ONE

Theoretical perspectives on
policy analysis

Policy analysis consists in the “study of the action of public authorities

within society” (Mény and Thoenig, 1989, p 9). In terms of disciplines,
a number of academic sectors have been and are associated with it. It
was adopted as early as 1979 by Wildavsky (1979, p 15) in his plea for
the development of this approach: “Policy analysis is an applied subfield
whose contents cannot be determined by disciplinary boundaries but
by whatever appears appropriate to the circumstances of the time and
the nature of the problem”. Similarly, Muller (1990, p 3) mentions
that “policy analysis is located at the junction of previously established
knowledge from which it borrows its principal concepts”.
We start by presenting a quick review of the literature from the
traditional policy analysis schools1 and then go on to examine the
specific theoretical framework adopted in this book.

1.1 Various currents in policy analysis

The main disciplines that can be observed within the different schools
define themselves in accordance with the theoretical and normative
perspectives, on which the positions of the different authors are based
and/or towards which they tend. Thus, after Mény and Thoenig (1989)
and Muller (1990, p 3), it is possible to identify three major currents
in policy analysis that reflect different aims without, however, being
mutually exclusive. These currents differ mainly in terms of their focuses
on specific fields of analysis.
Thus, we make distinctions between a first school of thought that
associates policy analysis and the theory of state, a second that explains
the way in which public action works and, finally, a third that focuses
on the evaluation of the results and effects of the latter.

1.1.1 Policy analysis based on the theories of state

For the first group of authors, policy analysis is a means of explaining

the actual essence of public action because policies are interpreted as
revealing its nature. This current, which the political sciences dominates

3
Public policy analysis

and lays claim to, in France in particular, attempts to link the policy
approach with political philosophy and major questions concerning
the theory of state. Thus, Mény and Thoenig (1989) define their
approach in terms of a contribution to questions concerning ‘the
emergence and nature of the state’ or to ‘the essence of politics’.
Similarly, Jobert and Muller locate their work on The state in action
(L’etat en action) in the context of “bridging the gap that today still
separates research on policies and the more general reflections on the
state in contemporary society” (Jobert and Muller 1987, p 9). It is
subdivided into different approaches that Mény and Thoenig (1989,
p 67) classify on the basis of three ‘theoretical models’:
 The first model is part of a pluralist approach that conceives the
state as a ‘service hatch’ whose purpose is to respond to social
demands. From this perspective, public policies are conceived as
responses to social demands and their analysis is in turn located
in a perspective based on the optimisation of collective choices,
the rationality of the decision-making processes and the behaviour
of ‘bureaucrats’ (‘public choice’ school2, theory of limited
rationality3). According to this concept, the lack of policies in
the area of sport, for example, is a reflection of the fact that there
is no public problem to be resolved. However, this absence could
also be interpreted as the result of corporate or private actions
that are aimed at controlling this sector despite the existence of
significant public problems (in particular drug use, corruption
etc).
 The second interpretative model places the emphasis on the
state as an instrument at the service of either a social class (neo-
Marxist approach4) or specific groups (neo-managerial approach5).
In this context, the analysis of public action makes it possible to
demonstrate the weak autonomy of the state with respect to
capitalist interests and/or with respect to the private actors and
organisations of which it consists. Seen in this way, a social
problem can only become a public problem if its resolution serves
the interests of the (economically) dominant classes. The neomanagerial
approach starts from a similar standpoint in that it
replaces the class concept with the concept of elites.
 Finally, the third model stresses the distribution of power and
interaction among and between actors, either through the
representation and organisation of different sector-based or
category-based interests (neo-corporatist6 approach), or through
the organisations and institutional rules that frame these

4
 Theoretical perspectives on policy analysis

interactions (neo-institutionalist7 approach). Seen from the neocorporatist

perspective, public sector employees are for the most
part ‘captured’ by the interest groups (‘clients’), with which they
maintain privileged and exclusive relationships in the exercising
of public power. In the UK a related approach involves emphasis
on the roles of policy networks and policy communities8. The
influence of agricultural interests has been analysed in this way
in various countries. In France, this analysis results in the
emphasising of the role of large state bodies and their privileged
relationships with their colleagues who work in the private sector
as a factor that explains the way the central administration works.

