The Dynamics of Public Policy

NEW HORIZONS IN PUBLIC POLICY

Series Editor: Wayne Parsons
Professor of Public Policy, Queen Mary and Westfield College,
University of London, UK
This series aims to explore the major issues facing academics and practitioners
working in the field of public policy at the dawn of a new millennium. It seeks to
reflect on where public policy has been, in both theoretical and practical terms, and
to prompt debate on where it is going. The series emphasizes the need to understand
public policy in the context of international developments and global change. New
Horizons in Public Policy publishes the latest research on the study of the
policymaking process and public management, and presents original and critical
thinking on the policy issues and problems facing modern and post-modern societies.
Titles in the series include:
Public Policy and Local Governance
Institutions in Postmodern Society
Peter Bogason
Implementing European Union Public Policy
Roger Levy
The Internationalization of Public Management
Reinventing the Third World State
Edited by Willy McCourt and Martin Minogue
Political Leadership
Howard Elcock
Success and Failure in Public Governance
A Comparative Analysis
Edited by Mark Bovens, Paul tHart and B. Guy Peters
Consensus, Cooperation and Conflict
The Policy Making Process in Denmark
Henning Jrgensen
Public Policy in Knowledge-Based Economies
Foundations and Frameworks
David Rooney, Greg Hearn, Thomas Mandeville and Richard Joseph
Modernizing Civil Societies
Edited by Tony Butcher and Andrew Massey
Public Policy and the New European Agendas
Edited by Fergus Carr and Andrew Massey
The Dynamics of Public Policy
Theory and Evidence
Adrian Kay

The Dynamics of Public

Policy
Theory and Evidence

Adrian Kay
Department of Politics and Public Policy, Griffith University,
Brisbane, Australia

NEW HORIZONS IN PUBLIC POLICY

Edward Elgar
Cheltenham, UK Northampton, MA, USA

 Adrian Kay 2006

All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system or transmitted in any form or by any means, electronic,
mechanical or photocopying, recording, or otherwise without the prior
permission of the publisher.
Published by
Edward Elgar Publishing Limited
Glensanda House
Montpellier Parade
Cheltenham
Glos GL50 1UA
UK
Edward Elgar Publishing, Inc.
Thomas Pratt House
9 Dewey Court
Northampton
Massachusetts 01060
USA

A catalogue record for this book

is available from the British Library

Library of Congress Cataloging in Publication Data

Kay, Adrian
The dynamics of public policy : theory and evidence / Adrian Kay.
p. cm. (New horizons in public policy series)
Includes bibliographical references and index.
1. Policy sicences. 2. Policy sciences Case studies. I. Title. II.
Series: New horizons in public policy.
H97.K39 2006
320.6dc22
2006012076

ISBN-13: 978 1 84542 105 2

ISBN-10: 1 84542 105 1
Printed and bound in Great Britain by MPG Books Ltd, Bodmin, Cornwall

Contents
vi
vii

Abbreviations
Preface and Acknowledgements
1

Introduction: why we need dynamic perspectives

PART I
2
3
4
5

6
7
8
9

THEORY

Policy dynamics and history

Path dependency
Evolutionary perspectives
Structured policy narratives

PART II

17
29
42
59

EVIDENCE

The development of the EU budget system

The Common Agricultural Policy 19772003
The GP fundholding scheme
UK pharmaceutical policy

77
90
104
118
131
143

References
Index

Abbreviations
ACTU
AFR
BBR
BMA
CAP
CCT
CE
CoAM
CRL
DHA
DoH
EU
FE
GATT
GDP
GNP
HI
IIA
NCE
NGO
NHS
NHSE
NICE
OECD
OR
PACT
PCGs
PCTs
PPRS
R&D
SFP
VIL
VPRS
WTO

Australian Council of Trade Unions

annual financial return
balanced budget rule
British Medical Association
Common Agricultural Policy
common customs tariff
compulsory expenditure
Council of Agricultural Ministers
Co-Responsibility Levy
District Health Authority
Department of Health
European Union
functional explanation
General Agreement on Tariffs and Trade
gross domestic product
gross national product
historical institutionalist
inter-institutional agreement
non-compulsory expenditure
non-governmental organization
National Health Service
National Health Service Executive
National Institute for Clinical Excellence
Organisation for Economic Co-operation and Development
own resources
prescription analysis and cost
primary care groups
primary care trusts
Pharmaceutical Price Regulation Scheme
research and development
single farm payment
variable import levy
Voluntary Price Regulation Scheme
World Trade Organization
vi

Preface and acknowledgements

During the writing of this book, I moved to a new job in Australia. This slowed
down the writing of the book significantly but prompted some new reflections
on policy dynamics and the incorporation of references to the situation in
Australia to complement the empirical chapters based on my previous work on
UK and EU public policy. If the reflective pause in writing ever tried the
patience of my publisher, Edward Elgar, it never showed: thank you to
Catherine and her colleagues for supporting this project through to
completion.
In trying to think about temporality in policy analysis I have benefited
greatly from many discussions with former colleagues at the School for Policy
Studies at the University of Bristol, as well as my new ones at the Department
of Politics and Public Policy at Griffith University.
I would like to thank the publishers of the following journals for permission
to draw on material, in heavily revised form, I have published previously:
(2003), Path dependency and the CAP, Journal of European Public Policy,
10(3), 40521; and (2001), Beyond policy community: the case of the GP
fundholding scheme, Public Administration, 79(3), 56177.
My greatest debt is to Siwan and I dedicate this book to her. Dw in dy garu
di.
Adrian Kay
Brisbane, March 2006

vii

1.

Introduction: why we need dynamic

perspectives

WHAT IS DYNAMIC ANALYSIS?

This is a book about the dynamics of public policy. Like many well-used and
widespread terms in the social sciences the idea of dynamic perspective or
analysis, whilst intuitive and appealing, is difficult to define precisely in a
manner that will cover all of its different uses. For example, even within the
field of economics and its commitment to a positivist science, the Nobel
Laureate Paul Samuelson (1947, p. 311) was able to say that: often in the
writings of economists the words dynamic and static are used as nothing
more than synonyms for good and bad, realistic and unrealistic, simple and
complex. In his Essays in Economic Semantics, the Austrian economist Fritz
Machlup (1975, p. 10) offered the view that: Typically, statics was what
those benighted opponents have been writing; dynamics was ones own
vastly superior theory.
Precise definitions do exist: for example, Samuelsons own formulation that
dynamic analysis refers to models in which time is an independent variable
would be recognized by students of economics. However, this limited
definition is of little utility outside the formal models of economics and does
not capture any substantial sense of the concept of dynamic analysis as it is
used variously in the social sciences. Instead this definition is an exemplar of
how economics forgot history, the title of Geoffrey Hodgsons investigation
of the long-standing difficulty of time and historical specificity in the social
sciences (Hodgson 2001).
It is precisely the difficulty that economists have had in modelling time that
reveals the value and importance of the dynamic perspective to the analysis of
public policy. Any policy process is a complex system and dynamic models of
complex systems are much more difficult to construct than static ones (which
is why dynamic models are also less well developed in biology and physics
(Auyang 1998)). The difficulty arises because there are typically several
processes with different speeds going on at the same time. This makes
separating different time scales such as the short, intermediate or long run as
essential to understanding and explanation as distinguishing between
different spatial scales, as in the macro, meso, micro and decision levels
1

The dynamics of public policy

common in policy studies (for example, Parsons 1995; Hudson and Lowe
2004).
Most importantly economics, like public policy, depends on human
decisions. All decisions are made in historical contexts; they are inevitably
influenced by the legacy of the past and the uncertainty of the future. Thus the
description of individual decisions requires a sense of memory and
expectation; these are subjective, personal and partial and therefore not suited
to the formalism of economic models, notwithstanding theoretical advances in
game theory on memory and expectation in repeated games.
One common and essential element in most writers use of the term policy
is purposiveness of some kind (Parsons 1995, pp. 1316). Policy expresses a
general set of objectives or a desired state of affairs. These are constrained by
a sense of possibility driven by legacies and forebodings. In a well-known
definition of policy, public policy is: anything a government chooses to do or
not to do (Dye 1972, p. 2). Policy is about choice: the choice of objectives;
the choice of reasons for (in)action; the choice of policy instruments; the
choice of how to respond to the consequences of policy outputs. These
choices, their consequences and subsequent choices unfold in a temporal
process in which uncertainty is a defining feature.
In this book dynamic analysis is not a conceptual framework or theory, nor
is it amenable to a precise, pithy definition that allows a succession of
theoretical corollaries in terms of policy dynamics. Rather it is a perspective
or a way of viewing the world more akin to a methodology in the sense of
prompting a series of questions about methods of inquiry than a substantive
theory. The idea of policy dynamics is not original; it is proclaimed in the titles
of Rose (1976), Harrison et al. (1990), and Baumgartner and Jones (2002).
However, the intellectual underpinnings of dynamic analysis remain
unexplored in the policy theory literature. The case for the importance of
understanding policy dynamics in the field of policy studies is put forward in
this chapter. This also gives context to the assessment in later chapters of the
different frameworks, concepts and theories that might be used for the purpose
of dynamic policy analysis.
In exploring the foundations of a social-scientific approach to dynamic
analysis, the issue of history in policy studies ineluctably emerges along with
the broader epistemological question of the relationship between socialscientific explanations and historical explanations. Indeed, this book aims to
make progress on the challenge set out by Pierson (2004, p. 5): The
declaration that history matters is often invoked, but rarely unpacked.
Something similar is repeated in Schwartz (2004), Thelen (2003) and
Mahoney (2003). In alternative terms, Reynolds (1999, p. 277) observes that,
without elaboration, the claim that to be properly understood things must be
considered within their historical context amounts to mundane historicism.

Introduction

The book addresses the notion of dynamics as a term that is widely, if

inconsistently, used within the social sciences, has been applied in the study of
policy, but is currently theoretically underdeveloped in policy terms. The
nature of what is being studied with a dynamic perspective temporality and
change at different scales demands notice of history qua academic
discipline, or more accurately historiography: dynamic analysis is the use of
concepts and theories to understand and explain longitudinal data of policy
development. I argue in Chapter 2 that there is a deep underlying common
structure involved in dynamic policy analysis and writing policy history;
however such a formulation faces the constraint that different disciplines have
different methods and analytical styles, as well as the acceptance of different
modes of explanation as valid. The second chapter establishes the narrative as
the appropriate methodology for policy dynamics; I go on to develop specific
standards of inquiry to construct and assess policy narratives in Chapter 5.
This chapter has three aims. The first is to set out the intellectual
foundations of dynamic analysis from a social science perspective in terms of
three notions: temporality, change, and different processes and scales. The
second aim is to establish public policy as a unit of dynamic analysis. Finally,
the structure of the book is outlined.

TEMPORALITY
There is a distinction between temporality and ordinary, common sense
conceptions of time. Broadly there are two categories of answers to
Heideggers famous question: what is time? The first has time as something
which exists independently of things and events; the second sees time as ideal,
in the sense that things are not temporal without temporal concepts. It is
beyond the scope of this book to investigate this question fully, however for
our purposes an answer does help for the understanding of what is meant by a
dynamic perspective. The assumption in this book is that time is an abstract,
imaginary notion; whereas temporality refers to how we make events or
experiences intelligible in terms of time. Time cannot be separated from
things, events, processes and is inherent in all empirical entities.
This allows two things to be set out immediately. First, any policy choice or
decision depends on knowledge of options and thus the concept of possibility
is fundamental. Next, possibility is inalienably linked to temporality: without
sense of a past, present or a future, the notion of possibility and thus choice
does not make sense. Possibility enters into most social scientific theories,
implicitly or explicitly, by the concept of state space. This is the structured
collection of all the possible momentary states of an individual thing. For
example: consumption and production sets in microeconomics; or the capacity

The dynamics of public policy

of a health system to produce certain health outputs; or the range of

predictions in a model of poverty over the life course. The individual thing in
this book is policy, so the term policy space is used interchangeably with
state space throughout the book. The concept of state space is the foundation
of dynamics and the use of dynamic perspectives; a dynamic perspective on
policy consists of examining successive states of a policy system and the
relationship between them.
This distinguishes dynamics from the comparative statics approach that
does compare certain states of a system but fails to consider the relationship
that links the states through time. Whilst it is possible to have a theory of
change within a comparative statics perspective that provides reasons why
some states will change to another, two important elements will be missing.
First, an account of the process of adjustment between the two states, and
some analysis of the connecting path. Second, temporality: the separation of
the two states is atemporal, as the snapshots (which may have temporal
identifiers to the extent that we can say one occurs before the other) are of two
states being compared for reasons other than temporality they may be salient
political events, for example. In a dynamic perspective on the other hand,
temporality is central. The purpose is to compare a system as it passes through
consecutive temporal states, and these consecutive states may come slowly or
quickly; thus a system that changes slowly would have different temporal
identifiers to one that changed quickly. And, of course, by making the claim
that something changes slowly, we are using some concept of time. This
establishes the point that just as we can separate scales of analysis in the policy
process, so we can show different rates of change. This is how I understand the
frequently cited aphorism that the social sciences should fulfil an ambition for
movies and not stills.
For Pierson (2004, p. 2) the key to temporal analysis is: systematically
[italics in the original] situating particular moments (including the present) in
a temporal sequence of events and processes stretching over extended
periods. However I argue that temporality involves more than just situating
different states; rather the ambition for movies is to trace a path between all
these different moments or states realized at various times. In principle these
paths may be described by deterministic dynamics; in formal social science
modelling this is usually in form of a set of functional equations in which time
is an independent variable. Alternatively there may be stochastic dynamics,
where the probabilistic element is limited by the set of all possible states in the
model or theory. However, the crucial point for this book and for policy
dynamics is that in the policy process the possibilities are so numerous and
heterogeneous, as well as unexpected and unintended, that they cannot be
circumscribed in state spaces. Therefore, I argue, deterministic or stochastic
models of policy dynamics are incapable of prediction, either prospectively or

Introduction

retrospectively, and thus fail one of the defining standards of social science
research.
Unpredictability is not, however, inexplicability; as explanation has the
benefit of hindsight. All the concepts, theories and frameworks of policy
dynamics covered in this book contribute to narrative explanation. Chapter 2
stresses that the notion of possibility, inextricably bound with temporality, is
central to narratives of policy dynamics. Most explanations of evolutionary
events are narrative and share with policy studies a focus on unique sequences
and processes; indeed, the limited applicability of state space indicates the
limit of generalization in policy studies. This is one of the reasons why
evolutionary approaches to policy dynamics are considered in Chapter 4, and
ideas of memory in Chapter 5.

THE NOTION OF CHANGE

There are four ideas involved in the notion of change: an enduring thing; its
various possible states; the identification of an initial and a final state by the
temporal index; and the characterization of these states. The logical structure
of change is of the following form: the thing changes from state S1 at time T1
to state S2 at time T2. T1 and T2 pinpoint, temporally, the states that S1 and S2
characterize. But without some idea of unity that binds the two states, we can
only say that they are different. We need the concept of the thing to bind the
two states, that is something, some element of the entity, must endure for
change to occur.
The characterization of the thing involves a type-value duality. A things
potential to change is limited by the range of possible states admissible for the
type set of which it is a member. If the thing is education policy, for example,
only certain policy states are possible for that type; that is, only certain things
can be education policy. If the boundary of possibility is overstepped, the thing
in question becomes another thing rather than a different value of the same
thing. There is a crucial distinction in logical terms between changes in things
and changes in kind.
How to make progress on this distinction in practical terms? I start with the
proposition that a thing changes substantively when its states at different times
have different characteristics. The successive states constitute the things
history. They can also be interpreted as the stages of the process the thing
undergoes, as in policy process where things change form, such as with the
generic stages model: from idea, to proposal, to legislation, to implementation
plan, to monitoring and evaluation framework. A thing need not change
substantively, but an unchanging thing still has a history and undergoes a
stationary process.

The dynamics of public policy

The endurance of a system is embodied in a path in state space that connects

the observed states of a system identified by a unique temporal reference
point. The temporal reference point establishes the order of the states in the
system and substantive differences between them. The identity of the thing
through time, the endurance, raises the troublesome philosophical question: if
a changing thing really changes, it cannot literally be one and the same thing
before and after the change; however, if a changing thing literally remains one
and the same thing (that is, it retains its identity) throughout the change, then
it cannot really have changed.
This is not an abstract irrelevance. The debate over public policy toward
foetal research, for example, emphasizes that debates about things and values
are politically contested: when does the thing, human life, begin and
therefore the legitimacy of scientific experimentation? The identity-throughtime difficulty is compounded where the thing is composite and its
constituents are in flux, as is the case where the thing is policy. For example,
is social policy under a Keynesian welfare state a different thing from neoliberal social policy, rather than a different value of the same thing? While the
economic policy of les trente glorieuses in France is obviously different from
that under the Maastricht Treaty process in the 1990s, important continuities
remain: it is a different thing with similarities to the previous thing. It will
emerge in the book that the notion path-breaking policy change, a change in
kind as well as value, forms an important part of debates in the field of policy
studies as in Chapter 8, for example, which looks at policy dynamics
following the break up of a well-established policy paradigm in UK health
care policy.

A MULTITUDE OF PROCESSES AND TEMPORAL SCALES

Robinson (1979, p. 286) states: Logical time can be traced from left to right
on the surface of a blackboard. Historical time moves from the dark past
behind it into the unknown future in front.As noted, time is the parameter that
distinguishes and identifies the various states of an enduring thing and is
inherent in the general concepts of endurance and the processes that occur.
However, the temporal parameter is defined individually for each thing or
process. There are therefore many kinds of time, and this is a characteristic
that can distinguish the different states of a system.
The major trend of a system is not the only ongoing process; there are a
multitude of processes proceeding together, each with its own pace and
temporal structure. So, economists talk about the short term, or intermediateterm adjustment, or the long term or cycles. Policy analysts in government
employing costbenefit techniques examine immediate and medium-term

Introduction

impacts. Wanna et al. (2000) discuss how the budget-making process at the
federal level in Australia includes requirements for 5-year fiscal impact
analyses of policy decisions and periodic intergenerational equity reports that
use a 30-year time horizon. The relative magnitudes of the different temporal
processes help in the understanding of which process dominates over a
particular timescale, which can be held stationary, or which might be
averaged. Whilst policy theory discusses time horizons, uses notions of time
consistency when critiquing policy design and contains discussions of term
limits for politicians, there is little systematic analysis of different rhythms,
cycles and process speeds in the policy process.
The notion of different temporal scales further limits the utility of static
analysis. Once time is frozen then it is impossible to observe different
velocities. The idea that what is observed from a snapshot has any relevance at
any time in the future is doubtful because the snapshot picture will change
very quickly when there are a multitude of temporal processes occurring; in
metaphorical terms, taking a snapshot of a kaleidoscope will not have much
value in trying to understand what preceded and succeeded the point at which
the picture was taken.
Another related limitation of static analysis is that it does not allow for the
idea of inertia, a key characteristic of policy change. A static approach
assumes away inertia and resistance once a factor driving change has been
nominated and an outcome position identified, but the dynamics question is:
how do you get from one to the other? What is the resistance involved in terms
of countervailing power, institutional inertia, bureaucratic entrenchment or the
reassertion of old thinking or analysis?

POLICY AS A TEMPORAL STATE

This brings us to the point that all books about policy must address: what is
policy? This is the thing that both changes and endures over time, and arouses
an interest in dynamic analysis. Hogwood and Gunn (1984, pp. 1119) discuss
a number of the common uses of the word policy: policy as a label for a field
of activity (for example, foreign policy); policy as an expression of general
purpose or the intended path towards a desired state of affairs; policy as a
specific proposal; policy as a decision of government; policy as a formal
authorization (for example, legislation); policy as a programme of activity;
policy as outputs or what governments actually deliver, as opposed to what it
is promised or authorized through legislation; policy as outcomes or what is
actually achieved; and policy as a theory or model (the notion that if we do X
then Y will follow).
Such definitional breadth is what attracts scholars to the brevity of Dyes

The dynamics of public policy

definition of public policy, or Howlett and Rameshs (2003, p. 3) offering that

Public policy is, at its most simple, a choice made by a government to
undertake some course of action. Stone (2001, p. 7) argues that definitions of
policy in terms of choice, or an action calculated to achieve a desired
objective, or the purposeful connecting of ends with means are premised on a
classical view of policy as the result of a rational process: the model of
policy making in the rationality project is a production model, where policy is
created in a fairly ordered sequence of stages, almost as if on an assembly line.
Of course, most policy textbooks readily acknowledge that this is an
oversimplified model: policies usually involve a series of interrelated
decisions; rather than a single decision-maker, many different people at
different levels and scattered throughout government organizations make
public policy decisions; policies are shaped by earlier policy decisions and
environmental factors; policies are mediated through their implementation;
policies involve both actions and inactions; policies cannot be analysed apart
from the policymaking process; policies have outcomes that may or may not
have been foreseen; policies are subjectively defined, and may be defined
retrospectively; policies extend beyond the formal records of decisions;
and policies need resources and action to be differentiated from political
rhetoric.
Nevertheless, despite the limitations of the classic view there is an
entrenched belief, particularly among policymakers, that policy should
achieve a desired change in the wider population. For example, as with the
evidence-based policymaking initiative in the UKs Cabinet Office where
Professor Ron Amman attempted to introduce something like a stages model
to improve policy-making (Centre for Management and Policy Studies
2001). Further, much of the recent policy-learning literature is premised,
usually implicitly, on some underlying rationality in the policy cycle where
policymakers use trial-and-error procedures in order to find the best solution
to a problem. First, policymakers develop a hypothesis about the best way to
achieve an objective (the causal model from inputs to outcomes noted
previously). They then test their hypothesis, that is, they implement their
policies and analyse their impact. From their analysis, policymakers can come
to a conclusion about how well their policies work and whether they should be
continued, improved, implemented in another way or terminated.
The notion of a policy cycle, prominent in the classical view, has its origin
in systems theory and the pioneering work by David Easton on political
systems (Easton 1965, 1966). According to Colebatch (1998) the policy cycle
imagines the policy process as an endless cycle of policy decisions,
implementation and performance assessment. Howlett and Ramesh (2003)
conceive of a similar cycle but with more steps: agenda setting (problem
recognition); policy formulation (proposal of a solution); decision-making

Introduction

(choice of a solution); policy implementation (putting the solution into effect);

and policy evaluation (monitoring results). Hogwood and Gunn (1984) also
envisage a cycle: issue search or agenda setting; issue filtration; issue
definition; forecasting; setting objectives and priorities; options analysis;
policy implementation; evaluation and review; and policy maintenance,
succession or termination.
This generic class of policy cycle models is idealized, but useful here for
showing what challenges and demands a dynamic perspective brings: all these
models of stages, cycles and learning, and variants thereof, even when they are
sophisticated enough to explain why particular decisions are made, do not
pinpoint or offer what drives policy from one stage to the next. In our terms
here, the notion of a cycle is an exercise of comparative statics rather than
dynamic analysis. The process bit of the policy process most often goes
un-theorized and is assumed away.
Policy cycle models fail to embrace the complexity of the policymaking
process and the reality that policy rarely, if ever, develops in a linear
progression. Stages are often skipped or compressed and the idiosyncrasies,
interests, preset dispositions, policy paradigms or mental maps of the actors
involved often usurp the sense of a smooth process. There are a multitude of
different processes at different scales and at different speeds occurring
simultaneously.
There is an alternative to the classical view of policy, that Colebatch (1998,
p. 102) labels the structured interaction model:
The structured interaction perspective does not assume a single decision-maker,
addressing a clear policy problem: it focuses on the range of participants in the
game, the diversity of their understandings of the situation and the problem, the
ways in which they interact with one another, and the outcomes of this interaction.
It does not assume that this pattern of activity is a collective effort to achieve known
and shared goals.

The interaction view recognizes that policy is an ongoing process with

many participants, most of whom do not have a formal or recognized role in
policymaking. They include ministers of state, their advisers, politicians,
public servants, party members, street level delivery staff, peak bodies,
interested members of the public, the media and academics. According to this
view, policy is not about the promulgation of formal statements but the
processes of negotiation and influence; indeed, much policy work is only
distantly connected to authorized statements about goals: it is concerned with
relating the activities of different bodies to one another, with stabilizing
practice and expectations across organizations, and with responding to
challenge, contest and uncertainty (p. 102).
Stone (2001, p. 208) provides the most memorable description of the

10

The dynamics of public policy

structured interactionist view, and one that emphasizes its corollary in terms of
the need for dynamic analysis:
policy is more like an endless game of Monopoly than sewing machine repair.
Hence the common complaint that policies never seem to solve anything. The
process of choosing and implementing the means of policy is political and
contentious. The actions we commonly call new policies are really somebodys
next move, and in politics, as in a good game, nobodys move completely
determines anybody elses future move.

THE POLICY SYSTEM

One consequence of accepting the complex interaction view of policy over the
classical view is that it presents policy as a multi-level phenomenon. There is
no unique level or scale but rather several levels that may be examined as
policy, as Heclo (1972, p. 84) puts it: As commonly used, the term policy
is usually considered to apply to something bigger than particular decisions,
but smaller than general social movements. Thus, policy, in terms of levels
of analysis, is a concept placed roughly in the middle range.
The first scale in this middle range is that of the policy system. In these
terms, policy is a whole, or system, as in discussions of health policy,
defence policy or housing policy. It is not just journalistic shorthand to talk
about policy development; there is a whole or a system at a policy level that
can be the subject of active and passive verbs and the object of empirical
investigation without stretching the limits of our imagination too far. A policy
system is a complex, composite variable consisting of many interrelated
elements. Within a policy system there may be several policy subsystems (or
elements), each with their set of actors, organizations, goals and instruments
(Baumgartner and Jones 2002, Chapter 1). For example, within the health
policy system there are inter alia the primary care policy subsystem, the
hospital policy subsystem and the public health policy subsystem. Using a
dynamic perspective, the development of policy subsystems may equally be
understood as the policy whole itself.
Rose and Davies (1994) argue the policy programme level should be the
basic unit of analysis in policy studies because it is most readily observable. A
policy programme refers to a specific combination of laws, commitments,
appropriations, organizations and personnel directed towards a more or less
clearly defined set of goals. In other terms, this is a policy instrument: an
identifiable tool or resource of government used for a specific set of purposes.
This is a more finely grained perspective than the policy subsystem and I
develop the argument in Chapter 5 that this should be the microfoundation for
policy narratives.

Introduction

11

One corollary of different policy scales is a lack of precise frameworks for

measuring policy, or alternatively locating policy continuity and policy
change. Judging policy change is difficult because even if at the macroscopic
or policy system level there is limited change and policy is considered stable,
there may be concurrently change observed at the policy subsystem level
or programme/policy instrument level. In the context of economic policy, Hall
(1993) sets out three orders of policy change: change in the level of policy
instruments (first order); change in the instruments actually used (second order
change); and change in the overall policy paradigm (third order change, which
occurs at the policy system level). Under this categorization, policy change
and policy stability may be simultaneously observed. Halls work is important
because it allows the crucial logical distinction between changes in things and
changes in kind to be made in terms of policy. It gives us categories of things
(policy instruments and their different levels) and kind (policy paradigm). This
allows us to assess whether an existing policy is changing or a new policy is
being introduced.
For Hall, a policy paradigm is an interpretative framework that operates in
the policymaking process. Specifically, it refers to the framework of ideas and
standards that specifies the goals, instruments and the very nature of a policy
issue. On this approach, how policymakers interpret and use evidence to
construct the notion of a policy, and options for reform, greatly affects policy
development. Such an interpretive framework or mental map may be path
dependent and sticky (Denzau and North 1994). Importantly, just as the
Kuhnian model challenges the conception of scientific activity approximating
to certain canons of pure rationality, so will any parallel model applied to the
policy case; thus policy paradigms belong outside the classical view of rational
policymaking.
Alternatively, Pritchard (2002) shows how it is possible to generalize away
from science by using Wittgensteins notion of hinge propositions. These are
the guiding assumptions of a certain activity; assumptions that inform and
restrict the choices taken. These guiding assumptions are forms of tacit
knowledge; it is not that agents do not know the assumption at issue this is
not a situation of incomplete information rather they do not recognize that
they are making an assumption. They are ingrained in how agents construct
their situations, their decisions and their actions. Neither policy paradigms nor
hinge propositions are evaluated or directly articulated in the policy process,
but they operate to reduce the range of possible alternative courses of action
that are scrutinized in the decision-making process. Situated agency in policy
studies is discussed in Chapter 5, but note here that I will use the concept of a
policy paradigm rather than a hinge proposition throughout the book, on the
grounds that this term is commonly used in the policy studies literature; in
many ways the latter notion is better however, because the implication of the

12

The dynamics of public policy

label of a policy paradigm is that these are the equivalent of Kuhnian scientific
paradigms in their completeness, internal coherence or resistance to
disconfirming evidence, whereas in Chapter 8 the argument is presented that
this is not always the case.
The emphasis on a multitude of processes at different speeds from a
dynamic perspective complements this multi-level view of policy. Throughout
the book I will use the terminology of events to describe abrupt changes of
things; with more gradual change labelled as a process. A process is of the
form S(t), where t is the temporal identifier or index, and S, the state of a
system. Thus, S varies in accordance with the variation in the temporal
identifier: a dynamic perspective. Following this notation the derivate, dS(t)/dt,
can be interpreted at a specific time as an event, the almost instantaneous
change in system.
The distinction between an event and a process helps avoid the dualism
between policy stability and policy change that sometimes affects policy
studies. From a dynamic perspective it is all change, because even where the
state of the system does not change in type or value, S at t1 is different from S
at t2, and so on. This analysis helps to get us beyond continuity (Streeck and
Thelen 2005) in our thinking about policy development. Instead the notions of
events and processes help decompose sequences of policy development into
temporal parts such as periods, stages, phases, movements or epochs. The
method of periodization raises questions for narrative explanation and the
construction of policy histories: these are dealt with in Chapters 2 and 5.

POLICY AS AN INSTITUTION
Institutions are central to contemporary social science theory. Indeed, an
acceptance of the importance of institutions for social and political
development is one of the few genuine cross-disciplinary agreements. This has
produced a variegated set of institutionalisms (see Hall and Taylor 1996, see
also contributions in Goodin 1998). The attraction of institutions, both to
policymakers and analysts, is that they help give a structure to a world that is
complex and in which there are a multitude of temporal processes underway
at different levels. Institutions are collective constraints; organized patterns of
socially constructed norms or roles with prescribed behaviours expected by
the occupants of these roles, which are created and recreated over time.
Institutions help provide a buffer against the uncertainty of interaction among
policy actors and perturbations external to the policy process. Institutions are
enduring, regular and tend to be difficult to change; as such they provide an
important part of the temporal context of policymakers.
As noted earlier, within the policy system whole there are various

Introduction

13

structures at different scales that act as institutions in shaping agents decisionmaking in the formulation and implementation of policy. These are not
reducible to individual level agents or elements in the policy process.
Examples of such policy institutions are budget rules, policy networks,
standard operating procedures in government departments, and agencies. Most
importantly in terms of understanding policy development, past policy
decisions are institutions in terms of current policy decisions: they act as
structures that can limit or shape current policy options. Institutionalism is an
important way of thinking about policy legacies, how policies accumulate and
gradually institutionalize. The ambition to understand how policy histories
affect policy in the present is what drives this project of dynamic policy
analysis: as Oakeshott says, we do not have a blank sheet of infinite
possibility in a policy area; the options for future policymakers are restricted
by past policy paths.
The conceptual distinction between a policy and an institution is significant;
to conflate the two would blur the object of analysis in policy studies.
Nevertheless, in certain circumstances a policy can act as an institution. As
Pierson (1993, p. 596) states: major public policies also constitute
important rules of the game, influencing the allocation of economic and
political resources, modifying the costs and benefits associated with
alternative political strategies, and consequently altering ensuing political
development.
In Pierson (2004, pp. 15051, 1656) policies are institutions, and although
they are less foundational than formal political institutions they can develop
in a manner complementary to and interlocking with those formal institutions.
Further, the application of the voluminous and important literature on
institutional development to the study of public policy represents a significant
research frontier (p. 165). I agree, and throughout the book will be using
concepts that have been developed for the analysis of institutional dynamics
for the purposes of understanding policy development, such as path
dependency for example, in Chapter 3.

OUTLINE OF THE BOOK

The book is in two parts. The first part consists of four chapters that take a
theoretical perspective on policy dynamics. In Chapter 2 the central role of
history in policy dynamics is established, both in terms of the effect on the
present of things that happened in the past and its place as an academic
discipline. In discussing the different modes of explanations in the social
sciences and historical disciplines, the argument is developed that narrative
explanations are the most appropriate means for explaining policy dynamics.

14

The dynamics of public policy

Further, the use of narratives in this context, far from being perceived as a
weakness or some lack of analytical capability (as some formal social
scientists maintain), should be viewed as the valid method of inquiry for
answering the questions raised by a dynamic perspective on policy
development. Chapters 3 and 4 assess critically two broad concepts as means
of organizing or structuring policy narratives. Chapter 3 looks at the
increasingly popular notion of path dependency and its application to policy.
What advantages and disadvantages does the concept have in understanding or
explaining chronicles of policy development? Chapter 4 subjects evolutionary
theory to similar scrutiny in a policy development context. On the basis of
these theoretical investigations, Chapter 5 sets out how to structure policy
narratives to explain policy dynamics. The chapter acts as both a summary of
the preceding analysis and a methodological guide for the empirical chapters
that follow: what is a narrative? How do you structure a narrative? How do you
evaluate a structured narrative? What makes it convincing or valid, successful
or true?
Chapters 69 constitute Part II of the book. These are all case studies of
particular policy dynamics. Theodore Lowis venerable, but imperfect, fourfold typology of constituent, redistributive, distributive and regulatory policies
has been used to select the cases. I make no claim that these are in any sense a
representative sample of policy dynamics, only that Lowis typology allows
four different policy contexts to be considered in the book. The EU budget is
an example of a constituent policy in Lowis terms in that it involved the EU
adopting a series of decision rules for subsequent policymaking. Chapter 6 will
consider the structure and history of the EU budget with particular attention to
the development of budget rules and their complex inter-dependency over time,
from the mid-1980s onwards. The EUs Common Agricultural Policy (CAP) is
described as a redistributive policy in Chapter 7, which will outline the
argument that although the CAP has been the subject of five reforms in 18 years
these reforms have for the most part been minor, and the key policy dynamic
that needs explaining is resistance to reform. Chapter 8 looks at UK primary
care between the late 1980s and 1997 as a distributive policy. The GP
fundholding scheme was introduced and repealed within seven years in the
1990s and the case illustrates the dynamics triggered by the collapse of a wellestablished policy paradigm and the subsequent instabilities where a major
reform initiative fails to be institutionalized in the policy system.
Pharmaceutical policy is considered in Chapter 9 as a regulatory policy in
Lowis terms. UK pharmaceutical policy has changed significantly since the
early 1990s, towards an emphasis on controlling NHS demand for medicines.
The policy dynamic is of an increasingly complex policy space with new
initiatives layered onto existing policies, creating new and potentially
contradictory interactions between demand-side and supply-side regulations.

PART I

Theory

2. Policy dynamics and history

In Chapter 1, the path linking successive temporal states of some entity was
defined as the history of that entity. When an entity, such as policy, is complex,
composite and polyvalent, there can be several different immanent histories.
This chapter aims to establish the role of history in policy dynamics, both in
terms of the effect of things that have happened in the past and of history as
an academic discipline. The first section is epistemological; it discusses what
counts as a valid explanation in historical and social-scientific disciplines. The
next section develops the argument that a dynamic perspective with socialscientific foundations, as outlined in Chapter 1, can draw on history qua
academic discipline to make a genuine and distinct contribution to policy
studies; in particular in emphasizing the utility of the narrative in
contradistinction to more formal models of dynamics that often fail social
science standards of good modelling. The third section illustrates this claim by
using a dynamic perspective to problematize the notion of policy choice,
suggesting that the narrative is the appropriate method to render the
complexities and conjunctural contingencies in the policy process tractable for
analysis.

HISTORY AND THE SOCIAL SCIENCES

Any claim of a historic turn in the social sciences, see for example McDonald
(1996), should be accompanied by the observation that this is a return rather
than a new trend; for example Hodgson (2001) charts the rise and fall of the
question of the role and nature of historical specificity in the social sciences.
Indeed, an emphasis on history is a defining feature of heterodox traditions in
economics, for example in the works of Georgescu-Roegen (1967, 1971) on
the possibility of an entropy law and Kaldor (1970) on cumulative causation.
Beyond economics and earlier, writers such as Marx, Hegel, Weber and
Toynbee all argued for some form of historicism in the sense of general laws,
rhythms or patterns to history that allow the social sciences to make
predictions, either prospectively or retrospectively.
It is in the ambition for some degree of generalizability that a potential
schism between a social-scientific view of society and a historical perspective
17

18

Theory

can be discerned. To assist exposition, this schism can be simplified and

slightly caricatured as follows: social science is essentially a nomothetic
exercise: it endeavours to establish laws or mechanisms through the scientific
method, based on the premise that social entities, like natural entities, are
repeatable and recurrent; in contrast, history does not repeat itself, however. It
occurs only once in a certain context and at a certain time. Thus, any
examination of history can only be idiographic, that is, a description of events
that have already occurred and are unique.
Although this is not a book about the epistemology of the social sciences,
in order to develop the place of dynamic analysis in public policy theory it is
necessary to consider the longstanding debate about the role of history in the
social sciences. This is revealed most clearly in the question of what counts as
a valid explanation. Van Parijs (1981, Chapter 1) submits that an explanation
of fact x consists of answering the question why x? in an appropriate way.
Significant differences exist within and between social-scientific and historical
approaches over what counts as appropriate. Progress on this question can be
initiated by asserting that the structure of any explanation is governed by two
formal conditions: a condition of causality and a condition of intelligibility.
The first condition is the assertion of a causal link or production relationship
between two facts. The second condition says that explanation requires
intelligibility by suggesting the mechanism through which the causal link is
produced. As long as the nature of the underlying mechanism is left
unspecified, the explanation remains a black box explanation; and the why
question is not answered because the fact being explained is not rendered
intelligible. In alternative terms, we cannot explain without understanding,
where this consists of imagining a plausible mechanism, through which the
fact to be explained (x) is brought about, created or produced.
In any given academic field only certain types of mechanisms are
considered plausible. Thus it is possible to reformulate the condition of
intelligibility in terms of any explanation conforming to one of the accepted
patterns of intelligibility, defined by the types of mechanisms recognized as
plausible within a particular academic community. At the deep structural level
of explanation outlined by Van Parijs, there is no idiographicnomothetic
schism; rather this divide emerges in the strong differences in epistemological
and ontological preferences within different social science disciplines about
which mechanisms can render phenomena or processes intelligible.
From a nomothetic perspective, one of the pre-eminent contemporary
writers on the nature of mechanisms and their role in social science
explanation is Jon Elster. For him mechanisms are frequently occurring
and generally recognizable causal patterns that are triggered under generally
unknown conditions or with indeterminate consequences (Elster 1998, p. 45).
The ubiquity of the language of mechanisms owes something to the ambition

Policy dynamics and history

19

for covering law explanations (or deductive-nomological explanations,

Hempel 1965) in the social sciences and among some historians (Roberts
1996; Roth 1999); that a phenomenon can be explained by being subsumed
under a covering law that is, the particular relationship can be deduced from
the general law that covers it. This ambition combines with a persistent failure
to uncover generalizations that might be formulated as a law in historical
social science to increase the attractiveness of the notion of mechanisms;
although not desirable in themselves they allow explanation when law-like
generalizations do not appear. To offer an explanation of an outcome or event
(the explanandum) requires the statement, after the event, of a set of initial
conditions and a mechanism that connects these initial conditions to an
outcome. This mechanism is more general than the phenomenon that it
subsumes and it is the reference to a more general category that renders
intelligibility. Goldstone (1998) emphasizes strongly this important point: the
adequacy of mechanisms in contributing to explanation relies on their
generality rather than their level; reduction to more micro-level phenomena is
not necessary for explanation.
Often, and especially in historical disciplines, why questions are answered
in a fully satisfactory way without there being any explicit general theory
from which the implied (but often unknown) mechanism can be deduced.
Instead, there are mechanisms that arise through thick historical description or
metaphors that provide reasons. This is often called narrative explanation. The
emphasis on a close connection between intelligibility and the asserted
mechanism remains in an idiographic perspective but what counts as
intelligible is deeply contextual, varying across time and place.

