Understanding Public Policy: Theories and Issues 2nd edition Paul Cairney

Chapter 4 Bounded Rationality and the Psychology of

Policymaking

Key themes of this chapter

 ‘Bounded rationality’ describes the limits to gathering and processing
policy-relevant information.
 It suggests that ‘evidence based policymaking’ is a misleading ideal.
Policymakers need to find ways to ignore most evidence.
 They use cognitive shortcuts to gather enough evidence and make what
they believe to be ‘good enough’ choices.
 Classic debates examined if bounded rationality would, and should,
produce incrementalism in political systems.
 Modern theories focus more on ‘fast thinking’ and its potential to
produce radical policy change
 The Narrative Policy Framework (NPF) describes actors telling stories
to set the agenda and influence policy change
 Social Construction and Policy Design (SCPD) shows how
policymakers use emotion and social stereotypes to reward some
populations and punish others

Chapter 3 suggests that some actors are powerful because they can exploit the ways in
which other people think. Therefore, studies of power link strongly to studies in
psychology about the cognitive ‘biases’ on which all people rely. In a complex world
beyond our understanding, we all have to rely on efficient ways of thinking, ignoring
almost all information available to us, to allow us to make choices decisively (Gigerenzer,
2001). Cognitive shortcuts can be described provocatively as ‘rational’, when we seek
systematic methods to identify the best information, and ‘irrational’, when we rely on gut
instinct and emotion to act almost instantly (Cairney and Kwiatkowski, 2017). However,
the use of emotion and morality to take care of one’s loved ones without weighing up the
costs and benefits is perhaps the most reasonable action of all! A more relevant worry is
that cognitive shortcuts leave us vulnerable to error and manipulation. Policymakers are
no exception to this problem.

1
 To understand policymaking, policy studies compare the classic concept
‘comprehensive rationality’ with bounded rationality in the real world.
Comprehensive rationality is an ideal-type, or unrealistic simplification of reality that we
can use to explore what really happens. For
Comprehensive rationality – an
ideal type of decision making in example, it is not realistic to expect
which policymakers translate their policymakers to (a) possess the organizational
values and aims into consistent and cognitive capacity to gather and process
policy choices following a all information relevant to their decisions then
comprehensive study of all choices (b) make clear, consistent, and well-ranked
and their effects. choices. Rather, they face bounded rationality,
Bounded rationality – a more
in which their possession and grasp of
realistic model which identifies the
factors – such as uncertain aims and evidence, and their ability to make and
limited information – that undermine implement consistent policy choices, is
comprehensive rationality. limited. They have to rely on cognitive and
organisational shortcuts.
Even though it is an impossibility, comprehensive rationality also seems to be an ideal
to which many people aspire. Many advocates of ‘evidence based policymaking’ seem to
criticize the need for policymakers to ignore most information when they make decisions.
Or, they display a passionate objection to any ‘irrational’ means to do so (Cairney,
2016a)! More broadly, comprehensive rationality sums up a common sense view of how
democracies should operate: elected policymakers employ reason to identify problems
and clarify their aims, produce and process evidence comprehensively to assess the
societal costs and benefits of solutions, use this evidence to make and legitimize their
choices, and evaluate policy success after a comprehensive search for evidence (Cairney,
2019a).
This double-ideal is what makes policymaker psychology and the debate on
‘incrementalism’ (Box 4.2) so interesting. It prompts us to think about the effect of
bounded rationality on system-wide policymaking and what we should do about it. Classic
accounts described different ways to adjust to bounded rationality. Simon focuses on the
need for policymaking organisations to ‘satisfice’, combining relatively systematic
searches for information with efficient procedural shortcuts to reach an acceptable
proximity to the comprehensive ideal. However, when discussing the nature of
policymaking systems, Lindblom appears to promote the practical and normative value
of departing from the ideal of comprehensive rationality. Incrementalism suggests that a
good strategy for boundedly rational policymakers with limited policymaking resources
is to make a succession of incremental changes to public policy based on the lessons of
past decisions. In a pluralist political system, it makes sense to limit information searches
only to politically feasible options, learn from trial-and-error, and not depart radically
from decisions based on a previous consensus (compare with Chapter 3 on pluralism).
Modern policy theories add two profound insights. First, many theories link major
2
policy change to bounded rationality (or they identify inertia that is not caused by
pluralism). For example, punctuated equilibrium theory identifies incrementalism
punctuated by rapid and profound shifts of policy (Chapter 9). Second, theories identify
new normative issues that arise from the use and abuse of cognitive shortcuts. In this
chapter, the Narrative Policy Framework identifies the ability of actors to tell stories to
influence their audience. Further, Social Construction and Policy Design shows how
policymakers draw on their own biases, and exploit social stereotypes, to make quick
choices that reward some populations and punish others. Such responses to bounded
rationality produce manipulative politics and ‘degenerative’ political systems, rather than
consensus-driven incrementalism. Overall, modern theories suggest that advances in
science and information technology have not solved the problem of bounded rationality.
Instead, politics and power will always matter.

Comprehensive and bounded rationality

A focus on comprehensive rationality prompts us to question our assumptions about the
power of the ‘centre’ to cause policy change:
 Do individual actors at the ‘top’ have the ability to research and articulate a
series of consistent policy aims, then make sure that they are carried out?
 Can an organization act in the same manner as a ‘rational’ individual?

Comprehensive rationality suggests that elected policymakers seek to translate their

