understanding Transformations in
Human and Natural Systems, Island Press, Washington, DC.
36. Hajer, M. (2003), ‘Policy without Polity; Policy analysis and the institutional void’, Policy
Sciences, 36, pp.175-95.
37. Hakkinen, L. (ed.) (1999), Regions – Cornerstone for Sustainable Development, Edita, Helsinki.
38. Healey, P. (1997) Collaborative Planning: Shaping Places in Fragmented Societies. Vancouver,
UBC Press.
39. Henderson, B. (2001), ‘Path dependence, escaping sustained yield’, Conservation Ecology, 5(1),
r3, [online] URL: http://www.consecol.org/vol5/iss1/resp3.
40. Hertin, J. and Berkhout, F. (2003), ‘Analysing Institutional Strategies for Environmental Policy
Integration: The Case of EU Enterprise Policy’, Journal of Environmental Policy and Planning,
5(1), pp.39-56.
41. Holland, J.H. (1995), Hidden Order: How Adaptation Builds Complexity, Helix, Reading, MA.
42. Holling, C.S. (1978) Adaptive Environmental Assessment and Management. New York, John
Wiley.
43. Holling, C.S. (1986), ‘The resilience of terrestrial ecosystems: local surprise and global change’,
in W.C. Clark, and R.E. Munn (eds), Sustainable Development of the Biosphere, Cambridge
University Press, Cambridge, pp. 292-317.
44. Holling, C.S., Berkes, F. and Folke, C. (1998), ‘Science, sustainability and resource
management’, in F. Berkes, and C. Folke (eds), Linking Social and Ecological Systems:
Management Practices and Social Mechanisms for Building Resilience, Cambridge University
Press, Cambridge, MA, pp. 342-62.
45. Holling, C.S. (2001), ‘Understanding the complexity of economic, ecological and social
systems’, Ecosystems, 4, pp. 390-405.
46. Hooghe, L. and Marks, G. (2001), Multi-level Governance and European Integration, Rowman
& Littlefield, Oxford.
47. Janssen, M., Walker, B.H., Langridge, J. and Abel, N. (2000), ‘An adaptive agent model for
analyzing co-evolution of management and policies in a complex rangeland system’, Ecological
Modeling, 131, p. 249-68.
48. Knill, C. and Lenschow, A. (eds) (2000), Implementing EU environmental policy. New
directions and old problems, Manchester University Press, Manchester/New York.
49. Lafferty, W.M. and E. Hovden (2002) Environmental Policy Integration: Towards an Analytical
Framework? PROSUS, Centre for Development and the Environment, University of Oslo, Oslo,
Report 7/02.
50. Langton, C.G. (1990), ‘Computation at the Edge of Chaos: phase transitions and emergent
computation’, Physica D, 42, 1-3, pp. 12-37.
51. Lee, K.N. (1993), Compass and gyroscope: Integrating science and politics for the environment,
Island Press, Washington, DC.
52. Lenschow, A. and A. Zito (1998). “Blurring or shifting of policy frames? Institutionalization of
the economic-environmental policy linkage in the European Community.” Governance 11(4):
415-442.

Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
28
53. Lenschow, A., ed. (2002a) Environmental Policy Integration: Greening sectoral Policies in
Europe, London: Earthscan.
54. Lenschow, A., ed. (2002b) Dynamics in a multilevel polity: Greening the EU regional and
Cohesion Funds. In Environmental Policy Integration: Greening sectoral Policies in Europe, A.
Lenschow, ed. London: Earthscan, pp. 193-215.
55. Levin, S. (1999), Fragile dominion: Complexity and the commons, Perseus Publishing,
Cambridge, MA.
56. Lewanski, R. (2002), ‘Environmental policy integration in Italy: I a green government enough?
Evidence from the Italian case.’ in A. Lenschow, (ed.), Environmental Policy Integration:
Greening Sectoral Policies in Europe, Earthscan, London, pp. 78-100.
