7 Collective Action Problen1s in

Public Policy

Key themes of this chapter:

• Political science models and policy theories help us understand collective action
problems.
• Rational choice theory and the Institutional Analysis and Development(IAD)
framework help us identify why people may, or may not, cooperate with each
other.
Simple examples of game theory describe individual action undermining
•

collective action, such as when 'free-riding' or contributing to the 'tragedy of the

commons'.
• The IAD helps compare the effectiveness of institutions designed to foster
collective action.
• Most famously, it identifies the conditions under which 'common pool resource'
tragedies can be averted by communal rather than market or state action.
• A key theme is institutional complexityand the overlaps between many different
rules. 'Polycentric' governance describes many decision-making centres, respon-
sible for their own action, and cooperating or competing with others.
• These approaches provide important ways of thinking about policy problems
rather than universal solutions or explanations.

Political systems often represent the main venue for collective action. For example, gov-
ernment often involves coercion, through policy tools such as regulation and taxation
(Chapter 2). Or, governmentsuse 'behavioural public policy' which incorporates psycho-
logical insights to 'nudge' people into changing their behaviour (Box 7.1). In this chapter,
we take a step back from such specific measures, to consider more generally why people act

individually and collectively. Why do people agree, or refuse, to cooperate with each other
to produce rules to solve policy problems? What are the consequences for policymaking?
Should market, state, or communal practices drive collective action?
In that context, rational choice theory and game theory provide important ways of
thinking. Through simple models and thought experiments,they help identify collective
action problems: the potential for choices made by individuals to have an adverse societal
effect when there is an absence of trust, obligation, or other incentives to cooperate.
People may have collective aims or interests that require cooperation but individual incen-
tives to defect. While the action of one individual makes little difference, the sum total of

110
 Collective Action Problems in Public Policy 111

individual actions may be catastrophic. It may cause (1) 'market

Game theory the use -

failure' when there is scope to free-ride and no incentive to of simple mathematical

purchase or contribute to public goods (such as constructing models to compare how
national defences or maintaining clean air) and (2) the 'trag- individuals interact
edy of the commons' when individual actions combine to ruin with each other when
common pool resources (CPRs) such as land, forests, and fish faced with different
stocks (Hardin, 1968; Heikkila and Carter, 2017). incentives in different
Government intervention is one solution to collective settings.
action problems: if the effect of non-cooperation is that Free ride to enjoy
-

everyone is worse off, let's make people cooperate. Yet such the benefits of a collec-
solutions are also problematic for several reasons. First, tive resource without
policymakers may act in their own interests rather than the paying; to benefit from
collective interest described by governments.Second, no gov- collective behaviour
ernment is able to make decisions that suit everyone. Rather, without engaging in
governments coerce everyone (through taxes) to contribute that behaviour.
to policies that benefit only some, and regulate one group to
Public goods -

benefit another (Wilson, 1980: 419-22). Third, government

collective resources
solutions may be expensive and produce unintended conse- of which (1) no one
quences. Fourth, policymaking systems contain many venues can be excluded from
with policymakingauthority 'polycentric' or 'multi-centric'
enjoying their benefits
-

governance
-
which also need to find ways to cooperate with ('non-excludable') and
each other (Cairney et al., 2019a). Overall, if we need rules to (2) their use by one
encourage individual citizens to cooperate, we also need rules to actor does not diminish
encourage policymakers. their value to another
These issues are taken forward by modern policy theories ('non-rival').
that describe a combination of important factors: (1) bounded
Common pool
rationality (Chapter 4); (2) trust and norms; (3) institutional resources
solutions to collective action problems; (4) many layers of (CPRs) collective
-

rules to encourage cooperation among many actors and resources of which (1)
'centres' of government; and (5) the complexity and unin- it is difficult to exclude
tended consequences of such activities. Most importantly, actors from enjoying
the Institutional Analysis and Development(IAD) framework their benefits and (2)
their use by one does
provides a language to study these dynamics. It focuses on
institutions (Chapter 5) as vehicles for problem-solvingand (usually) diminish their
value to another.
collective action. Actors develop and modify rules to aid col-
lective action. These rules can be primarily state, market, or
communal but are often a complicated mix of all three. Indeed, one aim is to compare
different types of institutions and evaluate their relative success, using criteria such as
economic efficiency and equity.
Elinor Ostrom won the Nobel Prize in economics for her applicationof the IAD to the
study of CPRs, in which she often found that communal arrangements were more success-
ful than purely state or market solutions. However, these conclusions come with a mas-
sive dose of caution against drawing too many simple conclusions from many different
and complex contexts. A huge number of studies of CPR management suggest that certain
'design principles' can help produce effective communal collective action but do not pro-
vide a blueprint. The IAD also contains a 'family of theories' that can be applied to related
112 Understanding Public Policy

Box 7.1 Nudge theory and behavioural public policy

'Nudge' sums up the idea that we can take insights from psychology (Chapter 4)
to influence human behaviour. Rather than simply coercing people to change their
own behaviour, or cooperate with others, we can shape the environment in which
they make choices and reduce the cognitive burden of the choice we would like
people to make (Thaler and Sunstein, 2008; compare with Box 4.3). It has very
-

quickly become a 'mainstream' idea in economics and an essential 'policy tool'

-

(Chapter 2) for many governments (John, 2018: 49, 6-7). The superficial attraction
is clear: many politicians like the idea of policy change without direct imposition

(coercion can be electorally damaging), and citizens often 'resist such commands
and find ways to avoid regulations'or 'resist messages that come from authority'
(2018: 8, 29; compare with Jones et al., 2013). Consequently, many governments
established their own 'nudge units' and despite some concerns
-
about the ethics or
effectiveness of many interventions -
there has been a major increase in the use of
'behavioural public policy' (John, 2018: 82). Still, there remains uncertainty about
how to use it (2018: 145-6; compare with Strassheim, 2019). For some, it provides
subtle manipulations by reducing the cognitive load of certain options. For others,
it encourages the conditions under which people can process information efficiently,
and deliberate together, to allow them to think about and co-produce choices.

topics, such as the sustainability of Social-Ecological Systems (SES Framework), how to

deal with 'polycentric governance' (Institutional Collective Action), and how to study and
manage institutional complexity(Ecology of Games).
Overall, these approaches prompt the big normative questions in public policy: is there
a common good, and how should people cooperate to achieve it? Domestically, we might
ask: should governmentscoerce citizens to pay for public goods, do governments perform
better than markets, and do we need big or small, centralized or polycentric, government?
However, the solutions to many problems -
such as climate change and global public
health are not in the gift of single countries. They take the possibility of central coercion
-

out of the equation, prompting us to wonder if citizens and governments across the globe
can solve collective action problems togetheror are doomed to catastrophic failure. These

approaches also highlight an intellectual trade-off between (1) the search for parsimonious
explanationsof many cases and (2) the acceptance that the world is complicated and each
case is unique. Can we find a happy medium between too-simpleand too-complex analysis
of collective action problems?