values into policy, aided by organizations which operate in a ‘logical, reasoned and
neutral way’ (John, 1998: 33). The model includes a series of assumptions:
1 Organizations can separate values (required by policymakers to identify
their aims) from facts (required by organizations to assess the best way to
achieve those aims).
2 Organizations and policymakers can produce consistent policy preferences,
and rank them, to help maximize societal gain (in the same way that an
individual ranks preferences to help ‘maximize utility’).
3 Policy is made in a linear fashion, as in a policy cycle. First, policy aims
are identified in terms of the values of the policymaker. Second, all means
to achieve those aims are identified. Finally, the best means are selected.
There are clear-cut stages to the process – such as between agenda-setting
(identifying aims), formulation (identifying choices and making
decisions) and implementation (carrying them out).
4 Analysis of the decision-making context is comprehensive. All relevant
factors and possibilities have been explored, and all theories regarding
how the policy process works have been considered (Simon, 1976;
Lindblom, 1959; Jordan and Richardson, 1987: 9–10; John, 1998: 33; Hill,
2005: 146).
3
Since this is an idealized version of the policy process, the assumptions are deliberately
unrealistic. Most approaches begin their analysis by identifying the limitations to
comprehensive rationality and the presence of bounded rationality. First, it is impossible
to separate facts from values in such an artificial way. Simon (1983: 8) argued that the
best way to demonstrate this point is to read the alleged ‘facts’ in Hitler’s Mein Kampf
(see also Brinkmann, 2008; Douglas, 2009; Etzioni, 1967: 386). More generally, the cause
of policy problems is always subject to interpretation and debate, and there is no objective
way to evaluate the success of a policy solution (McConnell, 2010).
Second, policymakers have multiple, and often unclear, objectives which are
difficult to rank in any meaningful way. Therefore, they tend to pursue a small number of
those aims which command their attention at any one time. Policy aims can also be
contradictory: choosing a policy to address aim A may mean undermining a policy to
address aim B, producing clear winners and losers from the policy process. These
problems are multiplied when extended to organizations. Governments contain a mass of
organizations pursuing policy aims relatively independently of each other, with little
regard to the idea of centralized and ranked preferences or the trade-offs in policy choices
(there is no ‘holistic’ or ‘joined-up government’).
Third, the policy process is not linear and it is difficult to separate the policy cycle
into discrete stages. The ‘garbage can’ model and ‘multiple streams’ analysis (Chapter
11) present the most significant departure from an assumption of linearity, suggesting that
three processes – problem definition, solution, choice – appear to act almost
independently, with the potential for a completely different order. Policymaking often
begins with solutions that ‘chase problems’.
Finally, policymakers are faced with incomplete knowledge of the policy
environment and the likely consequences of their solutions, and cognitive and time
constraints which limit their ability to consider and understand every possible solution.
Organizations do not have the capacity to consider every fact and solution; the cost of
research forces them to set priorities. The search for theories to explain policy problems
is limited by the values and emotions of policymakers (which predispose them to consider
only some solutions) and organizational rules based on past experiences.
Consequently, Simon (1976: xxviii) presents a more realistic view of policymakers
and policymaking organisations based on bounded rationality:
1 Individuals and organizations cannot ‘maximize’ their utility; instead, they
‘satisfice’, or seek ‘a course of action that is satisfactory or “good enough”’.
2 They have neither the ability nor the inclination to consider all facts. Instead,
they combine intense searches for information with efficient approximations to
focus on the factors considered to be most relevant and important.
Simon expressed the aim to, as far as possible, use ‘administrative science’ (or, in
Lasswell’s language, the ‘policy sciences’) to reduce the problem of bounded rationality
within organizations: training officials in policy analysis to open their mind to new
4
possibilities; fostering the development of specialization and expertise in information
processing; teaching officials the most appropriate informational shortcuts to make
organizations more effective; and supporting the long-term goals of an organization by
providing the right incentives (1976: 242–3; 1960). His term ‘satisfice’ sums up the need
to be pragmatic when facing limits to the cognitive power of individuals and coordinative
capacity of organisations (Hill, 2005: 146–7; Parsons, 1995: 273–84).

Bounded rationality is more important than ever

We might be tempted – wrongly - to think that bounded rationality is less of a problem
now than when first identified in the policy sciences (Cairney and Weible, 2017). Our
‘evidence-based age’, involving major advances in scientific research methods and
information technology, has given ‘a new breath of life’ to hopes for comprehensive
rationality and evidence-based policymaking (Botterill and Hindmoor, 2012: 367).
Yet, when we revisit the four conditions for comprehensive rationality, we see
that the accumulation of new evidence does not come close to solving the problem of
bounded rationality (Cairney, 2016: 14). It addresses the problem of gathering policy
relevant knowledge somewhat (point 4 above), but the most pressing policy issues are
characterized by ‘radically uncertainty’: their complexity makes them difficult to define
and manage, partly because so many aspects are interdependent, producing too many
possible outcomes to predict (Tuckett and Nicolic, 2017), Further, more evidence does
not help us adjudicate between unclear preferences (point 2) or simplify the policy process
in which they are considered (point 3). Indeed, Chapter 11 suggests that the garbage can
model is as relevant as ever.
Modern science remains value-laden (point 1) even when so many people employ
so many systematic methods to increase the ‘replicability’ of research and reduce the
reliance of evidence on individual scientists. The role of values is fundamental: anyone
engaging in research uses professional and personal values and beliefs to: decide which
research methods to gather evidence are the best; generate research questions, concepts
and measures; evaluate the impact and policy relevance of the results; decide which issues
are important problems; and, assess the relative weight of ‘the evidence’ on policy
effectiveness (Cairney, 2016a: 16; 2019a). We cannot simply focus on ‘what works’ to
solve a problem without considering how we used our values to identify a problem in the
first place (Botterill and Hindmoor 2012). It is also impossible in practice to separate two
choices: (a) how to gather the best evidence, and (b) to centralise or localise policymaking
(see Cairney, 2017b; Cairney and Oliver, 2017). Most importantly, as Chapter 3 suggests,
the assertion that ‘my knowledge claim is superior to yours’ symbolises one of the most
worrying exercises of power. We may decide to favour some forms of evidence over
others, but the choice is value-laden and political rather than objective and innocuous
(compare Douglas, 2009; Pielke, 2007; Jasanoff, 2008).
Further, our growing knowledge of psychology has reinforced the importance of

5
bounded rationality and challenged a lot of the mid-20th-century optimism about the
prospect of comprehensive searches for information: “People are ‘cognitive misers’
(Kam, 2005), using informational shortcuts and heuristics to gather just enough
information to make decisions” (Cairney and Kwiatkowski, 2017: 2; Cairney and Weible,
2017; Vis, 2019). Modern studies of psychology suggest that:

People use short cuts to gather enough information to make decisions quickly: the
‘rational’, by pursuing clear goals and prioritising certain kinds of information,
and the ‘irrational’, by drawing on emotions, gut feelings, values, beliefs, habits,
and the familiar, to make decisions quickly (Cairney and Kwiatkowski, 2017: 2).

All people – including policymakers – address bounded rationality by thinking ‘fast and
slow’: ‘System 1 operates automatically and quickly, with little or no effort and no sense
of voluntary control. System 2 allocates attention to the effortful mental activities that
demand it, including complex computations’ (Kahneman, 2012: 20). Haidt (2001: 818;
2007; 2012) distinguishes between the ‘intuitive system’ and ‘reasoning system’ to
explain ‘moral reasoning’ as a retrospective process. In theory, moral reasoning could
represent an archetype of ‘rational’ judgement to determine what is right or wrong in a
society. However, in practice, people often engage with policy issues using gut instinct
and ‘when faced with a social demand for a verbal justification, one becomes a lawyer
trying to build a case rather than a judge searching for the truth’ (2001: 814). Box 4.1
identifies how people limit their exposure to information before making choices.