In these theoretical models policy analysis is treated as a way of verifying

the hypotheses underlying the favoured model. In summary, the main
characteristic of the first school is that it does not focus on policies in
themselves but as a way of understanding the role of the public sector
in society and its evolution in time, which results in the introduction
‘of policy’ into the empirical analysis of public action and organisations
and in the focusing of the analysis on this interface.
The status of the thinking, trends and claims of this school is reflected
in French work such as the fourth volume of the Traité de science politique
(Grawitz and Leca, 1985) and in Mény and Thoenig’s (1989) book on
policy. More recent and more informal references can be found in the
debates organised by the ‘policy’ group of the French Association of
the Political Sciences, whose findings are published in the Revue
Française de Sciences Politiques (see, in particular, Majone, 1996; Muller
et al, 1996). Hill explores the Anglo-American literature of a similar
kind in The public policy process (2005, chapters 2 to 5).
Our own approach is partly rooted in this perspective, because many
of the authors who inspired it belong to this school and also because
the work we have done in this area facilitates, in part, a real interpretation
of the role of the state in society without making this the primary
aim. However, our approach borrows more strongly from the second
and third schools.

1.1.2 Explaining how public action functions

The second school aims to explain the way public action works. Thus,
in this context, the function of policy analysis is not to explain the
general functioning of the political system but to act as a way of
understanding the operational modes and logic of public action (see

5
Public policy analysis

Dente, 1985, 1989; Dente and Fareri, 1993; Gomà and Subirats, 1998,
pp 21-36).
This approach does not exclude the adoption of a viewpoint based
on the above-presented theories, which explains why several authors
from this second group actually have a foot in both camps. Here,
however, the focus is not on the justification of a theory, but on the
demonstration of continuities, general rules of functioning that are
specific to public actions. In this context, policy analysis makes it
possible to understand how the state and, more broadly, public
authorities work.
This second approach actually constitutes the initial set of issues
tackled by policy analysts. Historically, the latter were strongly
influenced by North American political scientists, whose initial
considerations in this area emerged between 1950 and 1960 and were
linked with a context of the ‘rationalisation’ of public decision making
with a view to improving its efficacy. Lerner and Lasswell published
The policy sciences in the United States as early as 1951, thereby laying
the foundations for this approach.
However, this “unified approach to the study of public problems
and policy ... soon settled into two main approaches” (Parsons, 1995,
pp 18-19), one that endeavours to develop a better knowledge of the
policy formation and implementation processes (the analysis of policy),
while the other concentrates on developing knowledge that is usable
in and for the policy formation and implementation processes (analysis
in and for policy). It should be stressed, however, that the analyses
carried out by one school feed into the experiences of the other, and
vice versa. Thus, in their critique of this approach, Mény and Thoenig
(1989, p 65) make a distinction between the function of the scientist
who is interested in the progress of knowledge and learning and that
of the professional whose aim is to apply the sciences for the purpose
of action.
The second approach adopts its theoretical thrust from several
different scientific approaches: administrative science, the sciences of
complexity (particularly systems analysis), the sociology of (public)
decision making and, more generally, the sociology of collective action,
the economic sciences and the information sciences.
The emergence of this approach was dominated by four major figures
(Mény and Thoenig, 1989; Parsons, 1995). The first is the North
American political scientist Lasswell (1951) who was the movement’s
main source of inspiration and who adopted a completely ‘managerial’
approach: his work deliberately attempted to construct a dialogue
between social scientific researchers, economic circles and public