COLLECTIVE ACTION PROBLEMS IN RATIONAL CHOICE AND GAME

THEORY

Rational choice theory describes models and approaches which combine insights from
mathematics, economics, and political science (Hindmoor and Taylor, 2015). 'Rational'
describes the ability of individuals to apply reason while processing information. They

À
 Collective Action Problems in Public Policy 113

identify and prioritize their preferences, recognize the constraints of the environment
in which they operate, and make choices based on their beliefs about the best way to sat-
isfy their preferences. Many models are deliberately abstract and simplified to the point
of being unrealistic. Some assume 'perfect' rationality to predict the effect of different
rules or environments on collective action. Therefore, they differ from our discussion of
'comprehensive rationality' (Chapter 4) in which we automaticallytreat these assumptions
as false and examine the consequences of bounded rationality (Cairney, 2012a: 133-35).
Classic rational choice theory is based on three principles:

l. Deductive reasoning. Create models based on a small number of propositions ('axioms'

or 'assumptions')and examine their logical implications.
2. Methodological individualism. Explain political outcomes as the aggregate of the
intentional actions of individual actors.
3. Instrumental rationality. Individuals pursue the best means to fulfil their preferences
(which can include direct self-interest but also protecting their families or giving to
charity) (Shepsle and Bonchek, 1997: 16).

There are two main variants (Dowding and King, 1995: l;

Utility the satisfac-
-

Hindmoor, 2010; Hay, 2004a: SO). The first is an abstract tion gained from fulfill-
exercise to predict behaviour when individuals act 'optimally'
ing one's preferences
in different environments. For example, to maximize their (e.g. consuming a good
utility, actors would need to process all relevant informa- or service).
tion and be able to rank-order their preferences consistently
(Hindmoor, 2006a: 182; Griggs, 2007: 174; Dowding, 2010). The second involves more
detailed assumptionson the preferences of individuals and how they relate to institutions.
For example, much of 'institutional rational choice' (Chapter 5) is built on the premise that
(1) each decision-making situation provides particular incentives for individuals to act
and (2) institutions are the formal and informal rules that can be used to change those
incentives. Note the trade-offs between simplified models that produce potential generali-
zations about many instances and more accurate and detailed explanationsthat are more
difficult to apply generally (Dowding and King, 1995: 16; Dowding, 1995a: 49; Dowding,
1995c: 82).

The Prisoner's Dilemma, Logic of Collective Action, and Tragedy of the

Commons

Game theory examines the choices that actors make when situated within a strategic
decision-making environment, faced with a particular set of 'pay-offs' and anticipating
the choices of other actors. In the simplestgames, the assumptionsare that all 'players'
are instrumentally rational (focused on the means to a desired end), each player knows
that the others are instrumentally rational, everyone understands the rules of the game
and the pay-off from each choice, and that everyone would make the same choice in the
same circumstances (Hindmoor, 2006a: 106-7). Other games introduce greater levels of

uncertainty or different assumptions about the motivations of actors (Harsanyi, 1986:

90). A key aim is to identify points of equilibrium when actors make a choice and stick to
it, such as the 'Nash equilibrium' when players have made their best choice and there is
 l

114 Understanding Public Policy

Table 7.1 The prisoner's dilemma

Linda
Stay silent Confess

Stay silent P-1, L-1 P-10, LO

Paul
Confess P-0, L-10 P-8, L-8

Note: the confess-confess option represents the Nash equilibrium. The

numbers represent years lost to jail.

no incentive to change (Chwaszcza, 2008: 145). Three famous examples follow, which you
should treat as thought experimentsrather than actual predictionsof behaviour.
First, the 'prisoner's dilemma' (Laver, 1997: 45-46). Two people are caught red-handed
and arrested for a minor crime, placed in separate rooms, and invited to confess to a major
crime (they both did it and the police know it but cannot prove it). The pay-offs (Table 7.1)
are: if Paul confesses and Linda does not, then Paul walks free and Linda receives a ten-

year jail sentence (or vice versa); if both confess they receive a much higher sentence (eight
years) than if neither confesses (one year). Also assume that they derive no utility from
the shorter sentence of the other person (a non-cooperativegame). It demonstrates a collec-
tive action problem:althoughthe best outcome for the group requires that neither confess
(both would go to jail for one year), the actual outcome is that both confess (and spend
eight years each in jail). This point represents the Nash equilibrium since neither would
be better off by changing their strategy unilaterally(Linda assumes that if she stays silent,
and Paul confesses, Linda gets suckered into ten years). The effect of Paul and Linda acting
rationally as individuals is that they are worse off than if they had cooperated. Both 'defect'
when they should 'cooperate'.
The structure, pay-offs, and outcomes differ in other games. In the 'chicken game', in
which we imagine two people driving towards each other, there are greater costs when
both defect (death) and more likelihood that at least one will cooperate (by getting out
of the other's way). In many 'assurance games', there is a greater benefit to unilateral
cooperationand/ or a greater expectation by each player that the other will also cooperate
(McLean, 1987: 127; Ostrom, 1990: 42; Dowding and King, 1995: 8; John, 1998: 120-21;
Hindmoor, 2006a: 109-11). In other words, individuals do not simply cooperate; the situa-
tion must provide an incentive and pay-off.
Second, the 'logic of collective action'. Olson (1965; 1971) argues that, as the member-
ship of an interest group rises, so does (1) the belief among individuals that their contri-
bution to the group would make little difference and (2) their ability to free-ride. I may
applaudthe actions of a group but can enjoy the outcomes without leavingmy sofa, pay-
ing them, or worrying that they will fail without me or punish me for not getting involved
(Chwaszcza, 2008: 156-57). Therefore, 'unless the number of individuals is quite small,
...