Box 4.1 Cognitive biases and fast and frugal heuristics

Kahneman (2012) describes a series of cognitive shortcuts, many of which feature in
policy theories (Lewis, 2013: 7; Cairney and Kwiatkowski, 2017):
 Cognitive biases are at the heart of agenda setting, attention, and the competition
to resolve policy ambiguity: ‘framing effects’, based on emotional and moral
judgements, the ‘representativeness heuristic’, when people overestimate the
probability of events because of their vivid nature, and the ‘availability heuristic’
(or ‘processing fluency’), when people relate the size, frequency or probability of
a problem to how easy it is to remember or imagine (Chapters 9 and 11)
 ‘Prospect theory’ describes people valuing the losses they fear more than the
equivalent gains they might receive. It helps explain why advocacy coalitions
think that their competitors are more powerful than they are (Chapter 10)
 A ‘need for coherence’, to identify patterns and causal relationships, helps explain
the power of narratives with a simple hero and moral (Jones et al, 2014)
 Policymakers use exemplars of social groups to represent general experience and
describe why they reward and punish populations (Schneider et al, 2014)
 Status quo bias, the ‘sunk costs fallacy’, and ‘optimism bias’, in which unrealistic
6
 expectations about our plans working out well when we commit to them, all help
explain inertia and the endurance of institutions (Chapter 5)
 ‘Groupthink’ and other aspects of organizational psychology place crucial limits
on comprehensive searches for policy relevant information.
However, we are still scratching the surface of the links between psychology and
policymaking. For example, we know that ‘emotion and cognition are part of the same
internal mental process’ but have not considered how policymaking organisations might
use this insight (Cairney and Kwiatkowski, 2017: 3).
Nor do we agree on the positive or negative aspects of psychology. While it is
customary to describe cognitive ‘biases’ negatively, with reference to the comprehensive
ideal, Gigerenzer (2001: 37–38) provides a more positive description of ‘how actual
humans …make decisions, as opposed to heavenly beings being equipped with practically
unlimited time, knowledge, memory, and other unlimited resources’. ‘Fast and frugal
heuristics’ are the ‘computationally cheap’ methods people use to make choices. An
‘adaptive toolbox’ helps people ‘to experiment using trial and error, use emotions to limit
needless searches for new choices (such as considering the costs/benefits of keeping one’s
children), and make choices based on a small number of simple rules rather than trying
in vain to weigh all costs and benefits’ (Cairney and Kwiatkowski, 2017: 3; see also
Tuckett et al, 2015). One source of rules may be a core set of moral values in a political
system to which policymakers can refer. Such well-thought-out values are more difficult
to describe as ‘irrational’ compared to action driven by fire and fury. Or, we may find that
emotions are crucial to political mobilization, such as to maintain campaigns to challenge
injustice (Reger, 2004).

Further, the application of studies of psychology and cognition to policymaking systems

suggests that governments struggle to process policy-relevant information
proportionately (Koski and Workman, 2018). Policymakers set goals but ‘they are not
generally effective in judging the connections between [their] goals and the complex
reality they face’ (Jones and Thomas, 2017: 49). Individuals communicate their narrow
expertise within a system of which they have almost no knowledge (Sloman and
Fernbach, 2017). In any situation, ‘most members of the system are not paying attention
to most issues most of the time’ (Baumgartner, 2017: 72). This scarcity of attention helps
explain, for example, why policymakers ignore most issues in the absence of a ‘focusing
event’ (Shaffer, 2017), organizational searches for information miss key elements
routinely (Workman et al, 2017), organisations fail to respond to events or changing
circumstances proportionately (Epp, 2017), and the ‘centre’ will always lack coordinative
capacity and the ability to make policy from the top-down (Cairney et al, 2019a). This
description of policymaking must inform prescription: seeking comprehensive rationality
is like trying to fly unaided rather than designing and using a plane (Lindblom, 1964).

7
Incrementalism
We want to know how policymaking systems operate – and should operate – when
responding to bounded rationality. In that context, incrementalism is a key attempt to
consolidate empirical and normative aspects
Opportunity cost – the value of (Box 4.2). It describes the need to respond to
option B which is foregone when bounded rationality pragmatically, by (a)
we pursue option A. For example, gathering information efficiently (note the
the cost of a decision to study many opportunity cost of comprehensive research),
policy options is the foregone and (b) situating this activity within a real
chance to study fewer, more policymaking environment. When we focus on
realistic, options in more depth. comprehensive rationality and evidence-based
policymaking we should not forget that
decision-making is inherently political; it is about winners and losers. A policymaker has
to consider her values, the balance of power within the legislature, and the reaction to
policy change by interest groups. Policymaking is costly: it takes time and political will
to persuade political parties, vested interests and the public that major policy change is
appropriate (and to ensure that policy is implemented). It is also unpredictable:
policymakers often react to events or solve problems caused by previous polices rather
than devote the time to major new initiatives. Consequently, boundedly rational
policymakers are much more likely to introduce incremental policy changes – based on
learning from past experience and addressing the unintended consequences of previous
decisions – than introducing radically new policy initiatives.

Box 4.2. The meaning of incrementalism

The meaning of incrementalism is often unclear in three main ways. First, ‘incremental’
is often equated with small. However, Lindblom (1979: 517) argues that increments can
be large or small. The crucial distinction is between radical and non-radical change: does
it follow logically from existing policy or mark a significant departure? This argument
solves one problem but raises another: can we distinguish between radical and non-radical
change? Second, incrementalism can be a description of how policy is made or a
prescription of strategies to pursue (or a confusing mix of both). Third, incrementalism
can refer to analytical strategies, in which organizations try to overcome problems related
to comprehensive rationality:
 Simple incremental analysis. Analysis limited to a small number of policy choices
which diverge incrementally from the status quo; it is better to analyse those issues
comprehensively than seek comprehensive coverage of all issues.
 Disjointed incrementalism. The simplifying strategies used by organizations,
including simple incremental analysis, trial and error, parallel processing.

8
  Strategic analysis. Realistic policymaking strategies (including disjointed
incrementalism and ‘mixed scanning’) used as an alternative to the ‘futile attempt
at superhuman comprehensiveness’ (Lindblom, 1959: 88).
Or, it can relate to strategies to reach agreement. ‘Partisan mutual adjustment’ is a process
in which actors pursue their interests and respond to other actors doing the same (by
researching their positions, using persuasion, and seeking allies) (Lindblom, 1959: 85).