57. Liberatore, A. (1997). The integration of sustainable development objectives into EU policy-
making: Barriers and prospects. In The politics of sustainable development: Theory, policy and
practice within the European Union. S. Baker, M. Kousis, D. Richardson and S. Young, eds.
London, Routledge.
58. Liebowitz, S. and Margolis, S.E. (1995), ‘Path Dependence, Lock-In, and History’, Journal of
Law, Economics, and Organization, 11(1), pp. 205-26.
59. Liefferink, D. and Jordan A. (2002), ‘The Europeanisation of National Environmental Policy; A
Comparative Analysis’, University of Nijmegen, GAP, Working Paper Series 2002/14.
60. Limburg, K.E., O’Neill, R.V., Constanza, R. and Farber, S. (2002), ‘Complex systems and
valuation’, Ecological Economics, 41, pp. 409-20.
61. Lindblom, C. E. (1973) Incrementalism and environmentalism. In Managing the Environment.
Final Conference Report, Washington, DC, Environmental Research Center, US EPA.
62. Lowi, T.J. (1964), ‘American business, public policy, case studies, and political theory’, World
Politics, 16, pp. 677-715.
63. Manson, S.M. (2001), ‘Simplifying complexity: A review of Complexity Theory’, Geoforum,
32, pp. 405-414.
64. Marion, R. (1999), The Edge of Organization: Chaos and Complexity Theory of Formal Social
Systems, Sage, Thousand Oaks, CA.
65. Marks, G., Hooghe, L. and Blank, K. (1996), ‘European Integration from the 1980s: State-
Centric v. Multi-level Governance’, Journal of Common Market Studies, 34(3), pp. 341-78.
66. Metcalfe, L. (1994) “International policy coordination and public management reform.”
International Review of Administrative Sciences 60: 271-290.
67. Moss, T. (2003), ‘Regional Governance and the EU Water Framework Directive: a study of
institutional fit, scale and interplay’, Paper presented at the workshop of the EU Thematic
Network project REGIONET “Regional Sustainable Development – Strategies for Effective
Multi-Level Governance”, 29-31 January 2003, Lillehammer, Norway.
68. Moss, T. (2004), ‘Regional Sustainable development as a Cross Sectoral Task’, Discussion
paper presented at the REGIONET workshop “Cross fertilization and integration of results of
REGIONET”. Brussels, 14-16 January 2004.
69. OECD (1996a) Building policy coherence: Tools and tensions. Public management occasional
papers. No. 12. Organization for Economic Cooperation and Development, Paris.
70. OECD (1996b) Globalisation: What challenges and what opportunities for government?
Organization for Economic Cooperation and Development, Paris.
71. Opschoor, H. and J. van der Straaten (1993) Sustainable Development: An Institutionalist
Approach. Ecological Economics 7 (1993), pp. 203-222.
72. O’ Riordan and Voisey, 1998;
73. Ostrom, E. (1990) Governing the Commons: The Evolution of Institutions of Collective Action.
Cambridge, MA: Cambridge University Press.
74. Ostrom, E. (1998), ‘Scales, Polycentricity, and Incentives: Designing Complexity to Govern
Complexity’, in D.G. Lakshman and J.A. McNeely (eds), Protection of Global Biodiversity:
Converging Strategies, Duke University Press, Durham, NC., pp. 149-67.
75. Ostrom, E. (1999) Institutional rational choice: An assessment of the Institutional Analysis and
Development framework. In Theories of the Policy Process, P. Sabatier, ed. Boulder, Co.:
Westview Press, pp. 35-72.
76. Patton, C.V. and Sawicki, D. (1986), Basic methods of Policy Analysis and Planning, Prentice
Hall, Englewood Cliffs.
77. Persson, A. (2002) Environmental Policy Integration: An Introduction. PINTS - Policy
Integration for Sustainability, Background Paper (Draft). Stockholm Environment Institute.

Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
29
78. Peters, G. B. (1998) Managing Horizontal Government: The Politics of Coordination. Canadian
Centre for Management Development. Research Paper No. 21. Catalogue Number SC94-61/21-
1998, ISBN 0-662-62990-6.