6
 Theoretical perspectives on policy analysis

decision makers so as to improve the efficacy of public action. The

second is Simon (1957), whose work on human decision-making
processes directed this type of approach towards the analysis of public
decision-making processes (using the concept of ‘bounded rationality’).
Lindblom (1959) also impacted on the development of policy analysis
by concentrating the analysis on the limited room for manoeuvre at
the disposal of public decision makers (using the concept of
‘incrementalism’). Finally, there was Easton (1965), who was one of
the first political scientists to apply the science of systems analysis to
the policy world as a whole, and who made a significant contribution
to the development of the main concepts of contemporary policy
analysis.
All of these authors, who belonged to sometimes radically diverging
schools of thought, had an impact on the emergence of this approach,
and on the definition of the concepts used in this type of analysis and
discussed in this book. For them the state is no longer considered as a
single actor but as a complex and often heterogeneous politicaladministrative
system whose workings need to be understood to enable
the formulation of ‘predictions’ or ‘recommendations’. Nevertheless,
here again there are several different perspectives.
 Certain authors focused their analyses on the decision-making
process and actor strategies. This type of analysis is related on the
one hand to the work of sociologists of public organisations, whose
main representatives in France are Crozier and Friedberg (1977).
It is also connected with the application of systems analysis to
human decision making in the tradition of the work of Simon
(1957), Morin (1977, 1980) and Le Moigne (1990), and with
endeavours to engage in the analysis of systems actors or concrete
action systems. The professional consequence of this approach
was the emergence of ‘public management’ that was promoted,
in particular, by the work carried out by the Organisation for
Economic Co-operation and Development (OECD). Thus, the
OECD’s Public Management Service (PUMA) regularly
publishes literature on the role of ‘managers’ in public
organisations. This approach is, however, not very sensitive for
the analysis of the specific policies implemented by the analysed
administrations.
 Other works in this area are based on the tools and instruments
of public intervention. Economic approaches and, in particular,
research on the political economy predominate here. This work
analyses the modes of public action in terms of their efficacy

7
Public policy analysis

from either a macro-economic point of view (in the tradition of

Pareto, and Keynes (1936) and Musgrave (1959) or from a microeconomic
point of view (in particular, the approach adopted by
the clientele of public services at the centre of what is known as
‘New Public Management’9).
 Other works again focus on the structures, procedures and
institutional forms of public administration. This approach
constitutes the essence of the administrative sciences and
administrative law. It describes the way administrative institutions
work and, more generally, studies institutional policies as defined
by Quermonne10. In France, this approach mainly refers to
administrative reforms and, in particular, to the policies for the
decentralisation of power. This approach is of little interest for
the concrete policies implemented by the administrations studied.
However, some of the work that forms part of policy analysis is
published in the Revue Française d’Administration Publique11. In
Switzerland, the term ‘institutional policies’ generally refers to
the modes of control employed by the different types of public
administration, the questions of representativeness in the
composition of ministries (languages, sexes, age, political parties
etc), the civil servants’ statute, the (in)formal relationships between
the federal authorities and the cantons etc. In terms of policy,
the work carried out by the political scientist Germann (1996)
is closest to these analyses.
 Finally, a specific public policy approach has been emerging for
some years now. It is known as the cognitive approach, “which
attempts to understand public policies as cognitive and normative
matrices constituting systems of interpretation of reality, within
which the different public and private actors can register their
actions” (Muller and Surel, 1998, p 47). This approach stresses
the role of ideas and representations in the formation (and, in
particular, the definition of problems subject to public action)
and alteration of public policies. The distinctive element of this
approach is the emphasis it places on general principles, the
argumentation and values that define ‘a global vision’ that reflects
and/or produces the public policy.
In summary, the characteristic feature of this second school is the
concern to understand the complexity of the public decision making
processes by dividing the object of the analysis into different variables
(for example, actor rationality, internal decision-making processes in
organisations and so on). This approach was perpetuated, inter alia, in

8
 Theoretical perspectives on policy analysis

the public management and decision aid methods; however it differs

strongly from them in terms of its lack of direct operationalisation.
Our own approach is based on both a scientific and operational
perspective. As Friedberg (1993, p 22) reminds us, the analyst “now
has two interdependent tasks: on the one hand to produce a concrete
knowledge of the human reality underlying the context of action
analysed and, on the other, to help the interested parties to both position
themselves with respect to this knowledge and to draw consequences
from it and integrate these into their practices by modifying them”.