Policymakers do not begin by articulating their values, translating them into

policy aims in rank order, and seeking the best means to achieve them. Their willingness
to trade-off one aim for another only becomes clear when they make decisions. The
analysis is not comprehensive, considering all of the empirical and theoretical
implications (Lindblom, 1965). Organizations analyse the effects of incremental change
and ignore many important possible outcomes, alternative policies, theories and values.
Indeed, it is efficient for government organizations to spend most of their time focusing
on the effects of incremental departures from current policies. Taking the comprehensive
ideal seriously would ‘paralyse’ an organization by frustrating its staff and exhausting its
resources on many radical options that are rarely acceptable to major political parties
(Lindblom, 1964: 157; Etzioni, 1967: 386). It is also sensible for policy change to take
place through a series of steps, to reduce the chance of making ‘serious lasting mistakes’.
The effects of non-incremental decisions are unpredictable and difficult to solve (1959:
86).
Further, any test of ‘good’ policy shifts from the ability to satisfy a
comprehensively rational policymaker’s objectives, to whether or not it commands
agreement within the political system (Lindblom, 1959: 81). An incremental strategy
works in that context: if a previous policy commanded widespread respect then
policymakers recognise the analytical and political costs of a radical departure from it.
To change policy radically may be undemocratic. In a pluralistic system, existing policy
is based on a long-term process of negotiation, bargaining and adjustment. No single actor
commands the policy process as a whole (Lindblom’s comparator to the US was Soviet
command-and-control). Rather, multiple actors and agencies pursue their interests and
command their own powerful ‘watchdogs’ to anticipate or redress ‘damages done by
other agencies’ (1959: 85). It encourages ‘partisan mutual adjustment’ (Box 4.2).
These suggestions prompted one of the most significant post-war public policy
debates in the US (see Cairney, 2012: 100-4). Lindblom’s critics suggested that
incrementalism represents a more realistic model than comprehensive rationality, but has
less to offer as a policymaking ideal. First, it relies on pluralism or the assumption of an
equal balance of power to produce widely-supported policy. In unequal societies, non-
radical change preserves policies produced by elites (Dror, 1964; Etzioni, 1967: 387)
Lindblom (1977; 1979) accepts this problem in later work, but argues that ‘partisan
mutual adjustment’ may be a preferable response: ‘strong central authority can be – and

9
historically is, in case after case – an instrument for protecting historically inherited
inequalities’ (1979: 523). Indeed, the appearance of a comprehensively rational process
may be used to minimize public, parliamentary and pressure group attention to
inequalities. Putting power in the hands of the few does not guarantee that it will be used
wisely and in the spirit of benevolent neutrality that comprehensive rationality suggests.
Second, debates focused on how to manage more radical change for the long term
in ‘fundamental’ policy areas, by encouraging policymakers to plan their activities more
efficiently, clarify their values, and identify alternative policies and issues worthy of
comprehensive analysis (Dror, 1964: 156; Lindblom,1979). Lindblom’s proposal - ‘one
can make changes in the social structure as rapidly through a series of incremental steps
as through drastic – hence less frequent – alterations’ (1964: 157) – helped distinguish
incrementalism from inertia (Lindblom, 1979: 520). It also signaled a greater role for
central planning to ensure a consistent series of incremental steps towards a clear goal
(Etzioni, 1967: 387–8; Cairney, 2007a). However, it did not help us make a clear
distinction between incremental and radical change. Since the nature of policy problems
is open to interpretation and debate, there is no objective way to identify ‘fundamental’
decisions (Smith and May, 1993: 203; Hill, 2005: 151). The incremental/radical policy
change distinction seems obvious in theory, but where is the dividing line in practice?

Is incrementalism ‘universal’ and inevitable?

Incrementalism was based initially on a study of US politics, so how relevant is it to
policymaking in other countries? When making such comparisons, it is important to avoid
confusing the proposed advantages of incrementalism (the management of policy by
consensus and the minimization of unintended consequences) with forms of inertia made
more likely by veto points in particular political systems (Chapter 6). Hayes (2001)
suggests that incrementalist strategies are used in many political systems characterized
by: bargaining and compromise between actors who have different information, interests
and views; and, the need to build on past policies (Hayes, 2001:3). Indeed, the
identification of incrementalism as a routine ‘policy style’ – in majoritarian and
consensus democracies - was a key theme of Richardson’s (1982) Policy Styles in Western
Europe. Incrementalism transcends formal political structures because it results from
routine responses to bounded rationality (see chapter 5 on ‘network institutionalism’ and
chapter 9 on ‘policy communities’).
Consider, for example, how we might explain an incrementalist style in a system,
such as the UK, which enjoys a reputation for ‘majoritarian’ and ‘top down’ government
(Cairney, 2019b). Most policy decisions are beyond the reach of elected policymakers
such as UK ministers. The sheer size of government necessitates breaking policy down
into more manageable issues involving a smaller number of interested and knowledgeable
participants. Therefore, most public policy is conducted primarily through specialist
‘policy communities’ which process ‘technical’ issues at a level of government not

10
particularly visible to the public or Parliament, and with minimal ministerial involvement.
These arrangements exist because there is a logic to devolving decisions and consulting
with certain affected interests. Ministers rely on their officials for information and advice.
For specialist issues, those officials rely on specialist organizations. Organizations trade
information, advice and other resources (such as the ability to implement or ‘deliver’ a
large group membership) for access to, and influence within, government. Further, the
logic of this relationship holds regardless of the party of government, and we are unlikely
to witness radical policy change simply because there is a change of UK government
(Richardson and Jordan, 1979; Jordan and Richardson, 1982; Jordan and Maloney, 1997;
Jordan, 2005; Cairney, 2008; Jordan and Cairney, 2013).
When comparing political systems, commentators are often expressing unrealistic
expectations about how quickly policy can change, then complaining about unusually
high inertia compared to countries like the UK (Burns, 1963; Krauss and Pekkanen, 2004;
Fabrini and Gilbert, 2000: 28; Kitschelt and Streeck, 2003: 2). The implication is that a
majoritarian political system puts power in the hands of the few and gives them more
opportunity to pursue the comprehensive rationality ideal. Yet, as Chapters 5 and 6
describe, inertia is not caused solely by veto. Rather, the effect of decades of cumulative
policies is that newly elected policymakers inherit a huge government with massive
commitments (Rose, 1990; 1986; Rose and Davies, 1994). The size and scope of the state
is such that any ‘new’ policy is likely to be a minor revision of an old one, partly because
policy builds up its own clientele resistant to change (Hogwood and Peters, 1982; 1983;
Geva-May, 2004). Or, when a commitment to a policy has been established and resources
devoted to it, it becomes increasingly costly to choose a different path (Pierson, 2000a).
In other words, the abstract theme with which incrementalism engages - the lack
of comprehensive rationality and its impact on political systems - is universal (it applies
to all times and places). Further, incrementalism as a policy style can be found in many
systems, and the reasons for policy inertia are not restricted to systems with many veto
points. In general, all policymakers face bounded rationality and a policymaking
environment over which they have limited control. The differences associated with
individual political systems relate more to the specific ways in which actors make sense
of this problem. In some cases, policymakers appreciate the benefits of incrementalism;
in others, they seek quicker, more radical change.
However, modern theories suggest that incrementalism as a policy style is not
inevitable. Bounded rationality is universal, but it often helps to explain two forms of
activity that should not be confused with incrementalism.
1. Policy obstruction driven by fast thinking
Policy inertia often relates to the exercise of power to reinforce ‘fast’ responses to
bounded rationality. Actors rely on ‘quick gut, instinct, emotional, and moral choices’
and become ‘biased’ or ‘motivated’ reasoners devoted to the status quo despite evidence
that might support policy change (Cairney and Kwiatkowski, 2017: 3). For example,