or unless there is coercion or some other special device to make individuals act in their
common interest, rational, self-interested individuals will not act to achieve their common or
group interests' (Olson, 1971: 2; Hindmoor and Taylor, 2015: ch. 6).
Third, the 'tragedy of the commons' (Hardin, 1968). The scenario is that a group of
farmers share a piece of land which can only support so many cattle before deteriorat-
ing and becoming useless. Although each farmer recognizes the collective benefit to an
 Collective Action Problems in Public Policy 115

overall maximum number of cattle, each calculates that the marginalbenefit they derive
from one extra cow for themselves exceeds the marginalcost of overgrazing to the group.
Individuals place more value on the resources they can extract for themselves now than
the additional rewards they could all extract in the future. The tragedy is that, if all farm-
ers act on the same calculation, then the common resource will be destroyed (1968: 1244).
As described, the group is too large to track individual behaviour, individuals have 'high
discount rates' placing more value on current over future consumption and low mutual
- -

trust, with minimal motive and opportunity to produce and enforce binding agreements
(Ostrom, 1990: 183).
This 'tragedy' appears to reflect current anxieties about one of the big problems of our
time: global CPRs are scarce and the world's consumptionlevels are rising; there is no
magic solution; and collective action is necessary but not guaranteed. We may value sus-
tainable water, air, energy, forests, crops, and fishing stocks but find it difficult to imagine
how our small contribution to consumptionwill make much difference. As a group, we
may fear climate change and seek to change our ways but, as individuals, contribute to
the problem (Sandler, 1997; 2004; compare with profound critiques of Hardin -Aklin and
Mildenberge, 2018; Hale, 2018; Brinkley, 2019).
Overall, these simple scenarios suggest that instrumentally rational individuals have
weak incentives to cooperate even if it is in their interests and they agree to do so. Hardin
(1968: 1247) argued that collective action is unlikely without 'mutual coercion, mutually
agreed upon', amplifying a major and ongoing debate about the state's responsibility for
that coercion (Ostrom, 1990: 2; Boyd et al., 2018; Frischmann, 2018). These problems
appear to undermine the 'idea that cooperation naturally follows from mutual interests'
(John, 1998: 122), althoughthey are exaggerated by certain assumptions, such as that the
games are not repeated (Box 7.2).

Box 7 .2 Collective action during repeated games

The nature of collective action problems changes if we modify game theory assump-
tions. Games like the prisoner's dilemma are one-off and self-contained, but in many
games the players know that there are wider, longer-term consequences to defec-
tion (Ostrom, 2011: 14). Describing 'nested games', Tsebelis (1990) posits that the
behaviour of individuals often seems suboptimal in one game until we recognize
their involvement in a series of others. It may be optimal to act 'irrationally' in the
short term to support a longer-term strategy. These games play out differently if
they are connected to other games with different rules and pay-offs, and/or the par-
ticipants know they will be repeated. Long (1958: 251-53) describes an 'ecology of
games', in which:
•
many games take place at roughly the same time in local communities,
includingbuying and selling, banking, producing media, and 'civic organization'
(1958: 253)
'players in each game' know and follow their distinct rules, but also 'make use of
•

players in the others for their particular purpose' (1958: 253)

116 Understanding Public Policy

• these games overlap and interact to produce 'unintended but systemically

functional results for the ecology' (1958: 251)
key players foster 'a vague set of commonly shared values that promotes
•

co-operation in the system' (1958: 251).

Axelrod (1984; 1986: 1087) uses evolutionarygame theory to explore how behav-
iour changes over multiple games to reflect factors such as (1) feedback and learn-
ing from trial and error and (2) norms and norm enforcement (see Chapter 11 and
Box 7.3). For example, player 2 may pursue a tit-for-tat strategy. She cooperates at
first, then mimics the other player's previous choice: defecting, to punish the other
player's defection, or cooperating if the other player cooperated. Knowledge of this
strategy could provide player 1 with the incentive to cooperate. Axelrod's simula-
tions suggest that norms develop when players enforce and expect sanctions for
non-cooperation, and through socialization to discourage norm violation, while
some norms become laws. Ostrom (2011: 15) extends the analysis to real-world
settings, in which repeated face-to-face meetings affect how people make individual
calculations.

GOVERNMENT AS ONE INSTITUTIONAL SOLUTION TO COLLECTIVE

ACTION PROBLEMS

Collective action problems underpin a convincing argument for the role of 'institutions'.
However, institutions come in many forms and encourage many types of behaviour,
from the informal enforcement of norms to the governments with formal authority to
tax, redistribute income, and spend on public services (Dowding and King, 1995: 10).
Government can be a key solution to many collective action problems but is also problem-
atic for four reasons.
First, similar collective action problems relate to policymakers, who may act as indi-
viduals pursuing their own preferences, often at the expense of wider collective action
(Buchanan, 1988: 13; Ostrom, V. and Ostrom, E., 1971). Classic examples include:

• Bureaucrats more interested in themselves than policy outcomes (Niskanen, 1971;

Dunleavy, 1985, 2019).
•
'Rent-seeking', which relates to the 'investment of resources by firms and pressure
groups in the expectationof securing economic privileges' from politicians (Hindmoor,
2006b: 87).
• 'Political business cycles', or the (not necessarily strong) incentive of incumbent
governments to manipulatethe economic cycle to make it boom before an election
(Schultz, 1995: 81; Hindmoor,2006a: 46-47; Dubois, 2016).

Second, governments may coerce everyone to contribute to government resources (through

taxation) but cannot make choices that suit everyone. There is no such thing as a 'pub-
lic will' or 'national interest' that we can determine by merely adding up the opinions or
preferences of the population. There are always winners and losers when policymakers
 Collective Action Problems in Public Policy 117

make choices. One way to understand this point is via the idea
Intransitive -
when
that no voting system can aggregate individual preferences to
preferences are not
satisfy the collective will unproblematically(McLean, 1987: 10, consistent. If A is
26). Instead, there is continuous potential for instability related preferred to B and B to
to intransitive electoral results and policy decisions, or 'the C, A is not necessarily
selection of candidates or policies to which there are majority preferred to C.
preferred alternatives' (Riker, 1982; Hindmoor, 2006a: 80-87;
Pareto efficiency/
Ward, 2002: 66; Ward and Weale, 2010). The famous 'Arrow
optimality- a point at
problem' (or 'impossibility theorem') describes the impossibil- which no one is better
ity of action to make everyone better off. Arrow (1963) seeks off without someone
to establish a rule, to aggregate individual preferences into
being worse off.
social preferences, which can satisfy certain criteria-
achieving
Pareto efficiency without Kaldor-Hicks compensation, Kaldor-Hicks
with choice determined by all individual preferences rather compensation -

A gains more than B

than dictatorship -
and finds that no such rule exists
loses, and compensates
(Hindmoor and Taylor, 2015: ch. 5; Dewan et al., 2009: xxi).
B, so that the outcome
Third, government (or state intervention) may be expen- is Pareto efficient.
sive and produce unintended consequences, particularly
when policymakers lack enough information to make good choices (Ostrom, 1990: 11).
Consequently, we should keep an open mind about the relative benefits of many options
to solve collective action problems. Solutions may be achieved by 'communal property
arrangements'as well as state intervention (Feeny et al., 1990: 13). Fourth, policymaking
tends to be 'polycentric', producing the need for many policymakingorganizationsto solve
collective action problems across government (Cairney et al., 2019a). Overall, if we need
rules to encourage individual citizens to cooperate, we also need rules to encourage policy-
makers and governmentalbodies.