CROSSING THE IDIOGRAPHICNOMOTHETIC DIVIDE

One influential way of thinking about the idiographicnomothetic divide is to
imagine a spectrum bounded by two antithetical stances that disagree on the
conditions by which intelligibility, the second condition of explanation, is
fulfilled. These stances are pure, ideal types that allow the continuum to
function as a heuristic device to locate particular scholarly approaches. At one
end, the position that human beings are so smart and societies so complex that
scientific explanation and prediction by way of general laws is impossible. The
only thing that can be done from this perspective is to interpret the meaning of
particular historical events or conjunctures in isolation because meaningful
generalization across different situations is impossible. This approach records
detailed descriptions of settings or events and tries to understand and evoke the
mentalities of the participants. The position at the other extreme claims that
social science is basically like physics; it is possible to discover a set of

20

Theory

universal laws that explain human behaviour and the nature of all human
societies.
Of course, few social scientists or historians occupy either of the extreme
stances and the linear continuum acts to allow degrees of separation between
different positions. For example, many historians tend to the idiographic
stance; for example R.G. Collingwood in The Idea of History (1946) argued
that explanation in history requires understanding the intentions of agents,
their motives and the consequences of their actions for future events; it is a
hermeneutic activity. History is a complex and chaotic system. On the other
hand, it is usually sociologists, political scientists or anthropologists who are
sympathetic to the goals of explanation by reference to general mechanisms.
In practice, most nomological adherents believe, like Elster, that although a
fully elaborated (micro-)theory is not necessary for the condition of
intelligibility to be fulfilled, it helps. Indeed, it is true that whenever a microtheory from which the postulated causal link can be derived is available, the
explanandum is, ipso facto, made intelligible. This is where the drive for
generalizations is derived in the social sciences: to establish theories that can
articulate a pattern of intelligibility. The theory is not required to be
deterministic or stochastic, nor does it have to have any predictive power.
Instead, it has to be capable of being used deductively, or more realistically
(but more weakly), abductively. The concept of abduction is closely associated
with philosophical pragmatism in the philosophy of the social sciences, and of
C.S. Pierce in particular. The formal structure of abduction is: D is a collection
of data; H (a hypothesis) would, if true, explain D; no other hypothesis can
explain D as well as H does; therefore H is probably true.
As will be discussed in subsequent chapters, the historical institutionalist
(HI) school is an important and influential body of work about history in the
social sciences, and a source of ideas about dynamics in public policy. In
terms of the continuum, this school of thought can be located on this spectrum
toward, but still some distance from, the nomothetic extreme. For example
Hall (2003, p. 395), in his discussion of systematic process analysis, argues
that: systematic process analysis is an effort to elaborate and assess the
validity of theories capable of explaining a broad class of events or outcomes.
It seeks generalizations that are both more simple and more portable than
those at which historians typically aim. More circumspectly, another leading
figure in the HI school, Paul Pierson (2004, p. 6), holds that: Although
historically orientated scholars are (rightly) sceptical about the prospects for
generating anything like a general theory of politics, most social scientists
remain interested in developing at least limited generalizations arguments
that can travel in some form beyond a specific time and place.
While the linear continuum imagined between the two antithetical extremes
is an appropriate heuristic device for illuminating the idiographic and

Policy dynamics and history

21

nomothetic approaches to understanding and explaining historical development, it has significant limitations as a way of informing how we look at
policy histories or policy dynamics. Locating approaches or particular works
on a linear scale implies a fixed point; but while scholars may be at the same
abstract point on the continuum, they may differ significantly with regard to
which aspects of social reality they see as more patterned and able to be
generalized, versus those that are understood as primarily conjunctural or
unpredictable. Further, a position on the continuum may be contingent on a
particular descriptive scale. For example, a study may look at new public
management reforms at a nation state level and generalize in terms of
propositions about Organisation for Economic Co-operation and Development
(OECD) convergence; whilst at the same time explaining reforms in the
particular part of the civil service that deals with hospital management in terms
of a series of case-specific, contextual and unique factors.
Even within the nomothetic disposition of the social sciences, it is not
necessary to assert that all human behaviour is equally amenable to
explanation via nomothetic mechanisms. There may be whole sectors of
human action that are quite conjunctural and in which causality is so complex
and interactive that simplifications of the usual sort employed in socialscientific models are incapable of representing reality. The argument in this
book is that the complexity of government and the interaction of policy
systems are, at one scale of description, examples of these sectors. The
observed pattern of development of a particular policy is unique, and one
cannot construct generalities about it or build models in the sense of providing
explicit and deductively sound statements of theoretical arguments; this is
where policy is understood as a composite variable, as in British healthcare
policy or EU telecommunications policy.
Some argue that the antinomy in the divide between the idiographic and the
nomothetic is entirely false; indeed it is logically untenable and
methodologically impracticable (Bryant 2004, p. 452). In particular, to
conflate generality per se with explanatory power is to misconceive the
function of abstraction, and results in empirical content being sacrificed for the
illusory gains of formal parsimony and general scope; yet equally, to develop
interpretive frameworks without bracing them against some conceptual and
ideal-type heuristics confines analysis to atheoretical descriptions of particular
cases.
For Bryant (2004, p. 455) there are negative implications for accepting the
idiographicnomothetic divide:
Assigning separate and distinct epistemic responsibilities to different disciplines
a descriptive-narrational set for the temporal side of human affairs, and an
explanatory-theoretical responsibility for the structural side is to rend aspects of

22

Theory
social life that are mutually implicated and which, ipso facto, can be understood and
explained only by their relational immanence.

The open question for historical social science, of course, is how to

characterize this relational immanence. Indeed, the desire to achieve a
balance between the idiographic and the nomothetic in the analysis of policy
dynamics is a theme that runs through the book. Recent historically orientated
scholarship in the field of policy studies is sensitive to the need for such a
balance. For example, Rose and Davies (1994, p. 11) put it: A model that
treated 1790, 1900 and 1990 as equivalent because each could be labelled time
t would be abstract and academic in the pejorative sense. Further, the HI
school, while seeking some degree of portability for concepts, metaphors and
theories, emphasizes that institutions and their effects can only be understood
in particular spatio-temporal contexts.
This is the broad position taken in the book. However, it is far from being
uncontroversial in political science; Bthe (2002, p. 481) discusses the use of
historical narrative as data to test political science models, and holds that
although modelling history is difficult, it is important to remember the
particular strength and source of progress of American [italics in original)]
political science: the explicit modelling of the political phenomena we seek to
explain, so as to facilitate scrutiny of the deductive logic of explanation.
Deductive logic is ineluctably nomothetic logic; the general case must exist in
order to deduce the particular one. In Bthes terms, historical narratives are
data to test the general theory. From my point of view, such an approach runs
the real danger of producing unverifiable just-so stories; I develop this point in
Chapter 5 on the methodology of structuring narratives.

POLICY NARRATIVES
Where does this leave us? The claim is made that policy development is
beyond social-scientific modelling and prediction; this does not require us to
believe that there are no recurring causal tendencies in the processes of policy
development that may be formulated into general mechanisms. The question
remains: what aspects of social and political life are repeatable and
generalizable, and which are unique? The answer to this question is directly
connected with the temporal and spatial perspective adopted for any particular
set of explananda. I argue that at the policy level we cannot predict, either
retrospectively or prospectively, thus we cannot yield mechanisms that provide
intelligibility by deduction from models or theories.
Nevertheless it is possible to generalize, however cautiously or contingently,
and this contributes to the structuring of narratives as a form of explanation.

Policy dynamics and history

23

A narrative does more than recount events; it recounts events in a way that
renders them intelligible, thus conveying not just information but also
understanding. They therefore contribute to explanation. The purpose of a
narrative is to render various series of events into an intelligible whole;
viewing policy as a composite entity that endures over a significant duration is
one way of rendering the complex interactions of the policy process
intelligible. A narrative can be broken down into a sequence of events,
processes and tendencies in order that they may be individually explained.
This can sometimes be by reference to general, Elsterian mechanisms deduced
from theoretical or empirical frameworks or non-formal models; in other
cases, metaphors can be used to structure the social environment in which
agents are situated and to provide reasons for action. These may include
general, descriptive categories, concepts or frameworks that can be moved
across time and space and be applied inductively or abductively for the
purpose of narrative explanation.
Any historical narrative must simplify reality to render it intelligible by
designating some elements as salient and omitting many more as not
significant. They must also posit a causal relationship between factors and
events. Even within narrative explanation, models (formal or otherwise) can be
used: at certain times they suggest production relationships, which actors in
the narrative are important, and what the relationship is between their beliefs,
preferences and actions. The point remains that the overall sequence or
narrative is unique; at the level of UK pharmaceutical policy there is no
general class of UK pharmaceutical policies for the particular case to belong
to.
Much of the idiographicnomothetic balance is contingent on the scale of
perspective (both temporal and spatial). It may be valid to run historical
regressions of regime change on percentage of gross domestic product (GDP)
spent to establish general theories that predict with an acceptable degree of
accuracy regime or empire shifts over several centuries. Similarly, general
propositions may be uncovered at the level of welfare state regimes (as in the
body of work inspired by Esping-Anderson 1990) or the field of security
studies (as in Powell and Lake 1999 on UN mobilization of states for
peacekeeping). However, a different scale of perspective is involved when
looking at UK health policy developments over a 20-year period, or a series of
economic policy reforms within a particular governments lifetime. The
contextual moves to the foreground in detailed accounts of policy change over
time; this is the essence of policy studies. Explananda are therefore rich,
detailed and contextual whether they reflect a single case or part-comparative
study. Universal mechanisms of change may exist but they are unlikely to
yield an intelligible account of specific spatio-temporal contexts. Categories
and theories logically cannot take the form of unrestricted universals or of

24

Theory

deductive axioms unqualified by considerations of time and place. In

Chapter 5, the epistemological position of the structured narrative as a valid
form of explanation is developed into a methodological guide for the study of
policy dynamics as a structured narrative.

PROBLEMS WITH POLICY AS CHOICE

The themes of choice, intention and decision suffuse many concepts of policy.
For example under a subheading Policy as authoritative choice, Bridgman
and Davis (2004, p. 4) list inter alia the following characteristics of policy: it
is intentional, about making decisions, it is political and it is structured. Later
they summarize (p. 6): Public policy is ultimately about achieving objectives.
It is a means to an end. Policy is a course of action by governments designed
to achieve certain results. This is the locus classicus of the rationalist view of
policy, which as discussed in the previous chapter, remains important in policy
analysis and is certainly the prevailing view among policymakers.
It is an easy step from viewing policy as a composite whole to viewing
policy as a choice. This is because once you have simplified complex reality
to a single, composite variable, it is tempting to simplify the process by which
this variable has emerged to a situation where a single agent the government,
say exercises an authoritative choice over that variable. At a reasonably
aggregated level of analysis there is nothing wrong with such abstraction.
However, the purpose of this section is to show that a dynamic perspective
provides important counsel in analysing policy as a choice. The dynamics of
choice requires the tracing of a process from preference formation to intention
to action to consequences, and an examination of their interrelationship, rather
than simply assuming these are static points that are automatically connected,
as in some deterministic stages model of the policy process. Dynamic analysis
messes up the notion of choice by insisting that preference, intention, action
and consequences have different temporal identifiers, and need linking
together.
In terms of action and consequence, the sour law of unintended
consequences (Hennessy 1992, p. 453) is well established within the social
sciences; indeed Popper (1945) holds it as the defining feature of those
disciplines. For example, every economics student learns the paradox of thrift;
that saving a greater proportion of income reduces overall income and that
overall savings end up being less that they would otherwise have been. The
EUs CAP, discussed in Chapter 7, has had the unintended but foreseeable
consequence of capitalizing higher land values rather than raising net farm
incomes. Alternatively, the emergence of economic development in a region
may have the consequence of leading the government to act in a certain way

Policy dynamics and history

25

that can be ex post facto rationalized as an economic development policy, but

was not a choice in the sense of a path from preference to intention to action;
instead economic development caused economic development policy, as is
sometimes the case where particular industrial clusters emerge. Equally,
sophisticated policy evaluation work recognizes that policy actions can have
short-, medium- and long-term consequences; and that it is often difficult to
discern transitional from equilibrium-related effects.
Changes in preferences are often invoked to explain changes in action.
Indeed, this is the cornerstone of microeconomics and also within game
theory; changes in preferences alter the structure of the game and the rational
actions and strategies of players therein. The category of preferences refers to
the desires, interests and beliefs that allow the rank-ordering of different
outcomes. For Bthe (2002, p. 484) preferences must unambiguously yield a
rank-ordering of the outcomes that would result from the conceivable actions
in that situation. However, there is no regular law-like connection between
having certain preferences on the one hand, and performing a certain action on
the other. Indeed, in any kind of strategic reasoning there is a step between
holding a preference in terms of the ranking of different outcomes and forming
an intention to act. There are beliefs about which actions will lead to which
outcomes, and in a complex policy process this inevitably involves strategic
thinking about political opportunities and how the political planets are
aligned; in particular how other individuals or groups might respond to any
action. The dynamic perspective encourages the view that making the steps
preferenceintentionactionconsequence by assumption is inadequate: these
are temporally distinct and it is incumbent on scholars to study and theorize
the links. Once we put temporal identifiers on a government action or
behaviour, along with preference, intention and review of consequences, and
call this policy we have a dynamic thing consisting of four parts separated
temporally; and we have dynamic perspective of choice.
What is less emphasized in the literature, but of at least equal importance in
policy studies, is the relationship between intention and action. Preferences are
prior to intentions. The reasons for intending to do an act, A, are the agents
mental states and beliefs about the world, which rationalize and explain their
intentions. Are reasons for intending to do act A the same as the reasons for
doing it? Yes, for the most part. The reasons for both are desires for certain
ends, E, and include beliefs that doing A would further E, as well as the mental
states that provide reasons for holding such desires and beliefs. In these terms,
the policy process is simply about determining how to act.
From our dynamic perspective, however, intention and action are distinct
events. Thus it is credible to propose that reasons for intending to do A are
desires for ends E, and beliefs that now intending to do A would further E,
along with the mental states that provide reasons for holding such desires and

26

Theory

beliefs. Thus the policy process is also about what to intend. Such a
distinction, based on looking at the notion of choice from a dynamic
perspective, has implications for policy studies. Consider the following
hypothetical example: imagine a government who has to decide now whether
to attempt a controversial policy launch in six months time. The policy has
already been designed but the government has to decide now about attempting
the policy launch because now is the time to organize publicity, start playing
the political mood music, lay markers, call in favours and so on, for policies
that may be launched over the next year. The government needs to coordinate
any publicity it organizes now with any later attempt to launch the policy.
Timing is one of the great political skills, and affects the extent to which
policymakers can influence the external environment into which policy qua
action is launched. The mobilization of support for including a problem on the
agenda (prior to launching a policy) is a crucial component of models of
agenda setting that have been increasingly useful and sophisticated since
Baumgartner and Joness (1993) Agendas and Instability in American Politics.
The government needs, above all, to avoid uncoordinated combinations of
actions. It needs to avoid attempting the policy launch without having first
organized the grounds for the agenda a significant political risk or
organizing the agenda for a policy that is not launched, which would make it
look foolish. The government will not decide to attempt the policy launch
without also deciding to organize the agenda for it; and will not decide to
organize the agenda for the policy without deciding to launch it. The
government in question is at present risk-averse: having recently had a difficult
six months in office, it now prefers that, having first refrained from organizing
publicity it will not attempt the policy launch. But in the future, as the memory
of the difficulties fades, and this can be quick, the government may well
become less risk-averse. There is a possibility that a present decision not to
attempt the policy launch might later be abandoned by the government even
if, thanks to that decision, no agenda setting for the launch had been organized
beforehand. There is therefore a risk that a present decision not to attempt the
policy launch might lead to an unpublicized policy launch in the future and
so to a mis-coordination between policy and agenda setting.
The government will have reasons for intending to do A and a separate set
of reasons for doing A. This is not a distinction that I am aware of in the policy
studies literature. There may be different reasons in the separate sets that are a
function of time; their temporality affects how governments make decisions.
In Australia during 2005, there was a protracted and public battle over the
introduction of reforms to the industrial relations system. The Howard
government announced its intention to launch a reform policy in April but the
policy itself was only promulgated in October. The policy dynamics
established from announcing the intention for a policy launch ahead of the

Policy dynamics and history

27

policy were manifold, complex and contingent. For example, the absence of
any policy proposals allowed the Australian Council of Trade Unions (ACTU)
and the federal opposition, the Labor Party, to launch an AUS$8million scare
campaign on the reforms. The campaign was initially successful to the extent
that the governments opinion polls were adversely affected, industrial
relations reform dominated domestic politics for extended periods (crowding
out other agenda), and Prime Minister Howard was forced into a series of
public statements and announcements claiming that all the adverse publicity
about the reforms rested on unsubstantiated and misleading assertions about
the impact of what will be in legislation that was still months away from being
written. In December, the government had a reform package enacted because
it enjoyed a majority in the Upper House for the first time in 25 years.
However, the final reform was affected by the promulgation of the intention to
launch a policy well ahead of the policy: the political environment became
more hostile, public opinion was successfully mobilized by appeals to
insecurity in the absence of any policy details, and as a result the Australian
government was forced into a series of concessions to mollify its wavering
backbenchers and placate public opinion. This was a case where the step from
the intention to launch, to the launch itself, was not a straightforward matter
that might be ignored or assumed away in policy analysis.
The dynamic perspective on the policy level of description leads to analysis
of contingent conjunctures as opportunities, strategic actions, beliefs about
links between action and consequence, as well as preference formation and
strategic rationality. To understand and explain the sequence involved in a
policy choice in conditions of uncertainty generated by complexity requires a
structured narrative; formal models of policy choice that posit a coincidence
of self-interest of governments and particular actions are inadequate for
dynamic analysis.
Consider monetary policy in the UK in the 1980s. The Thatcher
administrations prior to 1985 believed that only by hitting the broad money
target could they achieve some further goal that they valued extremely highly
(low inflation). The belief and the desire were announced as a set of targets,
M3, a broad money measure. This though had the effect of making this
measure of broad money behave quite differently; indeed it seemed that the
targeting of this aggregate caused its behaviour to change. Several different
broad money targets were tried, but all proved impossible to control once they
had been targeted. This was memorably summed up in Charles Goodhardts
wallpaper bubble principle. If the government had cared less about hitting the
target, they might have succeeded more easily that is, they might have been
able to act in order to control it. The intention to hit the target had a series of
complex consequences for action, consequences and beliefs about the
relationship between actions and outcomes. Indeed, even as the professed

28

Theory

monetary targets had been overshot and changed many times, the economy
had been tightly even excessively squeezed, sterling rose strongly,
unemployment soared and inflation plummeted. Inflation was, for the
Conservative administrations of the 1980s, their judge and jury. The
targeting of the money supply was formally abandoned in the UK in 1985, to
be replaced by Chancellor Lawsons interest in fixed-exchange-rate regimes as
guarantors of low inflation.
It is possible to render this complex series of actions, consequences and
conjunctural contingencies intelligible in terms of a concept like monetary
policy, as in Kenway (1993) for example. As social scientists, we wish to add
something to the narrative by way of structure using portable and general
concepts, theories, metaphors and mechanisms. The methodology of
structured narratives is deferred until Chapter 5, for now we note the problems
of the general and the universal for issues of portability. This question is
highlighted here as the next two chapters look at two broad groups of
metaphors, theories and mechanisms that have the potential to be portable
across time and place in supporting structured narratives of policy dynamics:
path dependency (Chapter 3) and evolutionary theory (Chapter 4).
The general and the universal should be clearly distinguished: to make a
generalization is not to claim that the phenomenon or process occurs
frequently or in a large or universal range of cases; Goldstone (1998, p. 832)
argues that:
A general law is not general because it applies to a wide variety of different kinds
of cases and events what makes a general law general is that it applies to a
range of initial conditions and asserts a necessary or probable connection between
particular initial conditions and a subsequent event or events.

However, the nomothetic view of mechanisms is that they are a substitute

for laws of development when these cannot be uncovered, and returning to the
definition offered by Elster, mechanisms are frequently occurring and
generally recognisable. This raises an epistemological question to which I do
not know the answer: how many cases do we need in order to have a general
category? This is a variation on the theme of finding a balance between the
nomothetic and the idiographic. The more that the cases to which a general
mechanism applies are contextualized, described in detail, and contingent and
local factors are brought into the foreground, the less portable that general
mechanism becomes. Conversely, a mechanism that is specified with less
detail is potentially more portable across cases, but is likely to have less
analytical purchase in terms of explaining detailed historical phenomena.
There is a trade-off between the portability of concepts and their added value
in terms of a particular structured narrative making sense of a complex series
of events or processes.

3. Path dependency
The emergence of a new institutionalism across the social sciences has
coincided with the increased interest in temporality, change and history in
social and political analysis. Institutions are structures that trace a path
through state space; they endure, have a history and can be used to link
temporally events and processes. The concept of path dependency has been
used within political science almost exclusively within a broad institutionalist
framework. It is institutions that are path dependent; as Raadschelders (1998,
p. 569) states: whatever the discipline contemporary neo-institutional
analysis has one feature in common: the notion of path dependency. The
widespread and cross-disciplinary use of path dependency for the analysis of
institutional stickiness makes the concept an obvious starting point for the
examination of concepts, metaphors and theories that might be used to
structure narratives of policy dynamics.
Indeed, the concept appeals as a label for the simplest of policy dynamics:
that past policy decisions act as a constraint on the options available to current
policymakers; or to use the language of dynamics from Chapter 1, that past
policy decisions act to circumscribe or foreclose parts of policy space. This
basic dynamic raises the question of how robust paths are over time and
through changes in the policy environment, and supports the interest in
evolutionary thinking developed in Chapter 4. Further, the refinement of the
concept of path dependency in response to complaints of determinism and an
inability to accommodate change introduces the discussion of the
methodology of structuring narratives of policy dynamics, which is set out
more fully in Chapter 5.
A process is path dependent if initial moves in one direction elicit further
moves in that same direction; in other words, the order in which things happen
affects how they happen: the trajectory of change up to a certain point
constrains the trajectory after that point. As Douglass North (1990, pp. 989)
puts it, path dependency is a process that constrains future choice sets: At
every step along the way there are choices political and economic that
provide real alternatives. Path dependence is a way to narrow conceptually
the choice set and link decision-making through time. It is not a story of
inevitability in which the past neatly predicts the future.
The reference to choice sets and decision-making reveals the origins of the
concept in economics. Indeed, path dependency is problematic for that
29

30

Theory

discipline because it implies that decentralized interactions between

economically rational actors do not necessarily lead to efficient outcomes;
indeed, inefficient equilibria may be recognized as such but still persist.
The concept of path dependency is not a framework or theory or model in
the terms of Ostrom (1999, pp. 3941): it does not provide a general list of
variables that can be used to organize diagnostic and prescriptive inquiry, nor
does it provide hypotheses about specific links between variables or particular
parameters of those links. Instead, path dependency is an empirical category,
an organizing concept or metaphor which can be used to label a certain type
of temporal process. As Hall and Taylor (1996, p. 941) put it:
they [historical institutionalists] have been strong proponents of an image of social
causation that is path dependent in the sense that it rejects the traditional postulate
that the same operative forces will generate the same results everywhere in favour
of the view that the effect of such forces will be mediated by the contextual features
of a given situation often inherited from the past.

The application of this organizing concept or metaphor to a phenomenon is the

beginning of a form of explanation because it asserts a relationship between
the sequence of early events and the probability of later events. Importantly,
however, the concept of path dependency does not per se provide necessary or
sufficient conditions to understand or explain that which it labels: path
dependent processes, even when identified, require theorizing; it is the
mechanisms that connect decisions or actions across time that explain a path
dependent process.
Although both refer to mid-range phenomena, policy and institution are
not synonyms. As noted in Chapter 1, within the policy system there are
various structures at different scales that act as institutions in shaping agents
decision-making in the formulation, enactment and implementation of policy.
These are not reducible to individual level agents or single elements in the
policy process; the philosophers notion of supervenience is used to elucidate
the relationship between structure and agency for policy narratives in Chapter
5. Examples of policy institutions are budget rules, policy networks, and
standard operating procedures in government departments and agencies. Most
importantly in terms of understanding policy development as path dependent,
past policy decisions are institutions in terms of current policy decisions: they
can act as structures that can limit or shape current policy options.
The question of what about a policy is path dependent does not admit a
single, conclusive answer; rather it remains an open question for scholars
applying the concept with theoretical and empirical corollaries. If the policy
whole or system is path dependent, there may be several potential underlying
mechanisms operating, independently or in combinations. This property of
multiple realizability has theoretical implications in terms of the

Path dependency

31

microfoundations of path dependency and the spatial and temporal scales of

policy analysis. It is necessary when using the concept, either theoretically or
empirically, to be clear about the perspective being adopted. The development
of a policy may be labelled path dependent over some period, but the various
mechanisms that underlie that process remain unclear unless a more finegrained perspective is adopted. Without microfoundations, the value of the
concept in structuring narratives to make sense of policy dynamics is doubtful.
Indeed, one of the insights of a dynamic perspective is that path-dependent
processes may coexist with other types of processes within policy systems. An
adequately fine-grained perspective is essential to using path dependency to
structure narratives; when policies or elements of policies are seen as strongly
interrelated, or where our analytical lens shows policy institutions as deeply
interwoven, then a much clearer sense of the mechanisms that underlie
path-dependent processes is gained.
The first section of the chapter considers the application of path dependency
to the analysis of policy development and its potential advantages in
understanding the dynamics of that development. The next section considers
several criticisms of the concept: that it lacks a convincing account of
decision-making over time, both of the accumulation of constraints and of
context-bound rationality; it is incapable of dealing with policy change; and it
lacks a clear, normative focus. In the final section I argue that despite its
theoretical underdevelopment and relatively limited number of successful
empirical applications, the concept of path dependency does have potential
utility in the field of policy studies in terms of narrative explanations of why
policies might be difficult to reform, and also why they may tend to become
more complex over time.

BENEFITS OF PATH DEPENDENCY FOR STRUCTURING

POLICY NARRATIVES
Path dependency is an appealing concept for understanding public policy
development; it provides a label for the observations and intuitions that
policies, once established, can be difficult to change or reform. Recent
examples of the use of path dependency for understanding policy development
include health care policy in the USA (Hacker 1998, 2002; Wilsford 1994) and
the UK (Greener 2002); the reform of housing benefit in the UK (Kemp 2001);
UK pension policy (Pemberton 2003); and the CAP of the EU (Kay 2003).
Path dependency encapsulates the insight that policy decisions accumulate
over time; a process of accretion can occur in a policy area that restricts
options for future policymakers. In this sense, path dependency arguments can
provide an important caution against a too easy conclusion of the

32

Theory

inevitability, naturalness, or functionality of observed outcomes (Pierson

2000a, p. 252). For example, Pemberton (2003) argues that the pensions
crisis in the UK is not primarily demographic but rather due to a low savings
rate; further, this low rate is a function of the path dependency and increasing
complexity of pension policy. The system of pension provision in the UK has
shifted over the last 20 years from one dominated by state provision to one in
which the state pension plays a residual welfare role. Despite this large change
at the policy system level, there is evidence of path dependency in particular
policy subsystems. In the case of UK pensions, policy subsystems exist around
specific pension schemes. An individual contract established under a particular
pension scheme at a particular time is costly to change: there are large sunk
costs; increasing returns associated with rising numbers of contributors and
pensioners in a particular scheme; further, there may be significant learning
effects. All of these factors contribute to significant switching costs for the
abolition of one scheme and the transferring of that set of individual contracts
into a superseding scheme. Particular schemes are locked in for particular
individuals. Nonetheless, pension reform has been possible but change has
come in the form of the addition of new schemes or elements to the system.
This amounts not to a single, path-dependent policy trajectory but rather to a
widening array of locked in subsystems over time. This accretion of new
subsystems or schemes has lead to the increasing complexity of the overall
system of pensions and has raised questions of effectiveness at a policy
whole level. Kemp (2001) reports similar dynamics with regard to housing
benefit policy in the UK.
Path dependency can help separate not just different orders of policy change
as in Hall (1993), but different rates of policy change. As discussed in Chapter
1, one of the foundations of dynamic analysis is the assumption that there are
a multitude of temporal scales immanent in any system. Therefore within a
policy system, there may be some elements that are path dependent, and others
that are not. Further, there may be a relationship between the different
processes at different speeds. Later in the chapter, I develop the point that it is
the combinations of institutions and policies that provide the important
mechanisms underlying path dependency and also create the potential for
strategic action and policy innovation by agents.
The concept of path dependency has the additional advantage of flexibility
for policy scholars. In particular, because the concept does not contain within
it a fixed temporal or spatial scale of analysis, the insights of path dependency
can often complement rather than rival other accounts of policy change. Social
housing in the UK is an example of where path dependency can hold at the
subsystem level with interesting consequences, but where the policy system as
a whole has changed profoundly. Since the mid-1970s there has been a series
of failed initiatives by central government to directly control the rents charged

Path dependency

33

in the social housing sector and ensure equity between local authorities and
housing associations. This particular element of social housing policy is path
dependent; each local authority has an established policy for rent calculation,
often determined by initial decisions made in the immediate post-war era. That
these have proved resistant to central control or influence significantly
constrains the ability of central government to pursue some of its objectives
for social housing. However over the same period as this path-dependent
process, there has been a clear shift in the social housing policy paradigm. One
of the five pillars of the UK welfare state has been, to a large extent,
removed. Between 1971 and 2002 home ownership increased from 49 per cent
to 69 per cent, with most of the increase occurring in the 1980s. The right to
buy scheme introduced in the early 1980s contributed to the increase in home
ownership, as it allowed local authority tenants to buy their own home.
Corresponding to this, the percentage of householders renting council homes
increased from 31 per cent in 1971 to 34 per cent in 1981, but then gradually
declined during the 1980s to 24 per cent in 1991. This percentage has
continued to decrease and in 2002 14 per cent of all householders rented from
the council.

PROBLEMS OF PATH DEPENDENCY FOR STRUCTURING

POLICY NARRATIVES
Accounts of Decision-Making Over Time
The criticism that the concept of path dependency lacks explanatory power is
well expressed by Raadschelders (1998, p. 576): it is only by virtue of
retrospect that we are aware of stages or paths of development. Path
dependency refers to a string of related events: causality in retrospect. The
concept does not come even close to pinpointing a mechanism or the
mechanisms that propel social change.
The quotation contains two criticisms. The first is that the concept
cannot be used for current or future phenomena. This is, of course, not a
singular feature of path dependency but is common to all concepts that are
useful for structuring retrospective, thick historical descriptions to support
narrative explanations in the social sciences. Indeed, as discussed in the
previous two chapters, a dynamic perspective reveals that this is not an
appropriate standard for considering the utility of different concepts, theories
or metaphors.
The more important criticism is that even if one accepts path dependency as
a possible candidate for narrative explanation, it is unlikely to be convincing
because the notion does not provide any fine-grained mechanisms that might

34

Theory

provide necessary and sufficient conditions for the process observed. The
challenge for the use of path dependency in helping to structure narratives of
policy dynamics is the uncovering of mechanisms that can help make sense of
a path-dependent process. One influential strand of the literature on path
dependency has worked on the microfoundations of the concept using insights
from new institutional economics. Much of this literature, particularly by
US-based scholars, proceeds by analogy from technological development to
institutional development. In simple terms, imagine two technologies, A and
B, both of which are subject to increasing returns but there is uncertainty over
the rate of increasing returns. Initial adoptions of one technology, say B, that
may occur for a number of small or chance reasons, beget further adoptions of
B in the market because of increasing returns that is, it becomes cheaper for
future firms to adopt technology B rather than A. The interesting results from
the models built on these assumptions (for economists at least) are that you
might get inefficient technologies adopted by markets. The normative
implications of this borrowing from economics are considered in more detail
later.
Arthur (1994) states the circumstances in which path dependence as an
increasing returns process is likely: the presence of large fixed (and sunk)
costs; network effects; learning effects; and adaptive expectations. As noted,
these factors have been used at a macro, constitutional level to make
arguments about path dependency in institutional development (North 1990;
Pierson 2000a, b, c). Within this list of sources of increasing returns, it is
useful to distinguish between those factors that relate to the internal efficiency
of firms large fixed costs leading to declining average costs as production
increases and learning takes effect from those that are external to the firm;
in particular, network effects. The distinction is important because later in the
chapter path dependency is discussed in terms of the increasing returns
involved in combinations of institutions and policies, rather than increasing
returns as a property of the internal operations of firms. This complements the
insistence on a fine-grained perspective of policy systems developed in this
chapter.
A focus on increasing returns is only a partial interpretation of the
economics of path dependency. Increasing returns are sufficient but not
necessary for path dependency. As Arrow (2000) points out, the existence of
significant sunk costs along with sequencing arguments can support many of
the path dependency narratives of technological change. Although he does not
use the concept of path dependency, Arthur Stinchcombes (1968, Chapter 5)
celebrated work on constructing theories of historical causation emphasizes
the central importance of sunk costs. I submit that any decision that is difficult
to reverse and which has enduring and ongoing effects can be said to have
initiated a path-dependent process; and work on path-dependent processes

Path dependency

35

should not focus exclusively on increasing returns processes; it is moot

whether they should privilege this mechanism-type over other sources of path
dependency.
Contrarily, Schwartz (2004) argues that the combination of initial, small and
contingent steps with increasing returns defines a path-dependent process; if
there was a large initial cause that had significant and enduring effects on the
subsequent process then this historical cause would be salient in any
explanation of the process. The process would no longer be path dependent but
rather, as a policy dynamic, would be better characterized as the temporal
unravelling of the consequences of some initial event. This point is noteworthy
insofar as it highlights the variety of mechanisms that might underpin a pathdependent process, but for the sake of expositional clarity this chapter remains
with the definition of path dependency as a general metaphor for sticky policy
or institutional processes.
A number of non-increasing returns mechanisms have been suggested as
underlying path dependency in policy development: the effect of policy on
interest groups, as when policies constrain some groups and enable others
(Pierson 2000a); policies that involve investment or disinvestments in
administrative infrastructure, which transforms governmental capacity and the
set of possible future policies that may be enacted (Skocpol 1992); and
policies that involve the establishment of formal or informal contracts with
individuals (Pemberton 2003; Kay 2003; Kemp 2001), which are costly to
change. Further, there are network effects to types of contracts rather than the
number of signatories. Once a contract is established, the transaction costs of
agreeing another contract of that type in that area of public policy will be
considerably lower than any alternative contract.
All these policy-specific mechanisms are based on definite, conscious
choices, which have the foreseeable consequence of high future switching
costs; none relies on an increasing returns process. Nevertheless, there are
examples that suggest that increasing returns processes can occur in policy
development. In the structural reforms of the primary care sector in the UK
after 1997, a series of primary care models were piloted. By a series of chance
factors, a particular primary care trust model quickly became popular. This
model subsequently became the governments template for all future
combinations of primary care organizations. There was no particular feature to
this model to recommend it over any of the others that were piloted between
1997 and 2000, instead it was the case that this model was adopted early in the
governments reform process, which made it considerably easier (or cheaper)
for subsequent primary care groups to use, and with such a momentum became
the template adopted by the government for all primary care agglomerations.
At a general level, all metaphors of policy ideas or proposals emerging from
the policy soup or garbage can share a notion of a market place of ideas;

36

Theory

analogous to a market system, ideas compete for attention and influence.

Where the market structure produces increasing returns, policy ideas can
succeed into proposals and eventual enactment by an initial series of small,
contingent steps as early adopters of an idea increase the return to future
adopters.
Any borrowing from microeconomics, including the idea of increasing
returns, inevitably situates the agent in terms of responding to the costs and
benefits of different options in a manner consistent with straightforward
parametric rationality. The assumption of this type of rationality serves certain
purposes in formal economic modelling, but to use it in more informal,
intuitive and post-positivist accounts of path dependency in public policy is
problematic (Hay 2004). One response is to use the notion of context-bound
rationality in an account of decision-making in a path-dependent process.
Nooteboom (1997) describes the manner in which markets lock-in to certain
technologies in terms similar to how philosophers of science characterize the
entrenchment of scientific theories. Both can be path dependent. He cites
Kuhns famous account of how scientific theories develop according to
paradigms, a set of tacit and unarticulated guiding assumptions, rather than the
standard conceptions of pure rationality at the heart of a scientific approach.
Further, just as the Kuhnian model challenges the conception of scientific
activity approximating to certain canons of rationality, so will any parallel
model applied to the economic case. As discussed in Chapter 1, Hall (1993)
invokes the notion of a policy paradigm in a similar vein. A policy paradigm
is an interpretative framework that operates in the policymaking process;
specifically, it refers to the ideas and standards that specify the goals,
instruments and the very nature of a policy issue.
In these terms, the mechanism that underlies path dependency in the policy
process is a form of context-bound rationality among policy actors. The
current path-dependency literature is mostly developed around the following
two claims: (i) that the analogy from economics to institutions can be extended
to policy; and (ii) that microeconomics can be borrowed as the
microfoundations of path dependency in policy development. Importantly,
these claims require a rational choice actor for making the decisions. The
assumption of this type of rationality is a strict corollary of claims (i) and (ii).
This is problematic for policy studies; Chapter 1 outlined how a dynamic
perspective moves analysis away from a simple, atemporal notion of choice,
which underlies formal models in economics. Further, public policy theory has
begun to move away from a reliance on the simple postulates of instrumental
rationality to a more nuanced and contextualized views of rationality. In
Chapter 5, this point is developed in more detail and the concept of policy
memory is introduced as one way of introducing dynamics into the interpretive
tradition in public administration.