1.1.3 Evaluation of the effects of public action

The third school of thought tries to explain the results of public action
and its effect on society from the viewpoint of the objectives pursued
and/or in terms of indirect or undesirable effects. Compared with the
previous one, this approach is more evaluative than explanatory. For
the past 10 years or so, this approach has become particularly fashionable
in France and Switzerland where initiatives, symposiums and
publications on policy analysis are thriving12. In the UK the related
concern has been with ‘evidence’ for policy13.
It is possible to identify two main concerns in the context of
evaluation:
 The first of these involves the development of a methodological
approach and an evaluation ‘tool box’: thus, numerous studies
have undertaken to define evaluation methods that can be applied
to the non-market activities of the public sector. This work is
based on the statistical processing of quantitative data, multicriteria
analysis (Maystre et al, 1994); (quasi) experimental
comparison, cost-benefit analysis and so on. Extensive literature
has been published on this approach. In France, a presentation of
its ideas can be found in Deleau et al (1986) and in the annual
publications of the Conseil Scientifique de l’Évaluation. It is also
evident in the manuals for the evaluation of socio-economic
programmes recently published by the European Union with
the aim of facilitating the evaluation of programmes associated
with European structural funds.
 The second focuses on the process of evaluation and its
implementation in terms of improving public management and
influencing decision making. A significant number of North
American and, more recently, European authors have investigated
this question, including Rossi and Freeman (1993) for the US,

9
Public policy analysis

Monnier (1992) for France and the collective work of Bussmann

et al (1998) for Switzerland.
In Switzerland, policy evaluation has become almost a
profession: it is practised by academics at the universities, in private
consultancies and can also be found within the administrative
authorities themselves due to the creation of a parliamentary
and governmental evaluation service (Organe Parlementaire de
Contrôle de l’Administration, Service d’Evaluation du Conseil Fédéral).
A professional association (Société Suisse d’Évaluation, SEVAL)
monitors the quality of evaluations carried out (meta-evaluations
that monitor the adherence to quality, use and ethical standards
etc).
This trend is equally evident in France. It is associated with
the organisation of evaluation mechanisms at national and
regional level and with the development of European policies
that require the organisation of evaluation exercises during the
implementation of European structural funds, for example. A
Société Française de l’Évaluation was recently created with the aim
of rendering this activity more visible and improving its
organisation. However, this movement is finding it difficult to
become institutionalised as a standard practice in public
administration. The disappearance of the interministerial
mechanism that was introduced in 1990 is an indication of this,
similarly the absence of transparency in the work of the Office
Parlementaire des Choix Scientifiques et Techniques. However, it is
increasingly the subject of analysis by political scientists (Duran
and Monnier, 1992; Lascoumes and Setbon, 1996; Kessler et al,
1998). Thus, the French model for the evaluation of public policy
appears to be characterised by the weak involvement of the policy
actor and a very much reduced use of the results of evaluation in
the modification or conception of public policies, despite the
interest of actors in policy implementation.
In the UK the work of the Audit Commission and the
increasing use of quantitative performance indicators in education,
the health service and local government has had a similar impact
(Pollitt, 2003; Audit Commission, 2006).

This evaluative approach is generally accompanied by an explanatory

approach, even if it may be conceptually dissociated from it. It is an
inspiration for our own model in the sense that it is concerned with
the effects of public action – effects that are measured on the basis of
the collective problem that a policy tries to resolve.