INSTITUTIONAL ANALYSIS AND DEVELOPMENT (IAD) FRAMEWORK

The IAD provides a comprehensive guide to addressing collective action problems.

To begin with, it has a simple aim -
articulated and developedby Elinor and Vincent
Ostrom -
to understand 'the strengths and limitations of diverse forms of institutional
arrangementsin different settings' (Ostrom et al., 2014: 268; note that the Ostrom ref-
erences in this chapter are to Elinor). It focuses heavily on 'problem-solving', in which
actors produce, use, analyse, and refine institutions to aid effective collective action
and produce their preferred outcomes (2014: 269). However, the IAD requires a major
investment of time to understand its many components and relate them to specific
issues such as diagnosing collective action problems in relation to CPRs (Heikkila and
Andersson, 2018: 309).
To begin, consider key terms and insights described in Box 7.3
-
as part of a wider
-

framework. It promotes a common language, including a 'universal set of variables' and

collection of rules, to help ask the right questions during theory developmentand empiri-
cal analysis, on the basis that there are too many variables and too much complexity to
contain within a single theory or model (Ostrom, 1990: 214; Ostrom et al., 2014: 267-71;
Schlager and Cox, 2018: 216; McGinnis, 2011). Rather, a 'family of theories' can be used
to specify the relative importance of some elements to answer a research question, and
118 Understanding Public Policy

many models can provide many precise accounts of the interaction between key variables
(Ostrom, 2011: 9). As described, this approach seems to contrast with our initial focus on
very simple rational choice games, but think of these games as some of many ways to use
the concepts described by the IAD (2011: 8; Ostrom et al., 2014: 269).

Box 7.3 Key Institutional Analysis and Development (IAD) terms

Action arena and situation. The action arena contains action situations and actors.
Action situations are 'social spaces where individuals interact, exchange goods and
services, solve problems, dominate one another, or fight' (Ostrom et al., 2014: 271;
McGinnis, 2011: 172). Key questions include: who are the actors, what are their
positions, what action is allowed, how much information do they have, what are the
known costs and benefits of each action, and how does this action relate to the wider
'biophysical'and 'community'context? More philosophically, we may ask: under what
conditions do actors favour social solidarity over selfish action (Ostrom et al., 2014:
272- 73; Ostrom, 2011: 11-2; Heikkila and Andersson, 2018: 313)?

Institutions. The 'rules, norms, and shared strategies that structure human behav-
iour and choices' (2018: 310). They can be (1) formal, written, and widely under-
stood or (2) informal, unwritten, and difficult to detect (Chapter 5). We can be
describing the rules that are 'collectively created, adapted, monitored, and enforced'
(Heikkila and Andersson, 2018: 310) or understandings that 'exist in the minds of
the participants and sometimes are shared as implicit knowledgerather than in an
explicit and written form' (Ostrom, 2007: 23).
Rules. 'Shared prescriptions (must, must not, or may) that are mutually understood
and predictably enforced in particular situations by agents responsible for monitor-
ing conduct and for imposing sanctions' (Ostrom, 2007: 23). 'Working rules' are
common knowledge participantsuse them 'to explain and justify their actions to
-

fellow participants' and actors can enforce them on the assumptionthat all know
-

them (Ostrom, 1990: 51; 2011: 18). 'Rules in use' are 'followed and respected by
people', while 'rules in form' are 'formalized and written down' but not necessarily
respected (Heikkila and Andersson, 2018: 314).
Norms. 'Shared prescriptions that tend to be enforced by the participants them-
selves through internally and externally imposed costs and inducements' (2007:
23). Internal costs include feelings of guilt. External costs include 'social displeasure'
(Ostrom, 1990: 206).
Strategies. 'The regularised plans that individuals make within the structure
of incentives produced by rules, norms and expectations of the likely behaviour of
others' (Ostrom,2007: 23).
Common pool resources (CPRs). Generally refers to finite natural resources with two
properties: (1) it is difficult to exclude actors from enjoying their benefits and (2)
their useby one diminishes their value to another. Examples are groundwater
basins, fish stocks, forests, and land to grow food (Heikkila and Carter, 2017).
 CollectiveAction Problems in Public Policy 119

Evaluating outcomes. It is difficult to define success, and all measures come with
trade-offs. Possible criteria include that outcomes were (1) economically efficient, (2)
equitable (according to contribution/reward,or ability to pay), (3) in line with actors'
values on how to reward cooperation and punish defection, or conducive to ( 4) high-
trust/low-conflict relationships, (5) effective accountability mechanisms, or (6) CPR
sustainability(Ostrom, 2011: 16; Cox et al., 2010: 40; Baggio, 2016: 421).
Self-governance. People often have the ability to cooperate, to design institutions,
and to solve collective action problemswithout state intervention. The IAD analyses
the conditions under which 'bottom-up' action is more effective than 'command-
and-control' from central government,and how both approaches interact in practice
(Heikkila and Andersson, 2018: 311-19; Schlager, 2004: 159; Schlager and Cox,
2018: 216).

Polycentricitydescribes (1) 'many decision centers' with their own separate authority,
(2) 'operatingunder an overarching set of rules' but with (3) a sense of 'spontaneous
order' in which no single centre controls the rules or outcomes (Aligica and Tarko,
2012: 237, 254; Heikkila et al., 2011: 123).

Polycentric governance describes 'policymakingcentres with overlappingauthority;

they often work together to make decisions, but may also engage in competition or
conflict' (Cairney et al., 2019a: 2). Multiple centres can be found across many organi-
zations or within one. To understand a problem, who seeks to solve it, and how, we
need to examine 'multiple action arenas' containingdifferent actors (Ostrom et al.,
2014: 271).

We can usethe IAD to combine several elements of this book. Institutional rational choice
describes the rules that provide incentives and sanctions to guide individual and collective
practices (Chapter 5). Policymaking complexity describes the many actors, institutions, net-
works, ideas, and socioeconomic factors and events that combine to produce policy out-
comes (Chapter 6). Multi-level governance describes the absence of one single government
in control (Chapter 8) and the necessary and inevitable
- -
presence of 'polycentric' gov-
ernance (Box 7.3). Altogether, we can say that institutions might foster collective action,
but there are many available types, understood in different ways by many actors, and they
coexist in a complexpolicymaking environment; multiple layers and types of rules are
difficult to reconcile and use collectively. If so, how can people come together to generate the
sense of a common aim, and cooperate to produce the rules to shape their collective practices?