Path dependency

37

Policy Change and Stability

At the heart of any account of path dependency is stability: observations of
change challenge the notion. This is a common criticism of the HI school; in
its emphasis upon path dependence and historical legacies it is rather better at
explaining stability than change (Hay 2002, p. 15). Thelen (1999) argues that
path dependency is too deterministic in that once the initial choice is made,
then the argument for future development becomes mechanical. This sort of
claim is usually accompanied with reference to the Polya urn model of path
dependency: imagine an urn containing ten black balls and ten white balls. A
ball is pulled from the urn and then replaced by two balls of the same colour;
after a relatively small number of selections, the urn will be almost full of
either black or white balls. In terms of exemplifying path dependency, the
colour that dominates the urn depends on the random selections early in the
sequence.
In terms of policy studies, one possible counter-argument is based on the
interpretation of stability in path dependency. Specifically, the notion does
allow policy change; policy legacies constrain rather than determine current
policy. Policy does change but within a particular set for possibilities; and thus
the policy may be said to exhibit stability. There are two main implications of
the constrained change argument. The first is that these bounded possibility
sets may be large or the paths wide; and the wider they are, the less the notion
of path dependency can account for current policy development. The second is
that the weaker the echo of past policy developments in the present, the more
other concepts, framework and theories are required.
The notion of policy direction may assist with constrained change accounts
of policy development. A stable policy path when projected into policy space
may well imply significant cumulative policy change over time, or in other
terms, a significant distance from the initial position in policy space and time.
Rose and Davies (1994) show the importance of compounding effects as
small, incremental and constrained changes in annual budget allocations can
accumulate to significant policy shifts over a period of a decade or more.
Further, a change in direction may appear at one distance a small perturbation,
but by shifting the direction of the policy may turn out in retrospect to have
been a critical juncture and therefore a problem for the validity of path
dependency as a description. Chapter 7 presents the argument that the
development of the EUs CAP is an example of this possibility: although
considered minor at the time, the reforms of 1988 (rather than 1992), were the
critical juncture in shifting the direction of the CAP from a price-support to
direct-income-payment system.
Nonetheless, the limitations of path dependency as a conception of policy
change have been highlighted in recent empirical applications in public policy:

38

Theory

Kemp (2001) with respect to housing benefit reform in the UK; Pemberton
(2003) with regard to pensions; and Greener (2002) on the NHS. Each of these
studies finds path dependency in policy development alongside some policy
change. They consider change as a reaction to the unintended consequences or
side effects of policy, or from pressure for reform due to exogenous shifts in
the wider policy environment: for example, where the distribution of power
between interested groups has changed. Once a dualism between policy
stability and policy change is established, the notion of path dependency is
only useful for accounting for the former; indeed the purpose of the concept is
to aid understanding of policy stickiness and why actors do not change policy
across time.
However, the dualism between stability and change can be avoided by
considering the sedimentation of policy decisions or the growing complexity
of policy space that is implied by the notion of path dependency. The dynamics
of policy subsystem accumulation are theoretically underdeveloped but are
important for the use of path dependency in policy narratives. As discussed in
Chapter 9, the development of UK pharmaceutical policy since the 1980s
provides an example of new policies being added on as a patch or fix to
satisfy pressure to mitigate the consequences of the original policy. Relatively
high prices for medicines were agreed by the government to reward innovation
by the industry under the Pharmaceutical Price Regulation Scheme (PPRS);
this contributed to the rapid increases in public expenditure on medicines,
observed since the late 1980s. The PPRS remains unchanged and potentially
path dependent but its budgetary consequences have precipitated a series of
new policies aimed at controlling the demand for medicines in the NHS, such
as cash-limited prescribing budgets. The path dependency of a particular
policy subsystem is a contributory factor in the explanation of the
accumulation of these policy patches and the growing complexity of the policy
system, with potential consequences for the overall coherence and
effectiveness of policy.
To reprise, a key issue when using the concept of path dependency is the
granularity of the perspective. Much of the work within the HI literature uses
the concept at a macro-perspective, in which there is a single whole that allows
for discussion of an institutional setting or a policy. The path then refers to
the trajectory for that composite variable, the direction of which is reinforced
after early moves in the sequence. Whilst this is valid for some narratives, from
a more fine-grained perspective the issue is which elements of that composite
system are fixed or locked-in, and which are capable of being reformed.
Further, within the policy space occupied by the composite whole there may
be potential for the introduction of new institutions or policy subsystems.
Crucial to my analysis of the development of the EU budget system in
Chapter 6 are the incompleteness and limitations of the initial Treaty-based

Path dependency

39

budgetary institutions working themselves out over time through a series of

political and fiscal contexts. This developmental process involved the
accumulation of additional and complementary institutions, which has had the
net effect of significantly constraining expenditure-generating EU-level policy
activism; this is a quite different policy dynamic from a two-speed,
changestability pattern. An incomplete institution is a policy framework or an
institutional structure that is not singularly self-sustaining. For example, the
constitutional balanced budget rule in the Treaty of Rome has required a series
of other institutions in order to ensure that the budget balances in particular,
rules to coordinate spending and revenue decisions. The contingencies for
breaches of that budget rule enshrined in the initial design failed to address the
key issue in the validation of such a rule: who controls expenditure? The
contestation of this incomplete institutional space has led to a growing
complexity of budgetary policy, a key policy dynamic in the EU.
In more fine-grained analysis, increasing returns processes operate at the
level of sequences of institutional or policy choices: once an initial policy
framework is established, there are strong increasing returns involved in the
choice of new, supplementary policies within that framework. That is, an
increasing returns process explains policy change qua the introduction of new,
supplementary policies. As North (1990, p. 95) states, it is the interdependent
web of an institutional matrix that produces massive increasing returns. This
view of increasing returns helps avoid too sharp a distinction between stability
and change, as seen in on-path versus off-path change, or where stability is
followed by a path-breaking juncture and the introduction of a new
institutional or policy setting. At a more fine-grained perspective, institutions
exist in combinations: they are interdependent, with necessary and contingent
relationships. Thelen (2003, p. 233) describes examples where institutional
lock-in is combined with elements of institutional innovation that can push the
overall trajectory of policy and politics in a different direction; indeed to
understand how institutions evolve, it may be more fruitful to aim for a more
fine-grained analysis that seeks to identify what aspects of a specific
institutional configuration are (or are not) negotiable and under what
conditions.
In an important recent work pushing the concept of path dependency
beyond simply the understanding of continuity, Crouch and Farrell (2004)
consider how actors cope with exogenous changes in their environment. At
some point, the once reliably successful path no longer works, and even
though policy actors know this, they find it extremely difficult to change.
Simple path dependency has the actor trapped, in a strict sense the concept
does not admit any other possibility; however a more nuanced account would
look at how the perceived failure of a habitual path may lead to the search for
alternatives, but where that search process is itself path dependent.

40

Theory

The relatively informal models offered by Crouch and Farrell (2004) are
designed to address the apparent determinism of path dependency once a path
is selected. Change is explicitly modelled as the intentional adaptation of
agents to exogenous, environmental shifts. The emphasis is on the ability of
agents to reactivate redundant institutions, or convert existing institutions to
different purposes, or borrow wholly new institutions from elsewhere to tackle
exigencies. These models also prompt thinking about what has been called the
layering of institutions; in terms of policy the implication of this notion is
that at the level of the composite whole, policy systems cease to embody a
simple unique logic, but rather a complex bundle of different policy logics,
ideas and interests. Some of these may be dormant, unused or forgotten for
periods but are capable of being reactivated by strategic action by agents in
response to exogenous environmental shifts. Institutional layering is
considered in more detail in Chapter 6 on the EU budget.
The reassertion of the capacity of agents situated within path-dependent
processes, with increasing returns acting to change the direction of the path in
response to shifts in their environment, is important for how we structure
policy narratives; indeed, it contributes to my argument for methodological
localism as the appropriate foundation for a methodology of policy dynamics
in Chapter 5. Inheritance and policy legacies are not as hard or fixed or as
determined as some of the simple path dependency analysis may suggest; for
example, increasing returns processes in economics typically assume a static
environment, whereas changes in that environment can attenuate (or amplify)
feedback processes. In providing a set of mechanisms that may help to
structure a narrative in terms of transitions between paths, Crouch and Farrell
(2004) provide a service to the analysis of policy dynamics. This can
complement the emphasis on inter-policy and inter-institutional relationships,
in particular combinational effects, which imply that policy development
proceeds in a more subtle way than the two-speed view of policy development.
Normative Aspects of the Term
One of the consequences of constructing the explanatory foundations of path
dependency in public policy by analogy from the economics of technological
development is to import the normative result that inefficiencies can persist
in path-dependent processes. This is a powerful result for neo-classical
economics: certain historical factors can ensure that inefficiencies occur and
markets do not eliminate these over time. Efficiency is understood here as
social efficiency, that is, a situation where both technical and allocative
efficiency hold. The strength of this normative result depends on the
judgement as to whether the inefficiency could have been foreseen at some
point in the initial stages in the path-dependent process and corrected; and

Path dependency

41

second, whether the inefficiency remains remediable, as with a Pareto

improvement that can be identified and is achievable from the current
situation.
For some writers the question of whether path dependency implies the
persistence of inefficient institutions is an open and empirical one (Hay and
Wincott 1998, for example); for others path dependency is more clearly
something which inhibits the introduction of better, or perhaps more rational,
policy or organizational form (Greener 2002, for example). Overall, the
normative implications of path dependency are less pressing for scholars
outside the boundaries of neo-classical economics; it is generally accepted that
inefficient policies or institutions may persist. However, it is a much stronger
claim that policies in a path-dependent process are necessarily inefficient; or
alternatively, that contained within the concept is the imputation of
inefficiency. The claim is strong at a theoretical level as the concept would
require significant elaboration in terms of both policy design and the pressures
that sustain path-dependent and inefficient policy. There is also the problem of
indeterminism: path dependency emphasizes that policy paths are unique and
arrived at by a series of small and contingent moves. As such, it is difficult to
say that there exists another path that could have been arrived at that is more
efficient, and without such a relevant counter-fact it is difficult to accept the
imputation of inefficiency.
At an empirical level the claim that policies are necessarily inefficient is
also strong. Developments in performance measurement in the public sector
might allow arguments that, for example, health care or education policy are
better in one system than another. Despite this it is difficult to assert, within a
particular political system, that there exist policy options that represent a
welfare improvement over the current policy (the net of switching costs and
increased transaction costs) and there is widespread recognition of this by
policy actors. Without these two conditions holding, the normative
implications of path dependency in terms of public policy are attenuated.
Nonetheless recent empirical works on path dependency in policy
development seem willing to impute inefficiency to some degree (Pemberton
2003; Kay 2003; Greener 2002; Kemp 2001; Wilsford 1994). This is often not
so stark as labelling policies in terms of efficiency, and certainly involves no
quantitative analysis; however it is not an over-interpretation of these works to
tease out the implicit assumption that a policy would be better without a
path-dependent process acting as a barrier to effective reform.

4. Evolutionary perspectives
Although the terms evolution and evolutionary are widely used in the social
sciences, for the most part they are employed as convenient shorthand for slow
and gradual change over time. This chapter considers evolution in a more
substantive way; investigating the prospects for evolutionary theory
contributing to the ambition for more fully developed dynamic perspectives on
policy change.
The first section of the chapter will establish the distinction between
evolution as a process of change and evolution as a theory of change. This
distinction is a necessary start to understanding the terms on which one may
judge the added value of an evolutionary perspective to existing accounts of
the dynamics of public policy. Importantly, the distinction is relevant to the
following commonplace and important objection to evolutionary approaches
in the social sciences: that biological analogies in the social world are weak,
that there are no equivalents of natural selection operating in the social world
and that without selection, evolutionary theory has little explanatory power in
the social world. This is also known as the missing mechanism argument.
We can employ a two-part strategy to address this problem. The first part is
to understand evolutionary theory as a (realist) ontological proposition that is
valid for the analysis of the evolution of all open, complex systems including
natural ones (Campbell 1965). The next part argues that the correct analogy to
consider is not that between the natural and social worlds, but rather the
analogy between the problems of evolutionary theory in explaining natural
processes of change, and the problems of evolutionary theory in explaining
processes of social change in terms of issues such as time, history and
differential rates of change. It is from this analogy that policy studies can learn
from and adapt evolutionary theory for the purpose of understanding and
explaining dynamic processes of policy change. Indeed, the weaknesses of
adaptionist models in evolutionary biology provide further intellectual succour
for the narrative approach by buttressing the argument that relying on thick,
contextual and historical description structured by general concepts or
portable metaphors is not a failure of formal modelling, but rather the
appropriate response to what cannot be formalized into a model; nor is it
amenable to explanation by a general testable theory.
Section two considers the main arguments for scepticism about the value of
evolutionary theorizing in public policy. First, the close connection with
42

Evolutionary perspectives

43

functionalism at a time when anti-functionalism is in the ascendant in political

science and where functional explanations are seen to have dubious validity;
second, the importance of agency and intentionality in social science
explanations rather than selection and blind variation; third, the
interpretation of fitness and the charge of vacuity that has been levied at
evolutionary theory.
In section three, evolutionary theory is considered as a metaphor to advance
narrative explanations of policy change. This perspective can provide a
framework for: understanding policy development in terms of selforganization, emergence and selection pressures; examining different forms of
policy learning; and discerning different and competing rationalities for policy
action.

EVOLUTION AS PROCESS, EVOLUTION AS THEORY

Evolution as a Process: Incremental and Radical Change
Any gradual change. Organic evolution, often referred to as evolution for short, is
any genetic change in organisms from generation to generation, or more strictly a
change in gene frequencies within populations from generation to generation.
(Wilson 1975, p. 12)
Evolutionary change generally refers to change that depends heavily on the
structure of already existing institutional changes and rationalities. Hence, the
choices that actors make today bear heavily on future choices and change occurs
incrementally in a path dependent way. (Campbell 1997, p. 31)

We start with these definitions of evolution as a process of gradual change as

they capture the loose and more common use of the term evolution. The first
definition is from within the biological sciences, the latter from sociology.
They serve to highlight that gradualism has often been understood as a
defining element of evolution as a process. As Stuart Kaufman (1995, p. 151)
puts it: One of the most important presuppositions of Darwins entire thesis
is gradualism, the idea that mutations to the genome can cause minor
variations in the organisms properties, which can be accumulated piecemeal,
bit by bit, over the eons to create the complex order found in the organisms we
observe.
The assumption of gradualism in a process of change is closely connected
with reductionism in theories of change. If a composite whole changes only
gradually it is easy to view it as the sum of relatively independent parts, with
the corollary that these parts may be analysed independently without the need
to consider significant interaction effects on the whole. This is the ceteris

44

Theory

paribus method in comparative statics, as in the analysis of the effect on the

market price of a change in the conditions of demand. However, when every
component of an organism is strongly connected to all other components, a
minor mutation in one component influences all other components. The
density of interconnections among elements is a measure of the complexity of
a system. When this is high and the interactions are at a reasonably high speed,
it is easy to see that complexity in systems is the antonym of gradualism.
Gradualisms equivalent in policy studies, incrementalism, similarly allows
reductionism: macro-scale phenomena like policy, because they change
incrementally, can be reduced to the behaviours of independent micro-level
units of analysis, such as policymakers. Thus, incremental changes in policy
are understood in terms of the bounded rationality of policymakers or
institutionally constrained agents.
The assumption of gradualism is problematic for understanding evolution as
a process in the natural world because it is difficult to reconcile with the
observed stability of many life forms as well as disparate rates of evolution.
For instance, embryonic birds and mammals still have gill arches, which are
useless; or the case of cave-dwelling fish that have degenerate eyes that do not
function. Why are these vestiges not removed by natural selection? Auyang
(1998) notes that the banana was introduced into the Hawaiian islands about a
thousand years ago and already there are several species of moths feeding
exclusively on bananas; at the same time crocodiles in northern Queensland
have not changed for hundreds of millions of years. How are these different
tempos explained? The controversial and long-standing question in
evolutionary biology is whether natural history contains emergent
evolutionary changes, or novelties that are not merely the accumulation of
small, gradual and adaptive steps. In 1972, Niles Eldrege and Stephen Gould
whilst looking at fossil records found that morphological characters stay the
same for very long periods, occasionally punctuated by drastic change within
a short period of time.
The idea of punctuated change has had a strong grip on thinking about
evolution as a process in the political world. For example, recent popular
models of the dynamics of public policy by Kingdon (1984) and Baumgartner
and Jones (1993, 2002) have used the notion of punctuated change. Although
these do not represent a full evolutionary theory nor a detailed and
contextualized application of evolutionary concepts in the social sciences, they
do acknowledge their borrowings from evolutionary biology.
However the more general point, for our purpose here, is not the difference
between gradual or punctuated change, but rather whether interaction between
different evolutionary units of analysis can produce evolutionary effects in
terms of self-organization rather than the gradual grinding of selection
mechanisms. Within any system that has repeated interactions among its

Evolutionary perspectives

45

constituent elements with feedback (both positive and negative) and agents
capable of innovation, there is the possibility that internal order of a system
can increase, or self-organization can occur typically leading to emergent
properties independent of external selection pressures. If we admit such a
possibility then the perspective of evolution as a process moves away from
approximating the organism as the sum of independent genes, or policy as the
sum of independent actions by policymakers. Kauffman (1995) uses the idea
of complexity to doubt gradualism in evolutionary processes in the natural
world by establishing that in some complex systems any minor mutation
causes significant changes in the system due to interaction effects. The
influence of Kauffmans work is manifest in Robert Jerviss (1997) System
Effects: Complexity in Political and Social Life about international politics
(see for example pp. 13, 41, 48, 157).
Much of the recent work in evolutionary economics is also in this vein
(Potts 2000; Foster 1997); concerned with self-organization, how patterns of
collaboration and cooperation can emerge quickly out of the interactions of
agents within certain complex systems. The point for us here is that interaction
and relations among agents can make the whole (policy, for example) more
than the sum of its parts. The interactions and relationships in a policy system
form institutions, generate variety, and produce complexity in the interaction
of ideas, interests and material circumstances. This makes composition
important: these relations cannot simply be summed; rather they are dynamic,
complex and typically have emergent, self-organizing properties.
This challenge to reductionism should be distinguished from the emphasis
in much neo-institutional analysis on strategic agency and the need for firm
microfoundations for theories of institutional change. This is driven by a
proper concern that institutional analysis should not overdetermine
behaviour or actions; that the potential for creative, innovative and
entrepreneurial agency means institutions cannot explain uniquely, nor can
they be assumed to persist indefinitely. I accept this without hesitation; my
argument here, developed more fully in Chapter 5 on the methodology of
structuring policy narratives, is that this should not be taken as methodological
individualism. By always focusing on the individual agent and how their
decision-making is affected by institutional structure, evolutionary effects in
terms of the emergent properties of composite wholes are liable to be missed.
Institutions are collective: it is groups of agents following a rule that
constitutes an institution. Significant interaction effects include tipping points,
network effects, combinational effects, bandwagon effects, reinforcement,
emergence, learning and imitation. These are all concepts from the evolution
of complex systems that can be used to structure narratives of policy
dynamics.
Further, by acknowledging that there are properties of the whole that are not

46

Theory

reducible to its constituent elements, the analytical possibility of macroevolutionary effects is raised. This is where the causal mechanism runs from
macro-level to macro-level. For policy purposes, this means causal links at a
level higher than individual agency; as in situations where policies can cause
policy change, or policy institutions can cause institutional change. An
example would be policy reform packages; the Australian governments
decision to introduce a broad-based consumption tax, the Goods and Services
Tax (GST), in 1999 after 30 years of protracted and deadlocked political
debates over tax reform, caused other policy changes inter alia with:
intergovernmental fiscal relationships in Australia; small business support
schemes; and personal private investment policy.
The purpose of discussing evolution as a process is to establish first that the
explanandum of evolutionary theory is not just gradual change; and that looser
use of the term evolution misses a substantial part of the process of
evolutionary change that of as rapid bursts followed by periods of stability.
There are no analytical advantages to making an a priori assumption that
characterizes change as gradual or rapid; cumulative or equilibrating;
directional or non-directional; ergodic or nonergodic. Second, that viewing
evolution as a complex process with potential system effects renders the
gradualismreductionism analytical strategy invalid in evolutionary theory.
There are institution-to-institution causal links and policy-to-policy
production effects that cannot be reduced to individual agents and their
decisions. Instead the relationship is captured by the philosophers concept of
supervenience; Chapter 5 takes this point further.
Evolution as Theory
Any evolutionary theory supports a form of consequence explanation; social
phenomena are explained through their actual consequences. The central task
for an evolutionary theory supporting explanation-by-consequences is to
provide a mechanism by which the consequences uphold or maintain the
action or structure that one wants to explain. In the absence of some kind of
feedback from effect to cause, explanation by consequence remains unclear. A
functional explanation (FE) is a special class of consequence explanation
where the consequences of an institution or routine of behaviour are
favourable, or functional, for some agent or group who maintain that
institution or behaviour. Explaining why an institution emerges in terms of
its consequences reverses the temporal sequence of causes preceding
consequences. Thus it is highly problematic in a obvious sense that it is
metaphysically impossible for an event to be explained by another event that
occurs at a later time.
Usually the answer is to emphasize historical contingency, accident or

Evolutionary perspectives

47

random variation in terms of why an X occurs at some time; then to focus on

feedback loops from Y to X that explain why X continues to perform the
function (Elster 1983). The separation of analysis between the creation of
institutions and what sustains them is a characteristic of recent HI analysis
(Streeck and Thelen 2005). In the terms of this chapter, natural selection as a
feedback loop successfully supports a functional explanation. Elster (1983)
discusses why functionalism is a good strategy in biology by presenting the
traditional mutation-and-selection model, which says that functional
structures will lead to greater survival and replication of the genes that produce
them.
There are examples of FEs in the social sphere. As Cohen (1978, p. 272)
notes, the following Marxist argument is a FE: the bourgeois media report
industrial conflict in a style that favours the capitalist class because that style
of reporting has that tendency. However as Cohen (1978, 1980) and Elster
(1980) agreed in their exchange on FEs in the social sciences, successful FEs
require mechanisms by which the consequences uphold or maintain the
institution or behaviour that one wants to explain. Elster (1980, p. 127) argued
that even though an FE remained logically possible, no mechanisms to support
one actually exist: this is the missing mechanism argument noted earlier that
is often used to deny biological analogies in the social world. That is, there
is simply no equivalent of natural selection operating in the social sphere and
FEs in the social sciences can only work if supported by evidence of
institutional selection, or something of the kind.
In response to this general argument Campbell (1965, 1974) has asserted
that Darwinism contained a general theory of the evolution of all complex
systems, of which organic evolution was only one. This argument is prevalent
in contemporary evolutionary economics. Hodgson (2002, p. 270) argues for
a Universal Darwinism that upholds that there is a core set of general
Darwinian principles that, along with auxiliary explanations specific to each
specific domain, may apply to a wide range of phenomena. For example,
Nelson and Winter (1982) applied the principles of variation, inheritance and
selection (that underpin Universal Darwinism) to routines in firms.
Additionally, a tradition of evolutionary epistemology stretches back to
nineteenth-century American pragmatism, for example William James and
C.S. Pierce, and has more recently included Popper (1972) and Campbell
(1965, 1974), amongst others. This traces the development of scientific,
objective knowledge in terms of experimentation, trial-and-error learning and
selection. In asserting a general set of principles for the analysis of all complex
systems, evolutionary theory amounts to an ontological claim of the existence
of processes of variation, retention and selection, irrespective of the particular
mechanisms that might be plugged in for specific theoretical strata. For us, the
important point of the ontological claim that evolutionary mechanisms are

48

Theory

universal is that their instantiation in the social world need not be directly
analogous to genetic variation and selection.
The nature of ontological reasoning is notoriously elusive and, from a
policy studies viewpoint, prone to arcane abstraction. Nevertheless, the claim
here is problematic: what is the relationship of these general evolutionary
principles to some acceptable theoretical presuppositions about the policy
world, or empirically based understandings of policymaking? If the principles
of Universal Darwinism exist as logical possibilities in complex mathematical
models, what is the status of evolutionary theory in the social world? At least
formal models in economics are traceable back to assumptions about
economic behaviour. What would the patterns and predictions of complex
systems models mean in terms of explaining a policy process? In the terms that
have been used in this book the question can be rephrased as: does
evolutionary theory assist in providing intelligible mechanisms to link steps in
a narrative?
I will leave this question open for the moment but note that in accepting the
ontological claim that social systems are subject to evolutionary principles, the
ground by which to judge the validity or usefulness of evolutionary theory is
shifted. This does not of course avoid the missing mechanisms argument, but
rather changes its nature. The challenge is no longer to uncover a mechanism
of selection that is the equivalent of genetic selection in the natural world,
because the argument no longer proceeds by analogy from the natural to the
social world; instead the ontological assumption requires a search for
mechanisms of variation, retention and selection that are appropriate and
contextualized in the social world. The missing mechanism need not be
analogous to genetic selection, but it remains missing nevertheless. It is to this
challenge that the next section proceeds.

PROBLEMS OF EVOLUTIONARY THEORY IN THE

SOCIAL SCIENCES
Functionalism and Explanatory Power
For evolutionary theory to support FEs in the social sciences requires the view
that variation, retention and selection mechanisms operate such that
dysfunctional entities tend to disappear; alternatively, non-dysfunctional
institutions (including functional ones) maintain themselves over time because
they are not selected out. With this view, it is on the plausibility of variation,
retention and selection mechanisms providing the link from consequence to
the persistence of an institution or behaviour that evolutionary explanations
rest.

Evolutionary perspectives

49

In any given item of functional analysis, there are two groups of individuals
involved: those who engage in the practices to be explained and those who
benefit from these practices; that is, those for whom they are in some sense
functional. The question of intention arises only for the first group; the
question of recognition may arise for both. The two groups may coincide,
overlap or be totally separate. In the latter case, the possibility arises that the
effects are unintended by those who produce them but recognized by those
who benefit from them. However, identifying actors and beneficiaries is not
sufficient for an explanation of change. For example, in order to see how a
dysfunction can explain change, it is necessary that the victims need not only
perceive the problem but also correctly identify its cause. If the dysfunction is
not recognized, then the victims cannot act as a causal agent in social change.
Moreover, when the outcome but not the cause of a dysfunction is perceived,
as is often the case if the causal chains are protracted or complex, no pressure
for change should be expected; or, if there is pressure for change, it should not
be expected to be successful in doing away with the problem. Without the
agential link between consequences and future action, or an equivalent of
natural selection in the social sciences, then FEs are weak and rightly to be
avoided in the social science field. This is our preceding discussion: natural
selection supports FEs in the biological sciences, but without an equivalent
in the social world evolutionary theory cannot support FEs in the social
sciences.
What are the prospects for evolutionary theories having explanatory power
without functionalism? This requires the starting assumption that general
evolutionary principles operate, but not in a way that selects behaviours or
institutions that produce favourable consequences for some group or agent, as
that would amount to functionalism. This is the dilemma of wanting
evolutionary theory to have some explanatory power whilst avoiding
functionalism. Kerr (2002) asserts that institutions, behaviours or ideas that
have consequences that are beyond the strict limits or selective pressures
set by the environment, will not tend to occur; and this forces individual
actors or groups of agents to negotiate, and adapt to, the context in which
they are situated (p. 351).
His basic argument is that because dysfunctional elements tend to
disappear, persisting forms can be assumed to be adapted in the sense of not
being dysfunctional. It amounts essentially to the assertion that nondysfunctional institutions maintain themselves over time, because they do not
transgress environmental limits or constraints. Some might see this position as
vulnerable because there may be many non-dysfunctional alternatives to a
given dysfunctional institution. Without the ability to say which of them will
emerge and at what time, rather than simply that one of them ultimately will,
the predictive capacity of evolutionary theory is limited. This brings us back

50

Theory

to the earlier point that in terms of explanatory power, evolutionary theory is

anaemic without functionalism.
However, there is good reason for this lack of explanatory power. The
general presumption of this book is that dynamic perspectives introduce a
degree of complexity into the analysis that renders the ambition for covering
law explanations, manifest in some policy theory and institutionalist research,
invalid. The ambition for dynamic theories of public policy is for theories,
methods and concepts to structure narratives. In terms of evolutionary theory,
there may be different selection pressures operating at different levels and over
different time horizons. There is no single uniform and universal mechanism
that uniquely selects and that can be accounted for by a covering law theory.
Instead, there is a multitude of overlapping and potentially conflicting
environmental pressures over the short, intermediate and long term.
This notion of selection as environmental pressure, that produces a
disposition or a tendency, is closer to contemporary understanding of selection
in the field of complex systems and evolutionary economics (Potts 2000 is a
good example of this). In these terms, selection mechanisms are not universal
fields that operate consistently over time but rather are often local, relative and
operate over a specific period. Thus within the environment of the policy
system there are dispositions, tendencies and constraints that limit what policy,
policy proposal, idea or advocacy coalition (and so on) may be successful.
Consequences-based arguments are possible without functionalism; the
challenge from the previous discussions is whether such arguments are
convincing. Van Parijs (1981, pp. 2930) suggests that the problem of
consequences being used to explain causes can be overcome by saying that it
is not Y that causes X but rather the disposition of X has to cause Y that causes
X. Dispositions can exist through time, thus they can precede any particular
instance of Y. However, this argument is one of potentialities or dispositions:
why will X occur at any stage and second, why will it continue to have the
disposition to cause Y? One of the arguments of this book is that there is no
convincing general theory to answer these questions: selection mechanisms do
not select uniquely; so in terms of the narrative explaining what actually
occurred at certain times then selection pressures are part of the external
circumstances that may softly determine the process of change. The role of
historical conjunctures in providing windows of opportunity for policy
reform is well known to scholars of policy. The notion of a window of
opportunity and how it may assist in explaining policy dynamics is unpacked
in Chapter 5.
Agency, Intentionality and Evolutionary Theory
In a critical response to Johns (1999) arguments in favour of introducing

Evolutionary perspectives

51

evolutionary theories into political science, Dowding (2000) sets a distinction

between evolutionary and intentional explanations, and states (p. 75) that
specifying any non-intentional selection mechanism by which policies are
generated may be called evolutionary explanation. This distinction is too
severe for several reasons. First, once parsimony and formalism as features of
good theory are abandoned to adopt a dynamic perspective, then intentional
action as the centrepiece of an analytical strategy becomes more difficult. This
does not simply include the general problems of rational choice in public
policy theory (Hay 2004) but also the general question of rationality: of
connecting particular desires to particular intentions to particular behaviours
that was discussed in Chapter 1, and also as set out later, choosing between
different reasons for action. Further, intentional actions sit alongside
conjunctural contingencies, unintended consequences and environmental
selection pressures in terms of mechanisms to make sense of events and
processes in policy space.
One way to accommodate selection and intentionality is through the
concept of artificial selection, although on the Dowding distinction this is not
evolutionary. The essential characteristic of artificial as against natural
selection is that humans manipulate the criteria or environment of selection,
therefore it is intentional selection. However, the introduction of intentionality
in this way raises the question of rationality; the processes of cognitive
and cultural evolution that explain why agents come to act in the way that they
do.
An example of intentional selection is the evidence-based policymaking
initiative in the UK government since 1997. What matters is what works
is a New Labour mantra and has been the rationale for vastly expanded
research staffs and budgets in government departments and agencies, in the
NHS and in local authorities. Much of this effort is spent on ex post facto
evaluation or judgement by consequences; interim and formal evaluation
frameworks are routinely built into programme designs and budgets for
major programmes such as Sure Start, the New Deals and Neighbourhood
Renewal. Evidence-based policymaking in its use of pilots, monitoring,
performance measurement, full evaluations and its commitment to act on
the feedback of what works is, at least in an ideal form, a form of
evolutionary epistemology. The search for what works explicitly assumes that
policy should be developed in an evolutionary way accepting the insight from
Simon (1982) and Braybrooke and Lindblom (1963) that there are strong
limits on notions of a universal rationality in policymaking. Indeed,
myopia and the importance of random variation or designed experimentation
are at the essence of evolutionary theory; this is what distinguishes it from
other models of the policy process, or from different models in social
science.

52

Theory

Intentional selection mechanisms have been introduced into policy theory;

particularly in the voluminous body of work that pushes analysis beyond the
stages ideal of policymaking. In Kingdons (1984) influential policy streams
approach, different policy solutions enter at the initiation stage of policy, often
taken from the garbage can or policy soup, and are fitted by policy
entrepreneurs to problems that float by. There are selection pressures that
filter out or fail to select certain policies: from backbenchers in the legislature;
from different departments; from the electorate; from the media or policy
networks. This version of a policy window is a policy environment with a
confluence of different policy, political, economic and social trends occuring
over different temporal scales. At those times certain ideas and policies may
have their time (Kingdon 1984, Chapter 6). He sets out three main categories
that enhance the chances of an ideas survival: technical feasibility; value
acceptability; and anticipation of future constraints. The first is whether the
idea is fully worked out and, crucially, whether there is a practical and
achievable implementation plan. The second is concerned with its
compatibility with the values of the policy community. This is the equivalent
of the notion of a policy paradigm. Third is the anticipation of future
constraints, that the idea must be seen to have the potential, to have acceptable
budgetary costs and enjoy reasonable prospects of approval from politicians
and the public (mass, activist and media). These are all intentional selection
mechanisms.
However, for Kingdon these selection criteria only draw up a shortlist.
Which particular idea will emerge depends on a series of contingencies in
which tipping points and bandwagons can play a part: beyond a certain level
of popularity opposition to an idea ceases, or alternatives fade, and everybody
joins the bandwagon. This is the internal dynamics emphasized by the selforganization strand of evolutionary theory, which is developed further in the
third section of the chapter.
Fitness and the Charge of Vacuity
The technical definition of fitness is the relative rate of change in the number
of instances of a certain character within a population. It has no significance
itself, but rather only makes sense when compared with the fitness of other
organisms; or more correctly, only makes sense in the fitness spectrum for a
population of organisms. Why do characters grow at different rates? Natural
selection answers this question but raises the further issue: what are the causal
mechanisms responsible for the variation in growth rates of different
characters? Evolutionary biology as based on the survival of the fittest has
often been challenged on this question. Fitness is defined in terms of those
units that survive and reproduce successfully; in alternative terms, survival

Evolutionary perspectives

53

defines fitness. How to explain survival? According to natural selection, those

that are the fittest and most capable of reproducing successfully survive. So
fitness explains survival. This invites the criticism that evolutionary theory is
vacuous or circular. This is a weakness in evolutionary theory acknowledged
by both biologists (such as Wilson 1975; Mayr 1988) and philosophers (for
example Popper 1972).
The practical consequence of this circularity is that evolutionary theory may
be labelled descriptive in the sense of not allowing any counterfactuals or, in
Popperian terms, by being incapable of being falsified. Alternatively,
evolutionary theory is incapable of prediction either prospectively or
retrospectively, even in the statistical sense (Blaug 1986, p. 274). This is
important if you judge theory by its explanatory and predictive powers. It is
less of problem if you view the role of theory as contributing to narrative
explanation; with policy dynamics there are never counterfactuals other than
those imagined by scholars, and the scale of description is detailed, contextual
and fine-grained.
In evolutionary biology, adaptation models are aimed at providing some
answers to why some characters reproduce more successfully in certain
environments. They seek to explain an organisms ability to cope in a specific
environment. A typical model sets out possible behaviours, objective to be
optimized and a set of constraints, including environmental conditions.
The optimization procedure aims to find the behaviours, strategies or
characters that maximize the objective subject to the constraint. Oster and
Wilson (1978, p. 292) (quoted in Auyang 1998, p. 146) state that
Optimization arguments are the foundation upon which a great deal of
theoretical biology now rests. Indeed, biologists view natural selection as an
optimizing process virtually by definition. But as Oster and Wilson (1978) go
on to discuss, faith is required that equilibrium will be achieved. Optimization
models say how an organism is adapted (they predict an optimal equilibrium
to which the empirically observed behaviours or characters correspond) but
not how come. These are equilibrium models and suffer the problems of
comparative statics set out in earlier chapters. How do organisms approach
equilibrium? Analogously, how do market prices approach equilibrium? Some
major recent works of evolutionary biology, for example John Maynard
Smiths Evolution and the Theory of Games, William Hamiltons Narrow
Roads in Gene Land and Richard Dawkins Climbing Mount Improbable,
support the view that most biological models of evolution are of the
optimizationequilibrium type and bear strong resemblance to equilibrium
and optimization models that are used to underpin much of microeconomics.
In these models, it is assumed that equilibrium will be reached by natural
selection or the operation of the free market rather than by revealing the actual
dynamics of change.