11
Chapter 10 describes the ‘advocacy coalitions’ who romanticise their own cause,
demonise their opponents, and often try to obstruct policy change (Sabatier et al, 1987).
2. Incrementalism punctuated by sudden and profound policy change
Many theories highlight the links between bounded rationality and non-incremental
change. Punctuated equilibrium theory shows that political systems produce
‘hyperincremental’ and radical change. Boundedly rational policymakers have limited
resources (including time, knowledge and attention) and cannot deal with all policy
problems. So they ignore most and promote a few to the top of their agenda. This lack of
attention to most issues helps explain why most policies do not change dramatically,
while intense periods of attention to some issues may prompt new demands for change.
By ‘reframing’ issues, policy actors can draw the attention of policymakers to new ways
of seeing and solving problems. Successful persuasion prompts ‘positive feedback’, in
which policymakers pay disproportionate attention to the issue, and a ‘bandwagon effect’,
in which multiple actors in the policy process all suddenly pay attention to, and seek to
influence, the same issue (Chapter 9; Jones, 1999, 2003).
Multiple streams analysis (MSA) suggests that radical policy change happens
when there is a ‘window of opportunity’ for three processes: a new or reframed problem
gains attention, a feasible solution gains currency within policy networks, and
policymakers have the motive and opportunity to translate the idea into policy. MSA
predicts policy continuity and profound change (Chapter 11; Kingdon, 1984).
Studies of policy diffusion demonstrate that non-incremental change can happen
when one region emulates another on the assumption that it was successful. The policy
transfer literature suggests that some governments feel pressure to engage in non-
incremental policy change, often driven by anxiety that they are not keeping up with their
neighbours (Chapter 12; Berry and Berry, 2007; Dolowitz and Marsh, 2000).
In such cases of policy change, we can detect the role of cognitive misers driven
by fast thinking rather than comprehensive analysis. People can understand policy
problems in many different ways, but focus only on one, with the potential for their
attention to lurch dramatically following a disproportionate response to new information.
A key strategy is to encourage people to pay attention to policy problems and potential
solutions in a particular order (Box 4.3).

Box 4.3 Heresthetic

William Riker (1982; 1986) uses the term ‘heresthetic’ to describe the importance of a
particular kind of manipulation: ‘one can help produce a particular choice if one can
determine the context of, or order in which people make, choices’. In other words, we
may conclude from studies of psychology that we cannot change people’s minds when
they engage in fast thinking to entrench their positions. However, we can exploit the fact
that they possess many different – and often contradictory – preferences. If so, the aim is
to draw their full attention to one issue or preference at the expense of another. Additional

12
strategies may involve exploiting their cognitive miserliness to make sure that they do not
engage in a full enough analysis to compare their preferences systematically. For
example, one can present information in ways that increase the cognitive load and
dissuade people from analysis, such as by presenting abstract concepts or excessive detail
(Drew, 2019). Put simply, if you want to make something happen, it may be better to
influence the rules that people use to make decisions – such as the order in which we
consider each choice, or the ways in which we present limited choices – than change their
minds about their preferences. This strategy includes the restriction of options during ‘the
selection of … policies to which there are majority preferred alternatives’ (Riker, 1982;
Hindmoor, 2006a: 80–7; Ward, 2002: 66; McLean, 2002; Ward and Weale, 2010). Many
policy preferences are ‘intransitive’ or inconsistent: if A is preferred to B and B to C, A
is not necessarily preferred to C (Chapter 7). Therefore, the order of the vote is crucial.
This approach seems to contrast markedly with Aristotle’s logos, ethos, pathos
approach: there is minimal appeal to pathos to persuade, or ethos to follow rules on policy
debate. Instead, heresthetic is manipulation ‘concerned with the strategy value of
sentences’ (Riker, 1986: x). Similarly, it may contrast with modern trends in ‘evidence
based policy’ to communicate evidence as simply as possible. Herestheticians emphasise
the cognitive burden of wider choice, not the simplification of evidence based choices.

The Narrative Policy Framework

There is far more to bounded rationality than a technical response in which governments
seek ways to gather and process more evidence more effectively (Jones and Crow, 2017).
Rather, all people – including policymakers - ‘develop quick and easy emotional
renderings of the world that fit with who they think they are and what they know’ (Crow
and Jones, 2018: 217). If so, the shortcuts they use to limit their attention – and the small
number of issues to which they pay attention - really matter. Our focus on agenda setting
and ‘framing’ (Chapters 3 and 9) would be irrelevant if everyone had the willingness and
ability to understand and compare all possible frames of reference! Instead, their attention
to one way of seeing the world, at the expense of another, can prompt them to support
radically different policies than if they had full information. However, we can draw two
very different lessons from this starting point:
1. People may be highly vulnerable to manipulation if they rely on other actors to
help them understand the world.
2. People are resistant to a lot of manipulation because they do not pay attention to
or trust many sources of information.
Put more simply, the stories that we tell people can reinforce or compete with the stories
they tell to themselves (Tuckett and Nicolic, 2017). The same story can motivate some
audiences, if it chimes with their beliefs or pulls their heartstrings, but backfire in others,
if it grates with their view of the world. In that context, the Narrative Policy Framework
(NPF) identifies the strategies of actors to exploit other actors’ cognitive biases (described
13
in Box 4.1). Then, it tries to measure their success empirically (Box 4.4).
People use narratives to manipulate or persuade their audience to see a policy
problem and its solution in a particular way. Such strategies, and the context in which
they appear, can vary markedly in ‘rich and unique policy contexts’, but a narrative has
the same basic structure containing four key elements (Shanahan et al, 2018a: 176-7):
1. Setting. It relates to a specific policymaking context consisting of ‘policy
phenomena such as legal and constitutional parameters, geography,
evidence, economic conditions, norms’. The nature of some settings may
be taken for granted, while others are highly contested.
2. Characters. It contains at least one actor, such as a hero, villain,
beneficiary, or victim.
3. Plot. A plot establishes a timeline or story arc. Common arcs include
stories of: heroes going on a journey or facing and overcoming adversity,
often relating to villains causing trouble and preventing sympathetic
victims suffering tragedy.
4. Moral. A story’s simple take-home point describes the cause of, and hence
solution to, the policy problem. For example, we must protect the lives of
victims of mass shootings or the rights of gun owners to defend themselves
(Merry, 2018).