KEY APPROACHES IN THE IAD'S EXTENDED FAMILY

Since the IAD contains a 'family of theories' with a problem-solvingbent, let's try to
understand it through the lens of important problems, including how to manage CPRs
(Ostrom's most famous work, providing a valuable lens through which to understand the
IAD), polycentric governance, and institutional complexity. In the first two examples,
120 Understanding Public Policy

the research has been led by the IAD's matriarch, Elinor Ostrom. In others, such as
Institutional Collective Action and the Ecology of Games, there is a clear family connection
but more than one lineage.

Managing Common Pool Resources (CPRs) and Avoiding Tragedies

We can understand the IAD through the lens of Governing the Commons (Ostrom, 1990)
which explains how to challenge the 'tragedy of the commons' narrative and encourage
more effective CPR management(Heikkilaand Carter, 2017). The success of this approach
contributed to Elinor Ostrom winning the Nobel Prize in Economic Sciences in 2009:

It long unanimouslyheld among economists that natural resources that were col-
was

lectively used by their users would be over-exploitedand destroyed in the long-term.

Elinor Ostrom disproved this idea by conducting field studies on how people in small,
local communities manage shared natural resources, such as pastures, fishing waters,
and forests. She showed that when natural resources are jointly used by their users, in

time, rules are established for how these are to be cared for and used in a way that is
both economicallyand ecologically sustainable. (Nobel Media AB, 2009)

Ostrom (1990: 8-12) criticizes the uncritical use of rational choice games to conclude too -

quickly that disastrous collective action problems are inevitable and that we must secure
-

major government intervention or 'privatize' commons and assign propertyrights (Ostrom,

1990: 12). These recommendations seem to rely on too-simpleassumptions (Box 7.2) which
prompt people to make sweeping statements about the need for one solution (Ostrom,
1990: 8-14). They also assume a highly informed, efficient, fair, and reliable government
or market, and underestimate the unintended consequences of government intervention or
costs of privatizing resources (Schlager, 2004: 148-49). Yet empirical studies highlight many
examples of nationalized or privatizedCPRs with disastrous effects (Ostrom,1990: 23).
Alternatively, what if there is much evidence that people often work collectively and
effectively without extreme coercion? We can explore the relative benefits of more vol-
untarily cooperative arrangements, with people seeking agreements with each other that
could be enforced if necessary by CPR users or a private rather than state authority.
- -

The commons would remain common, users would monitor each other's behaviour, and,
in some cases, they would sanction defections (1990: 17). On that basis, Ostrom (1990:
20-22) describes interlinked aims:
• Theoretical. To identify key variables and the conditions that help some groups or
communities organize themselves to solve a collective action problem without state
coercion or completelyprivateproperty rights.
Empirical. To code and compare many examples to identify their common
•
elements (see
also National Research Council, 1986).

Think of an iterative process of using the IAD to analyse case studies and
using that analy-
sis to build insights to enhance the IAD. It helps produce a long list of variables to explain
success and failure in CPR management (Ostrom, 1990: 183). This approach requires
intense and careful study to (1) gather enough information about vital factors, such as the
nature of the CPR and 'culturallyacceptable rules', (2) understand how specific rules work
 Collective Action Problems in Public Policy 121

in specific contexts, and (3) account for inevitable variations in their impact on free-riding
and cooperation. Ostrom (1990: xv) describes analysing a subset of 5,000 'small-scale'
cases if they reported four factors: the 'structure of the resource system', the 'attributes'
and behaviour of the users ('appropriators') of CPRs, the rules they were using, and the
outcomes resulting from their behaviour.
In that context, the IAD serves initially as a framework, combining a language, way
of thinking, and collection of methods and insights to apply to empirical cases (Ostrom,
2011). Then, we can develop specific theories or models to apply to particular fields such
as CPR management. Ostrom (1990: 90; 140) describes eight (often mutually reinforcing)

'design principles' common to 'long-enduringCPR institutions', and further studies help

clarify the principles (Cox et al., 2010: 15; Baggio et al., 2016: 431):

l. CPRs have 'clearly defined boundaries'. Users (1) know what resource they are
managingand (2) can distinguish'legitimateusers and nonusers' to exclude 'outsiders'.
2. The rules are tailored to local conditions. Users know what they (1) are expected to
contribute to management and (2) receive from CPRs. They enjoy high certainty about
the likely costs and benefits of use.
3. The actors affected by the rules can also help shape them (preferably at low cost).
4. CPR monitors are users or accountable to users. They monitor (1) the conduct of users
and (2) the state of the CPR. The costs of mutual monitoringtend to be low, and their
consequences are felt relatively quickly, but such factors are contingent on context. For
example, it is easier to monitor resource extraction than pollution (Stem et al., 2002: 462).
New institutions often begin small and develop incrementallyto minimize early set-up costs.
5. The penaltiesfor rule-breaking are low if the choice is a one-off and understandable
under the circumstances (to avoid alienatingthe user to the extent that they no longer
believe in the rules). The penalties are high if the choice is part of a pattern which
makes other users feel they are being treated like 'suckers' for following the rules, or if
any rule-breakingwould be catastrophic to the CPR.
6. Conflict resolution is frequent, rapid, and low-cost (to address the inevitable ambiguity
of any rule which, if left unresolved, allows free-riding).
7. Users have the right to self-organize without external government intervention (or too
much interference from hubristic public managers Schlager, 2004: 167).
-

8. These rules and practices form part of 'multiple layers of nested enterprises'. Many
projects are connected geographically and at different scales local, regional, national
- -

hopefully in ways that do not undermine the function of individual projects (Heikkila
et al., 2011).

These design principles help explain why some communities manage CPRs successfully.
They allow users to 'adopt the same commitment' and expect the long-term benefits to
be relativelyworthwhile (Ostrom, 1990: 186; Schlager, 2004: 145-46). However, there are
three key reasons to not expect hard and fast conclusions to emerge.
First, design principles contribute to but do not determine
-
more
-
fundamental
requirements for collective action. For example, it requires the CPR itself to be conducive
to effective design: good managementwould improve its value to users, and users know
enough about the CPR to predict its contribution to their livelihood and the effect of their
action (Schlager, 2004: 151-52). Further, it requires the sense of a common goal to which
many actors will commit, driven by: tangible factors such as local knowledgeof likely
122 Understanding Public Policy

rewards; less tangible factors such as a sense of communityand local autonomy; and expe-
rience and leadership to promote common beliefs and interests, often when differences-

such as economic inequality might undermine solidarity (2004: 152, 168; Heikkila and
-

Andersson, 2018: 317).