54

Theory

PROSPECTS FOR EVOLUTIONARY THEORY IN POLICY

STUDIES
Evolutionary theory can act as a metaphor in policy studies as it does, for
example, in recent political economy work on how political and economic
crises often defy understanding: they generate change, but no necessary
remedial solutions. To complete this chapter, a general evolutionary metaphor
of policy development is developed in order to assess the potential for portable
concepts and frameworks: does the metaphor imply or hint at mechanisms that
may lend intelligibility to narratives of policy dynamics? Within the social
sciences, to label analysis as metaphorical or to claim that frameworks act as
a metaphor often implies weakness and a sense of a lack of substance. This is
where a metaphor is seen as rhetoric or a literary flourish without the requisite
rigour for the purposes of understanding society. This is a mistake. As recent
philosophers of social science have noted, metaphors have a deeply
constitutive and subterranean presence (Lakoff and Johnson 2003; Lewis
1996; Klamer and Leonard 1994). Indeed, metaphors are prior to any analogy;
they establish broad and general mappings across conceptual domains. There
is a thin line between metaphor and the structure of thought; metaphor is no
longer just a form of expression but a form of conception as well. They help
to structure understanding, assist the perception of connections between
different things, and are a way of supporting ontological claims (Lakoff and
Johnson 2003).
Popper (1972, 1984) developed the pragmatists idea that knowledge grows
by trial and error or, in alternative terms, by conjecture and refutation. The
starting point is the formulation of a problem situation confronting the
individual or group; the second step is the production of tentative solutions and
trial responses to the situation. The third stage is the process of error
elimination: weeding out those among the tentative solutions that do not work.
The fourth stage is the emergence of new problems, or perhaps the
reformulation of the original problem.
Evolutionary epistemology may be used as a metaphor for describing the
policymaking process to help develop dynamic policy analysis. The metaphor
is not perfect, as Popper discusses evolutionary epistemology in terms of the
advancement of objective, scientific knowledge, which is too limiting as a
description of the policy process but does serve to establish learning as an
evolutionary mechanism, subject to the general principles of variation,
retention and selection. Further, this metaphorical structure serves to raise
important questions about various forms of learning and competing
rationalities in dealing with the uncertain, complex and constantly shifting
environments in the policy process.
The evolutionary metaphor is a useful way of organizing thinking about

Evolutionary perspectives

55

policy learning by distinguishing adaptation (learning that affects calculations

about how to realize interests most effectively aka Bayesian rationality) from
more complex learning where interests, identities and institutions are learned
or constructed in the interaction of agents in the policy system. In this latter
sense of learning, the problem situation itself is constructed in the interactions
between agents; policy problems, policy solutions and the criteria of success,
or the intentional selection mechanisms that operate to weed out policy
failure, are all constructed and it is through this construction that policy
paradigms emerge and are institutionalized. As noted, Kauffman argues that
analysis of the internal dynamics of self-organization should complement
consideration of selection mechanisms in evolutionary theory (1995, p. 644):
I have tried to take steps toward characterizing the interaction of selection and
self-organization Evolution is not just chance caught on the wing. It is
not just tinkering of the ad hoc, of bricolage, of contraption. It is emergent
order honoured and honed by selection. One of the important recent insights
of the neo-institutionalist research programme is that there is the potential for
institutional change in terms of conversion, recombinations, layering and
activating redundancy (Crouch and Farrell 2004, Streeck and Thelen 2005;
Thelen 2003). These are all examples of evolutionary change as selforganization that may or may not become institutionalized over a period
through environmental selection pressures.
In Chapter 1 a dynamics perspective was used to problematize the notion of
choice. Preferences, intention, action and consequence were temporally
separated. It is possible to appeal to a broad notion of rational decision-making
to connect these different temporal states. Over the last 20 years, rationality
has been condemned in social theory on several grounds: as a homogenizing
force, as a logic of identity, as a mask for power, and as male and modernist.
Post-structuralist theorists have attacked its claims to impartiality and
universality, while the postmodern turn called for an abandonment of
rationality for an endless play of difference. From within the analytical
philosophy tradition authors such as John Searle have argued contra the
classical, scientific model of rationality that rational decision-making is often
about choosing between conflicting reasons for action rather than starting
from a consistent set of desires or preferences. In fact, humans are
distinguished by their ability to be rationally motivated by reasons for action
independent of desires or preferences (Searle 2001). Raz (1999) argues that an
account of rationality is an account of the capacity to perceive reasons and to
conform to them, of judgements of their appropriateness in different contexts
and of different modalities of conforming to them.
For some, particularly within the constructivist tradition, these theoretical
developments have meant an overturning of rationality; however for our
purposes these debates show the the idea of a single, unique and universal

56

Theory

rationality breaking down in favour of different types of rationality. The

parametric and generally non-dynamic rationality of rational choice theory has
a strong hold in parts of political science and policy studies; while elsewhere
the strict informational and cognitive assumptions of rational choice have been
attacked (see for example Hay 2004) and relaxed in the renewed growth of
interest in Simons (1957) original idea of bounded rationality (Gigerenzer and
Selten 2001). Also influential have been the ideas of Jurgen Habermas on
communicative rationality: that rationality comes out of a communicative
context but appeals to universal pragmatics and competences in
communication (Habermas 1984, 1987). Another version sees rationality as
wholly contextual and not just confined to explicit forms of communication in
speech, but to all aspects of interaction within social processes (Bridge 2001).
This leads to discussions of a possible transition from rationality and certainty
to uncertainty, reasonableness and openness to the other (Levinas 1994) that
comes from situation and context. Further, the notion of normativity in moral
philosophy, built around the concepts of ought and a reason, equally
supports arguments about the situatedness of particular rationalities: questions
of what counts as a reason, and what makes a reason appropriate in certain
circumstances and not others allow very different, and spatially and
temporally contingent, rationalities.
Importantly for the use of the evolutionary metaphor there are materialities
to these situations; rationality is not necessarily just ideational. Indeed, one of
the key concepts is appropriateness: how and why agents act in certain
situations. This is related to the issue of situational logic that Popper uses: how
are policy problems constructed within the interactions of the policy process?
This is how narratives are constructed, and the foundation of the interpretive
approach in political science advocated in recent work by Bevir and Rhodes
(2003, 2005). This insight is developed in detail in the next chapter.
In the evolutionary metaphor of policy development, policies, programmes
and projects are recognized as trial solutions to solve a problem. Situational
logic assists in understanding decision-making processes; it recognizes that
according to the logic of their situation, agents pursue certain goals or act for
certain reasons and they do this by assessing which is the best way of
achieving these goals within the given situation. In these terms the
evolutionary metaphor might be seen as privileging the objective elements of
a situation and demanding a formal rationalism whilst ignoring subjective,
psychological and non-conscious elements such as folk psychology, rules of
thumb and habits or routines. I think such a criticism would be unfair; the
evolutionary metaphor draws our attention to the possibility that a person or
institution may find that they need to choose between several different options
in their pursuit of certain aims, as in the classical view of policy; but also
between several different rationalities, as in the interactionist view of policy.

Evolutionary perspectives

57

This highlights the potential significance of meanings and emotions in

analysing the logic of the situation and understanding why particular decisions
are made. This is the basis on which claims to a universal rationality are being
disputed and where much of the recent interpretive work may be placed (for
example, Orr 2005).
The evolutionary metaphor helps suggest the idea of rationality as
contextualized and emerging from the interactions of agents in the policy
system. This is an important contribution to policy studies; back in Chapter 1
when discussing broad approaches to conceptualizing policy as a unit of
dynamic analysis, I preferred the structured interactionist view of policy in
contrast the classical, rational view. This was not a rejection of the importance
of rationality in policy; rather a rejection of a singular view of rationality,
favouring an emphasis on a plurality of rationalities that compete within the
policy system. Different types of rationality have been suggested theoretically
above so the empirical questions for policy scholars are: what type of
rationality informs policy? Are its assumptions strict and universalizing, or
more participative and contextual? Of particular interest will be the temporal
and spatial aspects: does policy cohere over different temporal scales and
different spatial scales of application? Are rational decisions made at one point
in time, or in a certain place, consistent with long-run and wide-reach
rationality assumptions?
At first glance, the evidence-based approach to policy advocated by the
UKs Cabinet Office seems to assume the neutral and scientific treatment of
evidence to inform policy intervention: an impartial and scientific rationality.
Yet the evidence base also informs policy in terms of what works (Davies et
al. 2000), a pragmatist conception that comes as much from particular context
as it does from the application of scientific principles. The significance of
particular context has been heightened in the debates in other policy contexts
over identity and diversity. The idea of rationality in policy analysis begs the
question: are there elements of universality that can be maintained intertemporally and across policy sectors, or does policy operate with different
conceptions of rationality depending on the context? This prompts the core
question for structured narratives of policy dynamics: what is the relationship
between context and rationality?
Where does the foregoing discussion leave us? In evolutionary biology, the
process of natural selection operates at a population level. This has been
criticized from within that discipline as bean-bag genetics, circular and
vacuous arguments about population statistics. In terms of the evolution of
individual species, evolutionary biology tends to rely on comparative static
models that are unable to account for actual processes of change. This
weakness is analogous to how economics, after the Marshallian marginalist
revolution, has periodically struggled with how prices adjust to equilibriate

58

Theory

demand and supply; but most economic textbooks assume that prices will
adjust, just as most evolutionary biology textbooks assume that evolution has
occurred. This may be an appropriate modelling device for certain purposes,
but for the dynamic analysis of actual processes of change, evolutionary theory
in biology must rely on structured narratives.
The evolutionary approach in policy studies shares this characteristic with
evolutionary biology. This is an important element in the value-added of
evolutionary theory for understanding policy dynamics by showing that
dynamic analysis in different fields favours narratives. In adopting a detailed,
contextual scale of description of individual policy development, policy
studies rely on narrative for explanation where mechanisms are expressed as
tendencies, dispositions and environmental limits combined with an emphasis
on conjunctural contingency, memory and history. The evolutionary approach,
while not at all ruling out intentional explanation and strategic agency,
certainly decentres the agent in the policy process; the metaphor puts selforganization and selection mechanisms alongside intentional action in
structuring narratives. Further, the evolutionary metaphor helps to introduce
policy learning in a constructivist sense, which is useful for understanding the
emergence and institutionalization of policy paradigms.

5. Structured policy narratives

What does a structured narrative consist of? The first section of the chapter
considers this question and how its answer contributes to understanding and
explaining policy dynamics. Section two is the major section of the chapter
and examines methods for the microanalysis of the policy dynamics in a
structured narrative. The final section investigates what makes a structured
narrative convincing or valid, successful or true. These three sections allow the
chapter to act both as a methodological guide for the empirical chapters that
follow and as a conclusion to the preceding theoretical chapters.

INGREDIENTS OF NARRATIVE EXPLANATION

The basic material of a narrative is a chronicle: a list of things that happened
in a chronological order. This list usually has some minimal organizing
principle, such as the fact that the events happened in a particular place or to
a particular organisation or person. Crucially, chronicles do not attempt to
make sense of what happened whereas narratives, in contrast, are a single
coherent story, albeit with subplots (Stone 2001, p. 74). Narratives vary in
ambition. A minimal narrative charts the sequence of events: how one leads to
another with appropriate details of accidents, coincidences, and
misapprehensions. The aim is to make the sequence, at some basic level,
intelligible or coherent. A structured narrative contains the ambition to make
sense of the sequence in terms of some greater interpretive scheme; for
example, a normative frame of success or failure; or in terms of the direction
and trajectory of policy; or type or value policy changes; or changes in the
political environment, and so on.
This interpretive scheme acts as the basic template or structure for the
narrative. This structure rests on two premises drawn from the discussion of
policy dynamics in previous chapters. First, policy as a concept is either an
ideal-type or a cluster concept; both permit us to classify a range of different
phenomena (both spatially and temporally) under a single concept. Many
social-scientific concepts are cluster concepts because they share some
amongst a cluster of properties (Putnam 1975, pp. 5054); an ideal-type
concept is a complex description of a group of social phenomena that
emphasizes some features and abstracts from others. Either of these two
59

60

Theory

allows policy to be used as a heuristic to facilitate empirical enquiry, rather

than conceived of as a natural category for investigation. In these terms, the
successive states of policy traced through time, a policy path, becomes the
basic structure of the narrative; as discussed in Chapter 1, this is the starting
point of dynamic policy analysis. Further, this is a non-essentialist ontological
claim, which has implications for the microanalysis of policy narratives that
are considered in the next section of the chapter.
The second proposition is that policy is a meso-level concept and policy
history is somewhere between the macro-scale, such as with accounts of
structural changes in the political economy or constitutional change, and
particular studies of precise events in deep historical detail, such as the
agreement of a particular budget package. This has several important
corollaries for the structure of the narrative and its microanalysis. First,
historical contingency becomes a guiding theme; at any given juncture there
are multiple outcomes that might have occurred. A sense of possibility is
essential to any narrative of policy dynamics. The meso-level of analysis is
fine-grained and so recognizes the role of agency in influencing the course of
development in particular historical contexts and demands analysis of the
multiplicity of causes that are at work in any temporal setting. In addition, at
the meso-level there are discernible structures, processes and constraints that
exist across time and recur in various historical settings; and these play a
causal role in the direction and pace of change. Therefore an important part of
policy narratives is the identification of these structures and the tracing of the
ways in which they constrain and motivate individuals in particular settings,
leading to outcomes that can be explained as the contingent results of
conjunctural moments. As Pierson (2005, p. 41) sets out:
In part because studies of policy enactment make it possible to examine moments
of change in fine detail, the role of particular actors in initiating such movements is
likely to be highlighted. Yet these studies have greater difficulty in identifying those
features that facilitate, impede or channel entrepreneurial activity. Broad, structural
features, as well as long, slow-moving processes, which may be crucial
preconditions for policy change, recede from view.

Policy narratives embrace the complexity of different processes of different

speeds and at different levels coexisting in the policy path; indeed, it is the aim
of the narrative to weave these together into a coherent story. Crucially for
explanation by narratives, it is only by virtue of hindsight and the analysis of
the conjunction of different processes that we can make any sense of which
process is dominant, which structures and constraints may have been operating
and the direction of their net effect. A structured policy narrative should
specify the institutions, structures and processes that are embodied in a given
historical setting; identify the possibilities and constraints that these structures

Structured policy narratives

61

create for agents within those settings; and construct explanations of outcomes
that link the causal properties of those structures to the processes of
development that are found in the historical record.
A policy path encompasses a series of steps in the development of policy
that are explained individually: there will be different explanations for
different steps in the sequence as there is no single causal mechanism to
encompass the whole policy narrative, nor is there a teleology or historicist
cause from which all else can be traced. These steps may be moments, events,
processes, periods, choices that are observed more generally, or are a typical
kind for which we might employ portable concepts, models or metaphors;
alternatively these may be salient events these are not necessarily large or
immediately recognizable, but are capable of being labelled in retrospect as
critical junctures in the development of a policy.
Pierson (2004) stresses that causes have their effects over different temporal
scales; just as the historian is counselled by the Mao Tse Tung view that the
twentieth century was too early to tell about the effects of the French
revolution. It is only by virtue of hindsight that one can judge salience;
additionally, different events acquire salience over time that is, our
judgement of salience is itself temporally distinct. The judgement of salience
relies on questions of potentiality and contingency: if possibilities are
foreclosed by a decision or action, or when the costs of reversing a decision or
action are high, this is a condition for salience. However it is not sufficient, as
there must be a supporting judgement that the decision or action had
significant or meaningful consequences in terms of policy development.
Unlike closed systems studied in theoretical models based on the ambition for
covering law explanations, policy processes sit in an open context. This is
what makes theories of the policy process so difficult: they are irreducibly
complex. This is a common predicament for professional historians; the
crossing of many causal paths drives events. Many of the steps in a structured
narrative occur at the nexus of contending forces. John Bury (quoted in
Oakeshott 1966, p. 201) argued that this confluence of paths was not governed
by laws and stressed the idea of contingency in historical analysis: it was the
conflux of coincidence that proved decisive.
In terms of narrative as a form of explanation, the key for any narrative is to
avoid being a Just So story. Such stories refer to Kiplings answer to how the
leopard got its spots and the rhino its wrinkled skin; because of the fanciful
natural history in these stories, the term came to be used in evolutionary
biology to refer to unnecessarily elaborate and speculative evolutionary
explanations that lacked any substantial empirical support. The term has come
to be used in the social sciences in the same way. In historical narratives,
theoretical models are used but they are local or contextual, and sometimes
limited to one specific, temporally distinct event within the narrative. Theory

62

Theory

is always subordinate to evidence. The burden of the narrative is to weigh

competing models, concepts or metaphors and show that one is the most
appropriate in view of the evidence. In discussing how to judge one narrative
explanation as being better than any other, the answer always refers back to the
evidence. In just so stories, by contrast, the evidence is subordinate to the
theory. There is a willingness to extrapolate basic models across whole areas
with limited evidential support. Applied rational choice theory has a tendency
to produce just so stories; Green and Shapiro (1994) note the poor empirical
support for such a widely used theory; and Elster (2000) argues that the Bates
et al. (1998) project for analytical narratives, aka rational choice history,
amounts to a series of just so stories.
Just so stories are not usually Kiplingesque in their absurdity; and indeed
almost all are the result of a genuine attempt to avoid ad hoc-ness in
historical narratives. For example, Goldstone (1998, p. 832) warns against Dr
Seuss-like explanatory principles, such as logic which suggests that events
are wholly contingent and unique and so just happened to happen this way and
are not very likely to happen that way again. The ambition to avoid the
imputation of ad hoc-ness is worthy but in the context of complex policy
systems it can lead to a reliance on general theories that can only ever be
idealistic, which approximate only for equilibrium conditions rather than how
and why equilibrium is reached, and for which empirical support is not general
but rather limited to particular cases. This book argues that just so stories are
the inevitable result of adopting the covering law view of explanation as the
basis for analysing complex and heterogenous policy dynamics.
For example, in a well-cited article Bthe (2002, p. 487) proposes that
structuring a narrative based on a model can allow scholars to treat them as
data on which to test the model of a general theory:
Beyond the elements identified in the model, however, additional context-specific
information should be minimised. Information that is extraneous to the model
should be provided only insofar as it affects salient elements and is needed either to
understand the relationship between these elements or to appreciate the
contingencies of a particular historical process.

In these terms, the model defines what is important in the narrative and thus
avoids the problem of ad hoc-ness that inductive narrative explanations, from
a social science perspective, may suffer from. However, the evidence that suits
the testing of the model is selected, which raises the possibility that the model
is true or correct in terms of confirming evidence, but inadequate in terms of
understanding or making sense of the overall phenomenon. In John Godfrey
Saxes fable of the six blind men confronting an elephant and touching
different parts of the animal, each of their different models was correct and
confirmed by the evidence from touching the elephant: an elephant was like a

Structured policy narratives

63

wall, a sharp spear, a huge fan, a giant snake, an old rope and an extremely
large cow. However, none of these models was remotely adequate for
characterizing an elephant.
From a dynamics point of view, the contextual elements of the narrative are
essential for making sense of the development of a complex, composite
variable over time. The thick, historical description, the emphasis on
conjunctural contingencies and strategic agency are the core elements of a
narrative; this is what needs to be made sense of rather than being stripped out
in the interests of lean modelling. Bthe (2002) is encouraging the analyst to
make the overall evidence subordinate to the theory (as expressed in the
model). We reject here the notion that narratives should be conceived as
testing the model, on the grounds that to do so would inevitably render the
narrative a just so story where features of the world that are essential and
causal in this context are ignored because they do not have, nor could they
have, a place in the general model because of the irreducible complexity that
characterizes policy processes.
In place of the goal of assisting the discovery-governing regularities by
testing models of general theories, the function of the narrative in policy
studies is to provide understanding and explanation of particular dynamics of
policy development. Because of contingency and the importance of possibility
in policy dynamics, there is no way to tell the trajectory or path of a policy
except by following it step by step. There is no covering law to be unveiled
here. When the contingent causal factors at one step are understood, then we
cannot call upon a dynamic rule to deduce what happens next; no such rule (or
theory or model) exists. Many causal factors pull and tug in different
directions in a historical process, and the policy analyst must compound them
in order to understand the process (it is not possible to strip them away into
component parts or use the ceteris paribus method). Complex systems
textbooks show the vast intellectual effort necessary to compound causes in a
theoretical model to deduce the progress of relatively simple systems; this
reveals the hopelessness of models aggregating micro-causes in complex and
heterogeneous policy systems.

MICROANALYSIS OF STRUCTURED POLICY

NARRATIVES
The approach to microanalysis in the book is premised on the claim that policy
is an ideal type that supervenes on the structured circumstances of the agents
who make up or constitute it. Policy supervenes on a set of properties or
individual behaviours when there can be no changes or differences in policy
without there being changes or differences in individual behaviours.

64

Theory

Philosophers use this term to describe a relationship between two levels of

analysis that is neither logical nor causal. Policy is not the logical identity of
the collective actions of agents largely due to considerations of multiple
realizability: policy cannot be the same thing as the behaviour of agents
because the same policy may be realized through a large variety of patterns of
agents behaviours; that is, it is not an identical thing just viewed at a more-orless grained perspective. Nor is the relationship between levels causal; the
behaviour of individually, socially situated agents does not cause the policy or
produce the effect of policy; rather it is collective behaviours that constitute
the policy.
Agents are situated in a nexus of structures: networks of other agents;
institutions; norms; as well as worldviews and paradigms; collective
memories; shared folk theorems, and so on. Little (1991) proposes a position
of methodological localism where individuals, socially situated in their local
contexts, constitute the basic unit of analysis of social phenomena. This level
of description is particularly relevant to policy studies, which usually stress the
relatively small number of elite actors that are involved in most policy
decisions. These individuals are regularly interviewed in case study analysis:
members of policy networks, occupants of institutional roles such as senior
bureaucrats and politicians, leading technical experts, and so on.
To stress, this is not methodological individualism. It invokes the social
in the definition of the position of the individual. It refers freely to norms,
networks, institutions, belief frameworks, and other supra-individual
constructs, such as policy or past policy decisions, or policy frames or
paradigms. Importantly, the social is local: individuals acquire their social
properties as a result of an actual history of interactions with particular
institutions, organizations, networks and other actors. This is what makes
methodological localism suitable for the dynamic analysis of the meso-level of
policymaking; it provides for the fine-grained, historical study of decisionmaking consisting of a limited number of agents within both particular
institutional roles and a wider policy environment.
The basis of the claim of supervenience between the policy level and
individual agents is that there is a complexity and looseness in inter-level
relationships that militates strongly against reduction to individuals because it
misses key inter-agent relational dynamics. The key to the looseness is the
human ability to create/imagine new forms of social interaction; to innovate
socially and collectively; and to defect from social expectations. In the terms
of Chapter 4, there is the potential for self-organization in the system. As a
result there are differential degrees of fit between individual action and
structures, institutions and norms. Meso-level structures may morph as agents
create policy, support the institutionalization of policy, adjust their behaviour
to the incentives and roles created by institutions but also defy or quietly

Structured policy narratives

65

defect from norms, act opportunistically or on principle, and forget policy

institutions. The relevant question is not whether policy and institutions
exercise autonomous and supra-individual causal primacy a version of
reductionism but rather: to what extent and through what sorts of
mechanisms do policy and institutions exert causal influence on individuals
and other structures?
Situated Agency
The concept of situated agency brings us back to the discussions in Chapters
1 and 2 about the nature of choice in policy studies. The basic unit of policy
analysis consists of agents situated within an institutional environment, where
institutions exist as constraints on the choices that may be made by agents. The
question of how and why such constraints operate has produced different
varieties of institutionalism rational choice, sociological or organizational,
for example. Works such as Hall and Taylor (1996, 1998), Hay and Wincott
(1998), Campbell (1997, 2004), Stacey and Rittborger (2003) and Wendt
(1999) demonstrate that the role of structure and agency within
institutionalism is an open and continuing question, with the interplay over
time between institutions as constraints (or enablers) and agents as capable of
strategic action, and the remaking or recasting of institutions a central part of
dynamic policy analysis.
The narrative form places an emphasis on agents motivations and
intentions; for some, this is what distinguishes the approach from other socialscientific methods. For example, Fischer (2003, p. 163) puts it: whereas the
scientific mode strives to identify stable, reproducible patterns of actions that
can be explained without reference to social intentions or purposes, the special
subject matter of the narrative form is the vicissitudes of human intention.
This is only true up to a point: the longer the narrative, the more complex the
conditions for the fulfilment of the intentions and designs of agents, and the
less likely it is that ongoing intentions, plans or designs can truly make sense
of the sequence. As noted previously, contingencies and conjunctures become
relevant in the dynamics of interactions between agents, and so too do
structures as buffers against some of the possible perturbations and related
complexity in analysis; this is why institutions are given a prominent role
within policy narratives.
Interpretivist policy analysis has made a significant recent contribution to
the question of how and why agents may act within certain institutional limits
or constraints. It should be noted that much of this work either implicitly or
explicitly draws on a much longer tradition within historically orientated
disciplines. Here a condition of understanding is that the actions we see in a
dynamic process make sense to us as analysts; or more accurately, make sense

66

Theory

to us that such actions should make sense to the agents in those circumstances.
If an agent acts for certain reasons, observers can be in a position to make
sense of that action in terms of those reasons. One way scholars can do this is
by thinking himself or herself into the position of the agent (their context,
outlook, preconceptions and memory). This deeply contextual form of
historical understanding was emphasized most famously in the works of R.G.
Collingwood (for example, Collingwood 1946).
This concern for the intentions and beliefs of agents has implications for the
research methods of policy dynamics. Direct evidence of mental states in
terms of interviews or other documentary evidence is a key aspect of analysing
choice; the tendency in some rational choice history (see Elster 2000, p. 693)
is to impute intentions and beliefs in order to explain actions. The interpretivist
approach holds that it is poor social science to construct a model in which
observed behaviour maximizes the interests of agents, and then assume that fit
between the interest and the behaviour explains the behaviour. It may just be
a coincidence; and this is where the historical form of understanding set out by
Collingwood can add such great value to policy narratives.
It is important to note what the interpretivist approach does not imply: it does
not mean that the observer identifies themselves with those reasons, or would
have acted similarly. Rather this refers to a feigned understanding of
the agent. Further, putting oneself in anothers shoes may provide an explanation
or understanding of the action observed, but it may not provide
the complete or adequate explanation. There are important social, psychological
and political questions that arise in why the agent should act on such reasons.
This detailed study of the context in which the agent is situated including the
actions or anticipated actions of others complements the analysis of the
motives and intentions of agents in narratives of policy development.
Neo-institutionalism tends to stress the importance of how agents construct
internally institutional constraints, and how they constitute themselves within
these institutional structures. This is how institutions constrain: it is only
through affecting the intentions of agents that institutions have any causal role
in explaining behaviour, including those behaviours that result in institutional
change. Fisher (2003, p. 28) summarizes the position:
It is not that institutions cause political action; rather, it is their discursive practices
that shape the behaviours of actors who do. Supplying them with regularised
behavioural rules, standards of assessment, and emotive commitments, institutions
influence political actors by structuring or shaping the political and social
interpretations of the problems they have to deal with and by limiting the choice of
policy solutions that might be implemented.

But what of institutions qua ideational structures, or more particularly,

conceptual maps, normative policy frames and policy paradigms? Do these

Structured policy narratives

67

sorts of institutions enter into the discussion of agent behaviour as conscious,

intentional constraints? For Hall (1993), policy paradigms are unconscious
and non-intentional in their influence on agents. An important distinction often
missing in neo-institutionalist analysis is that between institution-following
and institution-described behaviour. This distinction is pertinent in light of the
definitional breadth of the term institution: habits, norms and routines are
institutions for some, as are mental maps, frames and policy paradigms for
others, and as are budget rules, tax codes and electoral rules elsewhere. Some
institutions may be non-conscious and non-intentional, as when rules have
been fully internalized; in this case, institutions describe the behaviour rather
than having some effect on the intentions of agents. Other institutions,
however, may be interpreted and constructed by agents in a conscious manner
as, for example, in the regular and privileged access granted in a formalized
policy network to certain social policy groups. Elsewhere institutions are
tightly specified, codified and external to the agent, which does not permit
much space for individualized interpretive frames; instead there are real
penalties associated with certain forms of behaviour and agents may
consciously follow such rules, as in budget rules or electoral rules. The general
point is that the mechanisms that give causal effect to institutions can be quite
different according to different types of institutions and the local context in
which they are embedded.
One important dimension of all institutions that poses a problem for the
analysis of situated agency in a temporal process is that institutions are
collective: there must be sufficient number of agents who act in a certain
manner, whether consciously or non-consciously, for that manner to be
considered an institution. It is important to repeat the emphasis on
methodological localism at this stage; part of the context in which the agent is
situated is the behaviour (or expected behaviour) of other agents. This presents
a problem for narratives of policy development. First, being particularistic and
interpreting all individuals motivations and beliefs when acting in a certain
way introduces an intractable complexity into the analysis once the number of
agents goes beyond a small number. Alternatively, there are representative
agent models in which all agents act in such a manner that their cumulative
actions might as well be the actions of one agent maximizing its expected
utility function. These are popular in economics to deal with the issue of
aggregation from micro to macro. It is relatively simple to model the
behaviour of one person, given some assumptions about preferences and
constraints. Indeed, it may well be valid for certain agents in certain
circumstances. The problem is composition, as discussed in Chapter 1: how
can we aggregate this representative individuals behaviour to a higher-level
structure such as an entire economy? In formal modelling terms this is
straightforward, however the credibility of the aggregate result must be in

68

Theory

question for policy studies, given our continued emphasis on the complexity
of inter-agent dynamics in the policy process.
The notion of situated agency has a strong role in the scheme for policy
narratives in allowing agential space to make choices within institutional
contexts; and the capacity to innovate to make and remake institutions. But
this focus should not be at the expense of acknowledging institutions as
collective actions, choices or interpretations. How can groups of agents change
their institutional environment? This is a collective choice; and therefore
talking about agents choice within institutions is to miss much of the
dynamics at work around bandwagons, tipping points, threshold effects,
sufficiency parameters, and so on. Structured narratives of policy dynamics,
within a presumption that situated agents are the units of analysis, should be
cognisant of these emergent properties of relational dynamics.
Policy Memory
In a general sense, memory refers to the capacity for remembering, recalling,
recollecting or recognizing. The extent to which events, behaviours, routines
and institutions are remembered is an important, and relatively neglected, part
of policy dynamics. Memory is a mechanism that links past events and current
intentions, actions and behaviour. It is a function of time: it is always the
memory at a given moment in time of an event that occurred at some previous
moment in time. This is an important point for any dynamic analysis; policies
and institutions are reproduced and have local states at particular places and
particular times.
The idea of memory is well established in systems analysis. Cortes et al.
(1974, p. 3) define it as the impact of some event that happened in the past
upon the current response of the system. Memory is strongly related to the
concept of hysteresis that was first introduced into economics by GeorgescuRoegen (1967, 1971) in the study of consumer behaviour: individual utility at
a point of time is not simply a function of consumption at that time; rather it
is affected by the past consumption of that individual as well. It has
subsequently been used in a number of different economic applications,
perhaps most prominently in the idea that there is a mechanism whereby a rise
in unemployment increases the equilibrium (or natural rate) of unemployment.
Elster (1976) uses the notion of hysteresis more generally in the analysis of
social and political change, but the term memory is preferred here in order to
adumbrate the memory of individual agents as the object of interest.
Brain scientists and philosophers of the mind are a long way off being able
to predict what gets remembered and being able to explain why. In light of
this, the description of the policy memory mechanism here is appropriately
basic and focuses on two different aspects of memory. The first aspect is

Structured policy narratives

69

memory conceived as something analogous to the operation of computers;

while the second relates to how unconscious habits, routines and policy frames
are reproduced and remembered over time.
Computer memories store bytes of information and retain that information
over time, which can then be retrieved via well-designed algorithms. The
analogy to human memory is not perfect. Consider information retrieval,
where current experiences of current events produce memories of preceding
similar events, or where political actors use symbols and stories to deliberately
access agents memories. The content of such memory can be events or policy
decisions, as well as information and opinions and, importantly for the
computer analogy, this content becomes more inaccurate with time; it fades or
decays, and thus the effect of temporally distant events on current policy
processes will change with time. There is also emotional memory, which
refers to the intensity of feeling or the vividness and closeness of a particular
emotion connected with an object in the propositional memory. For example,
when policy has been enacted amid crisis as, for example, in the events that
produced the foot and mouth policy in the UK in 2001 (Taylor 2003), the
memory of those events will fade with time; both the propositional aspects (as
when senior policymakers forgot the lessons of the 1967 foot-and-mouth
outbreak) but also the vividness of the imagery of burning animal carcases.
This memory process may affect subsequent development of policy on animal
health security in the UK.
One way to consider policy memory at the unconscious level is to use
Bourdieus notion of an agents habitus; this is the active sediment of their
past that functions within their present, shaping their perception, thought and
action. It consists in dispositions, schemas, mental maps and competence, all
of which function below the threshold of consciousness. Bourdieu (1984, p.
466) puts it: the schemes of the habitus, the primary forms of classification,
owe their specific efficacy to the fact that they function below the level of
consciousness and language, beyond the reach of introspective scrutiny or
control by the will.
Each individual agents habitus will be different to some degree, as no two
biographies are exactly the same. Nevertheless, for Bourdieu individual
biographies are just strands in a collective history; individuals belong to a
group or variety of groups and develops their habits within these groups. This
is useful here for refining the notion of policy memory because it: (i) stresses
that is the shared or collective policy habitus that is of interest for policy
studies; and (ii) this memory may fade, drift or change through less-thanperfect reproduction over time.
The notion of memory affecting agent behaviour is difficult terrain for any
empirical social scientist, for the obvious reason of a lack of immediate
opportunities for description and observation (see, for example, the exchange

70

Theory

between Chalmers (1995) and Dennett (1996)). It also poses a substantial

theoretical hurdle for us here, in that throughout the book institutions,
including past policy decisions, are given a causal role only insofar as they
affect the intentions of agents, where agency is defined as the potential for
agents to have always acted differently in any given set of circumstances. The
intentional part of that formulation refers to agents acting for a reason. This is
fine when memory is conceived as something akin to a computers hard drive
with the agent consciously retrieving data, but the problem for unconscious
memory is: how do these affect the intentions of agents? If agents are not
responding to these unconscious institutions, how can they have a causal role
in that agents behaviour?
The policy memory argument can be developed to include the claim that
adherence with institutional rules can occur at a non-conscious level: the
following of institutional rules becomes second nature, like a habit rather than
an algorithm, and it is this second-nature adherence that helps explain the
reproduction of institutions through the actions of agents over time. Searles
(1995) idea of the Background that originates from his work on issues in the
philosophy of the mind and the philosophy of language can assist in developing
policy memory in terms of the reproduction of non-conscious, underlying
policy frames and other policy institutions that serve as background to agents
behaviour. The background is the set of non-intentional or pre-intentional
capacities that enable intentional states of function (Searle 1995, p. 129),
where non-intentional or pre-intentional capacities consist of abilities,
dispositions and other causal structures that ultimately function at some liminal
level (pp. 12930; see also Searle 1999, pp. 1078). Thus, the background is
the set of abilities, dispositions and memories that facilitate intentional human
agency; with the corollary that it can act as part of causal explanations.
This is how the notion of a policy paradigm, which is widely used in
contemporary policy studies, is understood in this book. Thus, it seems to be
a rule that only neo-classical economic analysis is allowed in the Central
Budget Agencies of OECD countries. However, this rule is rarely, if ever,
explicitly stated in introductory training. Instead, new entrants are simply
presented with a series of policy analysis cases and answers. Given this, as
time passes, they become more skilled in the application of such economic
analysis methods and they develop a natural disposition to think of policy
issues in these terms. At this point, adherence to the rule that only neoclassical policy analysis counts as policy analysis becomes widespread. Neoclassical economics becomes second nature; indeed Hall (1993) introduces the
notion of a policy paradigm with respect to economic thinking about
macroeconomic policy. The key is the propensity to apply this method of
thinking, this paradigm, when intentionally directing consciousness at a
particular issue: this is Searles Background.

Structured policy narratives

71

How then does a change in a policy paradigm come about? Because the
following of institutional rules is second nature, the rules themselves are
potentially recoverable by consciousness and it is this potential recoverability
that explains institutional change. In other words, institutional change results
when agents come to have reason to direct their consciousness at what has
become second nature. This reason may be shifts in the external policy
environment, new ideas, or the accumulation of feedback on the consequences
of the policy paradigm.
This notion of change bears a strong resemblance to John Deweys work on
habits and intelligence (Dewey 1922). For Dewey, awareness of the existence
of a particular institutional rule emerges, as it were, from the background to
the foreground of consciousness. It is at this point that intentional human
agency holds out the possibility of innovative and creative institutional change.
A strategic and conscious element enters into habitual action; although the
habitus or background may predispose agents to act in particular ways, it does
not reduce them to cultural dopes or inhibit their strategic capacities.
It is important to point out that for a particular policy paradigm to fade and,
hence, for institutional change to occur, it is not enough that a single agent
comes to recover into consciousness the rule and stop behaving in accordance
with it. Rather, the population of agents to which the rule applies must,
collectively, stop behaving in accordance with the rule. Of course, if an agent
that is pivotal to the institution, or perhaps society more generally, either
declares that people should no longer behave in accordance with a particular
institutional rule, or stops herself behaving in accordance with that rule, then
this can induce the collective intentionality required for institutional change.
However, even in this case, the collective intentionality is a necessary
condition for change. Thus, although the potential recoverability into
consciousness of institutional rules that have become second nature helps to
explain the potential for institutional change, the need for collective
intentionality to actually institute that change helps contribute to institutional
stability over, possibly long, periods of time.

EVALUATING NARRATIVES
A narrative is a selection of elements such as events, steps and processes from
a chronicle; the long sequence of things that happened, which can be
organized by some interpretive frame to make sense as an overall story. The
narrative has to make sense to at least the person who made the selection, so
is not entirely arbitrary in the manner suggested by a strong reading of the Dr
Seuss imputation of ad-hocness. Further, just like professional historians,
policy scholars are concerned with the discovery, validity and reliability of

72

Theory

evidence. Thus, in terms of selecting items for narratives there can be some
agreement on the basic facts of policy development, or in the terms here, the
chronicle on which the narrative draws. But crucially, there can be agreement
on facts but substantial disagreement on what makes sense of them to different
people. Very different narratives are possible, and indeed it may be that
whichever group or individual has the power over the writing of the narrative
determines what is accepted as making sense. This is what motivates the
reluctance of some to use narratives; they are always extemporized and
should be regarded as data sets upon which to trial different models of policy
change.
Indeed, it is difficult to produce an objective standard or set of criteria
against which to assess or test the validity of a narrative explanation. One
response is to stress the competition between different narratives, different
interpretations and different explanations of the same thing, and argue that this
competition may advance the literature toward something that is a better
explanation or perhaps closer to the truth. Williams (2002, p. 252) puts it in
terms of the marketplace of ideas: Continued inquiry eliminates earlier
interpretations, which become, in the face of further information and more
searching questions, indefensible. This is particularly pertinent for the
analysis of policy dynamics, which is history that ends in the present or near
past. New evidence of the consequences of policy decisions or of certain
actions emerges constantly.
Additionally, the policies themselves are the subject of continued political
contestation. For example, subsequent policy development can give new
meaning to the preceding policy history: does this policy mark an aberration
or permanent shift in direction? Can you see permanent interests in a policy,
beyond those produced by specific periods of electoral competition? Has
interest group power been realigned? Has a certain policy idea become
institutionally embedded, or is its influence more parlous? Is this policy
change an ephemeral response to short-term events, or something that is path
dependent? New evidence on questions of these types will confront the student
of policy dynamics at a rate that alleviates some of the difficulties that
confronted historians of the far past and increase the competition between
existing narrative explanations. The constant possibility of new disconfirming
evidence will, by one view, intensify the competition among different
narratives. This yardstick criteria of a successful narrative it is better than the
alternatives in explaining the evidence is similar to the abductive reasoning
emphasized by Pierce and other American pragmatists.
An alternative response to the challenge for criteria to assess the value of
narratives is to accept that there are variations in what makes sense to different
people (either at the same or different times). This invites the label relativist
and it is important for my argument to explore what this might mean here. The

Structured policy narratives

73

first point, for philosophers, is that the position is not relative to the truth.
There are historical facts in policy development: this legislation was enacted
on this date; the budget for 2004 allocated 10 million for this; the leader of
the Greens said this on this date, and so on. Relative, rather, refers to the
judgement involved in selecting from the chronicle the material to form the
narrative and the use of different concepts, metaphors or models to provide
intelligible mechanisms in the narrative. Different scholars may argue
different interpretations or use different metaphors or concepts, and may
contest other scholars judgements. There are different disciplines with
different standards of justification, and different standards for explanations: a
belief in causal explanation does not provide a unique or universal method of
justifying different causal explanations.
By denying a universal standard, Rorty (1982, p. 166) argues, does not have
the corollary that there is no preferred standard of truth: Except for the
occasional cooperative freshman, one cannot find anybody who says that two
incompatible opinions on an important topic are equally good. The
philosophers who get called relativists are those who say that the grounds
for choosing between such opinions are less algorithmic than had been
thought.
This book notes that by accepting relativism we are accepting that different
standards of justification of an explanation exist within the social sciences.
The standard of this book understanding and explaining policy dynamics
requires making sense of unique temporal sequences. The chapter has set out
the key foundational assumptions of policy narratives and how they allow
explanation of policy development:
l

Policy is a heuristic concept, which when applied with a dynamic

perspective allows a path to be drawn through policy space. This path,
which consists of a series of steps between different states of the policy
system, acts as structure for a narrative.
A policy narrative seeks to explain these steps and uses the socially
situated agent as its basic unit of analysis.
The microanalysis of the policy paths requires the investigation of the
dynamic interaction of the social situation and agency in the multitude of
contemporaneous processes immanent in the transition of policy systems
through successive temporal states.
Policy narratives highlight the conjunctural contingency of these
different processes at different speeds as the key driver of policy along
the path, step by step. In doing so, explanation in terms of the intentional
action of agents is either limited to relatively short temporal horizons; or
in some cases dynamic perspectives upset the entire notion of policy as
intentional choice.

74

Theory

These key principles act as the foundation of policy narratives; they inform
what should be included, which are the key stages, how those stages should be
analysed, and they push the analyst toward constructing a coherent narrative,
which makes sense of a unique temporal sequence of policy development. By
virtue of acting as the foundation of policy narratives, these principles also act
as a guide for the evaluation of narratives as explanations of policy dynamics.