Box 4.4 Narrative versus frames

Framing and narrative are not the same. This point is confusing, because: its authors did
not always make the distinction (Shanahan et al, 2018a: 204), and both terms are
ambiguous and overlap greatly (Weible and Schlager, 2014: 242). In policy studies,
framing is the definition of a policy’s image; how issues are portrayed and categorized
(Chapter 9). Yet, ‘framing’ is also a looser metaphor to describe the ways in which we
understand, and use language selectively. Many disciplines describe the metaphor
differently, but in two key ways. First, akin to a photo. The photographer decides what
you see (where the lens points, how much to zoom or Photoshop). Framing describes our
reliance on someone else to describe reality. Second, akin to a building’s timber frame,
providing essential structure. It describes less visible assumptions (Chapter 3). Combined,
framing sounds like storytelling. We limit information, to influence how people
understand the world, the issues to which they pay attention, and how they understand
policy problems. The difference comes from the way in which the NPF: (1) describes
narrative, to (2) conduct - generally quantitative - empirical research to test the effects of
strategies (Shanahan et al, 2013; 2018b). To count as narrative, it must have a structure
akin to a literary narrative: setting, character, plot, moral. In contrast, framing analysis
could quantify the frequency of keywords rather than whole phrases (although narratives
can contain few words (Merry, 2016) and NPF studies can use framing analysis - Lawlor
and Crow, 2018).

14
From this starting point, we can identify different strategies and factors which cause the
relative success of a narrative (Shanahan et al, 2018a: 177-8; 2014: 70). Strategies to
foster support from groups or coalitions include to: describe policy issues as a private
matter or an issue in which government should intervene (see Chapter 3 on
Schattschneider), establish the cause of a problem by (for example) assigning blame to a
character, or romanticize your own coalition’s aim while demonizing that of others (see
Chapter 10).
In each case, such strategies are constrained but not determined by the firmly-
held beliefs of the narrator’s audience (Chapters 10 and 11). NPF empirical studies
suggest that combinations of the following techniques or factors seem to improve
narrative success: using an audience’s fundamental belief (e.g. in the value of the market)
to influence a more malleable belief (e.g. business leaders will solve climate change) (Box
4.3); generating admiration for a hero in favour of policy change or empathy for a victim
of policy failure/ inaction; helping an audience imagine a problem and relate it to their
own lives; exploiting trust in the narrator; and, making sure that ‘micro level’ strategies
(aimed at individuals) and ‘meso level’ strategies (aimed at coalitions) relate to ‘macro
level’ or ‘grand narratives’ which capture fundamental societal beliefs, such as in the
relative merits of capitalism/socialism, or freedom/security (2018a: 188-95; Jones, 2014).
The latter strategy is crucial. It suggests that you need to know your audiences to
know how to influence them, and that this influence is bounded by context. In some cases,
there are only brief ‘windows of opportunity’ in which a narrative will have the desired
impact (Chapter 11). This window can relate to the emotional state of the audience or
series of events in political systems (Cairney and Kwiatkowski, 2017: 5). In other cases,
narrative success relates to the attitudes of left and right wing populations. Some
audiences are relatively open to government action to solve problems, while others only
accept arguments which emphasize individual responsibility. For example, the
Frameworks Institute experimented with a personal story of a woman to highlight the
difficulties of eating healthily as a family when poor, only to find that one audience
engaged emotionally and sympathized, while another judged the victim harshly and
assumed they could do better (Cairney 2016b). Jones and Song (2014) also find that, in
climate change debates, the stories with the biggest short term impact – ‘egalitarian’ (it is
caused by overconsumption) and ‘individualist’ (it is exaggerated by idealists and the
self-interested) – affected the audiences predisposed to accept them (compare with
Fadlallah et al’s evaluation of narrative in health policy).
The potential impact of NPF insights is not undermined by the fact that some
findings seem obvious. There is a big difference between assuming that some stories
matter and measuring their effect over the short and long term (Box 4.5). Further, it may
not seem important that stories have most impact when telling people what they already
think, but it could make the difference between thought and action, such as when people

15
turn out to vote or pay attention to one problem at the expense of the rest. We may struggle
to persuade people to change their minds, but we can encourage them to act by prioritizing
their attention to one belief over another (Box 4.3; Jones and Crow, 2017: 3).

Box 4.5. NPF positivism versus interpretivist postpositivism

To count as NPF study, research should be ‘clear enough to be wrong’ (Jones and
McBeth, 2010; Jones et al, 2014: 3-4). This ‘positivist’ approach to science is driven by
empiricism, hypothesis testing, generalizing findings and replicating research (Sabatier,
2007b: 326–31; Eller and Krutz, 2009: 1; Cairney, 2013: 10). It has provoked debate with
‘postpositivist’ scholars of narrative (Jones and Radaelli, 2015; 2016; Lejano, 2015)
(compare with ‘critical’ scholarship in Chapter 3). NPF authors hope that many people
can conduct their empirical studies systematically to accumulate knowledge (Pierce et al,
2014), unlike some other theories which lack coherence and coordination (Cairney and
Jones, 2016: 50). It requires replicability: two different people could use the same method
and get the same results (Chapter 13). Its critics suggest that the policymaking context is
too complex to boil down into such a basic description of narrative, and that researchers
are subject to the same cognitive biases that the NPF ascribes to policy actors . Therefore,
‘researchers cannot escape the world they inhabit to make sense from some external
position’; they tell one of many possible truths in each study rather than accumulate
knowledge of the truth over time (Dodge, 2015: 362; compare with Jones, 2018).
Yet, NPF studies may not always provoke such intense debate or require such a
narrow approach. Gray and Jones’ (2016: 199) call for more qualitative NPF studies takes
us from ‘clear enough to be wrong’ to consistent enough to seem credible to a critical
reader, partly by reflecting continuously and openly about the application of NPF
concepts (Jones, in correspondence). NPF articles for practitioners encourage scientists
to focus more on their narrative and less on the alleged superiority of their knowledge
(Jones and Crow, 2017; Crow and Jones, 2017; Jones and Petersen, 2017). Further,
McBeth’s (2014) story is that the NPF’s architects began as postmodernists ground down
by the US positivist academic system. Its international expansion (McBeth et al, 2018:
198) changes things.