Crucially, it requires high levels of trust to encourage norms of reciprocity (Poteete
et al., 2010: 226). Trust, and factors such as predictability and access to information, is
crucial to minimizing the costs of compliance monitoring and enforcement (Ostrom, 1990:
25-26). Trust may develop when participants communicate regularly, share an under-
standing of their common interests, reciprocate each other's cooperation, and have proven
reliable in the past (Ostrom, 1990: 183; Keohane and Ostrom, 1995: 6; Ostrom, 1998; see
also Putnam 2001 'social capital').
on

Many additional factors are key to developing trust and solidarity, such as the 'evolu-
tion' of behaviour under conditions of uncertaintyand complexity(Box 7.2). Actors have
often learned about rule efficacy to encourage cooperation and punish opportunism
- -

through trial and error over a long period, beginning with simple, low-cost operational
rules producing quick wins (1990: 14, 34, 140-42). They have also been motivated to
change rules over time, when the costs are low and the perceived benefits of improved
cooperationare high (1990: 211). The latter helps explain how actors solve 'first-order' col-
lective action problems, in which there are unclear incentives to invest resources to set up
institutions in the first place (1990: 182). Indeed, Ostrom's (1990: 59) discussion of actors
discovering rules together resembles (somewhat) complexitytheory's discussion of locally
'emergent' rules and outcomes which are not explained simply by cost-benefit calculus or
design factors such as the number of face-to-face meetings (Ostromet al., 2014: 275).
Second, the IAD is complicated because it analyses many types of rule-influenced activ-
ity. The IAD classifies rules according to multiple purposes, including how many actors
are part of an action situation and the role they might play ('positional');what they must/
must not do ('choice'); who is eligible to participate ('boundary'); who can move from one
position to another ('succession'); who controls membership and how ('entry and exit');
how many participants have to be involved in a choice and what will happen if there is no
agreement('aggregation'); how to manage and communicate information ('information');
the rewards or sanctions ('pay-off'); and the range of acceptable actions or outcomes from
action ('scope') (Ostrom and Crawford, 2005: 191-209; Ostrom et al., 2014: 278-79;
Schlager and Cox, 2018: 219-20). Good explanationrequires 'multiple levels of analysis'
to account for and help design the rules which influence the context of other rules, and
- -

are more or less costly to change:

•
'operational-choice' rules on the action 'on the ground' (including the pay-offs/
sanctions for behaviour);
• 'collective-choice' rules on how actors make and monitor operational-choice rules;
• 'constitutional-choice' rules on who performs legitimateroles, such as monitoringand
enforcement (and who can modify collective-choice rules); and
• 'metaconstitutional' analysis of how to define and modify constitutions with reference
to the wider political and social context (Ostrom, 1990: 51; Ostrom et al., 2014: 285).

Third, the IAD complicated because each case study of physical conditions and social
is

practices is different in some way (Stern et al., 2002: 446-47; Dietz et al., 2003: 1907).
We can use the IAD to compare experiences but accept that a profoundlysuccessful scheme
in one context may fail miserably in another (Ostrom, 2008: 16-17).
 Collective Action Problems in Public Policy 123

The complexity and global spread of CPR studies

Compared to most theories in this book, the international spread and scale of Ostrom's
(1990) study is immense. Ostrom (1990: xv) describes trying to make sense of 5,000
'small-scale' cases (containing 50 to 15,000 actors in a single country). The most ana-
lysed cases were in countries other than the USA, initially including Switzerland, Japan,
Spain, the Philippines, Sri Lanka, and Turkey. They were followed by new fieldwork on
multiple/larger-scaleprojects, including: irrigation projects in Nepal; a 17-country, four-
continent study of forest management; and land basin managementin Southern California
(Ostrom et al., 2014: 288; Blomquist, 1992). Crucially, the newer analysis suggested that
IAD insights held even when the projects were larger, with more actors, and even when
some users had advantageous user rights that could foster resentment (Heikkila and
Carter, 2017).
CPR scholarship has grown significantly to compare many different -
multi-level,
small- and large-scale, simple and complex settings (see Robson et al., 2014; Pacheco-
-

Vega, 2014 on countries including Mexico, Eritrea, and Belgium) and use multiple meth-
ods including interview-rich fieldwork and laboratoryexperimentsto explorenew game
theory scenarios (Ostrom et al., 2014: 286-87; Poteete et al., 2010). It has branched into
new fields such as how to share research knowledge, manage tourism, maintain housing

cooperatives, distribute international aid, avoid overusingantimicrobial drugs, and con-

sider the global commons (Heikkila and Carter, 2017; Boyd et al., 2018: 1240; Ostrom
et al., 2014: 289; Schlager and Cox, 2018: 242; Ostrom, 2008: 10). It also informs stud-
ies of the factors that influence collective action success, including the total number of
actors, the number required to make cooperation viable, the 'discount rate', 'similarities
of interests', and leadership. To this list, the IAD adds the value of learning (Box 12.2),
the role of external actors influencing collective-choice rules, and how people gather
information in real-world settings (Ostrom,1990: 188-92; Schlager, 2004: 163; Heikkila
and Gerlak, 2018).
With a research programme covering so many diverse experiences, using a rather com-
plicated approach, mixing methods, and identifying a huge number of potentially relevant
variables, it is inevitable to find variation in the way that people think about, research, and
produce conclusions on CPR management (Ostrom, 2009a). Different scholars also relate
insights to very different political contexts. We can use the IAD to analyse the relative
value of government,market, and communal action in each context, but know that atti-
tudes to (1) state versus market action, (2) local autonomy versus external 'command-and-
control', and (3) how to evaluate success vary widely across the globe (Schlager, 2004: 159,
169; Dietz et al., 2003: 1910; Stern et al., 2002: 457). This is true even when a key aim is to
produce commonlyagreed rules and methodologicalstandards (Ostrom, 2006: 5; Poteete
et al., 2010).