PART II

Evidence

6. The development of the EU budget

system
The historical development of the EU budget system has involved the
accumulation of a set of institutional constraints that limit the scope for EU
policy activism in terms of redistributive and distributive policies. This process
of EU integration is often underplayed relative to analyses of integration in
terms of big step Treaty agreements, such as the single market, single
currency and enlargements. This chapter provides a structured narrative of the
dynamics of institutional change in the EU budget system, where the
incompleteness and limitations of the initial Treaty-based budgetary
institutions have worked themselves out over time through a series of
historical, political and fiscal contexts.
This process of development has seen the accumulation of additional and
complementary budgetary institutions that have had the effect of significantly
constraining expenditure-generating EU-level policy activism. The granularity
of perspective adopted in the narrative helps to reveal the importance of
relationships between institutions; in isolation, institutions may be properly
labelled as either stable or changing over time, however once institutions are
viewed in combinations, or as part of systems of interdependent rules, the
question of institutional change is more complex, requiring description of the
conflux of interdependent causal relationships and some analysis of how they
are compounded into an overall net effect that connects the different temporal
states of the EU budget system.
The narrative presented includes an increasing returns process operating at
the level of sequences of institutional choice: once an initial institutional
framework is established, there are strong increasing returns involved in the
choice of new, supplementary institutions within that framework. That is,
an increasing returns process explains institutional change qua the
introduction of new, supplementary institutions. As North (1990, p. 95) states,
it is the interdependent web of an institutional matrix that produces massive
increasing returns through strong learning effects, coordination effects and
adaptive expectations. The analogue of the choice of technological standards
in a market from which the dominant strand of path dependency analysis
is derived is the sequence of institutional choices over time that increase
the pay-offs for certain choices further on in the sequence; it is not the
77

78

Evidence

choice of a single institution at one point in time that then persists or is stable.
This has an immediate obvious appeal in terms of explaining the
development of the EU budget system where the Balanced Budget Rule (BBR)
in the 1957 Treaty of Rome, and the 1970 Luxembourg Treaty distinction
between Compulsory Expenditure (CE) and Non-Compulsory Expenditure
(NCE), have persisted unchanged, consistent with a path-dependent process.
Further, one of the standard micro-mechanisms in an increasing returns
process is the existence of significant switching costs. Replacing the existing
constitutionally entrenched budgetary framework would be costly politically,
as budget reform tends to be a zero-sum game, or where switching costs are
significant, a negative-sum game. Since Treaty changes require unanimous
agreement by the member states, potential losers possess a veto.
If we adopt the perspective of the budget system as a single whole, this
expedites the use of path dependency as a concept that insists on an overall
trajectory for an institution and institutional configuration, the direction of
which is reinforced after early moves in the sequence. This is fine; however the
issue raised in Chapter 3 is that within a path-dependent system some elements
may be fixed or locked-in, while others are capable of being reformed. Also
within a path-dependent system, space may exist for the introduction of new
institutions. This is crucial to my analysis here: the initial budgetary
framework set in train path-dependent process in which institutional change
has been observed, and as will be shown, new budgetary institutions have been
made or remade for various reasons with a variety of effects but all have been
premised on the continuation of the initial budgetary framework; that is, all
change has occurred within that framework.
The danger of adopting an institutional configuration the budget system
as the unit of analysis is the creation of too sharp a distinction between
stability and change, as seen in on-path versus off-path change. This is a
function of the granularity of perspective and is a general charge against
historical institutionalism and path dependency. This unit of analysis leads
some to see 1988 as a path-breaking juncture in the EU budget system, with
the introduction of a new institutional setting: The comparison between the
final breakdown of the 1970 institutional setting and the continuous stability
of the 1988 institutional setting revealed that the specific combination of
reproduction mechanisms present in the 1988 setting was better equipped to
sustain stability (Lindner 2003, p. 932).
In the narrative presented here, institutions exist in combinations: they are
interdependent, with necessary and contingent relationships. Thelen (2003)
describes examples where institutional lock-in is combined with elements of
institutional innovation that can push the overall trajectory of policy and
politics in a different direction. This encourages a shift in the ground from a
sharp distinction between institutional persistence and institutional change

The development of the EU budget system

79

towards a position where (2003, p. 233) to understand how institutions

evolve, it may be more fruitful to aim for a more fine-grained analysis that
seeks to identify what aspects of a specific institutional configuration are (or
are not) negotiable and under what conditions.
While the concept of path dependency is a useful metaphor for the
dynamics of the EU budget system, the underlying increasing returns process
actually operates here at the level of combinations of institutions, or
interrelationships of institutions, rather than at the level of the individual
institution. The initial design of the EUs fiscal constitution did not establish
complete institutions; these are institutions that are singularly self-sustaining.
Instead, the initial Treaty-based institutions can be defined as a budget
framework or an institutional space in which other institutions could develop.
For example, the constitutional BBR in the Treaty of Rome has required a
series of other institutions in order to ensure that the budget balances in
particular, rules to coordinate spending and revenue decisions. The
contingency for breaches in the BBR that the European Court of Justice
declares the budget unconstitutional was insufficient to exercise direct and
immediate control over CE.
The narrative provides microfoundations to the institutional dynamics of the
development of the budget system in terms of situated agency: member states
as actors in the budget process have, as a minimum, two objectives. First, that
the EU avoids bankruptcy, and second that their net budget position remains
stable. Each actor would wish to maximize their net transfers but, given
national vetoes and the zero-sum nature of the budget game, stable net
positions (discussed below) represent the best outcome for the member states
collectively. The second objective dominates the politics of budget reform in
the EU in the absence of a threat of or an actual budget crisis. In the event of
the EU facing bankruptcy, member states preferences with regard to budget
share stability and respecting the BBR define a domain of feasible
compromise for the development of an additional or supplementary set of EU
budgetary institutions to confirm or validate the initial, Treaty-based
institutions.
The evolutionary metaphor discussed in Chapter 4 can be used to describe
the process of this search for budget institutions in response to the regular
budget crises that affected the first decade or so of the EU financial system.
The process was akin to a form of trial-and-error learning with slow feedback
mechanisms. Agreement on institutional reforms was mediated in periods of
fiscal crisis at significant political cost, which meant that they were difficult to
reverse; however the consequences of the reforms in correcting the tendency
of the EU budget system to crisis was only observed at some later point, that
is, when the threat of bankruptcy again emerged. The feedback was relatively
slow but the next crisis did not lead to the removal of the previous budget

80

Evidence

reform; it rather led to the introduction of another budget institution. Thus the
process was not classic trial and error, in the sense that budget institutions that
had not worked were not dismantled; they were instead added to. The process
of accumulation of additional and complementary institutions to avoid the
tendency to crisis is the dominant dynamic in the development of the EU
budget system.

THE INTERRELATIONSHIPS OF INSTITUTIONS

The notion of state space discussed in Chapter 1 as the foundation of
dynamics, implies both a temporal and spatial context to institutional
development. An appreciation of both is essential to examine the notion that a
multitude of institutional development processes may be simultaneous within
an institutional system in particular, to understand the claim that an
institutional setting is path dependent does not require that all institutions
contained therein are unchanged.
In spatial terms, the interrelationship between institutions may define a
space within which complementary institutions could develop. The most
prominent spatial interrelationship is between different levels of institutions.
Three levels or layers are regularly distinguished: the macro- or constitutional
level; the collective choice or policy decision level; and the operational level
of individual decisions. There is no simple structure to the relationship
between these three levels: constitutional rules affect the policy decision rules,
which in turn affect operational decisions, but some constitutional and
collective choice rules are subject to the control of operational-level decisions.
This has been a theme running through the theoretical chapters of the book;
to understand the layering of EU budgetary institutions requires
understanding the relationship between the initial Treaty-based rules and the
complementary institutions that have developed subsequently. In the narrative
presented here, this relationship is an increasing returns process; the new
institutions that have emerged complement and reinforce the initial budgetary
framework.
The temporal aspect of inter-institutional relationships is important because
it raises questions of inherited legacies and the extent to which institutions, as
artefacts of past decisions or actions, circumscribe or condition a particular
institutional space in which parallel or related institutions may develop. Whilst
agents are situated rather than determined, and remain capable of remaking or
reforming these inheritances, institutions tend to endure (almost
definitionally). Further, as dynamic analysis emphasizes, institutions are often
reproduced across time and may have consequences that bear little
relationship to their designers initial intentions.

The development of the EU budget system

81

Budgetary institutions are necessarily interdependent; they exist as a system

and evolve, co-evolve, complement and depend upon each other. Each
relationship between a constitutional rule and a policy-level rule cannot
accurately be studied independently of the other relationships in a multi-level
system, because a change in one affects the other two; that is, ceteris paribus
does not hold. The EU budget process is an emergent property of the
relationship between different actors and different budgetary institutions
across time. The challenge for all narratives of a complex system is to balance
microanalysis of the interactions of the constituents with a consideration of
system-level properties.

THE INCOMPLETE INITIAL DESIGN OF THE EU

BUDGET SYSTEM
In order to help structure a narrative of the complex interdependency of budget
institutions, I use Buchanan and Musgraves (1999, p. 118) identification of
two types of constraints affecting budgetary institutions: those affecting the
rules for reaching collective decisions are procedural constraints; whereas
domain constraints affect the set of permissible outcomes or solutions that
may be allowed under any agreed-on procedures. The 1957 Treaty of Rome
set out the basic framework for EU budget-making, containing the most
important domain constraint: the revenue and expenditure shown in the
budget shall be in balance each year (Article 199). This Balanced Budget Rule
(BBR) was included because the Treaty authors did not wish to offer the
Communities, and in particular the Commission, any easy solutions on
spending (Strasser 1992, p. 57). The development of a budget system that
produced budgets and budgetary outcomes consistent with, or which validated,
this constraint is the chapters narrative, for while the BBR denied easy
solutions it never assigned responsibility for hard choices to any individual
or organization in the budget process.
The EUs own resources system (OR) came into force in stages during the
1970s as first the Common Customs Tariff (CCT), then the Variable Import
Levy (VIL) of the CAP and finally, in 1979, a VAT-based contribution, which
passed to the financial authority of the EU. However, there were limits placed
on the ability of the EU to exercise control over this revenue base. Importantly,
VAT contributions are based on a two-stage calculation. First, the VAT base
is determined as the total revenue that would be collected in each country if
they adopted a common, hypothetical VAT system designed by the
Commission. Payments to the EU budget are then specified as a percentage of
the VAT base (the call-up rate). Crucially, this is subject to an annual
maximum (initially set at 1 per cent). This, combined with finite VIL and CCT

82

Evidence

revenues, imposes a domain constraint on total revenues that limits total EU

spending each year.
In terms of procedural constraints, the 1970 Treaty of Luxembourg
introduced a distinction between two classes of expenditure, differentiating the
powers over each that were granted to the European Parliament. Spending is
categorized either as CE defined in the Treaty as spending necessarily
resulting from this Treaty or from acts adopted in accordance herewith, or
NCE all other spending. The distinction has no administrative or accounting
logic but importantly the Parliament was granted greater powers over the latter
than the former, which in practice meant it was easier for Parliament to
propose changes to NCE than to CE and harder for the Council of Ministers
to reverse or otherwise alter changes to NCE.
In addition to limited powers over CE, Parliament faced other limits. First,
Article 203 imposed a domain constraint by limiting the annual growth rate of
NCE. The Commission fixed this with reference to trends in national GNP; the
average variation in member states budgets and the previous years rise in the
cost of living. Second, Parliaments de facto budgetary influence was limited
by the fact that, initially, NCE represented less than 5 per cent of total EU
spending (although the figure is now much closer to 50 per cent).
During the 1975 budget process, the Council defined CE as all expenditure
in respect of which, by virtue of existing enactments, no budgetary authority,
be it the Council or the European Parliament, has the right freely to determine
the appropriations (Strasser 1992, p. 176). Thus CE cannot be controlled
directly once the expenditure-generating policies and instruments are
agreed, all resulting expenditure obligations must be met. In practice, control
over the level of CE is greatest for the institution with the greatest say over the
setting of policy; this meant, before 1988, the CoAM (the Council of
Agricultural Ministers), which determined CAP Guarantee expenditures
(amounting in some years to almost 80 per cent of total expenditures).
The CoAM took decisions on the CAP without regard for the overall
balance of the budget. This, coupled with the open-ended nature of support,
discussed here, created this tendency to fiscal crisis. Periodic crises prompted
a search for other institutions to validate or enforce the initial BBR domain
constraint; I use the evolutionary metaphor of policy development set out in
Chapter 4 to describe this process. There was an overall budget system that
was path dependent but with the significant design flaw of lacking an EU-level
budget authority or anything in the EU budgeting system which required the
BBR constraint to be respected: even where the BBR was breached there were
no immediate contingencies for balancing the budget in the initial design. In a
form of trial-and-error learning, budget reforms were introduced for various
reasons at various times of budget crisis; the softer or more politically
palatable options were tried first, and when these failed to bring discipline to

The development of the EU budget system

83

the budget system (where failure was judged by the emergence of a new
crisis), new reforms were tried. The accumulation of these additional
budgetary institutions is the salient pattern of the institutional dynamics of the
EU budget system.
In understanding the evolution of the EU budgetary institutional matrix, it
is important to stress that the BBR, in combination with the OR system,
ensured that negotiations between member states in the Council on changes to
budget rules, or the introduction of new rules, were a zero-sum game. In
response to budget crises, member states had two main preferences for
institutional change: first, to optimize their net budgetary position and second,
to ensure their budget position was relatively stable across time. Given the
zero-sum nature of the negotiations, institutional change that satisfied the first
type of preference for all member states was infeasible. However, an
institution that achieved the second type of preferences was possible. The data
presented in Ackrill and Kay (2006) show a secular decrease in the volatility
of annual net budgetary positions for almost all member states. This was the
key microfoundation in the agreement of options for budget reform.

THE DEVELOPMENT OF THE EU BUDGET SYSTEM,

1970 TO PRESENT
The incomplete set of initial EU budget rules was exposed by the use of price
support instruments in the CAP that stimulated agricultural production and
caused surpluses, with the corollary of increased budget spending. The control
of CAP expenditure is crucial in the narrative of the development of the EU
budget system. The operation of the CAP precipitated regular crises in the
budget of the EU, and a series of institutional changes in the budget process
that failed to ameliorate the outstanding problem of the incomplete set of
initial budgetary institutions: the lack of a hard and enduring financial
constraint in the CAP decision-making process.
Price support, the dominant form of agricultural support until 1992,
operated by maintaining market prices at levels higher than those prevailing in
the rest of the world. The associated instruments had various consequences for
the EU budget. Cheaper imports below the threshold price were prevented
from entering the EU by means of the VIL, which then passed to the EU
budget. On the other hand, exports to third countries required subsidies (or
restitutions) to make them competitive with cheaper third country supplies.
These subsidies came from the EU budget.
As production rose, the EU market price was driven down. A system of
intervention storage was thus established, so that if the market price fell below
the target price set by the CoAM, farmers would receive a guaranteed

84

Evidence

minimum price by selling to the government instead. The associated costs

again came out of the EU budget. Moreover, intervention was key to farmers
having an incentive to produce ever greater quantities: the higher market price
would only translate through to higher revenues if farmers could sell their
produce. Intervention not only guaranteed a price above world levels, but also
offered a guaranteed buyer. Price support drove up production, which drove up
surpluses, which drove up budget costs. In addition, higher EU production
reduced imports, lowering VIL revenues to the EU budget.
Given the nature of CE, the EU could not stop exporting surpluses and/or
buying them into intervention just because spending was rising. The only way
to alter the trajectory of CAP spending, therefore, was to change the
expenditure-inducing policy instruments: that is, reform the CAP. Chapter 6
presented the dynamics of those reforms; here I examine them in terms of the
budget system.
In 1979, the first year the EUs OR system operated in full, spending
required a VAT call-up rate of 0.78 per cent. The Commission had already
begun discussing options for future financing, given that the CAP was driving
total EU spending towards the OR ceiling, as defined by the 1 per cent
maximum VAT call-up rate. The institutions governing own resources were
softer at this time, in the sense of being flexible in periods of fiscal
exigencies to accommodate the consequences of CoAM decision-making. For
example, by 1983 a combination of surpluses and low world prices saw claims
on the VAT resource rise to the maximum 1 per cent, but this limit was
respected only by carrying over ECU825 million of spending (more than 3 per
cent of the 1983 total) to the 1984 budget. This was unsustainable and in both
1984 and 1985 the EU budget, technically bankrupt, required additional
payments from the member states to keep operating.

THE FONTAINEBLEAU AGREEMENTS, 1984

Amidst this budget crisis, the first significant reform of the CAP was agreed,
and concentrated on the dairy sector, which was taking over 40 per cent of
CAP spending and 30 per cent of total EU spending. The Fontainebleau
European Council Summit of June 1984 also approved three key budgetary
measures. First, agreement was reached in principle to raise the VAT call-up
rate to 1.4 per cent. This was a budgetary institution that could be changed
relatively easily, as opposed to the BBR that was entrenched constitutionally.
Second, following years of dispute over the magnitude of the net contribution
paid by the UK (and four years of ad hoc compensation payments), a formula
was agreed to return back to the UK two-thirds of its net contribution each
year (implemented via national VAT contributions). Recognizing that

The development of the EU budget system

85

Germany was the largest net contributor, its contribution to the UK rebate was
reduced by one-third (with the additional cost shared among the other member
states). Third, in response to concerns over the unrestrained growth in
spending, general guidelines were produced for Budgetary Discipline,
transformed into several rules in December 1984. The most important of
which were that the Council of Finance Ministers should set a reference
framework for total expenditure, with other Councils asked to ensure that their
decisions respected this, and that the growth rate of CAP spending should not
exceed the growth rate of own resources.
This was the first public recognition of the design flaws in the initial
framework; that the BBR was an incomplete institution without the
complementary institutions to control expenditure. There was however no
incentive for other Councils, most notably CoAM, to respect the spending
guideline. As a domain constraint, Budgetary Discipline remained flawed:
there were no institutions proposed or agreed that gave effect to the agreed
spending limits. In terms of this development as an example of evolutionary
policy learning, this was the first of a series of reforms that failed to
institutionalize mechanisms to limit expenditure by Councils and respect the
fiscal constitution of the EU.

THE BRUSSELS AGREEMENTS, 1988

By 1987 deep concerns were emerging within the Commission about the
budget: the Community is at present faced with a budgetary situation which
can only be characterised as being on the brink of bankruptcy (Commission
of the European Communities 1987, p. 1). Prominent actors outside the CAP
system began to interpret the situation as undermining the ambition to
complete the Single European Market, limiting the scope for developing
existing, non-CAP policies (notably regional policy) and restricting new
measures (such as research and development). President Jacques Delors,
supported by an inner circle of the Budget and Agriculture Commissioners,
prepared a package of measures, agreement on which was forthcoming in
Brussels in February 1988.
Some of the measures addressed directly the shortcomings of the 1984
Budgetary Discipline agreement. The reference framework from 1984 was
hardened into a five-year Financial Perspective, which set out EU spending
in total and was disaggregated by main policy area. These have been agreed
subsequently every seven years, on average, and have become a key feature of
the EU budget process. Spending was allowed to rise to enable the EU to
accommodate its changing policy priorities. In particular, spending on regional
policy was to double, to 25 per cent of the total by 1992, whilst CAP spending,

86

Evidence

though allowed to rise in absolute terms, would fall as a percentage of the

total. This was all negotiated within a controlled rise in the total own resources
ceiling. Starting at 1.15 per cent of EU GNP in 1988, it rose to 1.20 per cent
in 1992 and continued through the second Financial Perspective, to reach 1.27
per cent of GNP in 1999. It has remained at this level ever since, although a
technical change was made recently, with this sum now expressed as 1.24 per
cent of Gross National Income. A fourth own resource, based on relative GNP,
was introduced and could be used to top total own-resources up to the
specified maximum.
Moreover, the Financial Perspective was presented through an Interinstitutional Agreement (IIA), a device that has become an increasingly
important element of budgetary planning and that was crucial in completing the
institutional matrix of the EU. A key feature of the IIA was that it was binding
on all members of the Budgetary Authority (Commission, Parliament and
Council). Any changes to the Financial Perspective have to be agreed by all
three other than annual technical adjustments to real-terms spending limits to
allow for inflation and GNP growth. Furthermore, the Council and Parliament
are bound by the maximum rates of increase for NCE laid down in the IIA.
Monar (1994, p. 698) argues: IIAs establish rules and principles which if
effectively complied with will limit the future freedom of action of the
institutions. The subjunctive is critical; problems remained, most notably with
the agricultural guideline, which was left as an incomplete institution after
both the 1984 and 1988 agreements on Budgetary Discipline. The 1984
agreement was ineffective as it failed to impose a binding obligation on CoAM
to respect the guideline. Even though the 1988 reform reduced the permitted
growth rate of agricultural spending to no more than 74 per cent of the growth
rate of EU GNP, once again it was not accompanied by the imposition of a
direct domain constraint, making its enforcement just as unlikely as the 1984
agreement.
There was also in 1988 a simultaneous attempt to control CAP spending
directly via the Stabilizers reform. The initial proposal was for price support
levels to be cut should spending exceed a certain trigger level; however, as
this proved unacceptable politically, a production trigger was agreed instead.
When production activated the trigger, that is, it exceeded a certain level
expressed in terms of total EU production, automatic support price cuts would
be imposed the following year. There are, however, various reasons why this
was inadequate, most notably because CAP support prices were typically
4060 per cent above world levels, yet the Stabilizer-induced price cut was
limited to a maximum of 3 per cent and for some commodities prices were
restored the following year if the production trigger was not reactivated. The
fundamental basis of support thus remained unchanged, as did the trajectory
of CAP spending. Thus the 1988 CAP reform too failed to impose an effective

The development of the EU budget system

87

domain constraint on CAP spending; in terms of the trial and error

development of the EU budget system, neither the budgetary nor CAP reforms
of the 1980s imposed an effective procedural constraint, nor a domain
constraint, on CAP spending. Indeed, the term domain is doubly appropriate
given that CAP spending remained the unchecked domain of CoAM, despite
the terms of the relevant Budgetary Discipline criterion. Thus, by 1992, the EU
was facing another budget crisis.

THE MACSHARRY CAP REFORM, 1992

This 1992 reform, named after the Agricultural Commissioner of the time, Ray
MacSharry, remains the most far-reaching CAP reform. It included a
substantial weakening of the production (and spending) incentives of price
support. Simultaneously, it imposed domain constraints on CE and the
budgetary freedom of CoAM. The timing and direction of reform were also
motivated by pressure on the EU through the Uruguay Round of GATT trade
talks. Under such pressure the decision was taken, first, to reduce high support
prices not by 3 per cent but by about 30 per cent. Second, replacing high prices
as the main instrument of farm income support would be direct payments to
farmers based on historical production levels. The payments for beef
production contained a limit on the total number of animals eligible for
support. For the arable area payments there were two limits. First, the yield
figure used to convert the per-tonne figure (derived from the initial price cut)
into a per-hectare payment was fixed in terms of data from an unchanging
reference period; there was also a limit on the area eligible for payments.
It was in the design of these payments that the key budgetary change was
made. A direct payment to farmers as compensation for price cuts allowed a
spending limit to be defined ex ante; this acted as a domain constraint on CAP
spending and for the first time in the history of the CAP meant that support
payments were closed-ended rather than open-ended. The two alternatives to
direct income payments as methods of imposing financial discipline carried
significant political costs: agreeing a new definition for CE required a Treaty
change and challenging the classification of CAP spending as Compulsory
would have ignited CouncilParliament feuds over budget control.
The MacSharry reforms stand out in the history of CAP reforms in finally
limiting the discretion of the CoAM to agree expenditure without reference to
a wider budget system; the direct income payments were set in the context of
the agricultural guideline, itself agreed in terms of the five-year Financial
Perspective premised on the BBR. The creation of a link between the BBR and
agricultural spending amounts to the completion of the institutions of the
initial budgetary framework.

88

Evidence

HARD, BINDING CONSTRAINTS?

EU spending (as a percentage of GNP) rose through the first two Financial
Perspectives. Since 1999, however, the spending ceiling has remained at 1.27
per cent of GNP (1.24 per cent of GNI). Wallace and Wallace (2000, p. 230)
suggest this is because further increases would have required ratification by
national parliaments and might have prompted acrimonious domestic debates
on future financing. Indeed, instead of seeking to spread a larger budget over
more policies for a given membership, the EU budget debate shifted to how to
keep the budget size and policy coverage constant whilst EU membership
rises. In agreeing the Financial Perspective for 20002006, the European
Council reduced the level of spending proposed by the Commission under
every heading (Wallace and Wallace 2000, p. 232). Moreover, whilst the own
resources ceiling was held at 1.27 per cent of GNP, the margin for unforeseen
spending, previously stable at about 0.03 per cent of GNP, ranged in the initial
Perspective for the EU15 between 0.08 per cent and 0.15 per cent of GNP.
Thus, as a percentage of GNP, planned spending fell significantly.
The 1999 IIA also changed the decision-making procedures to make it
harder for spending to be increased, even within the margin for unforeseen
expenditure. This limited still further the room for manoeuvre within the EU
budget system. For an increase of less than 0.03 per cent of GNP the Council
votes by qualified majority voting (QMV), with Parliamentary approval
needing a majority of members with three-fifths of votes cast. For spending
rises greater than 0.03 per cent of GNP (but still within the permitted margin)
Parliament and Council must agree, but with the Council acting unanimously.
Agreed as part of the 2003 CAP reform, and effective from the start of the
next Financial Perspective in 2007, a further domain constraint has been
imposed on CAP spending: if forecast CAP spending exceeds its guideline by
more than 300 million, the Commission must propose cuts to direct
payments to ensure the guideline is respected. This serves to harden the
constraint on CoAM by making explicit the contingency if the CoAM had
agreed a level of expenditure, which remains defined as Compulsory, above
the guideline. Under price support, this direct control of CAP spending had
been incompatible with the definition of Compulsory Expenditure.

SUMMARY
Institutional reform can have extremely high political costs, in terms of the
bargaining of agreement on a new institution, the switching costs from old to
new, and the transaction costs of adapting to and learning a new institution.
These costs are also subject to an increasing returns process; the more

The development of the EU budget system

89

institutions that are designed and agreed within an institutional framework or

configuration, the less costly it is for subsequent and additional institutions to
be agreed within that framework. In this chapter, the notion of path
dependency has been used to narrate the history of reforms to the EU budget
system from its foundational institutions as laid down in the 1957 Treaty of
Rome.
I have sought to draw upon theoretical developments in the first five
chapters of the book in order to provide a fine-grained analysis of the
development EU budgetary institutions. This has emphasized the stresses in
the initial design of the budget system working themselves out in different
fiscal and political contexts as the inter-temporal connection between reforms
of EU budgetary institutions. It is only by adopting a dynamic perspective that
it is possible to trace the emergence of the political significance of the gaps in
the initial institutional structures of the EU budget and how various budget
crises precipitated unsuccessful efforts at remedial action. I have characterized
this sequence of reforms as a form of evolutionary learning process that lead
to the accumulation of layers of institutions that only provided a complete
system of budget institutions when the politics of the CAP allowed the
agreement of a direct income payment policy instrument that could then be
used to impose a financial constraint on CoAM decision-making.
A key point in the narrative is that whilst the 1988 reforms to the EU
budgetary process were profound, it was not until the 1992 CAP reform that
the significant gap in the institutional framework of the EU budget was
plugged: the need to respect one of the original budgetary constraints set out
in 1957: a balanced budget for the EU. The new institutional framework has
created a much more stable budgetary process than was seen previously.
Despite this, the recent agreement on Financial Discipline as part of the 2003
CAP reform indicates that, in an evolving structure such as the EU, new
problems can emerge, requiring the introduction of yet more institutions. As
noted previously, policy history is history that ends in the present or near past.
This makes any narrative explanation presented especially vulnerable to the
emergence of disconfirming evidence. Thus, the structured narrative presented
here that concludes that the EU has a complete set of budgetary institutions
because it has a clear institutional link between the BBR and agricultural
spending may be premature.

7. The Common Agricultural Policy

19772003
The Common Agricultural Policy (CAP) has been a continuous presence since
the inception of the European Economic Community and in key aspects has
remained resistant to substantial reform. As a complex policy system, the CAP
provides a difficult empirical challenge to dynamic analysis. The preceding
chapters have established that viewing policy as a composite whole or system
is an appropriate theoretical starting point for a dynamic perspective on policy
development. The empirical challenge is that of microanalysis: how to analyse
the constituent elements and their interrelationships in the context of an overall
identity for the system, marked by a point position in policy space. In other
words, how is a multi-commodity, multi-level and multi-national policy
whole operationalized as a variable? What measures or indicators may be
used to trace or map a path through policy space that can be used to structure
a narrative?
This chapter aims to address these difficulties through examining a series of
CAP reform events, from 1977 to 2003, in terms of the CAP as a composite
variable. This structured narrative shows that reforms of the CAP over time,
when viewed at a system level, have occurred within a path-dependent
process; but simultaneously and within the CAP system over this period a
multitude of processes at different temporal scales can be observed. The
chapter suggests a microfoundation to the path-dependent process in terms of
situated agency: the protection of the net budget position of member states
with respect to the CAP as an objective for national governments bargaining
within different councils of ministers. The data on distribution of CAP-related
budgetary transfers across member states over time as the EU has enlarged and
the CAP reformed have shown great stability.

PATH DEPENDENCY AND THE CAP

As noted in Chapter 1, it is possible for a composite system to maintain its
identity even where some of its parts are changing or being replaced. Thus,
even if we decided that the CAP as a policy system was path dependent, this
would not preclude change in the constituents of the system or their
90

The Common Agricultural Policy 19772003

91

interaction. This reintroduces the philosophical question of changes in kinds

and changes in values: at what point do changes in the constituents mark a
change in the overall system? This open, and ultimately empirical question,
affects how a narrative is structured. I make a start by delineating a policy
system from its environment and, following economists, label the
environmental parameters as exogenous, distinct from the endogenous
variables that are determined by the internal workings of a system. The
distinction is pragmatic and synthetic; imposed by the scholar structuring the
narrative to facilitate analysis. It points out at least one way of identifying a
change at the policy system level, or a change in kind: by establishing a
synthetic endogenous/exogenous barrier, any temporal point at which
environmental factors that were previously assumed in the analysis to be
relatively fixed and external to the policy system become part of the policy
system marks a change in that system. In economists terms, the
endogenization of a variable necessarily involves system-level change. The
endogenous/exogenous barrier varies temporally and any structured narrative
should reflect this, as where the financial actors in the CAP policy system have
had an epiphenomenal influence in terms of temporally specific reform
episodes but were not institutionalized as a permanent interest in the CAP
policy system until much later.
There are several dimensions to the CAP: its objectives; its effects; the
different instruments and their levels; the different commodities; and its policy
process. There is no straightforward or automatic method for constructing a
single, holistic variable the CAP from these different dimensions. Indeed,
analysts will weight different components of the CAP differently, when
forming an overall CAP-level view. This is why, for example, some authors
(such as Ackrill 2000a) consider the 1992 reform as radical, whilst others
(such as Kay 1998; Daugbjerg 1999) suggest it falls some way short of that.
In understanding the composite variable, the CAP, as path dependent the
starting point is to consider how the policy, as initially established, limits
future policy options by creating a particular incentive structure that influences
political and economic behaviour. I do not detail the establishment of the CAP
here, but note that the CAP took shape under various pressures, including postwar food shortages and constitutional foundations laid down in Articles 38 to
43 of the Treaty of Rome, which included three pillars of the CAP: a single
market, community preference and financial solidarity. The various reasons
why price support emerged as the dominant policy instrument are connected
with the problems of the main alternatives, especially direct payments:
political (opposition to the CAP offering social payments), administrative
(there were many more farmers in the late 1950s) and economic (the EU
budget, at the time, could not have coped with the cost).
Chapter 4 provided the insight that narratives of policy development should

92

Evidence

focus on the internal dynamics of a policy system as well as the systems

interaction with the external environment. The development of the CAP has
been driven by the consequences of certain design features of the price support
policy, all of which were foreseen at the time (European Commission 1958,
p. 70): the direct link between production and support stimulated production
and, through high prices, contributed to the suppression of demand growth. As
surpluses emerged so the impacts of price support on the EU budget became
apparent, in the form of the revenues from variable import levies (later import
tariffs) and the costs of dealing with surpluses export subsidies and
intervention storage. Moreover, a certain policy-contingent pattern began to
develop in the distribution of these budgetary transfers across member states:
net exporters gained from the surplus-related transfers from the EU budget,
whilst net importers faced greater transfers to the EU budget (through the
levy/tariff). Further, as the underlying basis of support remained unchanged
until at least 1992, the consequence was a continual rise in budget costs that
have been observed as policy encouraging higher production and (sooner or
later) surpluses. Those initially gaining from the budget continued to do so
(net exporters did not suddenly become net importers), but more countries
joined them in becoming net exporters.
CAP reforms have traditionally been agreed unanimously. The ability of a
single member state to veto a reform proposal tends to favour the status quo,
whilst the willingness to support change will be influenced, in part, by how EU
member states view the CAP transfers they receive from the EU budget. Later
in the chapter the view of member states as agents adopting positions in EU
negotiations solely in terms of the net transfer they obtain through the EU
budget is examined. Net budgetary beneficiaries will often oppose CAP
reform, whilst even those seeking reform will not pursue options that see their
receipts fall. Total CAP spending will, therefore, tend to be maintained even if
some elements of the policy are altered.

SELECTED KEY FEATURES OF CAP REFORMS

Since 1977 CAP reforms have generally been the result of pressures from one
or both of two sources the EU budget and, more recently, international trade
negotiations. In these cases, environmental parameters that had been
previously fixed and external to the CAP system for the purposes of analysis
become crucial influences in the CAP policy process and associated with
system-level change. With the former, the problem lies with most CAP
spending being defined as Compulsory, meaning all CAP expenditure
obligations must be met, regardless of the overall situation facing the EU
budget. This would not of itself be a problem (ignoring the issue of judging an

The Common Agricultural Policy 19772003

93

appropriate level of spending) were it not for the presence of a balanced

budget rule. The effect of this rule is to mark a point, defined in terms of
revenues, beyond which spending cannot rise.
However, price support generated rising export refund and intervention
storage costs because the wedge between EU and world prices encourages
higher production and, sooner or later, surpluses (an incentive structure that is
unaffected by marginal changes to the size of the price wedge). Central to the
incentive to produce is the role played by the intervention system in
guaranteeing farmers an outlet for surplus production, however much is
produced. As spending rises, the BBR imposes an effective limit. Given the
nature of CE (and the centrality of high prices and intervention to the
execution of Article 39), the necessary response to rising spending is a change
in the policy the tap of CE cannot simply be turned off if the revenue limit
is reached. Price support thus contained an open-ended commitment.
The budget pressure on the CAP therefore arises from a combination of the
budgetary consequences of price support and two endogenous budget-related
rules agreed by the member states the definition of CE and the BBR. The
second pressure on the CAP comes from other countries through the
GATT/WTO where some, but not all, forms of agricultural support policies
have been challenged. Policies which distort trade (including price support)
are opposed, whilst other policies that support farming incomes but as far as
possible de-couple support and production are acceptable. This issue is
considered later.

REFORMS BEFORE 1992

The reform options facing the EU were various: leaving price support intact,
supplementing it with extra policy instruments; reducing price support levels
with larger cuts accompanied by alternative income-support instruments;
through to replacing price support totally. In this section we look at CAP
reforms in the context of the EU budget reforms of 1984 and 1988, which, as
noted in the previous chapter, altered the environment in which CAP-related
decisions were taken. This was the first occasion where actors concerned with
the budgetary consequences of the CAP entered into the CAP policy system.
However, their influence was limited to specific reform episodes and was not
permanently institutionalized.
Principles of Support
The basic principle of the CAP, the constitutional commitment to support
agriculture set out in Article 39 of the Treaty of Rome, has remained

94

Evidence

unchanged since 1958. Moreover, most reforms prior to 1992 left the
underlying principles of price support intact. The one notable change to price
support (agreed in 1988) was an automatic price cut if production exceeded a
Maximum Guaranteed Quantity. Some of the reforms shown in Table 7.1
merely limited the extent of price rises (notably Guarantee Thresholds), whilst
even when support prices were cut (for example, through the cereals
Co-Responsibility Levy (CRL) and Stabilizers) the reduction was modest.
Given that the margin of EU prices over world prices was often in excess of
50 per cent, the fundamental basis of policy and the resulting production
incentives remained unchallenged.
Table 7.1 CAP reforms before 1992 a summary of key features
Date

Reform

Description

Pressure Binding?1

1977

Dairy CoResponsibility
Levy (CRL)
Guarantee
thresholds
(multicommodity)

Producer levy (max. 3% of

target price)

Budget

No

Negotiated cut (max. 5%) in

support prices if production
exceeds specified amount
after normal price rise is
agreed
100% super-levy if
production exceeds specified
amount
Producer levy (3% of
intervention price)
Automatic cut (max. 3%) in
support prices if production
exceeds specified amount
(plus additional CRL)

Budget

No

Budget

Yes

Budget

No

Budget

No

1982

1984

1986

Dairy
production
quotas
Cereals CRL

1988

Stabilizers

Notes:
1 A constraint is defined as binding if the need for reform is immediate.

Of the reforms shown in Table 7.1, dairy quotas stand out as the only one to
limit spending by building some kind of constraint (production) into CAP
policy instruments. In 1988 the original stabilizer proposal had been for a
spending-based trigger, but this direct control on spending proved
unacceptable politically. Thus, other than for the dairy regime, CAP support
has remained open-ended. As set out in Table 7.1, dairy quotas were the only

The Common Agricultural Policy 19772003

95

reform enacted under binding pressure, created by the BBR. Spending had
exceeded the revenue ceiling in 1983 but over ECU800 million of CAP
spending and the ensuing crisis were merely delayed by transferring them
to the 1984 budget.
Path dependency can be used to structure a narrative of these changes in
terms of the price support system limiting the options for reform. Moreover,
the changes that have been introduced have reinforced the development of the
CAP along its existing, particular path. Notably, the underlying operating
principle of quotas production in excess of a certain quantity triggering a
financial penalty was the same as Guarantee Thresholds, the difference being
the size of the penalty. This also implied continuity in administration, with
some tasks also devolved to the member states, or even to individual dairies.
Moreover, quotas were chosen over two alternatives, a rise in the CRL and a
cut in the intervention price, on the basis that these would have undermined
the principle of price support.
Support prices remained high with quotas. Indeed, domestic politics,
especially in Germany, meant that retaining high prices was a prerequisite for
securing the 1984 reform agreement. By containing spending quotas also
targeted the one binding constraint, thus winning support from those countries
concerned by rising CAP spending. Quotas thus reconciled member states
whose positions were otherwise mutually exclusive. Although spending
growth was contained by quotas, spending levels were only reduced
subsequently as quota levels were reduced. In the meantime the spending limit
was again breached in 1985, the deficit covered by additional payments from
the member states. From 1986, a previously agreed rise in the spending limit
helped restore budgetary balance.
A feature common to both the 1984 and 1988 reforms was the nonexclusivity of the policy network driving change. Until then, the perception
was that the Commission and CoAM worked in a closed system with farmers
and agri-business interests to shape policy. Further, as noted in Chapter 6, the
financial system of the EU worked in a manner that accommodated the
financial consequences of CoAM decisions. As the BBR was threatened then
breached, the influence of the Budget Commissioner and national Finance
Ministers became prominent in the CAP policy system. Moyer and Josling
(1990, p. 70) described an inner circle of Commissioners, appointed in 1985,
who drove the 1988 reforms Commission President Jacques Delors,
Agriculture Commissioner Frans Andriessen and Budget Commissioner
Henning Christophersen (and their cabinets).
Given this composition and the role the budget played in forcing CAP
reform, it was no coincidence that 1984 and 1988 also saw changes to the
budget and budget process. In 1984 it was agreed to limit annual CAP
spending growth to 2 per cent. This formal institution could have represented

96

Evidence

a significant change in the direction of the CAP, but without a formal

mechanism to enforce this, spending rose by an average of 18 per cent a year
between 1984 and 1987. In terms of the narrative, the position of the Budget
Commissioner and national finance ministers in the CAP policy process was
epiphenomenal. In 1988 a wide-ranging budget reform was agreed, which
included the introduction of multi-annual Financial Perspectives. This made
spending limits more transparent, but again there was no new institution to
help contain CAP spending. A more detailed narrative than the one provided
here would explore why the position of budget actors was not institutionalized,
but for our purposes the path dependency concept in which the
institutionalized price support system drives policy dynamics can make sense
of the evidence.