10
 Theoretical perspectives on policy analysis

1.2 Policy analysis as a ‘science of action’

Rather than fitting perfectly within the framework of one of the abovedescribed
schools, our analysis borrows from all three. It is our ambition
to establish a diagnostic approach that demonstrates the factors that
explain the ‘good’ and ‘bad’ functioning of public policies in terms of
public administration production and with respect to the efficacy of
its policies and their products. This type of analysis ultimately leads to
describing, understanding and explaining the workings of the politicaladministrative
system as a whole and its interactions with private actors.
Thus, our approach is mainly based on the explanation of the products
or services provided by public administration that are traditionally
referred to as ‘outputs’, and on the explanation of the effects produced
by these services on social groups (‘impacts and outcomes’) that cause
and/or are affected by a particular collective problem.
To the extent that it aims to understand the ‘logic’ of public actions
by reconstructing the hypotheses on which public authorities
(sometimes implicitly) base their thinking for the resolution of public
problems, our intellectual reasoning belongs within the framework of
the action sciences.
More precisely, the majority of the concepts presented here are
derived from the publications of the Centre de Sociologie des Organisations
(Crozier, 1963, 1991; Crozier and Friedberg, 1977; Friedberg, 1993)
as well as the work of the German social and political scientists of the
1970s (that is, the Frankfurt School), who, in turn, were strongly
influenced by neo-Marxism (Offe, 1972; Habermas, 1973; Grottian,
1974). However, this influence is limited to the individual heuristic
contributions that are particularly well developed by these authors.
This enables us to identify the actors, their networks and their modes
of interaction. As opposed to the ‘systemic forces’ often favoured by
these authors, in our approach, the retraceable strategies, ideas, interests
and actor behaviour essentially depend on factors connected to their
resources and their ‘institutional framework’ and must all be observed
empirically. In this sense, our reasoning strongly resembles actor-centred
institutionalism as presented by Scharpf (1997).
The concept of public policy adopted here (as well as most of the
definitions and terms used in this book) originate in part from the
work carried out in Germany by the Forschungsverbund: Implementation
politischer Programme at the Deutsche Forschungsgemeinschaft between 1976
and 1981 under the direction of Mayntz, Scharpf, Kaufmann and
Wollmann14; one of the authors of this book was associated with this
work. It is also based on texts on the implementation of public policy15.

11
Public policy analysis

The approach presented has been particularly strongly influenced by

the international comparison of public policies, in particular because
the comparative approach leads to the definition of a common analytical
frame that is applicable to different countries and institutional regimes
and that constitutes the essential aim of our approach.
As discussed in detail in the following chapters, our approach is
mainly characterised by the fact that it:
• tackles a policy from the angle of its ‘action logic’, thus its starting
point constitutes the arena of the political-administrative action
and social actors who interact in a defined sector;
• integrates the influence of institutions on the behaviour of these
actors and on the substantive results of the public action (while
the first generation of policy analysis tended to neglect the
institutional variables);
• pays particular attention to the resources mobilised by these actors
in order to assert their interests (which facilitates the combination
of policy analysis and public management).

Finally, our approach, which is essentially based on a retraceable

interpretation of empirical data, differs from other contemporary
research currents, in particular:

• neo-Weberianism that supposes that bureaucratic actors benefit

from the rigidity or at least inertia of certain structures and
administrative rules and try above all to obtain secure incomes at
the cost of the content of the policies for which they are
responsible;
• neo-Marxism which, despite claiming with justification that in
addition to its primary democratic legitimacy the state must
ensure a secondary legitimacy through the approbation of the
‘quality’ of its public policies by actors powerful in resources,
interprets the latter as acts of domination of one social class or
group over the other. The state and its policies are reduced to an
instrument of power and repression controlled by the minority
of these powerful actors. We, on the other hand, believe that the
public actors have a certain margin for manoeuvre in their
choices;
• the theories of rational choice (‘public choice’, ‘game theory’)
that assume that for political parties and bureaucrats the only
value that policies have is as currency during electoral calculations
and/or in the appropriation of personal advantages (material and