Social Construction and Policy Design

Social Construction and Policy Design (SCPD) identifies the profoundly ‘degenerative’
effect of narratives and other ways to exploit people’s beliefs to make policy.
Policymakers describe and use their value judgements to make fundamental choices about
which social groups should be treated positively or negatively by government bodies.
When addressing highly politicised issues, they seek to reward ‘good’ groups with
government support and punish ‘bad’ groups with sanctions (Schneider et al, 2014). In
psychology, such ‘moral reasoning’ (Haidt, 2001) is ‘motivated reasoning’: policymakers
make quick, biased, emotional judgements, then seek to back up their actions with
16
selective facts:

Likes and dislikes are not the result of individual or collective reason and
deliberation but mainly the product of emotion and heuristics … judgments begin
with emotional reactions … and reason is used mainly to justify initial emotion
responses (Schneider et al, 2014: 106).

Yet, social constructions can also be based on conscious bias and the strategic
exploitation of other people’s emotions and stereotypes for political gain. Policies reflect
the goal-driven use of constructions, ‘strategically manipulated for political gain … to
create political opportunities and avoid political risks’ or, at least, an anxiety by
politicians ‘not to be caught in opposition to prevailing values’ if it affects their
performance in election (Schneider and Ingram, 1997: 6; 192). They seek support from
the populations they describe as ‘deserving’, as well as a wider public satisfied with
describing others as ‘undeserving’ (1997: 6).
These judgements have an enduring ‘feed-forward’ effect (Ingram et al, 2007:
112). Choices based on values are reproduced in ‘policy designs’, as the ‘content or
substance of public policy’:

Policy designs are observable phenomena found in statutes, administrative

guidelines, court decrees, programs, and even the practices and procedures of
street level bureaucrats … [they] contain specific observable elements such as
target populations (the recipients of policy benefits or burdens), goals or
problems to be solved (the values to be distributed), rules (that guide or constrain
action), rationales (that explain or legitimate the policy), and assumptions (logical
connections that tie the other elements together) (Schneider and Ingram, 1997: 2).

Examples of positive feed-forward effects include policy designs that signal that ‘elderly
citizens are worthy of respect and deserving of the funds they receive’, and prompt ‘a
level of political participation rivaled by no other group’ (Schneider and Sidney, 2009:
110). Negative actions include signaling to welfare recipients that they have themselves
to blame and deserve minimal support; and, restricting voting rights directly or
introducing convoluted rules to diminish participation (2009: 111).
Policy designs based on moral choices often become routine and questioned rarely
in government because they are ‘automatic rather than thought through’ (Schneider et al,
2014: 124). Emotional assignments of ‘deservingness’ act as important ‘decision
heuristics’ because this process is ‘easy to use and recall and hard to change’ (2014: 124).
In this respect, SCPD builds on classic discussions of actors exercising power to reinforce
or challenge attitudes (Chapter 3; Pierce et al, 2014). Policy designs are difficult to
overcome, because a sequence of previous policies, based on a particular framing of target

17
populations, helps produce ‘hegemony’: the public, media and/ or policymakers take this
set of values for granted, as normal or natural, and rarely question them when engaging
in politics (Pierce et al, 2014). For example, if most people assume that people in poverty
deserve little government help, because they are largely responsible for their own fate,
policymakers have little incentive to intervene. In such cases, powerlessness relates to the
inability of disadvantaged groups to persuade the public, media and/ or government that
there is a reason to make policy or a problem to be solved. Or, people may take for granted
that criminals should be punished because they engage in deviant behaviour.
Since policy design sends a powerful signal to the recipients of benefits or
punishments, target populations participate more or less according to how they are
characterised by government (Schneider and Ingram, 1993: 334). To challenge policy
designs, groups often have to challenge fundamental public assumptions, reinforced by
government policy, regarding what constitutes normal and deviant behaviour. Yet, only
some have the resources to mobilise to influence the way they are described by
policymakers (Schneider and Ingram, 1997: 21-4; 2005b: 444; Pierce et al, 2014), or to
persuade the public, media and/ or government that there is a reason to act on their behalf.
Social constructions are – in theory – ‘inherently unstable’ but few groups are
categorized differently over time, at least in the absence of long term change in social
attitudes (Ingram and Schneider, 2005a: 10). For example, it can follow a major external
event such as an economic crisis or game-changing election, exploited by ‘moral
entrepreneurs’ to influence the way that policymakers view particular groups (Ingram and
Schneider, 2005: 10-11; Nicholson-Crotty and Meier, 2005). Or, it can be prompted by
policy designs which suit powerful populations but produce spillover effects for the
powerless. For example, a shift to drug treatment as an alternative to incarceration,
designed for white people, benefits people of colour unintentionally (Schneider and
Ingram, 2005a: 639).
Ingram et al (2007: 102) depict this dynamic with a table (4.1) in which there are
two spectrums. One describes the positive or negative ways in which groups are portrayed
by policymakers. The other describes the resources available to groups to challenge or
reinforce that image. There are four categories of target population: the advantaged are
treated positively and are powerful, so they receive benefits publicly. Deviants are their
polar opposite. Dependents are treated positively in public but do not have the power to
exploit their image, while contenders are treated negatively in public but powerful enough
to secure benefits behind the scenes.

Table 4.1 Social Construction and Power

Described positively Described negatively
High advantaged, treated contenders, treated
power positively in public and negatively in public but
receiving benefits publicly negotiating benefits
privately
18
 Low dependents, treated deviants, treated negatively
power positively in public but in public and punished by
unable to mobilise to policy
negotiate benefits
Source: adapted from Schneider and Ingram (1997)