THE 1992 CAP REFORM

The main features of the 1992 reform were substantial cuts in the level of price
support, with the negative impact on farmers incomes offset by direct
payments that were partially de-coupled from production. Large arable
farmers were also asked to set aside part of their land: they were not obliged
to, but if they did not then they would not receive the direct payments.
This reform does question the idea of the CAP as path dependent; for the first
time there was a significant reduction in prices. However, the option of
totally replacing price support in 1992 was opposed by some member states on
the grounds that it would represent the ending of one of the three founding
pillars of the CAP Community preference. Although not a formal de jure
legal requirement for the CAP, it has de facto been treated as one by these
member states. The alternative of reducing price levels significantly, but
accompanied by compensation payments, was the politically feasible
option.
In the 1984 reforms, there was one binding pressure (budget costs rising
beyond the spending limit) and quotas addressed only that all other elements
of the CAP remained fundamentally unchanged. The 1992 reform can be
understood in a similar manner, this time in terms of two binding constraints,
the confluence of which created a policy window. Key actors in the policy
system agreed that change was required, but a combination of external and
internal forces directed that change in one of a limited range of directions.
Once again budget costs were threatening the spending limit. In addition, there
was pressure through the GATT to address trade-distorting agricultural
policies. In terms of the budget constraint, despite the 1988 reforms to budget
and CAP, there was nothing explicit in the latter reform to contain
CAP spending. Price support cuts of 3 per cent left the basic policy

The Common Agricultural Policy 19772003

97

fundamentally unaltered, maintaining upward pressure on surpluses and

expenditure.
However, the international trade constraint was new and focused on the
trade-distorting elements of the CAP. The main focus of international pressure
for CAP reform was the high level of support prices, and this acted as a
constraint that proscribed maintaining the status quo of EU prices significantly
above the world price. This set the basic environmental parameters within
which reform could be enacted. The response to the budget pressure was the
large (30 per cent) cut in support prices and the introduction of compensation
payments. Crucially, these payments could build-in both spending limits and a
degree of de-coupling of support from production. Indeed, it was the way in
which the spending limits were designed that contributed to the (partial) decoupling of the payments. The arable area payments used a fixed yield figure
based on past yields and imposed a limit on the area of land eligible for
support. The beef payments were limited in terms of the number of animals
eligible for support. In this way, the reform curtailed considerably the degree
to which CAP support remained open-ended.
Despite this, the new payments were not without their problems. Replacing
price support with direct payments inevitably increases the budget cost of the
policy, as the high prices paid by consumers are replaced by taxpayer (budget)
costs. However, the shift to direct payments led to a one-off rise in the budget
cost of the CAP, but subsequent spending levels have been much more stable
(due to the changed basis of support and built-in spending limits) and,
moreover, remained within the constraints set out in the Financial
Perspectives. As such, support for this reform could be gained from countries
concerned about the budget costs of the CAP.
The main trade concern was not the level of spending on the CAP per se,
but on the trade-distorting elements of the policy. Even though the new
payments broke some of the links with production, the separation was only
partial. As a result, it is interesting to note that the new payments did not
conform to the then GATT definition of (fully de-coupled) Green Box policies
that would automatically shield the payments from reduction. Instead, it
required bilateral talks between the EU and USA to develop a completely new
definition for partially de-coupled support (the so-called Blue Box) to
protect new payments.
A further benefit of direct payments was the previous experience of the EU
in managing this type of policy instrument. As the reform proposals were
being discussed, the EU was required to reform the oilseeds regime, as a result
of an adverse GATT ruling on the trade-distorting features of the previous
policy under extant GATT rules. The EU replaced this with area-based direct
payments to producers. The reformed policy was not entirely to the liking of
GATT but the revised version, tweaked slightly and incorporated into the

98

Evidence

wider arable regime in 1992, was accepted. The EU was, in effect, able to
use the oilseeds reform as a test bed and precedent for the arable sector
reform.
The 1992 set-aside provision also developed from pre-existing CAP
instruments. A voluntary set-aside scheme was first introduced in 1988. It was
then extended by the 199192 price package, when (large) cereals farmers
could gain exemption from a 5 per cent CRL if they set aside 15 per cent of
their land. Both the 1991 and 1992 changes kept set-aside participation
voluntary but they progressively raised the cost of non-participation.

REFORMS SINCE 1992

The income support element of the 1999 reform built upon the 1992 reform
to reduce further the level of support prices, and raise higher the level of direct
payments. The latter change, however, covered only 50 per cent of the price
cut, mitigating some of the budgetary effects of the change. A second element
of the 1999 reform drew together the disparate components of agricultural
Guidance policy into a new Rural Development Pillar II of the CAP. This
was given 10 per cent of the CAP budget and, for the first time, Guarantee
money could be used for Guidance policies. For example, member states
could opt to top-slice some of the direct income payments and recycle the
money into Rural Development (modulation). Since Pillar II requires
national co-financing, modulation is a way of raising CAP spending without
threatening the EU spending limit. The combination of modulation and the
partial compensation of price cuts also broke the link between specific price
cuts and specific Compensation. As such, the support system introduced in
1992, based on direct payments, was confirmed as a general shift in the way
the CAP supported farm incomes.
The main element of the 2003 reform was the introduction of the decoupled Single Farm Payment (SFP). Agreement was not forthcoming to
convert all the pre-existing direct payments into SFPs, but initial estimates
suggest about three-quarters will be converted. The timing of these reforms is
important, occurring during the Doha Round of WTO talks. Whilst other
countries in those talks would prefer all payments to be de-coupled, this
reform, adapting the majority of CAP Blue Box payments, offers the basis
for a compromise agreement. On the other hand, one proposal from 2003 that
was not accepted was the final 5 per cent cut in cereals support prices, being
a step towards the final removal of Community Preference. Given the
strength of other countries opposition to the continued use of price support, it
is this element of the CAP that continues to raise serious concerns for
agreement in the Doha Round.

The Common Agricultural Policy 19772003

99

MICROFOUNDATIONS OF CAP DEVELOPMENT:

THE DISTRIBUTION OF CAP SPENDING
Commitment to the CAP and its principles has remained unaltered since the
1950s. Article 39 is unchanged and two of the three original pillars of the CAP
have remained unchallenged. Community Preference has been eroded but it
still survives and, for some commodities, at very high levels. The commodity
coverage of Guarantee support has also been maintained, with support for
sheep production introduced in 1980. Moreover, until 2004 each new country
joining the EU had the CAP extended to it in full, in terms of both the support
instruments and level of support. With the latest enlargement, ten countries
joined but the total EU spending limit has remained unchanged. The new
member states have received all CAP instruments but not the full level of
support. Price support was extended in full, given its relevance to the Single
European Market, but direct payments have been extended initially at reduced
levels (to be phased in over ten years) given EU15 budget concerns, whilst
Rural Development support is effectively treated as NCE, extended to the new
member states at a level that ensures the overall EU25 budget spending limit
is respected. As a result, whilst the CAP remains Common in terms of
support instruments offered, it is not so regarding levels.
Earlier it was noted that the 1992 reform is viewed as radical by some
observers but not by others. Those in the former group who have so judged this
reform have based this opinion largely on two aspects: the huge reduction in
the level of the preceding support instrument (price support) and the closing of
open-ended support. Although the 1992 reform applied principally to just two
sectors (arable and beef), their combined cost represented over 60 per cent of
total CAP costs in 2001, with over 90 per cent of those costs being direct
payments. The narratives that do not judge the 1992 reform as radical have
focused more on the continued level of support under the CAP. Daugbjerg
(1999, pp. 41516) notes, for example, a large share of farming incomes still
derived from government policies. Other empirical works show not only how
little the level of support changed with the 1992 reform but also the marginal
shift in the distribution of support by member state, the extent of the change to
support instruments notwithstanding (Ackrill et al. 1997).
The importance of the budget in the narrative encompassing both the
occurrence of reforms and the timing of reforms has been highlighted in this
chapter; it provides the situational context in which to interpret the actions by
member states in the CAP policy process. This prominence is a result of the
combination of the BBR, the definition of CAP spending as Compulsory and
the budgetary impact of certain income support policies. This offers an insight
into path dependency as an organizing concept for structuring the narrative of
CAP development.

100

Evidence

Most of the empirical applications of path dependency to policy or

institutional development view changes in kind, where a path breaks or varies
significantly in direction, as the result of external shocks where environmental
parameters change dramatically. With the CAP, however, change is forced by
endogenous factors the coupling of support and production under price
support increasing EU expenditure, that at certain temporal points has
threatened the viability of the EU budget process and brought Budget
Commissioners and national finance ministers into the CAP policy system.
The original CAP thus created periodic bursts of crisis almost by design.
Even so, whilst the only notable reforms to the CAP have been implemented
in times of crisis, defined as times when pressures for reform have been
binding, removing the status quo as an option, the policy is sufficiently multidimensional that much of the CAP has remained unaltered. For example, the
1984 and 1992 reforms both addressed specific pressures yet left support
levels generally unchanged.
Once the CAP was established with price support as the primary tool, a
certain pattern of budgetary transfers emerged. Ackrill and Kay (2005) present
a simple analysis of the data on the distribution of support across member
states. Since then, a working hypothesis is that member states that defend the
CAP typically do well financially from the policy. For example, France has
consistently received in excess of 20 per cent of total CAP Guarantee
expenditures. Per capita transfers are substantial for both France and Ireland.
For Spain there are less data to analyse, but by 2001 only France was receiving
more money from the CAP, and their per capita transfers were similar.
Among the pro-reform countries both the UK and the Netherlands receive
modest sums per capita from the CAP and the EU budget as a whole.
Germany follows the same pattern, although the German pursuit of CAP
reform has been less consistent over time, balanced as it sometimes has been
against preservation of the Franco-German axis. On the basis of CAP
transfers, one outlier amongst pro-reform countries is Denmark, a rich country
that does extremely well out of the CAP. If, however, net total budgetary
transfers are considered, the set of pro-reform countries (now including
Sweden) is somewhat clearer.
Ackrill and Kay (2005) examine in detail the stability of countries spending
shares. This is a proxy for the dynamics of the CAP policy system. Briefly
summarized, countries opposing CAP reform typically receive substantial
absolute and/or per capita budgetary transfers. These countries, but especially
France and Ireland, have not only defended their shares of CAP spending, but
also reduced the variability of transfers over time. Those seeking CAP reform
have, however, generally either faced declining shares of CAP spending and/or
are significant net contributors to the EU budget and the CAP. In most cases
they have also faced rising or unstable variability of transfers. The domestic

The Common Agricultural Policy 19772003

101

politics and history of Germany make for a complex attitude towards CAP
reform (Wilson and Wilson 2001).
During the negotiations of the 2003 reform, countries opposing or
supporting reform generally matched the foregoing classification closely.
Moreover, when presenting the proposals to the European Parliament in 2002,
Commissioner Fischler stated that by these measures we intend to give
consumers and taxpayers a recognisable service in return for the tax they pay
(speech 02/330) thereby seeking to justify the current level of transfers rather
than seeking lower spending. Indeed the proposals were drawn up with a view
to achieving the objective of stabilising agricultural expenditure in real terms
(European Commission 2002, p. 29, emphasis added).
As regards sums received by member states, the SFP is based on actual
direct payment receipts over the period 200002, a move that will sustain the
existing inter-country distribution of expenditures. Moreover one of the
concerns expressed by member states over modulation, including pro-reform
countries such as the UK, was that the recycling of money could also result
in a redistribution between member states. Whilst the final agreement did not
rule out redistribution, it did limit it. The first percentage point of the 5 per cent
of direct payments to be modulated must remain in the source country and
overall a minimum of 80 per cent of total recycled funds must remain at
home.

A STRUCTURED NARRATIVE OF THE CAP

This chapter has presented a structured narrative of CAP dynamics in terms of
how its origins have acted to constrain its subsequent development. An
examination of several reform events from 1977 to 2003 in terms of the CAP
as multi-component policy demonstrates that the fundamental principles of the
CAP have remained unchanged, and that many of the reforms amounted to
marginal changes in the value of the main policy instruments. In these terms,
it is accurate to apply the path dependency label to the CAP as a single,
composite variable with a particular direction of development over time.
On the other hand, it has also been noted that the combination of an
accumulation of modest changes and the addition of extra elements to the set
of policy instruments can create the conditions for a critical juncture, and
possibly path-changing policy change. Examples of CAP changes in this
regard highlighted in the chapter include: the progression from elements of the
CRL and Guarantee Thresholds to the agreement on dairy quotas; the shift
from negotiated price cuts under Guarantee Thresholds to automatic price
cuts under Stabilizers to the pressure for reform in 1992; the move from setaside exempting farmers from the cereals CRL in 1991 to direct payment

102

Evidence

cross-compliance in 1992; and from the reform of oilseeds direct payments to

agreement on the new arable regime.
Typically, the literature on path dependence sees changes of kind (those that
mark the end of or shift in path-dependent processes) as being driven by
external shocks. The notable feature of the CAP policy system is that in the
initial stages the EU budget constraint was exogenous to the CAP and the
influence of budget actors in the CAP system was epiphenomenal, periodically
triggered by budgetary crises caused by the price support, harvest yields and
world markets bumping up against the fiscal constitution of the EU. Even after
the involvement of the European Council in the reforms of 1984, 1988 and
1992 and the agreement of medium-term financial frameworks, there was no
institutionalization of the budget interest in the CAP policy system. Only
recently has this design fault been corrected with recent CAP reforms that
have built in spending limits.
But still a simple analysis of member states shares of CAP spending as a
measure of path dependency shows they have been very stable over time,
despite changes to the CAP. Two countries most closely associated with
opposition to CAP reform, France and Ireland, stand out not only in terms of
their gains from the CAP (in total and/or in per capita terms) but also in their
ability to defend their spending shares even as the EU has enlarged, whilst
managing to increase the stability of those shares over time (for example, the
year-on-year data on relative shares have been much less volatile). Countries
who support CAP reform tend to have more modest spending shares (in total
or per capita terms) and also are characterized by shares that are falling or
unstable over time. This has by no means precluded reform as the various
reform events discussed in this chapter, especially that of 1992, have shown.
What it has done, however, has imposed a quite specific and durable constraint
upon the parameters within which reform can take place. This is the key
situational logic through which member states intrepret their roles as agents in
the CAP policy process. Moreover, this is consistent with the observation of
the breakdown of exclusive agricultural policy networks in several member
states. Whilst the newcomers to the reform negotiations are non-agricultural,
their interests are often financial and they too have an interest in maintaining
CAP receipts.
The greatest threat to the spending shares of the EU15 came with
negotiations for the 2004 EU enlargement. The outcome of the negotiations
was inconsistent with many previous predictions but entirely consistent with
the analysis presented in this chapter. The overall EU budget spending limit
was left unaltered, despite ten new entrants. CAP spending in the new member
states had, in effect, to be adjusted to fit into the residual left once EU15
transfers had been accounted for and effectively ring-fenced. CAP transfers
in the new member states are, de facto, being treated as non-compulsory

The Common Agricultural Policy 19772003

103

spending, a move made easier by the switch from price support to direct
payments. Some observers viewed EU enlargement as a way of forcing radical
CAP reform because of the budgetary pressure that would result. The latest
negotiations showed how wrong that view was. Moreover, the analysis here
helps explain why the existing member states presented the accession package
they did. Maintaining the existing distribution of CAP transfers for the EU15
was more important than maintaining the commonality of the CAP across all
member states.

SUMMARY
I have presented a narrative in terms of the price support mechanisms of the
CAP, and the financial and international trade consequences structured by the
concept of path dependency. This is a single, coherent story that makes sense
of the evidence of the budgetary consequences of the CAP and periods of CAP
reform. However, the narrative misses other aspects of the development of the
CAP, for example the increasing pressure at national and EU levels to take
account of the deleterious environmental consequences of the CAP in policy
considerations; or NGO campaigns around the disastrous consequences of the
CAP for many developing countries reliant on primary commodity exports.
The question not addressed is whether these political discourses around the
policy system are having the effect of gradually eroding the CAP policy
paradigm. These gaps are in the nature of any structured narrative of the
dynamics of a complex, multi-dimensional and multi-national policy such as
the CAP; the constituents in the CAP system are in constant flux, its
consequences are multi-faceted and spread out over a significant temporal
scale. As established in preceding chapters, it is the nature of narratives of
dynamics to be backward looking: changes in the CAP policy system are best
judged retrospectively; and obviously the emergence of confirming evidence
of the effect of environmental or development actors on the CAP policy
system will affect the extent to which the structured narrative I have presented
here is held as a valid or true explanation of the contemporary policy dynamics
of the CAP.

8. The GP fundholding scheme

The Conservative government in the UK introduced a range of reforms in April
1991 to address a number of perceived weaknesses in the National Health
Service (NHS) management structure; in particular, an allocation of resources
that had been determined largely by the sum of the individualistic
behaviour of individual doctors rather than through a hierarchical process of
resource management (Wistow 1992, p. 59). The government attempted to
compensate for the lack of correct economic signals by introducing new
incentive structures designed to encourage provider units and NHS staff to
meet the limitless demand for state-funded health care within a cash-limited,
public budget (Thatcher 1993, p. 606). To this end the government introduced
an internal market for NHS services, which separated the purchasing and
provision of health care interventions and allowed general practitioners (GPs)
to elect to hold a cash-limited budget for the purchase of a limited range of
secondary care, staffing and pharmaceutical services.
This chapter examines the conception, implementation and abolition of the
GP fundholding scheme element of these reforms as an example of policy
dynamics after the breakdown of a long-established policy community. This
exemplifies a more general trend in the 1990s across several OECD countries
of the break up of well-developed policy networks dominated by industrial
actors, professional elites and bureaucrats, ushering in periods of instability in
policy areas such as transport, agriculture, food safety and nuclear power.
Health care policymaking between 1948 and the mid-1980s was a worked
example of a policy community according to the widely-used typology
provided in Marsh and Rhodes (1992, p. 25): there were two policy actors;
membership was stable across time; there was a shared ideology expressed in
an established policy paradigm around the values of the NHS; and other
potential policy actors were excluded. This exclusion occurred institutionally
through the Department of Health (DoH) giving formal recognition to the
interests of doctors, and through a policy paradigm that stressed clinical
autonomy in the implementation of health care policy and the delivery of
health care services. Within the health care policy community,
conventional wisdom has emphasized the dominant nature of professional
rather than political or managerial influences (Wistow 1992, p. 51).
A defining characteristic of policy communities is stability over time. They
are strongly institutionalized networks that structure the policy process in
104

The GP fundholding scheme

105

terms of the rules of the game, privileging certain interests and supporting a
dominant view of the world. Although the enactment of the 1991 NHS
reforms, and GP fundholding in particular, is a useful temporal identifier for
the end of the policy community in health care, the breakdown of a policy
community was not a single event but rather a process of transition from a
strongly institutionalized policy system to one that is less structured, more
contested and prone to instability. Importantly, after the breakdown of a policy
community there is no automatic institutionalization of a new policy
community or the establishment of a fresh policy paradigm.
In the structured narrative presented in the chapter, the policy community
and its ideational base disintegrated in a conflux of internal and external
processes: budgetary pressures in the NHS; a perceived inability to satisfy
citizens rising demands for health care; an intense politicization of health in
electoral terms in 1987; waning of trust among members; along with reduced
expectations that the institutions of the policy community would be respected.
The dynamics of the breakdown of the health policy paradigm in the UK was
not a Kuhnian scientific revolution in the sense of an existing paradigm no
longer explaining the facts, and where the weight of disconfirming evidence
for a particular world view reaches a critical level where the paradigm
collapses. Indeed Hall (1993, p. 291), who introduced the term policy
paradigm, admitted that the notion of a narrow, cognitive frame that sets strict
constraints on policy thinking is not universal: only in some cases is it
appropriate to speak of a fully elaborated policy paradigm In others, the
web of ideas will be looser and subject to more frequent variation.
Nevertheless, even in this latter sense, the concept of a policy paradigm is
useful in structuring the narrative; it is the fragmentation of the policy
paradigm after 1987 that triggered some of the key dynamics of policy
development in the 1990s; a process of competing problems, ideas and
policies.
In Chapter 4, it was suggested that the evolutionary metaphor is a useful
way of organizing thinking about policy learning by distinguishing adaptation
(learning that affects calculations about how to realize interests within the
institutional structure of the policy community) from more complex learning
where interests, identities or institutions are learned or constructed in the
interaction of agents in the policy system. In the second sense of learning the
problem situation is constructed in the interactions between agents, the
negotiations of policy problems, policy solutions and criteria of success. In
alternative terms, learning is the process of agreeing standards for the
intentional selection mechanisms that operate to weed out policy failure; it is
through this construction that policy paradigms, however narrow or loose,
emerge and are institutionalized. This concept of policy learning is used to
structure the description of the process of health care policymaking after 1989

106

Evidence

and the introduction, implementation and eventual abolition of the GP

fundholding scheme. There was no ready-made replacement for the health
care policy paradigm that had underpinned the health policy community from
the inception of the NHS in 1948 until the late 1980s. Without the policy
community institutions, including the policy paradigm, the policy process was
more fluid, more contested politically and ideationally; and the GP
fundholding scheme episode resembles a mode of evolutionary learning
process in this latter sense, rather than a technical, deliberative and evidencebased process based on shared values of the purpose of health care reform.
The chapter is in three parts. The first charts the fracturing of the health care
policy community from the mid-1980s up to the introduction of GP
fundholding; in particular how the institutions dissolved along with the trust
between the government and the British Medical Association (BMA). The
second section highlights the policy process around the GP fundholding
scheme post-policy paradigm, emphasizing the ideological and political nature
of policy contestation; there was no shift to evidence-based policymaking that
might have revealed a coalescing of policy actors around a public management
policy paradigm to replace the previous medicalprofessional one that had
fractured in the late 1980s. As Hunter (1998, p. 133) puts it, Health policy has
been driven by a mix of ideology, fashion and pragmatism but never by
evidence. This continued ideological contest was an important factor in why
the GP fundholding policy was never institutionalized in its early period and
subsequently proved vulnerable to reform and abolition. The third part charts
the abolition of the scheme by the Labour government after 1997 and the
transition to health care policy process under New Labour.

THE INTRODUCTION OF GP FUNDHOLDING

The ideas of Professor Alain Enthoven from Stanford University Business
School and Alan Maynard, Professor of Health Economics at the University of
York, were influential in the 1991 reforms. Neither had been part of the
previous health care policy community and they viewed the NHS from an
economics perspective. As a result of a visit to the UK during the mid-1980s,
Enthoven argued that the funding arrangements for NHS hospitals lacked any
incentives for them to improve the quality or the efficiency of the services that
they provided (Enthoven 1985). Based upon this finding, the Conservative
government attempted to improve the incentives for NHS hospitals by
introducing an internal market into the NHS, as part of the 1991 reforms. At
a meeting arranged by the Office of Health Economics in 1985, Professor
Maynard argued that NHS general practice also lacked the incentives to
encourage GPs to control NHS costs and to make efficient use of public

The GP fundholding scheme

107

resources. In response, he suggested that a system of budgets and associated

incentives should be introduced into NHS general practice for a range of
secondary care and pharmaceutical services. Although Enthoven did not
envisage GP budget-holding as part of his internal market plans, Maynards
ideas were championed by Kenneth Clarke, the Secretary of State for Health,
and consequently became a late addition to a package of NHS reforms
(Lawson 1992; Klein 1995; Webster 1998a).
Details of the fundholding scheme were first published in the NHS White
Paper Working for Patients, and the accompanying working paper, Practice
Budgets for General Medical Practitioners (DoH, 1989a; DoH, 1989b). These
documents outlined the basic regulations of the scheme and the entry criteria
that practices wishing to attain fundholding status had to satisfy. Under the
initial regulations, practices that attained fundholding status were allocated a
cash-limited budget, which contained an element for: (i) the purchase of a
defined group of elective hospital services (including diagnostic tests); (ii) the
salaries of non-medical staff; and (iii) prescribing. The scheme allowed
participating practices to vire funds between the various elements of the
budget and to generate a surplus, which could be spent on improvements to
practice services, staffing or premises. However, fundholders that failed to
manage their budgets appropriately could be removed from the scheme and
become subject to the same regulations as other non-fundholding practices.
Mrs Thatcher led the NHS Review that worked during 1988 to produce
Working for Patients in January 1989. The Cabinet Committee that steered the
Review was small and consisted of a few trusted ministers, advisers and civil
servants. There was no representation of either the BMA or Royal Colleges
(Klein and Day 1992). Further, there was no Royal Commission, which had
been the traditional instrument of ensuring consensus between the government
and medical profession. Klein (1995, p. 184) argues that in its style,
though not in its outcome, the Review marked a brutal break with the past
and it was the deliberate exclusion of the medical profession in the policy
process that was the brutal break. As Smith (1993, p. 179) puts it, Doctors
were no longer an essential partner in the development of health policy but a
vested interest that had to be challenged.
The immediate consequence was that the BMA launched a major publicity
campaign during the summer of 1989 to discredit general practice fundholding
and the opposition (the Labour Party) vowed to dismantle the scheme if
elected into power. During the period 198991, the Labour Party and BMA
worked in a curious and unspoken alliance (Klein 1995, p. 192) in opposition
to the proposals. The intensity of this political conflict between the BMA and
the government has been compared with that involved in the setting up of the
NHS between 1946 and 1948 (Wistow 1992; Klein and Day 1992). Despite the
BMAs campaign, the GP fundholding scheme came into being in April 1991.

108

Evidence

There was no compromise or negotiation between the government and the

BMA over the operational details or implementation of the scheme (in contrast
to the battles of 194648). The outcome of the political conflict is evidence
that the medical profession had lost their ability to veto change in the NHS
(Klein 1995, p. 199).
The first part of this narrative of the GP fundholding scheme concerns how
the policy community of the preceding 40 years had unravelled to the point
where one of its foundational institutions the privileged position of the
medical profession collapsed. The corollary of this was the unvarying
political conflict throughout the history of the GP fundholding scheme,
explored later in the chapter: its enactment, implementation, evaluation and
eventual abolition.

THE POLICY COMMUNITY IN HEALTH CARE UP TO

THE MID-1980S
One of the aspects that distinguishes a policy community as a network form is
the existence of trust between its members; in particular, each actor trusts all
other actors to observe its institutions or the rules of the game. This is a key
theme in the policy community literature generally and analyses of the UK
health care policy community in particular (Smith 1993; Wistow 1992;
Hindmoor 1998). Hindmoor (1998) describes the development of trust
between the BMA and the post-war Labour Government. He uses a simple
Prisoners Dilemma (PD) game to illustrate how this trust supported a bargain
between the two actors on the rules of the game. The 1946 National Health
Service Act that set up the up the NHS provided an enabling framework for the
establishment of a publicly owned, publicly funded health service. The Labour
Government of 194550 relied on the medical profession for the
implementation of the new NHS, and was required to concede to certain BMA
demands because of the political and policy resources they enjoyed. The
medical profession was regarded as the custodian of technical and specialist
knowledge that was essential for the well-being of society, and the BMA was
seen as apolitical (in party-political terms at least), helping to give the views
of the BMA a moral authority and make them, potentially, highly influential
in public opinion (Klein 1995). There was a clear interdependence between the
Labour Government and the BMA in the NHS and the possibility for mutually
beneficial political exchange.
In the familiar one-shot PD game, the Nash equilibrium point is (Defect,
Defect), which means that neither the government nor the BMA cooperate
with one another in terms of establishing the NHS. The problem is that there
exists the outcome (Cooperate, Cooperate), which is better for both players

The GP fundholding scheme

109

than (D, D). However, there are large transaction costs associated with
negotiating, implementing and monitoring the agreement for both actors to
cooperate. The emergence of trust between actors helps ameliorate these
transaction costs as informal rules and tacit understanding emerge to support
a (C, C) equilibrium. Hindmoor (1998) argues that the negotiations between
the BMA and the government between 1946 and 1948 were a repeated PD
game, in which trust emerged through the development of informal rules and
tacit understandings to support a stable (C, C) equilibrium. A stable (C, C)
equilibrium can be interpreted, in terms of the health care policy community,
as each actor trusting the other actor to observe its emergent institutions over
time.
There were two institutions that underpinned the operation of the health
care policy community between 1948 and the mid-1980s. The first was that
each actor should trust the other on both the process and substance of
policymaking. Hindmoor (1998) and Wistow (1992) talk of trust becoming
embedded in the health care policy community. This trust supported a
consensus on how business was to be conducted, as Jordan and Richardson
(1987, p. 101) describe it the process by which and the atmosphere within
which policymaking is decided. In particular, this meant the exclusion of
other potential interests in health care policy, for example public opinion,
Parliament, the rest of Whitehall and hospital managers. Ham (1992), Webster
(1988), and Klein (1995) describe how this exclusion became heavily
institutionalized after 1948. The BMA had privileged access and a central role
at every stage of the policymaking process.
The second rule of the game was an implicit contract between the
government and the medical profession. The contract was that the former
respected clinical autonomy in how to use resources but the latter accepted that
the decision on the overall level of resources, the budgetary constraint, was a
matter for the government. The observance of these two rules marked the
policy community period in UK health care policy.
Within this well-established policy community, there was policy learning
and limited policy change. For example, Royal Commissions were the main
conduits for the introduction of ideas and evidence into health care
policymaking during the policy community period. They received
submissions, interviewed experts and interrogated advice. Their reports, along
with the input of the comprehensive central expert advisory machinery which
was built up between 1948 and 1979 (Webster 1998a, p. 29) constituted the
evidence that informed most of the structural changes to the NHS prior to 1991
(Webster 1988, 1996, 1998a).
Policy changes before the late 1980s reflected the policy community that
introduced them. They were never radical, they concentrated on the structure
of the NHS as a way of influencing the behaviour of staff and they were

110

Evidence

introduced incrementally with long lead times after exhaustive consultations

within the policy community. The early period of the Thatcher administration
was characterized by this type of policy learning in the health care policy
community. For example, in July 1979, the Royal Commission on the NHS
established by the previous Labour administration and chaired by
A.W. Merrison produced its final report (Royal Commission, 1979). Based on
the reports recommendations, in 1982 the Conservative government
reorganized the administrative structure of the NHS, replacing Area Health
Authorities and districts with 192 District Health Authorities (DHAs) and
placing Family Practitioner Committees (FPCs) under the direct authority of
the Department of Health and Social Security (DHSS).

THE BREAKDOWN OF THE HEALTH CARE POLICY

COMMUNITY
A conflux of contingent factors caused the emergence of distrust in relations
between the government and the BMA. This emergence was temporally
intertwined with the fading of the institutions of the policy community. Two
events are salient: the introduction of a limit on the range of drugs that GPs
could prescribe on the NHS, and a ferocious and public conflict in 1987 about
the level of NHS funding. The plans for the first event were announced in
November 1984 and the General Medical Services Committee of the BMA
came out strongly against the governments intention, with several of its
members predicting that a two-tier NHS would be created (Anon. 1984). In
response to the BMAs opposition and campaigning by the pharmaceutical
industry, the government decided to reduce its plans. In consequence, the
limited list that was introduced in April 1985 only restricted the medicines
that could be prescribed in seven major therapeutic categories, producing
estimated savings of 75 million (rather than the originally estimated 100
million) in the first year (Harris 1996).
Webster (1998, p. 40) argued that the limited list episode was an apparent
humiliation for a government increasingly renowned for its success in
confrontations with corporate interests. However, it seems to have furnished
the Thatcher administration with the conviction that radical policy initiatives
could be introduced without prior consultation with the medical profession, or
the establishment of a formal investigation or Royal Commission. As Webster
(1998b, p. 40) states, the episode constituted a firm platform for further and
more audacious forays into the primary-care policy field.
Financial pressure has been an ever-present feature of the history of the
NHS, and pre-dated the Thatcher administrations. The desire for macro-level
cost containment and fiscal prudence was prominent in the Labour

The GP fundholding scheme

111

Government of 197479. This had had the corollary of raising policymakers

interest in the efficiency and effectiveness of the use of existing resources.
However by 1987, it was the willingness of the medical and nursing
profession to accept those pressures which appears to have diminished
dramatically (Klein 1995, p. 183). This conflict took both a public and partypolitical form. Wistow (1992) and Klein (1995) observed a widespread
perception, presented by medical professions through the media, that the NHS
was on the point of collapse. Health was a prominent and strongly contested
issue in the 1987 general election, and despite the Conservatives comfortable
victory had been an issue on which Labour held a substantial lead in the polls.
In the six months after the election, Webster (1998a, p. 22) states that, It is
well known that the government was blown off-course by the financial crisis
affecting the health service in the autumn of 1987, with the result that it was
forced to address as a matter of urgency the problems of additional resources
and cost containment.
Importantly for the GP fundholding policy process, Mrs Thatcher believed
that before and after the 1987 election the BMA had entered the party-political
debate over the level of NHS funding (Thatcher 1993). The BMA had abused
the medical professions authority with the public and broken one of the rules
of the policy community game, contributing to the fact that the public
obstinately continued to see the health service as a casualty of the Thatcher
administrations parsimony (Klein 1995, p. 178).
In the repeated PD game, once cooperation breaks down the dominant
strategy for each player is to continue to withdraw cooperation. The PD game
then reverts to the Nash equilibrium of the one-shot game, (D, D), as Defect
becomes the dominant strategy for both parties. In policy community terms
this is the non-observance of the rules of the game described above. Both the
government and the BMA infringed the rules that had supported a stable (C,
C) equilibrium since 1948. In the context of the introduction of GP
fundholding, the lack of a policy community meant that the transaction costs
of political exchange or negotiation between government and the BMA
reached a critical level where the former simply imposed a policy and one
of the casualties was precisely any willingness to discuss the nuts and bolts
of implementation (Klein 1995, p. 195). The imposition of such significant
change, as with the GP fundholding scheme, so quickly and without any pilot
scheme, appraisal or Royal Commission represented a new period in the
development of health care policy in the UK.

POLICY LEARNING WITHOUT A PARADIGM

A defining characteristic of the development of the GP fundholding scheme

112

Evidence

between 1989 and 1997, the immediate post-policy community period, was
the absence of a shared policy paradigm through which incremental, evidencebased policy learning could take place. The two established mechanisms for
introducing evidence into policymaking did not operate. There was no Royal
Commission set up, and the central advisory machinery had largely been
abolished by the Thatcher administration. Importantly, there was no formal
appraisal or evaluation of the effect of GP fundholding on health care or
health. This contradicts the view that the reforms of 1991 represented a victory
for the new public sector management in the NHS. Instead, the reforms
marked the end of the health care policy community and the start of health care
policy being driven by political conflict and the struggle to institutionalize
guiding assumptions and values about health care policy.
The evidence on which policy development was based under the policy
community was undoubtedly biased towards the medical professions view of
the health care system, as this was the policy paradigm. As Webster (1988,
1996 and 1998a) discusses, Royal Commissions collected evidence from the
medical profession about how the NHS was working and how any reform
proposals might work. The nature of that evidence reflected the
professionalized network that controlled its collation and reporting.
Without a policy community, there was no consensus on the introduction of
GP fundholding, and therefore there were no common values against which to
evaluate the scheme. Both Nigel Lawson (1992) and Mrs Thatcher (1993)
expressed the considerable worries that the Conservative government had
about the BMA exploiting any policy appraisal or evaluation of the GP
fundholding scheme: any formal policy evaluation would not have been
evidence to inform rational policy development but instead would have been
appropriated for political advantage. As public opinion had become a factor in
health care policy there was no time for the government or BMA to wait and
see the effects of any pilot scheme or policy assessment. Instead, each interest
articulated a public position immediately; policy analysis and public positions
were based on folk theorems or common beliefs about the traits,
characteristics or dispositions of a policy without any reliable and verified
source of evidence.
In the absence of trust, debates about the consequences of the GP
fundholding scheme were highly political. For example, in July 1995 the
Labour Party restated its criticisms of the scheme (Labour Party 1995): their
position was that fundholding had increased NHS management costs,
introduced financial pressures into the doctorpatient relationship and had
created a two-tier health service. These were claims based on folk theorems,
as there was no obvious evidence for any of them. Some politicians may have
calculated that vilifying the scheme was the best political response, as any
initiative that attempted to introduce financial management into the NHS was

The GP fundholding scheme

113

likely to be unpopular with voters. The absence of a formal evaluation of the

scheme allowed myths about fundholding and fundholders to remain in the
general political discourse, without any independent, formal mechanism for
testing their claims.