Social-Ecological Systems (SES) Framework

The SES framework sums up this scientific dilemma. On the one hand, Ostrom (2009b:
419) describes it as a way to address the lack of a common interdisciplinary framework to
study 'complex social-ecological systems (SESs)' and it is a 'direct descendent' of CPR stud-
ies (Ostrom et al., 2014: 291). On the other, it is a yet more complicated framework with -

a high cognitive barrier to entry -

to reflect Ostrom's collaboration with natural scientists
(2014: 289-91) and connections to the wider literature on complexity theory (Preiser
124 Understanding Public Policy

Table 7.2 Factors essential to self-organization in social-ecological systems

'Second-level'variables Summary of their indicative effect*

Size of resource system If too large, thecosts of establishing boundaries are too high. If too
small, there insufficient benefits to prompt collective action.
are

Productivityof system If too abundant, users will see little reason to worry. If too depleted, it
will seem too late to manage for the long term.
Predictability of system Users need a clear sense of the effect of their rules, which is easier in

dynamics; Resource unit static forests than fisheries, water, and mobile wildlife.
mobility
Number of users Large numbers of actors are difficult to manage but may be needed to
manage large resource systems.
Leadership Self-organization is more likely when guided by actors with entrepre-
neurial skills and local respect.
Norms/social capital Shared ethics and norms of reciprocity reduce the costs of agreement
and monitoring compliance

Knowledge of the SES Users will perceive the costs of self-organizing to be lower if they
understand the resource system and the likely effect of their actions.
Importance of resource Users are more likely to self-organize if they depend on the resource
to users units for their livelihood.
Collective-choice rules Users are more likely to self-organize if they have high autonomy to
create and monitor their own rules.

Source: adapted from Ostrom (2009b: 420-21).

*
They are indicative since, 'As in most complex systems, the variables interact in a nonlinear fashion ...
Simple
blueprint policies do not work' (2009b: 421).

2018). Put simply, 'the ecological and social sciences have developed independently
et al.,
and do not combine easily' (Ostrom, 2009b: 419).
The result is a framework that resembles CPR studies in key respects, with visual
emphasis on the interactions between 'first-level' concepts including users, their govern-
ance system, system (such as a protected park), and resource units (such as its trees)
resource

(2009b: 420). It also raises similar questions such as 'When will the users of a resource
-

invest time and energy to avert "a tragedy of the commons"?' and answers them with -

reference to Ostrom's 'second-level' concepts (or 'variables') describing factors that encour-
age users to (1) value long-termsustainabilityand (2) self-organize to secure this outcome
(Table 7.2).
While the SES has inspired many studies (particularly on fishing), the project
is less organized than CPR research, partly because new aspects and terms such as -

'governance system' are less well defined than terms in the IAD (Ostrom et al., 2014:
-

294-96). It remains a promising way to 'diagnose' the sustainability of social-ecolog-

ical systems (Partelow, 2018) and address problems with self-organizationto manage
resources, but largely by people who are already immersed in the IAD, or scholars using
the factors in Table 7.2 as a checklist to structure discussion (Schlager and Cox, 2018:
237-44).
 CollectiveAction Problems in Public Policy 125

Polycentric Governance and Institutional Collective Action

For our purposes, the potential benefit of polycentricity in policymaking(Box 7.3) relates
to the best of both worlds: (1) the freedom of many actors to act according to their exper-
tise and judgementbut (2) the sense of a 'common ideal', combined with some agreement
to contribute to a common aim (Aligica and Tarko, 2012: 238). As such, it could be rela-
tively effective if it avoids associated with more impositional mono-centric arrange-
costs
ments, such as reduced freedom, high costs of finding information about the system, and
a tendency not to consult all actors when making decisions (2012: 240). For example, a

polycentric system may have a 'built-in mechanism of self-correction' as many actors inno-
vate and adjust through trial and error, during an 'evolutionarycompetition between the
complementaryideas and methods of those different decision centers' (2012: 240; 246;
compare with 'adaptivecapacity' in Pahl-Wostl and Knieper, 2014, and 'The evolutionary
metaphor' in Chapter 6, p.101).
Polycentric governance studies pre-date the IAD's CPR research and describe governmen-
tal and non-governmentalcooperation more widely. Heikkila (in Cairney et al., 2019a:
16-18) links its early study from the 1960s 'to the work of Elinor and Vincent Ostrom and
the "Bloomington School" of institutional analysis' in relation to US metropolitan govern-
ance, and describes the elements that became central to the IAD framework:

• Theoretical discussions emphasize (1) the ability of actors to self-organize to produce

institutions to help them solve collective action problems and (2) the fact that public
goods need not be delivered by a single central government. Many governments can
deliver many different types of goods and service at different levels or scale, while
working with each other and non-governmentalactors. They operate under a 'shared
system of laws' (such as providedby a constitution) and produce 'interorganizational
agreements' to aid cooperation.
• The empirical aim is to identify the conditions under which polycentric governance
arrangements, held together by a collection of nested rules, provide more efficient,
equitable, democratic, or sustainable arrangements during the provision of public
goods.

This work challenged the idea that metropolitan regions would necessarily be inefficient
simply because they contained many semi-autonomous and overlapping jurisdictions
(Aligica and Tarko, 2012: 241-44). Greater efficiency is not a given since multi-centric
-

governance could produce damaging practices associated with competition, duplication,

and unclear accountability but has been documented in an impressive number of cases
-

(Aligica and Tarko, 2012: 243).

Institutional Collective Action (ICA)

Polycentric governance may help address 'institutional collective action' (ICA) dilemmas
when policy problems transcend the responsibility of one government (Feiock, 2013). It
is difficult to match the perceived 'scale and nature of policy problems'(including how to
tax, deliver services, and allocate land use) to the 'scale and coerciveness of policy inter-
vention' (2013: 397). Citizens demand services, jurisdictions exist, and each service has its
126 Understanding Public Policy

own economies of scale. However, it is common for all three

Economies of scale -

when the marginal cost

to diverge
-
when governments are too small, or government
of producing a good or responsibilitiesoverlap too when there is insufficient
much -

service drops as you collaboration. Many policymakers, often at different levels

increase production of government,seek to coordinate their response, produce an
(reducing the average agreement on how to act collectively, ensure that each actor
cost). trusts and maintains their agreement, and therefore minimize
the transactions costs of ICA. Otherwise, their own action
Transactions costs -

will produce negative externalities for others (2013: 398;

the costs of 'doing
business' with oth-
Marks and Hooghe, 2000). Swann and Kim (2018: 273) sum-
marize the conditions under which such 'semiautonomous'
ers, including costs
of searching for governing bodies 'overcome barriers to collective action and
information on others, reduce the risk and uncertaintyof collaborative arrangements'.
professional fees, and The most effective solutions appear to be:
the setting up and
enforcing of contracts. •
participating heavily in informal social networks and

the speaking directly with people, to help build 'social capital'

Externality -

and share information effectively (2018: 281-82)

(often unintended)
'spillover' effects of the
•
tailoring strategies to the nature of the task, such as
actions of one actor on to maintain closed networks between a small number
another. of organizations when you need to 'defend against
opportunism',or flexible arrangements when you need to
act quickly and reduce the burden of agreement (2018: 283)
•
working with organizations who already share or collaborating with them to
-

encourage similar values, ideals, policy preferences,

-
and goals (2018: 284)
•
keeping the costs and other barriers to participation low, to maintain high levels of
'transparency, accountability and equity' (2018: 285)
•
producing 'small wins' then workingincrementallyto reduce the burden of cooperation
on ill-resourced organizations (2018: 286; compare with Ansell and Gash, 2008).