79. Pressman, J. L. & A. Wildavsky (1992) Implementation: How Great Expectations in
Washington Are Dashed in Oakland. Berkeley, CA, University of California Press.
80. Reynolds, J.F. and Stafford-Smith, M. (2002), ‘Do Humans cause deserts?’ in J.F. Reynolds and
M. Stafford-Smith (eds), Global Desertification: Do Humans Cause Deserts?, Dahlem
University Press, Berlin, pp.1-22.
81. Richardson, J.J. (1982), Policy styles in Western Europe, Allen and Unwin, Hemel Hempstead.
82. Robert, J., Stumm, T., de Vet, J.M., Reincke, C.J., Hollanders, M. and Figueiredo, M.A. (2001),
Spatial Impacts of Community Policies and the Costs of Non-Coordination, EC, DG-Regional
Policy, ERDF Contract 99.00.27.156, Brussels.
83. Sabatier, P., ed. (1999), Theories of the Policy Process. Boulder, Co.: Westview Press.
84. Sanderson, I. (2000), Evaluation in complex policy systems, Evaluation, 6(4), pp. 433-454.
85. Science, (1999), Complex Systems, Vol. 284, pp. 79-109 (April 2, 1999).
86. Shannon, M.A. (2002) Theoretical Approaches to Understanding Intersectoral Policy
Integration. Paper presented at the Finland COST Action meeting (European Forest Institute).
87. Thomas, D.S.G. (1997), ‘Science and the desertification debate’, Journal of Arid Environments,
377, pp. 599-608.
88. Thomas, E. (2003), ‘Sustainable development, market paradigms and policy integration’,
Journal of Environmental Policy and Planning, 5(2), pp. 201-16.
89. True, J.L., Jones, B.D. and Baumgartner, F.R. (1999), ‘Punctuated-equilibrium theory:
Explaining stability and change in American policymaking’, in P.A. Sabatier (ed.), Theories of
the Policy Process, Westview Press, Boulder, CO., pp. 97-116.
90. UNCCD (1994), United Nations Convention to Combat Desertification and Drought
(www.unccd.int).
91. Underdal, A. (1980) Integrated Marine Policy: What? Why? How? Marine Policy 4 (3): 159-
169.
92. Vogel, (1986), National Styles of Regulation: Environmental Policy in Great Britain and the
United States, Cornell University Press, Ithaca, NY.
93. Waldrop, M.M. (1992), Complexity; The Emerging Science at the Edge of Order and Chaos,
Penguin Books, London.
94. Weale, A. (1996) Environmental rules and rule-making in the European Union. Journal of
European Public Policy, 3(4): 149-167.
95. Wilson, J. (2002), ‘Scientific uncertainty, complex systems and the design of common-pool
institutions’, in E. Ostrom, T. Dietz, N. Dolsak, P. Stern, S. Stonich and E. Weber (eds), The
Drama of the Commons, Committee on the Human Dimensions of Global Change, National
Academy Press, Washington, DC, pp. 327-60.
96. Wiman, B.L.B. (1991), ‘Implications of environmental complexity for science and policy’,
Global Environmental Change, June 1991, pp. 235-47.
97. Wittrock, B. and de Leon, P. (1986), ‘Policy as moving target: A call for conceptual realism’,
Policy Studies Review, 6(1), pp. 44-60.
98. Wollin, A. (1999), ‘Punctuated Equilibrium: Reconciling Theory of Evolutionary and
Incremental Change’, Systems Research and Behavior Science, 16, pp.359-67.
99. Zahariadis, N. (1999), ‘Ambiguity, time and multiple streams’ in P.A. Sabatier, (ed.), Theories
of the Policy Process, Westview Press, Boulder, CO., pp. 73-95.
100. Zahariadis, N. (2003), ‘Complexity, coupling, and the future of European integration’, Review of
Policy Research, 20(2), pp. 285-310.

Paper presented at the 2004 Berlin Conference “Greening of Policies: Interlinkages and Policy Integration”
Berlin, December 3-4, 2004.
30