OFFICIAL AND EMPIRICAL STUDIES OF GP

FUNDHOLDING
In June 1995, the House of Commons Committee of Public Accounts
published a report on the first three years of the operation of the fundholding
scheme in England (Committee of Public Accounts, 1995). Based on evidence
provided by the NHS Executive (NHSE), the report detailed the effects of the
scheme on the provision of patient services, practice-level prescribing patterns
and on the generation and use of budget surpluses. Under the first heading, the
NHSEs evidence suggested that fundholders had achieved a faster rate of
delivery of secondary care services, whilst securing reductions in waiting
times, improvements in access and a widening of the range of services
available to their patients. However, the NHSE argued that the scheme had not
affected the care given to patients of non-fundholding practices, as
fundholders had secured many of these improvements by purchasing
previously unused hospital capacity.
Although evidence suggested that fundholding had helped to improve the
services that participating practices offered to their patients, the Public
Accounts Committee had two main concerns about the operation of the
scheme. First, the committee suggested that the scheme had relatively high
running costs, with cumulative expenditures on management allowances and
computer purchases during the first three years of the scheme being 147
million. Second, the committee noted that between fiscal 1991 and 1993 a
significant number of patients had been removed from their practices at the
request of their GPs. For example, during fiscal 1992, 78 000 patients had been
asked to leave their practices at the request of their doctors. However,
according to the Public Accounts Committee there was no clear evidence that
this occurred on cost grounds, or that patients were more likely to be expelled
from fundholding practices.
Although the government did not commission an independent evaluation of
the effects of fundholding, the external auditor for the NHS, the Audit
Commission, produced some reports on the initiative. For example, in 1995
the commission published a briefing that described the scope of the
fundholding scheme, charted its growth and analysed the budgets that
fundholders managed (Audit Commission, 1995). In 1996 the commission
published a report on the first five years of the operation of the scheme,

114

Evidence

entitled What the Doctor Ordered: A Study of GP Fundholding in England and

Wales (Audit Commission, 1996a). To accompany its report, the commission
published a digest of information from the survey of fundholding practices,
which had been undertaken for the main study (Audit Commission, 1996b).
What the Doctor Ordered was one of the most comprehensive reports
published on the management, operation and effects of the fundholding
scheme. As part of the study, the Audit Commission examined the
demographic and organizational characteristics of participating practices. In
relation to the former the commission found that, initially, fewer practices in
inner-city areas had become fundholders. As a result, participating practices
tended to have more affluent and less socially deprived patients. Significant
regional variations in fundholding coverage were evident, with the scheme
achieving proportionately lower coverage in some inner-city and/or deprived
areas. In relation to their organizational characteristics, the Audit Commission
reported that fundholding practices tended to be relatively large, often housed
in purpose-built premises, with more support staff and equipment, including
computers. However, they differed from equally large non-fundholding
practices, as they had more of the features normally associated with high
standards and better quality in general practice (for example, higher rates of
childhood immunization and accreditation to provide minor surgery and/or
postgraduate training). Therefore, the Audit Commissions work suggests that
the fundholding screening process ensured that only larger, well-organized
practices were initially allowed into the scheme.
Following the introduction of the 1991 reforms, a number of independent
academics also examined the effects of the fundholding scheme on GP
referrals, prescribing and other activities in NHS general practice. Although
they varied in terms of years studied, the number of practices examined and
the areas analysed, these studies helped compensate for the absence of a
government-sponsored evaluation (or piloting) of the schemes. These suggest
that evidence on the desirability and effectiveness of the scheme was both
limited and equivocal when fundholding was abolished by the new Labour
government in March 1999. For example, Coulter (1995, p. 233) concluded
that there were extensive gaps in current knowledge about the impact of the
scheme and that the claims that GP fundholding has resulted in
improvements in efficiency, responsiveness and quality of care are in general
not supported by the evidence. In a similar vein, Petchey (1995, p. 1139)
argued that few reliable conclusions about fundholding, either positive or
negative, can be drawn from existing research. Gosden and Torgenson (1997,
p. 103) also concluded that there was a dearth of high quality evidence on
many aspects of the fundholding scheme, particularly in relation to referral
rates, patient outcomes and service quality.
In a review of the schemes effects on prescribing, Baines et al. (1997)

The GP fundholding scheme

115

found that, in the short term, many early-wave fundholders had managed to
secure economies in their prescribing by switching to cheaper, generic drugs.
However, in the longer term, such savings may not have been sustainable. In
one of the last reviews of the literature before abolition, Smith and Wilton
(1998, p. 1253) concluded that evidence concerning the success or otherwise
of general practice fundholding over the last six years is incomplete and
mixed and, unless further research was undertaken, the jury will have to
remain out on whether fundholding has secured improved efficiency in the
delivery of health care.
Both the Conservative government and the Labour Party (and subsequently
Labour government) had clear public positions on the effect of GP
fundholding. These positions did not have a basis in evidence; one of the
dynamics set off by the collapse of a policy paradigm is a contestation of the
basic values and assumptions that will be used to consider policy development.
Without this information, there is no basis on which evidence-based
policymaking can take place; instead the benefits and costs of a particular
policy is the subject of strong political contestation in which policy analysis is
often a case of the assertion of various folk theorems.

THE ABOLITION OF THE FUNDHOLDING SCHEME

Soon after winning the May 1997 general election the new Labour government
suspended entry into the fundholding scheme and began a debate on
replacement commissioning models for the NHS (NHSE, 1997). After seven
months in office, the new administration announced plans to replace the
Conservative governments reforms in England in the White Paper, The New
NHS: Modern, Dependable (DoH, 1997). In keeping with Labours preelection promises, the White Paper stated that, at the end of 1998/99, both the
NHS internal market and the fundholding scheme would be abolished. In their
place an integrated care framework would be introduced, which would
separate the planning of hospital services from their provision. Under these
arrangements, 500 Primary Care Groups (PCGs), each serving a population of
approximately 100 000 people, were established from April 1999 onwards
(these subsequently became 300 Primary Care Trusts (PCTs) in April
2003).
Under the new arrangements for the NHS, the Labour governments White
Paper announced that each PCG would be allocated a cash-limited budget for
hospital and community health services, prescribing and general practice
infrastructure for the patients that they serve. Once established, the document
stated that the groups would be expected to subdivide their allocations
amongst local practices in the form of indicative budgets that cover all of the

116

Evidence

aforementioned services. However, in the short-term, the White Paper

announced that every practice will have a prescribing budget, as most do now
(DoH 1997, Chapter 5, p. 8). The practice-level budgets allocated by PCGs
were similar to those assigned under the fundholding scheme. Indeed, The
New NHS White Paper announced that the Government wants to keep what
has worked about fundholding, but discard what has not (p. 33). As a result,
PCGs were left to determine, for themselves, what aspects of the fundholding
scheme should be employed when devising their local, practice-level budget
schemes.
Although Labours plans allowed fundholding to be abolished without, in
principle, discarding the effective aspects of the scheme, the absence of an
independent evaluation meant that conclusive evidence on what aspects of the
initiative actually worked was not available to the government or PCGs.
Indeed, little evidence was produced on whether the fundholding budget itself,
the extra resources given to participating practices, or the difference in
property rights assigned to non-fundholding practices determined the
outcomes observed amongst the schemes incumbents. Moreover, it was not
clear whether the types of practices that elected to join the scheme influenced
its effects, or whether fundholding would have been equally effective amongst
all practices.
It is possible to interpret The New NHS as an attempt to re-establish trust
with the medical profession and configure a new policy community. Two
aspects stand out, the repeated emphasis on the need for cooperation with the
medical profession, and second, the commitment that a Labour government
would not change the PCG structures for at least ten years. This commitment
to policy stability in the primary care sector is a basic requirement for the reestablishment of trust between the government and the BMA after the
turbulence of the 1990s. Alongside this commitment to a policy lock-in is the
signal of a more evidence-based approach to policymaking. For example, the
White Paper states that there will be piloting and a proper evaluation of
Primary Care Trusts before their universal application.
However, to build trust takes time and other parts of the White Paper
explicitly limited clinical autonomy, for example the National Institute for
Clinical Excellence (NICE) and the national performance frameworks. The
subsequent path of the health care policy system is not the subject of this
chapter, but in the 1999 Labour Party conference there was an organized
message that the government believed the forces of conservatism were
hindering the modernization of public services, and that staff in those services
were part of the problem. Despite the 2000 NHS Plan and record increases in
NHS funding, contemporary health politics have taken place in a context of
frequent ministerial frustration, sometimes expressed publicly, at the failure of
the health care system to show demonstrable improvements on various output

The GP fundholding scheme

117

measures revealing that a high-trust, closed policy community has not

coalesced in this policy space.

SUMMARY
This chapter has set out the history of the GP fundholding scheme as an
example of policy dynamics after policy communities collapse and where the
government has an ambition for substantial policy change. The absence of
trust between the major interests means that policy tends to be formulated
without valid or reliable evidence on the effects and cost-effectiveness of
existing initiatives and any proposed new initiatives. In this sense, the 1991
NHS reforms were not a victory for the new public sector management at the
expense of clinical autonomy, but rather marked the beginning of a period of
turbulence in health care policy between 1991 and 1997 (and subsequently).
The GP fundholding scheme is a case study in how health care policy in the
1990s was driven by folk theorems and political competition; health care
problems and policy solutions were learned or constructed in the interaction of
agents in the policy system. In this sense of policy learning, the problem
situation is constructed in the interactions between agents; policy problems,
policy solutions and the criteria of success are bargained and different
attempts are made to constitute institutions within the policy system.
The structured narrative in the chapter is inconclusive on the question of
whether the period 199197 was an aberration in the style of health care
policymaking. The existence of a publicly funded and publicly owned NHS
had seemed to afford doctors a privileged position relative to other groups.
Many of the institutionalized forms of access for the medical profession of the
policy community era have remained. However, Ham (1999, p. 1092) notes the
Labour governments apparent willingness to challenge the power of its
traditional support base in the trade unions and entrenched interests of the
health professionals, including doctors. The 15 years since the break up of the
post-war health care policy paradigm have been marked by a fluid dynamics
of shifting balances and patterns of influence and values within the NHS,
without a particular policy paradigm becoming institutionalized.

9. UK pharmaceutical policy
The tendency of doctors to overprescribe medicines because they do not bear
the cost of the decision to prescribe, and the monopoly that pharmaceutical
companies enjoy in the production of certain medicines under patent, are both
long-standing justifications for pharmaceutical public policy to regulate the
price of medicines (Bloom and Van Reenen, 1998). All OECD countries have
some form of regulation of the pharmaceutical industry and mechanisms to
control public expenditure on medicines. These policies have been under
budgetary pressure in most countries since the 1980s due to the combination
of an ageing population and technological development. NHS expenditure on
prescription medicines increased by almost 10 per cent per annum during the
1990s (OHE, 2002). This fiscal pressure has produced a series of policy
dynamics that exemplify how policy processes are as much about choosing
between different reasons for action or different values, as they are about how
to achieve particular values in isolation. This insight complements the point
made in the theoretical section of the book: a dynamic perspective raises
serious doubts about the instrumental, parametric version of rationality that
exists in rational choice theory by challenging the notion of a straightforward
policy choice.
Public policy towards pharmaceuticals can be divided into those aimed at
influencing the demand for medicines, chiefly to increase the sensitivity of
GPs to the cost of medicines, and those aimed at regulating the supply side,
the price paid for medicines by public authorities. In the UK, since the mid
1980s there have been a series of initiatives to influence GP-prescribing
behaviour. On the supply side, the Pharmaceutical Price Regulation Scheme
(PPRS) is an agreement between the pharmaceutical industry and the
government that has been renegotiated roughly every six years since 1957; the
latest agreement came into force on 1 January 2005. The PPRS is unique in
the EU pharmaceutical policy area in regulating drug company profits rather
than prices directly, although relatively small one-off reductions in the average
price of the portfolio of drugs supplied by a company to the NHS were agreed
in 1993, 1999 and 2005.
The PPRS has a dual identity; it is both an industrial policy to support
pharmaceutical companies and the mechanism by which the NHS procures
drugs. There is an ineluctable conflict between those two identities. They
represent different values or rationalities for public policy that are directly
118

UK pharmaceutical policy

119

contradictory: the support of a research and development (R&D) intensive,

high value-added successful industry versus maximizing the health benefits of
a finite amount of public expenditure in the health care system. It is through
policies and policy paradigms that basic values are enacted and
institutionalized in the policy system. This chapter provides a structured
narrative of the dynamics of this conflict of values since 1957 when the
Voluntary Price Regulation Scheme (VPRS), which subsequently became the
PPRS in 1978, was first introduced. The narrative reveals how multiple
rationalities have become institutionalized in the pharmaceutical policy
system and that pharmaceutical policy overall is complex and inconsistent.

THE TRADE-OFF PROBLEM

When values are in conflict, as in the case of pharmaceutical policy, it is
common within policy studies to conceptualize the policy process as balancing
these values or providing for a trade-off between the values. This is always the
case with economic policy analysis; indeed, the whole notion of a social
welfare function, the central plank of public sector economics, is premised on
the view that different values can be differentially weighted in a common
function, and those relative weights represent the result of the trade-off. The
notion of a trade-off can be criticized philosophically because it requires the
assumption of commensurability: that different values may be converted to a
common metric (utility, for example) and thus balanced against each other.
However, for the purposes here the problem of a trade-off is one of
dynamics; from a dynamic point of view, policy systems do not remain at
some static equilibrium, balancing two conflicting values. Instead, there are
sequences of policy decisions: each decision at a particular temporal point
may embody particular reasons for action and particular values. Policy
decisions may be institutionalized and consequently the values that they
embody endure in the policy system. Further, the reasons for action or the
values that are being enacted in a policy may vary over time, but they do not
vary independently of each other. Rather an initial policy decision may set in
train a set of different, often countervailing, effects as policies adopted in
pursuit of one value or for a particular reason have unintentional consequences
for another interest or value, whose relative importance in the situational
context at the time of the initial policy may not have been considered. These
consequences may prompt the introduction of another policy as a patch to
ameliorate those policy effects. Complex policy systems may contain several
different rationalities and different values that have accumulated over time; a
key question is the extent to which the system can maintain coherence and
stability with such immanent accumulations.

120

Evidence

In terms of policy dynamics, there is no definitive and enduring trade-off;

there is no reason to expect that the policy process will bring both values
together and make a definite political choice between them. Instead policy
actors focus on each value sequentially, emphasizing one value until the
negative consequences for other values reach a critical level, at which they are
asserted contrarily. Thacher and Rein (2004, p. 465) label this as a process of
cycling: To say that policy responds to conflicting values through cycling
means that the attention public institutions pay to each value varies over time,
shifting back and forth between opposites most commonly, subordinating
one-half of a dilemma and then the other. UK pharmaceutical policy is a good
example of this policy dynamic: the Department of Health (DoH) has a Pricing
and Supply branch that administers the PPRS but is also responsible for the
budgets of PCTs and Hospital Trusts that purchase medicines in the NHS. The
narrative presented in this chapter shows how the cycling process can lead to
the steady accumulation of policy decisions and policy subsystems to mitigate
the negative consequences of one policy for another policy value, and the
consequences in terms of the overall coherence and effectiveness of policy in
an area.
One effect of temporal cycling between values in pharmaceutical policy is
a functional separation of different parts of policy. This is where different
policies are established for different values at different times. In one sense, a
functional separation of parts of policy space is foundational to bureaucracies:
a health department, a treasury, a foreign affairs department, and so on.
Policies established to mitigate the consequences of other polices may exist as
a separate policy subsystem, with different actors, institutions and values. The
joined-up government initiative is, of course, an attempt to remedy the adverse
consequences of this separation, and indeed in Chapter 7 on the CAP, several
spillovers between different parts of policy space were described.
The theme of the narrative presented here is that the functional separation
within pharmaceutical policy between the demand and supply side of the
regulation of purchasing medicines has become unstable. The cycling process
has resulted in the introduction of new, additional policies, which intermesh
and interact in an increasingly crowded policy space with original policies, as
first the budget of the NHS for medicines is emphasized and then the needs of
the pharmaceutical industry. This is a type of policy inheritance process where
different parts of policies enacted at different times get institutionalized
consecutively. The present mixture of different values and multiple
rationalities from different times exacerbates complexity in the policy system.
I describe the emerging complexity in the sections below but stress that the
immutable conflict of the two values remains. As Wildavsky (1964, pp. 1289)
observed: the fundamentals of any budget problem, which pharmaceutical
policy is, always concern value choice.

UK pharmaceutical policy

121

THE STRUCTURED NARRATIVE OF UK

PHARMACEUTICAL POLICY DYNAMICS
The narrative is broken down into two periods. Between 1957 and the mid1980s, pharmaceutical policy consisted of one instrument, the PPRS, and the
mode of regulation was as a gentlemans agreement; since the mid-1980s,
however, the objective of the NHS in controlling the level and nature of public
expenditure on medicines has been periodically asserted, in a number of policy
initiatives, in a process of cycling with the pharmaceutical industry
development value.
Thacher and Rein (2004) present a case that, from a practical reasoning
point of view, cycling may be effective; it is not a failure of rationality or the
triumph of short-term, political interests over longer-term, evidence-based
strategic analysis. They concede, however, that it is an open and empirical
question whether, in practice, cycling can manage the conflict of values in a
consistent and effective manner. The dominant theme of the narrative of UK
pharmaceutical policy is the increase in the number of policy instruments
within pharmaceutical policy space. Different demand- and supply-side
policies have been enacted separately, at different times and for different
reasons; yet their material consequences make them strongly intertwined in a
manner that has left pharmaceutical policy overall as inconsistent, incoherent
and ineffective. In particular, it has proved impossible to isolate particular
domains within the pharmaceutical policy area; new policy instruments have
had complex spillover effects into other domains of pharmaceutical policy. No
institution or policy mechanism has emerged to manage, in a stable and
effective manner, the conflict of the two values and the linkages and
interconnections between different pharmaceutical policy domains.
The format and rules of the PPRS, and its predecessor the VPRS, have not
changed significantly since 1957. The PPRS does not regulate prices directly
but is instead a form of rate-of-return regulation. There are profit targets,
measured in terms of the return on capital that companies can make on their
sales to NHS. A company launching a new drug has discretion over the initial
price but a second part of the PPRS strictly controls any subsequent price
increases. A third part of the scheme regulates the amount of advertising
expenditure that is an allowable cost in the profit calculations, and a fourth part
controls the amount of R&D expenditure that can be set against sales to the
NHS in the same calculation. Pharmaceutical companies provide the DoH
with this information in an annual financial return (AFR). The scheme allows
UK-based pharmaceutical companies to make a 21 per cent return on capital
on their sales of medicines to the NHS. The scheme allows an expense of up
to 23 per cent of the total value of sales and 9 per cent of the same figure to be
set against the profit calculation for R&D and marketing respectively.

122

Evidence

The PPRS is, de facto, an industrial policy; it affects the profitability and
structure of the UK pharmaceutical industry. Further, the industrial policy
element is explicit in the objectives of the scheme. As most recently stated
(DoH, 2004), they are to: (i) secure the provision of reasonably priced, safe
and effective medicines to the NHS; (ii) promote a strong and profitable
pharmaceutical sector capable of such sustained R&D expenditure as should
lead to the future availability of new and improved medicines; and (iii)
encourage the efficient and competitive supply of medicines to pharmaceutical
markets in this and other countries.
As previously indicated, the PPRS as a mode of regulation is akin to a
gentlemans agreement. The metaphor is a useful description because it
captures the high-trust, personal relationships that have existed between the
government and the pharmaceutical industry away from public scrutiny.
Common to all regulatory systems is an information problem; the regulator
cannot directly observe the cost structure of the regulated company. The PPRS
has dealt with this problem by allowing the industry a large degree of selfregulation on a voluntary basis. The branch of the DoH that operates the PPRS
consists of only 20 people (DoH, 2004). The majority of these will be in
clerical grades and the Branch Head is a Grade 7, a middle-management grade
that is not part of the senior civil service structure. Therefore, it is a reasonable
assumption that this branch cannot engage in detailed scrutiny of the
information provided by companies in the PPRS. Rather, the level of resources
that the DoH commits to the scheme is only sufficient to ensure the schemes
routine administrative functioning. The DoH must take on trust the validity of
the large majority of the information provided by the pharmaceutical
companies in their AFRs. In these terms, the PPRS is a relationship-based
rather than a rule-based regulation and the metaphor of a gentlemans
agreement expresses the nature of that relationship.
In contrast, UK utility regulators are independent agencies with a large staff
and substantial resources. For example, OFTEL has a staff of over 160
(OFTEL, 2004). They play a quasi-judicial role in protecting customers and
exist in an adversarial relationship with the regulated company. The approach
to the information problem is to pursue detailed analysis and scrutiny of the
information revealed by the company, their accounts and forecasts, in order to
formulate an optimal price that the company can charge.
The gentlemans agreement mode of regulation seems to have been
successful in preventing a hold-up problem in the relationship between the
government and the pharmaceutical industry. A hold-up problem occurs in a
contract between two parties where one of the parties has to make some kind
of investment that is irreversible. In this case, it is R&D expenditures by
pharmaceutical companies that are irreversible. The party that makes such an
investment becomes vulnerable to the demands of the other party to

UK pharmaceutical policy

123

renegotiate the contract; in this case, the government imposing drug prices that
are close to marginal cost and substantially lower than had been anticipated at
the time of the R&D investment. Given the frequent cash crises in the history
of the NHS (Webster 1988, 1996) and the continually rising public
expenditure on drugs, there always exists an incentive for governments to
impose a low price regime on the pharmaceutical industry. In game theory
terms, the governments promise to keep prices at a level to reward investment
in R&D is not time consistent (Dixit 1996).
This potential threat of government expropriating the returns for investment
in future time periods should tend to lead to severe under-investment by
pharmaceutical companies. However, there is considerable evidence of the
long-run success of the UK pharmaceutical industry (Maynard and Bloor
1997; Martin 1995; Earl-Slater 1997) and in particular, its high rates of R&D
expenditure (Bloom and Van Reenen 1998). The gentlemans agreement
established a level of trust between the industry and the government that has
given credibility to commitments not to cut future prices. The PPRS, in terms
of the value of supporting the development of high value-added industry in the
UK economy, has been successful. However, as previously noted this conflicts
with the value of maximizing the health benefits of a limited public budget for
medicines.

VALUE CYCLING AND A MORE CROWDED

PHARMACEUTICAL POLICY SPACE
The PPRS remained the sole instrument for regulating expenditure on
pharmaceuticals from its inception in 1957 until 1985, when the Conservative
government specified, for the first time, a limited list of medicines that could
be prescribed in certain therapeutic categories at the expense of the NHS. The
list was extended to several further therapeutic categories in 1992. This
marked a shift in pharmaceutical policy in the UK towards an emphasis on
controlling the demand for medicines. Further demand-side policies followed:
the setting of practice-level indicative prescribing budgets using Prescription
Analysis and Cost (PACT) data in the late 1980s, the introduction of GP
fundholding in 1991 and the successful encouragement of increased rates of
generic prescribing. These developments represent the assertion of the value
of health policy against that of industrial policy in the purchasing of medicines
by the NHS. This sequence of policy developments has been in a synchronized
but anti-phase pattern with the sequence of renegotiations of an essentially
unchanged PPRS; a policy decision asserting one value produces another
policy decision that counter-asserts an alternative value in a process of cycling.
The Conservative governments decision to take generics out of the PPRS

124

Evidence

in 1986 (Luce, 1987) marks the beginning of the increase in the number of
pharmaceutical policy instruments. Any number of pharmaceutical companies
can supply a generic drug because the patent guaranteeing a monopoly in the
production of that drug has expired. Generics trade under the British Approved
Name of a drug rather than a brand name. They were removed from the PPRS
in order to encourage a market with price competition. As generic medicines
are usually considerably cheaper than their branded equivalent, it was hoped
that GPs would prescribe, where available, a generic drug rather than its
branded alternative, thereby securing economies in the NHS pharmaceuticals
budget. The various generic prescribing initiatives that followed (see for
example Audit Commission 1994) have successfully raised the rate of
prescribing generically in England and Wales to close to 70 per cent of items
(from 40 per cent prior to 1986), representing around a fifth of the total value
of prescription medicines.
The development of a generics market came alongside the introduction of
the GP fundholding scheme in 1991 (see Chapter 8). This scheme allowed
participating practices to keep any savings they could make on their purchases
of drugs. These initiatives provided a financial incentive to GPs to limit the
effect of their prescribing on pharmaceutical expenditure. However, any
control of NHS expenditure on pharmaceuticals is a control on pharmaceutical
companies revenues. The combination of competition through generic
substitution and a system of price regulation in the market for non-branded
drugs has had the effect of reducing the life cycle of branded products as the
revenues expected by a branded manufacturer post-patent expiry have been
reduced. This means that branded manufacturers have been forced to make
their return on R&D investment over the period of the patent instead of over a
longer period.
The freedom given by the PPRS to companies in the pricing of new drugs
means that pharmaceutical policies to control the growth of drug costs will
tend to be undone by the ability of the drug companies to increase their
revenues (NHS drug expenditure) through the launch price of new drugs. The
average price of branded medicines prescribed increased by about 47 per cent
in real terms over the period 198898, whilst the average price of generic
medicines has remained about the same in real terms over the same period
(Kay 2002). However, under the PPRS the price of existing medicines can
only be increased by application to the DoH. There are very strict criteria for
allowing a price increase (DoH 2004) and the industry view is that such
applications tend to be unsuccessful (Lawton 1999). It therefore seems
reasonable to infer that the major determinant of the steep rise observed in the
price of branded medicines is the increased average price of new medicines
launched.
The data tend to show that these developments in GP prescribing policy

UK pharmaceutical policy

125

over the last 20 or so years have been successful in controlling the volume of
items prescribed. The growth in volume of items has been roughly half that of
the average price of an item. However, the policies to influence GP prescribing
had a greater objective than simply influencing the prescribe/not prescribe
decision; they were supposed to make GPs more price-sensitive when taking
the decision to prescribe a particular drug. The demand-side policies since the
1980s were designed, all other things being equal, to reduce the average price
of a prescription item. However, since the mid-1980s all other things have not
been equal. As noted, supply-side regulations have operated in a contradictory
way to demand-side policies by allowing industry profits on sales to the NHS
to remain at historic levels because the initial price of new medicines is not
controlled under the PPRS. The introduction of NICE discussed in the
following section is the latest attempt to close this persistent gap in the
regulatory framework of the PPRS and another stage of value cycling within
pharmaceutical policy.

1999 AND 2005 PPRS AGREEMENTS

Unlike previous PPRS agreements, the 1999 negotiations were both
contentious and a matter of public debate. The government had publicly stated
its intention to save the NHS money (The Guardian, 24 September 1998) and
industry representatives warned that companies would start to locate
production abroad (Financial Times, 20 October 1998). Despite this acrimony,
the PPRS was renegotiated for another six years. The new scheme had a basic
architecture similar to previous agreements; the objectives of the PPRS
remained the same with only marginal changes to the mechanics of its
operation (Lawton 1999). The headline change was an agreed average price
reduction (across all products supplied to the NHS) of 4.5 per cent, compared
to 2.5 per cent in 1993.
One part of the agreement with potentially significant longer-term
consequences for the future of the PPRS and the nature of the regulatory
relationship was the 1999 Health Act. This gave the scheme and its rules a
statutory base for the first time. For example, the government had complained
that some PPRS companies had abused the spirit of the agreement at best by
selling the brand name of some low-cost but essential products to smaller
rivals, who then steeply increased the price. The government could now
enforce legally the PPRS rules on the selling on of old brands. This change in
the legal status of the PPRS had little impact on the regulatory style of the
scheme; it still relies on a trust-based relationship between the government and
the industry. It remains soft law, in the sense of being a system of selfregulation where the rules that govern behaviour in medicines sales to the

126

Evidence

NHS are administered and enforced by the industry itself. There was no
increase in the resources available for regulatory scrutiny to the levels usually
associated with utility regulation. There are still only 20 officials operating the
scheme in the DoH and although the details of the operation of the scheme
have been reported to Parliament annually since 1996, the scheme remains
opaque from the outside; in particular, there is no evidence of a
confrontational, hard law regulatory style with frequent recourse to the
statutory courts to enforce the rules of the scheme.
Independently of, but contemporaneous with, the 1999 PPRS negotiations,
NICE was established with the objective of producing clinical guidelines on
the clinical and cost effectiveness of medical interventions. It replaced a
system in which individual health authorities took their own decisions on the
funding for new drugs and technologies. This system had resulted in variations
in the patterns of commissioning across the country, with varying access to
specific treatments and allegations of rationing by postcode. These decisions
had generally been taken in response to resource pressures and financial
reductions rather than research into the cost-effectiveness of medical
interventions.
As stated, the PPRS does not directly regulate prices but instead controls
profit, with pricing freedom given to companies when launching branded
medicines. NICE, however, is a form of price regulation; it judges the cost
effectiveness of a medicine. However, the cost-effectiveness methodology that
NICE use to make comparative judgements across diseases and for clinical and
technology appraisal benchmarks are opaque and open to dispute (Williams
2004); implicit is the recommendation of a price that provides a benefit to cost
ratio that is acceptable for public expenditure on health care. There is therefore
an inconsistency between Labour opting to renegotiate the PPRS in 1999 and
also introducing NICE. The price of branded drugs is a policy variable
performing two potentially contradictory regulatory functions: rewarding
innovation and providing cost-effective treatment. Prior to NICE, the price of
drugs in the PPRS was the outcome of political negotiations for a reasonable
level for the NHS and the industry. The requirement for NICE to reach an
opinion on the cost-effectiveness of a drug may lead to a more formal and
confrontational regulatory relationship between the government and the
industry as public recommendations, with adverse profit consequences. The
relationship will be further complicated by the trend in the demand for
medicines: this is no longer simply a function of physicians prescribing habits
when acting as agents of patients, but rather is patient-led, where responsibility
for health care is increasingly individualized and where certain groups and
interests are organizing successfully for public subsidy of their medicines.
The PPRS agreement covering 200510 was the subject of far less public
debate than in 1999. There have been a few very minor changes, but the

UK pharmaceutical policy

127

scheme remains essentially the same as that under the 1999 agreement: the
objectives remain, its statutory basis under the 1999 Act is untouched, and
there was an agreed average price reduction of 7 per cent this time (compared
to 4.5 per cent in 1999). In order to limit room for manoeuvre in complying
with price cut commitments and increase transparency and competition in the
generics market, branded generics have been excluded from the PPRS. These
are out-of-patent products to which suppliers, who did not ever hold the patent,
have applied a brand name. Any encouragement of the generics as set against
the branded industry is an assertion of the objective of securing value for
money in public spending on medicines.

THE COMPLEX SPILLOVER EFFECTS IN A CROWDED

POLICY SPACE
The previous section described a more crowded pharmaceutical policy area in
the 1990s and identified some spillovers between different policy domains.
The development of a generics market has affected the expected profitability
of new branded medicines, and this seems to be linked with an increase in the
average price of branded medicines. In addition, the work of NICE affects the
expected sales revenues of companies in the PPRS by insisting that a branded
medicine be cost-effective treatment for the NHS. However, the PPRS has
recently been renegotiated, with the objectives of supporting innovation in the
pharmaceutical industry unchanged and the launch price of medicines to
remain at the discretion of companies.
The overall NHS drugs budget increased by almost 10 per cent per annum
during the 1990s. This resulted in expenditure on pharmaceuticals accounting
for over 15 per cent of total NHS expenditure by 2001; between 1969 and
1989 it had been a relatively stable share that averaged 8.6 per cent per annum
(OHE, 2002). In response to the rising drugs budget, successive governments
have introduced various measures to control the demand for medicines by
NHS (such as GP fundholding, encouraging a generics market and NICE). The
overall NHS demand for medicines is regulated by different institutions, as
well as separately from the supply of medicines. However, all need to be
considered in conjunction in order to control the growth of expenditure on
pharmaceuticals, but the spillovers between policy domains are not the
responsibility of any policy or institution. As Walley et al. (2000) put it, The
pharmaceutical arena is now far too complex to be enveloped in the current
balkanised approach to policy. Without policy spillovers being managed or
controlled there is a potential instability in pharmaceutical policy. Such
instability would affect both the pharmaceutical industry and the NHS:
long-term R&D investment in the UK by the pharmaceutical industry would

128

Evidence

be threatened and the ability of the NHS to predict and control overall drug
expenditure would be diminished (ultimately adversely affecting patients
through unplanned and random rationing).
The combination of complex policy spillovers suggests the need for policy
or institution to regulate the governmentpharmaceutical industry relationship
and manage the value conflict that has produced the process of cycling,
described in the previous sections, with the adverse consequences for policy
coherence. There is no institution or policy independent of the DoH, which
takes an overall view of the regulation of the pharmaceutical industry, to in
particular take account of the spillovers between different policy instruments.
In particular, an OFDRUG, equivalent of OFTEL or OFWAT, could help to
ensure that there is a consistent and stable approach to regulation. One of the
rationales for an independent regulatory agency for the privatized utilities was
the necessity for a stable regulatory environment to allow for long-term
resource planning away from short-term political pressures. In the case of
pharmaceutical regulation, both the NHS and the pharmaceutical industry
wish to plan resource allocation over a long time horizon.
There are two practical roles an OFDRUG could adopt to ensure this
stability. The first would be to ensure that NHS expenditure on
pharmaceuticals is cost-effective. The separate and political question of the
affordability of a medicine or the overall size of the pharmaceutical budget
would remain for the DoH. A second and related role would be for an
OFDRUG to control both the recommendation of new medicines for use and
the funding for their implementation; currently Health Authorities may ignore
NICE guidance on the financial grounds that the introduction of a new
medicine would compromise its locally agreed priorities.
The low level of administrative resources dedicated to the PPRS by the DoH
was sustainable in the period of relationship-based regulation with high trust
between the regulators and the regulated. However, the robust policing of the
PPRS, the regulation of the generics market and administration of NICE rules,
as well as the required analysis of the extent of spillovers, would require
substantial expertise and resources. The question for the future development of
UK pharmaceutical policy is: whether any such single body can be designed
to manage the unavoidable value conflict; whether the complexities and
inconsistencies precipitate a crisis and radical reform; and whether the policy
system can self-organize under the selection pressures of fiscal constraints and
global industry demands.

SUMMARY
For many years drug price regulation posed few public management problems.

UK pharmaceutical policy

129

The growth in NHS expenditure on medicines was contained to a relatively

stable share of overall NHS expenditure, the pharmaceutical industry was
successful, the PPRS was not the subject of significant public scrutiny and the
associated transaction costs were relatively low. However, the growth in the
number of pharmaceutical policy instruments during the 1980s and 1990s has
introduced an element of instability into the regulation of the industry.
Instability is a problem that affects NHS financial programming as well as
R&D investment decisions by companies.
The narrative of this increased complexity and inconsistency in
pharmaceutical policy involves value cycling. The PPRS was introduced as an
industrial policy to assist the development of the infant pharmaceutical
industry in the UK in the 1950s. However, by the 1980s the consequences of
the PPRS for health spending began to attract the attention of health
policymakers. The objective of value for money in public expenditure on
medicines has been enacted in a series of demand-side policies from the 1986
reform, to PPRS, over generics with the associated prescribing initiatives, to
the introduction of NICE. However, the assertion of the health budget value in
the pharmaceutical policy system has been concurrent with a periodic
agreement of an unchanged PPRS and support by the DoH for the value of
supporting an innovative, R&D-intensive, high valued-added sector of the
economy.

References
Ackrill, Robert W. (2000), The Common Agricultural Policy, Sheffield:
Sheffield Academic Press, for the University Association for Contemporary
European Studies.
Ackrill, Robert W. and A. Kay (2005), Measuring path dependency in the
CAP, unpublished manuscript.
Ackrill, Robert W., R.C. Hine, A.J. Rayner and M. Suardi (1997), Member
states and the preferential trade and budget effects of the 1992 CAP reform:
a note, Journal of Agricultural Economics, 48, 93100.
Anonymous (1984), Two tier service feared, British Medical Journal, 289,
14689.
Arrow, K.J. (2000), Increasing returns: historiographic issues and path
dependence, The European Journal of the History of Economic Thought, 7
(1), 17180.
Arthur, W. Brian (1988), Self-reinforcing mechanisms in economics, in P.W.
Anderson, K.J. Arrow and D. Pines (eds), The Economy as an Evolving
Complex System, Reading, MA: Addison Wesley, pp. 122.
Arthur, W. Brian (1994), Increasing Returns and Path Dependence in the
Economy, Ann Arbor, MI: University of Michigan Press.
Audit Commission (1994), A Prescription for Improvement: Towards More
Rational Prescribing in General Practice, London: HMSO.
Audit Commission (1995), Briefing on GP Fundholding, London: HMSO.
Audit Commission (1996a), Fundholding Facts: A Digest of Information
About Practices Within the Scheme During the First Five Years, London:
HMSO.
Audit Commission (1996b), What the Doctor Ordered: A Study of GP
Fundholders in England and Wales, London: HMSO.
Auyang, Sunny Y. (1998), Foundations of Complex-System Theories,
Cambridge: Cambridge University Press.
Baines, Darrin, Keith Tolley and David Whynes (1997), Prescribing, Budgets
and Fundholding in General Practice, London: Office of Health
Economics.
Bates, Robert, Avner Grief, Margaret Levi, Jean-Laurent Rosenthal and Barry
Weingast (1998), Analytical Narratives, Princeton, NJ: Princeton
University Press.
Baumgartner, Frank and Bryan Jones (1993), Agendas and Instability in
131

132

The dynamics of public policy

American Politics, Chicago: Chicago University Press.

Baumgartner, Frank and Bryan Jones (eds) (2002), Policy Dynamics, London:
University of Chicago Press.
Bevir, M. and R.A.W. Rhodes (2003), Interpreting British Governance,
London: Routledge.
Bevir, M. and R.A.W. Rhodes (2005), Intrepretation and its others,
Australian Journal of Political Science, 40 (2), 16988.
Blaug, Marc (1986), Economic History and the History of Economics, New
York: NYU Press.
Bloom, N. and J. Van Reenen (1998), Regulating drugs prices: where do we
go from here?, Fiscal Studies, 19 (3), 32142.
Bourdieu, Pierre (1984), Distinction: A Social Critique of the Judgement of
Taste, London: Routledge.
Braybrooke, D. and C.E. Lindblom (1963), A Strategy of Decision, New York:
The Free Press.
Bridge, G. (2001), Rationality, ethics and space: on situated universalism and
the self-interested acknowledgement of difference, Environment and
Planning D: Society and Space, 18, 51935.
Bridgman, Peter and Glyn Davis (2004), Australian Policy Handbook, 2nd
edition, Sydney: Allen and Unwin.
Bryant, J.M. (2004), An evolutionary social science? A skeptics brief,
theoretical and substantive, Philosophy of the Social Sciences, 34 (4),
45192.
Buchanan, James M. and Richard A. Musgrave (1999), Public Finance and
Public Choice: Two Contrasting Visions of the State, Cambridge, MA: MIT
Press.
Bthe, T. (2002), Taking temporality seriously: modelling history and the use
of narratives as history, American Political Science Review, 96 (3), 48193.
Campbell, Donald T. (1965), Variation and selective retention in sociocultural evolution, in H.R. Barringer, G.I. Blanksten and R.W. Mack (eds),
Social Change in Developing Areas: A Reinterpretation of Evolutionary
Theory, Cambridge, MA: Schenkman Publishing Company, pp. 1949.
Campbell, Donald T. (1974), Evolutionary epistemology, in Paul Arthur
Schilpp (ed.), The Philosophy of Karl Popper, The Library of Living
Philosophers Volume XIV, La Salle, IL: Open Court, pp. 413643.
Campbell, John L. (1997), Mechanisms of evolutionary change in economic
governance, in Lars Magnusson and Jan Ottosson (eds), Evolutionary
Economics and Path Dependence, Cheltenham, UK and Lyme, USA:
Edward Elgar, pp. 1033.
Campbell, John L. (2004), Institutional Change and Globalization, Princeton,
NJ: Princeton University Press.
Centre for Management and Policy Studies (2001), Better Policy-Making,

References

133

London: UK Cabinet Office.