12
 Theoretical perspectives on policy analysis

immaterial). Much empirical data shows that this theory is

strongly reductionist;
• neo-corporatism and network theory that suggests that the
political-administrative apparatus is in the thrall of organised
sectoral groups and that, as a result, the state remains unable to
develop and implement redistributive interventions for the benefit
of non-organised social groups;
• classical pluralism that defends a vision of the ‘the state as service
hatch’ that is attentive to all social claims and demands and whose
public policies reflect the priorities for action emerging from all
of the members of civil society. However, as experience shows,
numerous social problems are never politically acknowledged as
being worthy of a public policy;
• simple systematisation that does not grant policy actors the
appropriate autonomy and intentionality. The latter’s behaviour
is simply seen as a function of the role assigned to them by their
direct organisational environment. However, social reality is full
of examples that demonstrate the importance of actors even when
the scope for manoeuvre is theoretically very limited;
• the comparative approach in terms of political systems
(‘comparative politics’), which is based on the comparison of
statistics and structural data about political systems and different
public authorities without really analysing the process and more
qualitative aspects of the actual content of public policies
(Hofferbert, 1974). We believe these dimensions are too unrefined
to analyse the substantive policies that interest us;
• the critical approach that refuses to consider any positive or
rationalistic approach to policy analysis and concentrates on
underlining the power and domination dimensions implicitly
associated with concrete public actions (Fischer and Forester,
1993; Fox and Miller, 1995; Fischer, 2003).

In actively refraining from subscribing to one or other of these

theoretical concepts of the state, ‘society’ or any other ‘system’ (merely
touched on here), our approach remains completely open to all of the
hypotheses originating from these theories. The analysis model
presented in this book aims to remain as neutral as possible with respect
to specific theories so as to be able to accommodate – in terms of
working hypotheses to be empirically verified – the broadest possible
range of theories developed in trends as divergent as neo-Marxism,
neo-liberalism and neo-corporatism, on condition that the researchers

13
Public policy analysis

take the trouble to use the concepts in accordance with the basic
dimensions proposed with respect to their empirical field testing.

Notes
1 This analysis is adopted in part from that presented in the work Analyser les
politiques publiques d’environnement (Larrue, 2000).

2 The ‘public choice’ school is based on the work of Buchanan and Tullock
(1962). A critical review of the main principles of this school can be found in
Self (1993).

3 See Simon (1957); Lindblom (1959).

4 The neo-Marxist approach was mainly developed in the 1970s by urban
sociologists like Castells and Godard (1974) and German sociologists, such as
Offe (1972) and Habermas (1973).

5 The neo-managerial approach is based, for example, on the sociology of

administrative elites or, more broadly, the sociology of organisations (Crozier
and Friedberg, 1977).

6 For France, see, in particular, the work of Jobert and Muller (1987) and, for
Germany, that of Lehmbruch and Schmitter (1982).

7 See, in particular, the work of March and Olsen (1984) and our own approach
(Chapter Five, this volume).

8 See Jordan and Richardson (1987); Marsh and Rhodes (1992); Smith (1993).

9 See, in particular, the review of this phenomenon by Emery (1995), Hood

(1995), Pollitt and Bouckaert (2000) and Pollitt (2003).

10 Quermonne defines institutional policies as policies whose main object is

“the production, transformation or decline of public or private institutions”
(Quermonne, 1985, p 62); see also Germann (1996, pp 5-6).

11 See Chevallier’s reflections (1981).

12 Since 1983, policy evaluation has developed significantly in France, from

both an institutional and scientific perspective, for example, the article by Duran
(1993) and more recently Kessler et al (1998). For Switzerland, refer to the

14
 Theoretical perspectives on policy analysis

work of Bussmann et al (1998), which summarises the main message of Swiss

National Research Programme no 27 on ‘the effects of state measures’.

13 See Davies et al (2000) and the emergence in 2005 of the journal Evidence &
Policy (see https://www.policypress.org.uk/journals/evidence_policy/).
14 See Mayntz (1980, 1983).

15 Lester et al (1987) present a good synthesis of the implementation analysis

models developed in the 1970s and 1980s in the US and in Europe. See also
Bohnert and Klitzsch (1980); Parsons (1995); Hill and Hupe (2002).

15