Most issues are not politicised in this way, because people can only pay attention to a
small number of issues (Chapter 9). Yet, low salience can exacerbate problems of citizen
exclusion. Policies dominated by bureaucratic interests often alienate citizens receiving
services (Schneider and Ingram, 1997: 79). Or, the pursuit of ‘evidence based
policymaking’ (Box 2.2) helps experts dominate policies, and many government
agencies, when there is high scientific agreement and wide acceptance that the 'public
interest’ is served largely through the production and use of evidence. The process does
not include ordinary citizens routinely. Rather, ‘experts with scientific credentials aid and
abet the disappearance of the public sphere’ (1997: 153). Issues ‘with important social
value implications’ transform into ‘a matter of elite scientific and professional concern’,
such as when official calculations of economic activity override personal experiences
(1997: 153; 167).
Overall, SCPD describes a political system with major potential to diminish
democracy. Politicians politicize issues to reward or punish populations or depoliticize
issues with reference to science and objectivity. Policy designs are not informed by
routine citizen participation. Schneider and Ingram (1997: 3) argue that, although the US
political system may ‘meet some standard of fairness or openness’, the policies they
produce may not be ‘conducive to democracy’. They describe an increasingly
individualistic US system with declining rates of collective political participation (in
elections), a tendency for actors to seek benefits for their own populations, and
‘degenerative’ policy which produces major inequalities along sex, race, and ethnicity
lines (Ingram and Schneider, 2005: 22-6; Soss, 2005). Public policies have failed to solve
major problems – including inequality, poverty, crime, racism, sexism, and unequal
access to effective healthcare and education – and policy failure contributes to the sense
that the political process serves special interests at the expense of the general public
(Schneider and Ingram, 1997: 4-7). Policy designs ‘are strongly implicated in the current
crisis of democracy’ because they have failed, and they discourage many target
populations (the ‘undeserving’, ‘deviant’, or ‘demons’) from public participation:

These designs send messages, teach lessons, and allocate values that exacerbate
injustice, trivialize citizenship, fail to solve problems, and undermine institutional
cultures that might be more supportive of democratic designs (1997: 5-6; 192).

Is social construction and policy design ‘universal’ and inevitable?

SCPD began as a study of US politics, and there are few direct international comparisons,
19
but we can see that US-style ‘degenerative’ politics is often less apparent in (for example,
multi-party) systems with contrasting policy ‘styles’. The ‘majoritarian’ and ‘adversarial’
UK contains comparable processes. Hunter and Nixon’s (1999: 166) study of housing
policy identifies the same ‘advantaged’ status for homeowners as in the US. Media, court
judges, and politicians express sympathy for owner-occupiers affected by a system over
which they had minimal control, but describe disadvantaged social housing tenants as
feckless individuals. In contrast, in their study of poverty policies in two provinces in
Canada, Mondou and Montpetit (2010) argue that the worst excesses of degenerative
politics are associated with adversarial contexts. Consensual policy styles, associated with
proportional electoral systems and multi-party politics, produce less incentive to reward
advantaged and demonise deviant groups.
Such applications help us reflect on the ways in which abstract concepts have
‘universal’ relevance but produce very different applications. In the abstract, key SCPD
processes are likely to be found anywhere: policymakers combine cognition and emotion
to define highly salient problems and populations, their choices are reflected in policy
design, and policy design sends signals to target populations, only some of which can
mobilise to exploit or challenge their portrayal to seek government rewards. Further, the
tendency to process most other policies out of the public spotlight is a key feature of
political systems (Chapter 9). However, different political systems and policymaking
environments constrain and facilitate such behaviour, producing often different results.
Compared to the NPF and most other theories in this book, we may be relatively
likely to find SCPD-style applications, in studies of other countries, which do not cite the
foundational authors (Schneider and Ingram). For example, a large literature on the social
construction of ‘problem’ families in the UK has direct parallels with SCPD and its
depiction of degenerative US politics. Politicians have long blamed parents or an
‘underclass’ for family breakdown or dysfunction and its relationship to social problems
such as school truancy, anti-social behaviour, crime, and low employability (Crossley,
2017). This overlap of study, but not citation, partly reflects SCPD’s intellectual
connection to ‘critical’ scholarship (Chapter 3) which is relatively diverse and does not
require the same reference to a foundational text (compare with Chapter 11 on Kingdon).

Conclusion
Comprehensive rationality is an ideal type. The task is to identify its assumptions and
consider the implications when these conditions are not met. Policy theories build on the
concept of bounded rationality which highlights: the inability of organizations to separate
facts from values; unclear and conflicting political objectives; a non-linear decision-
making process; and, an incomplete search for knowledge combined with limited
resources, time and cognitive abilities. From this model we then explore what happens
(description) and what should happen (prescription).
For example, policymakers ‘satisfice’ and use informational and cognitive
20
shortcuts to make their task more manageable. While Simon uses this discussion to seek
improvements to policymaking, Lindblom extends the analysis to a political context in
which policymakers bargain with other actors and treat high levels of agreement in the
political system as an indicator of good policy. Lindblom advocates this approach because
it reduces the chances of governments making big mistakes that are relatively difficult to
reverse, and ensures that they do not use resources – including political will – unwisely.
The argument that incrementalism does and should happen produced a US debate
focused on unequal power and inertia. Yet, key themes are universal. The use of
incrementalist strategies to address the limits to comprehensive rationality is common to
many political systems. Most political systems face some need for bargaining and
compromise and to build on past policies. More generally, a focus on comprehensive
rationality helps us focus on the big technical questions of political science: if
comprehensive rationality cannot be achieved, what happens and what should happen?
Should we spread analytical resources in an attempt to cover most areas, or focus those
resources on the issues most likely to receive policymaker attention? How do we identify
the issues to which everyone should pay most attention versus those we leave to the
experts?
Modern theories still focus on bounded rationality but have taken forward two
profoundly important insights. First, when policymakers respond to bounded rationality
it produces a combination of incremental and radical policy change. Second, all people –
including policymakers and influencers – combine cognition and emotion to make
decisions. Both points undermine the idea that we can get closer to the comprehensive
ideal summed up in contemporary mythical discussions of ‘evidence based
policymaking’, or that policymaking is simply incremental.
Incrementalism may seem appropriate to actors who want to be able to reverse
mistakes more easily, establish ‘stable expectations in a complex and uncertain
environment’, and address the spirit of compromise, particularly in political systems
which have a formal separation of powers and ‘overlapping, conflicting and interacting
institutions’ (Jones and Baumgartner, 2005: 119). Non-radical policy change may also be
the norm if all policy actors must ignore almost all policy issues almost all of the time.
Yet, the desire by some for consensus-driven policy does not ensure that
governments will act that way. Emotionally or ideologically driven governments may still
make radical policy choices even if their ability to produce a consistent and well-
researched plan is out of reach (Hill, 2005: 152). Radical change may follow a window
of opportunity, disproportionate attention to one issue at the expense of others, or
international emulation and coercion to transfer policy, all of which suggest that bounded
rationality should not be equated with policymaking stability, policy continuity, or
incremental change. Further, the NPF suggests that, although people are most susceptible
to stories that reinforce their beliefs, an effective narrative can mobilize policymakers to
act to protect or challenge the status quo. More worryingly, the SCPD equates policy

21
inertia with a combination of emotional policymaking and profound inequalities of
power. Lindblom’s image of pluralistic and consensus-driven incrementalism compares
with Schneider and Ingram’s elitist and degenerative politics. If so, we can no longer fool
ourselves that policymakers routinely produce good policymaking when dealing with
bounded rationality.

22