Chalmers, D.J. (1995), Facing up to the problem of consciousness, Journal
of Consciousness Studies, 2 (3), 200219.
Cohen, G.A. (1978), Karl Marxs Theory of History: A Defence, Oxford:
Oxford University Press.
Cohen, G.A. (1980), Functional explanation: reply to Elster, Political
Studies, XXVIII (1), 12935.
Colebatch, Hal K. (1998), Policy, Buckingham: Open University Press.
Collingwood, R.G. (1946), The Idea of History, Oxford: Clarendon Press.
Commission of the European Communities (1987), On the Financing of the
Community Budget, COM(87)101, Brussels, 28 February.
Committee of Public Accounts (1995), General Practitioner Fundholding in
England, London: HMSO.
Cortes, F., A. Przeworski and J. Sprague (1974), Systems Analysis for Social
Scientists, New York: Wiley.
Coulter, A. (1995), Shifting the balance from secondary to primary care,
British Medical Journal, 311, 14478.
Crouch, C. and H. Farrell (2004), Breaking the path of institutional
development? Alternatives to the new determinism, Rationality and
Society, 16 (1), 543.
Daugbjerg, C. (1999), Reforming the CAP: policy networks and broader
institutional structures, Journal of Common Market Studies, 37, 40728.
Davies H.T.O, S. Nutley and P.C. Smith (2000), What Works? Evidence-based
Policy and Practice in Public Services, Bristol: The Policy Press.
Dawkins, R. (1996), Climbing Mount Improbable, London: Viking.
Dennett, D.C. (1996), Facing backwards on the problem of consciousness,
Journal of Consciousness Studies, 3 (1), 46.
Denzau, A.D. and D.C. North (1994), Shared mental models: ideologies and
institutions, Kyklos, 47 (1), 331.
Department of Health (DoH) (1989a), Working for Patients, London: HMSO.
DoH (1989b), Working Paper 8: Implications for Family Practitioner
Committees, London: HMSO.
DoH (1997), The New NHS: Modern, Dependable, London: DoH.
DoH (2004), The Pharmaceutical Price Regulation Scheme, London: DoH.
Dewey, John (1922), Human Nature and Conduct: An Introduction to Social
Psychology, New York: Modern Library.
Dixit, Avanish (1996), The Making of Economic Policy: A Transaction-cost
Politics Perspective, London: MIT Press.
Dowding, K. (2000), How not to use evolutionary theory in politics: a critique
of Peter John, British Journal of Politics and International Relations, 2 (1),
7280.
Dye, Thomas R. (1972), Understanding Public Policy, Engelwood Cliffs, NJ:

134

The dynamics of public policy

Prentice Hall.
Earl-Slater, A. (1997), Regulating the price of the UKs drugs: second
thoughts after the governments first report, British Medical Journal, 314,
3658.
Easton, David (1965), A Framework for Political Analysis, New York: Prentice
Hall.
Easton, David (1966), A Systems Approach to Political Life, Indiana: Purdue
University Press.
Elster, Jon (1976), A note on hysteresis in the social sciences, Synthese, 33,
37191.
Elster, Jon (1980), Cohen on Marxs theory of history, Political Studies,
XXVIII (1), 12128.
Elster, Jon (1983), Sour Grapes: Studies in the Subversion of Rationality,
Cambridge: Cambridge University Press.
Elster, Jon (1998), A plea for mechanisms, in Peter Hedstrom and Richard
Swedberg (eds), Social Mechanisms: An Analytical Approach to Social
Theory, Cambridge: Cambridge University Press, Chapter 3.
Elster, Jon (2000), Rational choice history: a case of excessive ambition,
American Political Science Review, 94 (3), 68594.
Enthoven, Alan (1985), Reflections on the Management of the NHS, London:
Nuffield Provincial Hospital Trust.
Esping-Anderson, Gosta (1990), The Three Worlds of Welfare Capitalism,
Princeton, NJ: Princeton University Press.
European Commission (1958), First General Report on the Activities of the
Community, Brussels: Publications Department of the European
Communities.
European Commission (2002), Mid-term review of the Common Agricultural
Policy, Communication from the Commission to the Council and the
European Parliament, COM(2002)394 final, Brussels, 10 July.
Fischer, Frank (2003), Reframing Public Policy, Oxford: Oxford University
Press.
Foster, John (1997), The analytical foundations of evolutionary economics:
from biological analogy to economic self-organisation, Structural Change
and Economic Dynamics, 8, 42751.
Georgescu-Roegen, Nicholas (1967), Analytical Economics, Cambridge, MA:
Harvard University Press.
Georgescu-Roegen, Nicholas, (1971), The Entropy Law and the Economic
Process, Cambridge, MA: Harvard University Press.
Gigerenzer, G. and R. Selten (eds) (2001), Bounded Rationality: The Adaptive
Toolbox, Boston: MIT Press.
Goldstone, J.A. (1998), Initial conditions, general laws, path dependence, and
explanation in historical sociology, American Journal of Sociology, 104

References

135

(3), 829-45.
Goodin, Robert E. (1998), Institutions and their design, in Robert E. Goodin
(ed.), The Theory of Institutional Design, Cambridge: Cambridge
University Press, Chapter 1.
Gosden, T. and D. Torgenson (1997), The effect of fundholding on
prescribing and referral costs: a review of the evidence, Health Policy, 40,
10314.
Green, Donald P. and Ian Shapiro (1994), Pathologies of Rational Choice
Theory: A Critique of Applications in Political Science, New Haven: Yale
University Press.
Greener, I. (2002), Understanding NHS reform: the policy-transfer, social
learning, and path dependency, Governance, 15 (2), 16183.
Habermas, Jurgen (1984), The Theory of Communicative Action, Volume 1:
Reason and the Rationalisation of Society, London: Heinemann.
Habermas, Jurgen (1987), The Theory of Communicative Action, Volume 2: A
Critique of Functionalist Reason, Cambridge: Polity Press.
Hacker, J.S. (1998), The historical logic of National Health Insurance:
structure and sequence in the development of British, Canadian, and US
medical policy, Studies in American Political Development, 12, 57130.
Hacker, J.S. (2002), The Divided Welfare State: The Battle Over Public and
Private Social Benefits in the United States, Cambridge, MA: Cambridge
University Press.
Hall, Peter A. (1993), Policy paradigms, social learning, and the state,
Comparative Politics, 25 (3), 27596.
Hall, Peter A. (2003), Aligning ontology and methodology in comparative
historical analysis, in James Mahoney and Dietrich Rueschmeyer (eds),
Comparative Historical Analysis in the Social Sciences, Cambridge:
Cambridge University Press, pp. 373406.
Hall, Peter A. and R.C.R. Taylor (1996), Political science and the three new
institutionalisms, Political Studies, 44 (4), 93657.
Hall, Peter A. and R.C.R. Taylor (1998), The potential of historical
institutionalism: a response to Hay and Wincott, Political Studies, 46 (5),
95862.
Ham, Christopher (1992), Health Policy in Britain, Basingstoke: The
MacMillan Press Ltd.
Ham, Christopher (1999), New Labour and the NHS, British Medical
Journal, 318, 1092.
Hamilton, W. (1996), Narrow Roads in Gene Land, vol: Evolution of Social
Behaviour, New York: Oxford University Press.
Harris, Conrad (ed.) (1996), Prescribing in General Practice, Oxford:
Radcliffe Medical Press.
Harrison, Stephen, David J. Hunter and Christopher Pollitt (1990), The

136

The dynamics of public policy

Dynamics of British Health Policy, London: Unwin Hyman.

Hay, Colin (2002), Political Analysis: A Critical Introduction. Basingstoke:
Palgrave.
Hay, Colin (2004), Theory, stylized heuristic or self-fulfilling prophecy? The
status of rational choice theory in public administration, Public
Administration, 82 (1), 3962.
Hay, Colin and D. Wincott (1998), Structure, agency and historical
institutionalism, Political Studies, 46 (5), 9517.
Heclo, H. (1972), Review article: policy analysis, British Journal of Political
Science, 2, 83108.
Hempel, Carl G. (1965), Aspects of Scientific Explanation, New York: Free
Press.
Hennessy, Peter (1992), Britain: Never Again, 194551, London: Jonathan
Cape.
Hindmoor, A. (1998), The importance of being trusted: transaction costs and
policy network theory, Public Administration, 76, 2543.
Hodgson, Geoffrey M. (2001), How Economics Forgot History: The Problem
of Historical Specificity in Social Science, London: Routledge.
Hodgson, Geoffrey M. (2002), Darwinism in economics: from analogy to
ontology, Journal of Evolutionary Economics, 12, 25981.
Hogwood, Brian W. and Lewis A. Gunn (1984), Policy Analysis for the Real
World, Oxford: Oxford University Press.
Howlett, Michael and M. Ramesh (2003), Studying Public Policy: Policy
Cycles and Policy Subsystems, Oxford: Oxford University Press.
Hudson, John and Stuart Lowe (2004), Understanding the Policy Process:
Analysing Welfare Policy and Practice, Bristol: The Policy Press.
Hunter, D. (1998), Review of Martin Powell, Evaluating the National Health
Service, Journal of Social Policy, 27, 13234.
Jervis, R. (1997), System Effects: Complexity in Social and Political Life,
Princeton, NJ: Princeton University Press.
John, P. (1999), Ideas and interests; agendas and implementation: an
evolutionary explanation of policy change in British local government
finance, British Journal of Politics and International Relations, 1 (1),
3962.
Jordan, A.G. and J.J. Richardson (1987), Government and Pressure Groups in
Britain, Oxford: Clarendon Press.
Kaldor, N. (1970), The case for regional policies, Scottish Journal of
Political Economy, 18, 33748.
Kauffman, Stuart A. (1995), At Home in the Universe. The Search for Laws of
Self-Organization and Complexity, New York: Oxford University Press.
Kay, Adrian (1998), The Reform of the Common Agricultural Policy: The Case
of the MacSharry Reforms, Wallingford: CAB International.

References

137

Kay, Adrian (2002), Devolution and drugs: the case for Welsh autonomy,
Policy and Politics, 30 (2), 18798.
Kay, Adrian (2003), Path dependency and the CAP, Journal of European
Public Policy, 10 (3), 40521.
Kay, A. and Robert W. Ackrill (2006), Historical institutionalist perspectives
on the development of the EU budget system, Journal of European Public
Policy, 13 (1), 11333.
Kemp, P. (2001), Housing benefit and welfare state retrenchment in Britain,
Journal of Social Policy, 29 (2), 26379.
Kenway, Peter (1993), From Keynesianism to Monetarism: Evolution of the
UK Macroeconometric Models, London: Routledge.
Kerr, P. (2002), Saved from extinction: evolutionary theorising, politics and
the state, British Journal of Politics and International Relations, 4 (2),
33058.
Kingdon, John W. (1984), Agendas, Alternatives and Public Policies, New
York: Longman.
Klamer, Arjo and Thomas C. Leonard (1994), So whats an economic
metaphor?, in P. Mirowski (ed.), Natural Images in Economics,
Cambridge: Cambridge University Press, pp. 2055.
Klein, Rudolf (1995), The New Politics of the NHS, London: Longman.
Klein, Rudolf and P. Day (1992), Constitutional and distributional conflict in
British medical politics: the case of general practice 19111991, Political
Studies, XL, 46278.
Labour Party (1995), Renewing the NHS: Labours Agenda for a Healthier
Britain, London: Labour Party.
Lakoff, George and Mark Johnson (2003), Metaphors We Live By, Chicago:
Chicago University Press.
Lawson, Nigel (1992), The View from No.11: Memoirs of a Tory Radical,
London: Transworld Publishers Ltd.
Lawton, V. (1999), Pharmaceuticals: from regulations to competition, Public
Money and Management, 19 (4), 46.
Levinas, Emmanuel (1994), Outside the Subject, Stanford, CA: Stanford
University Press.
Lewis, P. (1996), Metaphor and critical realism, Review of Social Economy,
54, 487506.
Lindner, J. (2003), Institutional stability and change: two sides of the same
coin, Journal of European Public Policy, 10 (6), 91235.
Lipsky, Michael (1980), Street-Level Bureaucracy: Dilemmas of the
Individual in Public Services, New York: Russell Sage Foundation.
Little, Daniel A. (1991), Varieties of Social Explanation: An Introduction to
the Philosophy of Social Science, Boulder, CO: Westview Press.
Luce, T.R.U. (1987), Thirty years of pharmaceutical price regulation:

138

The dynamics of public policy

developments in the NHS price regulation scheme since 1957, in G.

Teeling-Smith (ed.), Costs and Benefits of Pharmaceutical Research,
London: Office of Health Economics, pp. 1228.
McDonald, T.J. (ed.) (1996), The Historic Turn in the Human Sciences, Ann
Arbor, MI: University of Michigan Press.
Mahoney, James (2003), Strategies of causal assessment in comparative
historical studies, in James Mahoney and Dietrich Rueschmeyer (eds),
Comparative Historical Analysis in the Social Sciences, Cambridge:
Cambridge University Press, pp. 33772.
Marsh, D. and R.A.W. Rhodes (1992), Policy communities and issue
networks: beyond typology, in David Marsh and R.A.W. Rhodes (eds),
Policy Networks in British Government, Oxford: Clarendon Press, pp.
24968.
Martin, S. (1995), The Pharmaceutical Price Regulation Schemes: An
Economic Evaluation, York: University of York.
Maynard, A. and K. Bloor (1997), Regulating the pharmaceutical industry,
British Medical Journal, 315, 143840.
Maynard Smith, J. (1982), Evolution and the Theory of Games, Cambridge:
Cambridge University Press.
Mayr, Ernst (1988), Toward a New Philosophy of Biology, Cambridge, MA:
Harvard University Press.
Monar, J. (1994), Interinstitutional agreements: the phenomenon and its new
dynamics after Maastricht, Common Market Law Review, 31 (4), 693719.
Moyer, Wayne and Tim E. Josling (1990), Agricultural Policy Reform: Politics
and Process in the EC and USA, Hemel Hempstead: Harvester Wheatsheaf.
National Health Service Executive (1997), Changing the Internal Market EL
(97) 33, Leeds: DoH.
Nelson, Richard R. and Sidney G. Winter (1982), An Evolutionary Theory of
Economic Change, Cambridge, MA: Harvard University Press.
Nooteboom, Bart (1997), Path dependence of knowledge: implications for the
theory of the firm, in Lars Magnusson and Jan Ottosson (eds),
Evolutionary Economics and Path Dependence, Cheltenham, UK and
Lyme, USA: Edward Elgar, pp. 5779.
North, Douglass C. (1990), Institutions, Institutional Change and Economic
Performance, Cambridge: Cambridge University Press.
Oakshott, M. (1966), Historical continuity and causality analysis, in W.H.
Dray (ed.), Philosophical Analysis in History, New York: Harper and Row,
pp. 199202.
Office of Health Economics (2002), Compendium of Health Statistics, 14th
edn, London: OHE.
Office of Telecommunications (2004), accessed 28 September at
www.oftel.gov.uk.

References

139

Orr, K. (2005), Interpreting narratives of local government change under the

Conservatives and New Labour, British Journal of Politics and
International Relations, 7 (3), 37185.
Oster, G.F. and E.O. Wilson (1978), Caste and Ecology in the Social Insects,
Princeton, NJ: Princeton University Press.
Ostrom, Elinor (1999), Institutional rational choice: an assessment of the
institutional analysis and development framework, in Paul Sabatier (ed.),
Theories of the Policy Process, Boulder, CO: Westview, pp. 3572.
Parsons, Wayne (1995), Public Policy, Aldershot, UK and Brookfield, US:
Edward Elgar.
Pemberton, H. (2003), Lock-in or lock-out? Path dependency and British
pensions, Paper presented to the Political Science Association, Leicester,
April 2003.
Petchey, R. (1995), GP fundholding: weighing the evidence, Lancet, 346,
113942.
Pierson, Paul (1993), When effect becomes cause: policy feedback and
political change, World Politics, 45, 595628.
Pierson, Paul (2000a), Increasing returns, path dependence, and the study of
politics, American Political Science Review, 94 (2), 25167.
Pierson, Paul (2000b), Not just what but when: timing and sequencing in
political processes, Studies in American Political Development, 14 (1),
7292.
Pierson, Paul (2000c), The limits of design: explaining institutional origins
and change, Governance, 13 (4), 47499.
Pierson, Paul (2004), Politics in Time: History, Institutions, and Social
Analysis, Princeton, NJ: Princeton University Press.
Pierson, Paul (2005), The study of policy development, Journal of Policy
History, 17 (1), 3461.
Popper, Karl R. (1945), Open Society and Its Enemies, Vol. 1, Princeton, NJ:
Princeton University Press.
Popper, Karl R. (1972), Objective Knowledge: An Evolutionary Approach,
Oxford: The Clarendon Press.
Popper, Karl R. (1984), Evolutionary epistemology, in J.W. Pollard (ed.),
Evolutionary Theory: Paths into the Future, London: John Wiley & Sons
Ltd, pp. 23955.
Potts, Jason D. (2000), The New Evolutionary Microeconomics, Cheltenham,
UK and Northampton, MA, USA: Edward Elgar.
Powell, Robert and David A. Lake (1999), International relations: a strategic
choice approach, in Robert Powell and David A. Lake (eds), Strategic
Choice in International Relations, Princeton, NJ: Princeton University
Press, pp. 131.
Pritchard, D. (2002), Are economic decisions rational? Path dependence,

140

The dynamics of public policy

lock-in and hinge propositions, Reason in Practice, 2 (3), 3145.

Putnam, Hilary (1975), The meaning of meaning, in K. Gunderson (ed.),
Language, Mind, and Knowledge: Minnesota Studies in the Philosophy of
Science, Minneapolis, MN: University of Minnesota Press, pp. 13193.
Raadschelders, J.C.N. (1998), Evolution, institutional analysis and path
dependency: an administrativehistory perspective on fashionable
approaches and concepts, International Review of Administrative Sciences,
64 (4), 56582.
Raz, J. (1999) Explaining normativity: on rationality and the justification of
reason, Ratio, XII (4), 35479.
Reynolds, A. (1999), What is historicism?, International Studies in the
Philosophy of Science, 13 (3), 27587.
Roberts, Clayton (1996), The Logic of Historical Explanation, Philadelphia,
PA: Penn State University Press.
Robinson, Joan (1979), Collected Economic Papers of Joan Robinson,
Oxford: Blackwell.
Rorty, Richard (1982), Consequences of Pragmatism Essays 19721980,
Minneapolis, MN: University of Minnesota Press.
Rose, Richard (1976), The Dynamics of Public Policy, Beverly Hills, CA and
London: Sage Publications.
Rose, Richard and Phillip L. Davies (1994), Inheritance in Public Policy:
Change Without Choice in Britain, London: Yale University Press.
Roth, P.A. (1999), The full Hempel: a review essay of Clayton Roberts, The
Logic of Historical Explanation, History and Theory, 38, 24963.
Royal Commission on the NHS (1979), Report of the Royal Commission on
the National Health Service, London: HMSO.
Samuelson, Paul A. (1947), Foundations of Economic Analysis, Cambridge,
MA: Harvard University Press.
Schwartz, Herman (2004), Down the wrong path: path dependence,
increasing returns, and historical institutionalism, accessed September
2005 at www.people.virginia.edu/~hms2f/Path.pdf.
Searle, John R. (1995), The Construction of Social Reality, London: Penguin.
Searle, John R. (1999), The Mystery of Consciousness, New York: A New York
Review Book.
Searle, John R. (2001), Rationality in Action, Cambridge, MA: MIT Press.
Simon, Herbert (1957), Models of Man: Social and Rational, New York:
Wiley.
Simon, Herbert (1982), Models of Bounded Rationality. Behavioral
Economics and Business Organization, Cambridge, MA: MIT Press.
Skocpol, Theda (1992), Protecting Soldiers and Mothers: the Political Origins
of Social Policy in the United States, Cambridge, MA: Belknap Press of
Harvard.

References

141

Smith, Martin (1993), Pressure Power and Policy, London: Harvester

Wheatsheaf.
Smith, R. and P. Wilton (1998), General practice fundholding: progress to
date, British Journal of General Practice, 48, 12537.
Stacey, J. and Rittberger, B. (2003), Dynamics of formal and informal
institutional change in the EU, Journal of European Public Policy, 10 (6),
85883.
Stinchcome, Arthur L. (1968), Constructing Social Theories, London:
University of Chicago Press.
Stone, Deborah (2001) Policy Paradox and Political Reason, 3rd edn, New
York: W.W. Norton & Company.
Strasser, Daniel (1992), The Finances of Europe, 7th edn, Luxembourg: Office
for Official Publications of the European Communities.
Streeck, Wolfgang and Kathleen Thelen (2004), Introduction: institutional
change in advanced political economies, in Wolfgang Streeck and Kathleen
Thelen (eds), Beyond Continuity, Oxford: Oxford University Press, pp.
139.
Taylor, I. (2003), Policy on the hoof: the handling of the foot and mouth
disease outbreak in the UK, Policy and Politics, 31 (4), 53546.
Thacher, D. and M. Rein (2004) Managing value conflict in public policy,
Governance: An International Journal of Policy, Administration, and
Institutions, 17 (4), 45786.
Thatcher, Margaret (1993), The Downing Street Years, London: Harper
Collins.
Thelen, Kathleen (1999), Historical institutionalism in comparative politics,
Annual Review of Political Science, 2, 369404.
Thelen, Kathleen (2003), How institutions evolve, in James Mahoney and
Dietrich Rueschmeyer (eds), Comparative Historical Analysis in the Social
Sciences, Cambridge: Cambridge University Press, pp. 20841.
Van Parijs, Phillipe (1981), Evolutionary Explanation in the Social Sciences:
An Emerging Paradigm, Totowa, NJ: Rowman and Littlefield.
Wanna, John, Joanne Kelly and John Forster (2000), Managing Public
Expenditure in Australia, Sydney: Allen & Unwin.
Wallace, Helen and William Wallace (2000), Policy-Making in the European
Union, 4th edn, Oxford: Oxford University Press.
Walley, T., A. Earl-Slater, A. Haycox and A. Bagust (2000), An integrated
national pharmaceutical policy for the United Kingdom?, British Medical
Journal, 321, 15236.
Webster, Charles (1988), The Health Services Since the War: Vol I, Problems
of Health Care, the NHS Before 1957, London: HMSO.
Webster, Charles (1996), The Health Services Since the War: Vol II,
Government and Health Care, the British NHS, 19581979, London:

142

The dynamics of public policy

HMSO.
Webster, Charles (1998a), NHS Reorganisation: Learning from History,
London: Office of Health Economics.
Webster, Charles (1998b), The Politics of General Practice, in I. Louden, J.
Horder and C. Webster (eds), General Practice Under the NHS 19481997,
London: Clarendon Press.
Wendt, Alexander (1999), Social Theory of International Politics, Cambridge,
MA: Cambridge University Press.
Wildavsky, A. (1964), The Politics of the Budgetary Process, Boston, MA:
Little, Brown.
Williams, Alan (2004), What could be nicer than NICE?, Office of Health
Economics, Annual Lecture.
Williams, Bernard (2002), Truth and Truthfulness: An Essay in Genealogy,
Princeton, NJ: Princeton University Press.
Wilsford, D. (1994), Path dependency, or why history makes it difficult but
not impossible to reform health care services in a big way, Journal of
Public Policy, 14, 25183.
Wilson, Edward O. (1975), Sociobiology: A New Synthesis, Cambridge, MA:
Harvard University Press.
Wilson, Geoff A. and Olivia J. Wilson (2001), German Agriculture in
Transition: Society, Policies and Environment in a Changing Europe,
Basingstoke: Palgrave.
Wistow, Gerald (1992), The health service policy community, professionals
pre-eminent or under challenge?, in D. Marsh and R. Rhodes (eds), Policy
Networks in British Government, Oxford: Clarendon Press, pp. 21935.

Index
agent behaviour
EU member states 102
in institutions 3940, 44, 45, 49, 52,
667, 70, 71
path dependency 3940
in policy paradigms 64, 71
public policy 32, 567, 61, 648,
6970, 96, 102, 105, 1089, 120
representative agent models 67
Australia
budget-making process 7
Goods and Services tax 46
industrial relations reforms 267
bandwagon effects 45, 52, 68
Baumgartner, Frank and Bryan Jones 2,
10, 26, 44
Brussels Agreements 857
budget systems see under EU
change 56, 11, 12, 91
incremental 436
radical 436
reductionism 43, 45, 65
and stability 3740
see also evolution
Collingwood, R.G. 20, 66
Crouch, C. and H. Farrell 3940, 55
Darwin, Charles 43, 47, 48
decision-making
EU budget system 88
and path dependency 336
public policy 11, 13, 256, 29, 55,
567, 88
Denmark, and EU CAP 100
dynamic analysis
and agents actions 567, 648,
6970, 96, 102, 105, 1089, 120
definition of 13
and history in social sciences 18, 20,

21
interactionist view 910, 56, 57
methodological localism 40, 64, 67
narrative elements 60, 63
policy as choice 249, 55, 65, 68, 118
and punctuated change 44
in social sciences 3, 45
and state space 45, 29, 8081
see also individual case studies
Easton, David 8
economics
evolutionary see evolution
hysteresis 68
increasing returns processes 40
neoclassical 4041
path dependency 2930
representative agent models 67
thrift paradox 24
Elster, Jon 1819, 20, 23, 28, 47, 62, 68
EU
and CAP see EU CAP
and GATT 87, 93, 96, 978
and Single European Market 85, 91,
99
telecommunications policy 21
EU budget system
Balanced Budget Rule (BBR) 78, 79,
815, 87, 89, 93, 95
bankruptcy avoidance 79, 85
Brussels Agreements 857
Budgetary Discipline 85, 86
and CAP see EU CAP
Compulsory Expenditure (CE) 78, 82,
84, 87, 88, 923
crises 7980, 82, 83, 84, 85, 87, 95,
100, 102
decision-making procedures 88
development 834
domain constraints 78, 82, 86, 87, 88,
96, 97, 102
143

144

The dynamics of public policy

EU25 budget spending 99, 1023

evolutionary metaphor 7980, 82, 83
Financial Perspective 856, 87, 88,
96, 97
Fontainebleau Agreements 845
incomplete initial design 813, 85
increasing returns process 77, 78, 79,
80, 889
institutional change 77, 7980, 823
institutional layering 77, 80, 83, 85,
89
Inter-institutional Agreement (IIA) 86,
88
Luxembourg Treaty 78, 82
Non-Compulsory Expenditure (NCE)
78, 82, 86
and path dependency 389, 78, 79, 82,
89
regional spending 85
situated agency 79
spending ceiling 88
switching costs 78
temporal states 77, 80
trial and error development 87
UK net contributions rebate 845
VAT system 812, 84
EU CAP
arable set aside 96, 978, 99, 1012
beef production 87, 97, 99
Blue Box payments 97, 98
cereal production 87, 94, 101
Co-Responsibility Levy (CRL) 94, 95,
98, 101
and CoAM 82, 84, 85, 86, 87, 88, 95
Commissioners, inner circle of 956,
102
Common Customs Tariff (CCT) 812
Community Preference 96, 98, 99
compensation payments 97, 98
dairy industry 945, 101
demand growth suppression 92
and Denmark 100
dimensions to 91
direct payments 88, 91, 96, 97, 98, 99,
1012
environmental parameters 92
export subsidies 92, 93
Financial Discipline 89
and France 100, 102
and Germany 95, 100, 101

Guarantee Thresholds 94, 95, 99, 100,

101
import tariffs 92
incentive structure 91, 93
and international trade negotiations
92, 93, 96, 978, 102
intervention storage 834, 92, 93, 95
and Ireland 100, 102
land values 24
MacSharry reform 87
Maximum Guaranteed Quantity 94
member states role as agents 102
microfoundations 99101
modulation 98, 101
and Netherlands 100
oilseeds reform 978, 102
and path dependency 31, 9092, 95,
96, 99, 100, 1012
price support instruments 834, 868,
914, 967, 98, 99, 100, 102
quotas 945, 101
reform (1977) 92, 94
reform (1982) 94
reform (1984) 84, 93, 94, 956, 100,
102
reform (1986) 94, 95
reform (1988) 37, 867, 93, 94, 956,
102
reform (1992) 91, 968, 99, 100,
1012
reform (1999) 98
reform (2003) 98, 101
Rural Development policy 98, 99
sheep production 99
Single Farm Payment (SFP) 98, 101
and Spain 100
spending growth rate 85, 867, 88,
956, 97
Stabilizers reform 86, 94, 101
support principles 936
surpluses 92, 93
and Sweden 100
three pillars of 91, 96, 98, 99
trade-distorting elements 97
and UK 100, 101
USA, trade talks with 97
Variable Import Levy (VIL) 812, 83,
84
and WTO Doha Round 98
see also EU budget system

Index
EU Treaty of Rome 39, 78, 79, 81
Articles 38 to 43 91
Article 39 934, 99
Article 199 81
Article 203 82
evolution
artificial selection 51
bandwagon effects 45, 52, 68
biological 43, 44, 45, 47, 49, 50, 51,
523, 578
fitness 523
functional explanations 47, 4850
incrementalism 436, 55
and intentionality 5052
metaphor 545, 567, 7980, 82, 83,
105
metaphor in EU 7980, 82, 83
myopia 51
narrative explanation 5, 28, 29
natural selection 47, 49, 50, 523, 57
network effects 45
optimization models 53
perspectives 4258
in policy studies 548
as a process 436
random variation 51
in social sciences 4753
as theory 468
tipping points 45, 52, 68
see also path dependency; public
policy
firms
efficiency of 34
routines in 47
see also institutions
Fontainebleau Agreements 845
France, and EU CAP 100, 102
game theory 2, 25, 105
Prisoners Dilemma 1089, 111
GATT
Blue Box policies 97, 98
and EU 87, 93, 96, 978
Green Box policies 97
Germany, and EU CAP 95, 100, 101
Habermas, Jurgen 56
Hall, Peter A. 11, 12, 20, 30, 32, 36, 65,
67, 70, 105

145

Hay, Colin 36, 37, 41, 51, 56, 65

health policy see under UK
historical institutionalist (HI) school 20,
22, 30, 37, 38, 47
institutions
agents capacity 3940, 44, 45, 49, 52,
667, 70, 71
budget rules as 13, 30
change potential 55, 66, 71, 7780
collective 67, 71
conceptual maps 667, 69
constraint of 66, 77
described behaviour 67
EU budget system, change in 77,
7980, 823
EU Inter-institutional Agreement (IIA)
86, 88
formation of new 38, 39, 78, 80
historical institutionalist (HI) school
20, 22, 30, 37, 38, 47
incomplete 39, 85
inefficient 41
inherited legacies 80
innovation 39, 45, 78
Inter-institutional Agreements (IIA) 86
interdependence of 39, 8081
layering of 40, 55, 77, 80, 83, 85, 89,
121
lock-in 39
and path dependency 29, 30, 32, 34,
37, 80
and policy, conceptual distinction 13
and policy paradigms 55, 667, 71,
1056
strategic agency 45, 65
structure and agency in 65
and technological development 34
see also firms; public policy
interest groups 35, 72, 126
Ireland, and EU CAP 100, 102
Kay, Adrian 31, 35, 41, 83, 91, 124
Kemp, P. 31, 32, 35, 38, 41
Kingdon, John W. 44, 52
Kuhnian scientific paradigms 12, 36, 105
layering of institutions 40, 55, 77, 80,
83, 85, 89, 121
Lowi, Theodore 14

146

The dynamics of public policy

Luxembourg Treaty 78, 82

MacSharry reform, EU CAP 87
metaphors
gentlemans agreement in UK
pharmaceutical policy 121, 122,
123
public policy 23, 28, 30, 356, 545,
567, 61, 73, 79, 122
in social sciences 54
narratives, policy
ad hoc-ness avoidance 62, 71
causal relationships 19, 22, 23, 46
Dr Seuss-like explanatory principles
62, 71
elephant fable 623
evaluation of 714
and evolutionary metaphor 545,
567, 7980, 82, 105
ingredients of 45, 5963
and institutions see institutions
interpretivist policy 667
just so story avoidance 612
metaphors, use of 23, 28, 30
microanalysis of 6371
path dependency 28, 29, 3141
perspective scale 23
and policy memory 6871
relativist 723
salience judgement 61
situated agency 658, 79
strategic agency 63
supervenience 30, 46, 64
see also individual case studies; path
dependency; public policy
Netherlands, and EU CAP 100
North, Douglass 29, 34, 39, 77
OECD
Central Budget Agencies 70
neoclassical policy analysis in 70
pharmaceutical policy 118
path dependency
adaptive expectations 34, 55
agents capacity 3940
confluence of paths 61
and constituent changes 9091
decision-making over time 336

in economics 25, 2930, 77

EU budget system 389, 78, 79, 82,
89
EU CAP 31, 9092, 95, 96, 99, 100,
1012
and exogenous changes 3940, 9091,
1012
as increasing returns process 345, 39
inefficiency in 4041
and institutions 29, 30, 32, 34, 37, 80
learning costs 34
locked-in elements 38, 78
mechanisms 334, 35, 61, 65
microfoundations of 31, 79
multiple realizability 3031
network effects 34
normative aspects 4041, 667
on-path versus off-path change 39, 78
perspective granularity 38, 39
policy change and stability 11, 12,
3740, 9091
and policy narratives 28, 29, 3141
Polya urn model 37
scientific theories 36
sunk costs 34
temporal process 30, 32, 35
Pierce, C.S. 20, 47, 72
Pierson, Paul 2, 4, 13, 20, 32, 34, 35, 61
policy paradigms
agents in 64, 71
definitions of 1112, 36, 52, 70
institutionalization of 55, 667, 71,
1056
lack of, in UK GP fundholding
scheme 104, 1056, 11113, 115
and UK pharmaceutical policy 119
Popper, Karl R. 24, 47, 53, 54
Prisoners Dilemma game theory 1089,
111
public policy
accumulation 13, 119, 120
action and consequences 24, 25, 278,
38, 49, 50, 55, 11920
actors, elite 64, 96
administrative infrastructure
investment 35
agents actions 32, 567, 61, 648,
6970, 96, 102, 105, 1089, 120
change 11, 12, 3740, 9091
as choice 249, 55, 65, 68, 118

147

Index
classical view 89, 11
cluster concept 5960
combinational effects 40
as composite variable 21, 67, 90
contestation, continued 72
contingent conjunctures 278, 36, 60,
63, 65
cost-benefit analysis 67
crises 69, 7980, 82, 83, 84, 85, 87,
95, 100, 102
cycle 89
decision-making process 11, 13, 256,
29, 55, 567
definition of 2, 710
dynamic analysis see dynamic
analysis
and electoral competition 72
and evolution see evolution
failure 55, 105
folk theorems 112, 115
future constraints, anticipation of 52
generic stages model 5
government department operating
procedures 13, 30
government selfinterest 27
hinge propositions 1112
historical institutionalist (HI) school
20, 30, 37, 38, 47
history 23, 67, 21, 60, 72, 73, 120
ideal-type 5960
increasing returns processes 356
incrementalism 44
individual contracts 35
inefficient 41
and inertia 7
institutions see institutions
intention and action, distinction
between 257, 52, 55
interactionist view 910, 56, 57
and interest groups 35, 72
learning 55, 1056
memory 6871
meso-level concept 60, 64
metaphors 23, 28, 30, 356, 545,
567, 61, 73, 79, 122
methodological localism 64, 67
narratives see narratives, policy
neoclassical analysis 70
networks 13, 30, 35, 64, 108
paradigms see policy paradigms

patches 38, 119

path dependency see path dependency
performance measurement in 41
policy soup 356, 52
and political risk 26
preference changes 25, 55
punctuated change 44
rationality in 8, 245, 36, 44, 51,
557, 62, 65, 118, 119, 120,
121
selection mechanisms 52, 55
situated agency 11, 658
social welfare function 119
stability 11, 12, 3740, 1045
strategic actions 27, 32
subsystems 10, 11, 323, 38, 120
system 1012
temporal cycling 120, 121
and temporality 310, 25, 267, 52,
55, 57, 73
and uncertainty 2, 12, 27, 56
windows 52, 96
see also individual case studies
reductionism 43, 45, 65
Simon, Herbert 51, 56
social sciences
abduction 20, 72
biological analogies 478
cluster concepts 59
dynamics in 3, 45
evolutionary theory in 4753
functional explanations 1819, 20, 28,
30, 47, 4850
historical 1719, 20, 22
historical institutionalist (HI) school
20, 22, 30, 37, 38, 47
idiographic-nomothetic divide 1822,
23, 28
mechanisms, role of, and explanations
1819, 28, 30
metaphors in 54
Spain, and EU CAP 100
Stinchcombe, Arthur 34
Sweden, and EU CAP 100
temporality
and public policy 310, 25, 267, 52,
55, 57, 73

148

The dynamics of public policy

temporal cycling 23, 120, 121, 1235,

128
Thelen, Kathleen 2, 12, 37, 39, 47, 55,
78
tipping points 45, 52, 68
UK
and EU CAP 100, 101
EU contributions rebate 845
evidence-based policymaking (New
Labour) 51
foot and mouth policy 69
health policy see UK GP fundholding
scheme; UK health policy; UK
pharmaceutical policy
housing benefit reform 31, 32, 38
inflation 278
monetary policy, 1980s 278
national performance frameworks 116
Neighbourhood Renewal 51
New Deals 51
OFTEL 122
pension policy 31, 32, 38
policy-making improvements 8, 57
social housing 323
Sure Start 51
UK GP fundholding scheme
abolition of 114, 11517
and Audit Commission 11314
BMA representation, lack of 1078,
111, 112
budget allocation 107
and generic drugs 115
hospital services 107
in inner-city areas 114
introduction of 1068
Labour Party opposition to 107
lack of evidence to support 11415
management allowances 107, 113
non-fundholding practices 113
official and empirical studies of
11315
patient removal 113
policy paradigm, lack of 104, 1056,
11113, 115
Practice Budgets for General Medical
Practitioners 107
prescribing practices 107, 113,
11415, 118, 125
Royal Commission, lack of 107, 110,

111, 112
secondary care services 113
What the Doctor Ordered 114
Working for Patients 107
UK health policy 21, 23, 31, 38, 51, 104
BMA 1079, 110, 111, 112
community breakdown 11011, 112
Health Act (1999) 125, 127
Hospital Trusts 120
National Health Service Act 1946 108
National Institute for Clinical
Excellence (NICE) 116, 125,
126, 128
The New NHS: Modern, Dependable
11516
NHS funding 105, 11011, 116, 123
NHS internal market 106
NHS Plan (2000) 116
NHS policy paradigm 104
NHS reforms (1991) 105, 106, 114
NHS, two-tier 110
paradigm, breakdown of 1056
policy learning in healthcare 1056
politicization of 105, 1078
Primary Care Trusts 11516, 120
public demand for health care 105,
112
Royal Commissions 107, 109, 110,
111
trust in 109
UK pharmaceutical policy
annual financial return (AFR) 121,
122
and Audit Commission 124
brand name selling 125
branded products 124, 125, 126, 127
companies, support for 11819
demand-side policies 123, 125
DoH Pricing and Supply branch 120,
126
generic subscribing 115, 1234, 127,
128
and gentlemans agreement regulation
metaphor 121, 122, 123
and GP fundholding see UK GP
fundholding
interest groups 126
layering 121
limited list prescribing 110
marketing 121

Index
NHS drug procurement 11819
and NICE 125, 126, 128
overprescription of drugs 118
Pharmaceutical Price Regulation
Scheme (PPRS) 118, 120, 121,
122, 1234, 1257, 128
policy inheritance 120
policy instruments, increase in 121,
124
and policy paradigm 119
postcode rationing 126, 128
Prescription Analysis and Cost Data
(PACT) 123
price competition 124, 125
price regulation 126
profit targets 121, 122
public demand for drugs 126

149

R&D investment 119, 121, 1223,

124, 126
rationality in 118, 119, 120, 121
regulatory body, need for 122, 127
spillover effects 1278
subsystems 120
temporal cycling 23, 120, 121, 1235,
128
trade-off problem 38, 11920
and value choice 120
Voluntary Price Regulation Scheme
(VPRS) 119, 121
USA, health care policy 31
wallpaper bubble principle 27
WTO Doha Round 98