However, as in the IAD more generally, the overall message is that no single arrangement
provides blueprint
a or panacea for collective action (Cairney et al., 2019a: 18).

Institutional Complexity and the Ecology of Games

The 'Ecology of Games Framework' (EG) combines insights from many approaches to
analyse 'institutional complexity' or 'complex institutional systems' (Lubell, 2013: 538).
The focus is on actors learning how to secure 'mutuallybeneficial outcomes', cooperating to
produce and deliver agreed solutions, and bargaining 'over the distribution of the costs and
benefits' within a system over which no actor has control (Lubell, 2015: 42).
First, EG connects Long's (1958) 'ecology of games' (Box 7.2), the IAD, and 'a complex
adaptive system perspective' to reinforce the idea that institutional arrangements are
not simple and orderly. In simple games, we need only analyse the interaction between a
small number of actors with reference to one set of self-contained rules providing clear
sanctions or pay-offs. However, 'governance involves multiple policy games operating
simultaneously' and over time (Lubell, 2013: 538). For example, Lubell and colleagues
 CollectiveAction Problems in Public Policy 127

identified 'over 100' different venues 'operatingsimultaneously'in a single case study of

San Francisco Bay water policy (Lubell, 2013: 537; 2015: 42).
In that context, policy games may be contained within a geographical area, but there are
no self-contained collective action problems. For example, 'biodiversity policy' is a collec-
tion of interdependentpolicies relating to issues like local planning, protected species,
and water management (2013: 541; see also Lubell et al., 2010: 292-94). Many games
associated with different policy issues -
such as 'water supply, water pollution, air pollu-
tion, traffic congestion, or loss of biodiversity' interact with each other, such as when
-

the same actor participates in multiple games subject to different rules, or each game
produces 'externalities' which affect the 'pay-offs' from others (2013: 540-41). A focus on
complex adaptive systems suggests that central governmentsdo not have the resources to
control or even understand fully interaction at this frequency and scale. Rather, drivers
- -

to change behaviour and outcomes are:

endogenous (internal), when actors (1) follow and shape the rules of each institution
•

and (2) learn when they 'try out different strategies'; and
•
exogenous (external), when physical resources (such as water basins) change, or other
levels of government change the resources of local actors (2013: 542).

Second, EG takes insights from the wider literature to analyse specific aspects of this
process, such as actors forming networks with their allies to get what they want from dif-
ferent games (see 'advocacy coalitions' in Chapter 10) and engaging in 'policy learning'
(Chapter 12) update their knowledge (2013: 543). Actors deal with bounded rational-
to
ity primarily with reference to their 'social tribal instincts', with EG emphasizing: 'the
importance of emotions, instinctual "fast thinking," cooperation, in-group biases, and
social learning [producing] some individuals who are altruistic, others who condition-
...

ally cooperate, and a minority who pursue their self-interest regardless of the welfare of
others' (2013: 544-45). Actors form alliances based on strongly held beliefs, process new
information to refine their beliefs and strategies, but 'never have complete knowledge'of
the environment in which they operate (2012: 544). Consequently, we have come a long
way from the 'utility maximizing'approach outlined at the start of this chapter.
EG also emphasizes the relative 'fragmentation'of wider policymakingarrangements,
negative effects of political competition,and disproportionatepower of some actors even
in a 'self-organizing system' (2013: 546). For example, 'pay-off externalities' occur when
organizationswith overlappingjurisdictions produce policies that undermine each other,
and 'strategy externalities' describe an actor developing behaviour that is useful in one
game but damaging in another (2013: 546).
Yet, as with the IAD, the theoretical emphasis is on solutions that come from initia-
tives consistent with self-organizing systems -
such as 'collaborative governance' (Ansell
and Gash, 2008) rather than 'consolidatingall decision-making authority into a single
-

institution ...[because] it is naïve to believe that an optimal system could be exogenously

designed' (2013: 547). In that context, the ongoing empirical agenda is to identify how
'collaborative partnershipspromote cooperation, learning, and bargaining', such as when
actors recognize the benefit of 'participating in a portfolio of forums' to reduce strategy
externalities (Lubell, 2015: 45; Mewhirter et al., 2018: 304). This approach must be flex-
ible, because there are too many different -
and often-changing venues, institutions, and
-

actors (with different motives)to produce a one-size-fits-all solution.

128 Understanding Public Policy

CONCLUSION

It may seemstrange to relate rational choice theory to public policy because it is difficult
to apply parsimoniousexplanationsbased on a small number of variables to a field with
'inherently messy circumstances' (John, 1998: 117; Hay, 2004a: 46-50; Hindess, 1988:
115 with Laver, 1997: 11-12). Yet the issues raised by rational choice and game theory
inform many empirical studies, prompting questions such as: how can people cooperate
effectively, and what should be the role of government? There will always be winners and
losers when policymakers make choices, and it is unrealistic to seek a rational-synoptic or
technical solution to conflicts of preferences within society. Further, governmentsolutions
may be expensiveand produce unintended consequences. We need institutions to produce
more convergence between individual incentives and collective action, but their appropri-
ate nature is an open question.
You may have detected different assumptions about human behaviour and levels of
expectations about these dilemmas from the stories people tell. For example, the 'pris-
oner's dilemma' and 'tragedy of the commons' thought experiments seem to present a
bleak view of human behaviour, in which it would be surprising to find much evidence
of cooperation in the absence of coercion, even when the fate of the world is in their hands.
In the bleakest scenarios, people act independently and make calculations with insufficient
thought for other people or the future. Further, many oppose government intervention
because they argue that politiciansare just as bad and would only make things worse.
In contrast, IAD empirical scholarship describes a long history of cooperation to maintain
CPRs. In that context, the IAD presents a more positive account of non-market and non-
governmental solutions to collective action problems based on rules, norms, trust, and
the monitoring and enforcement of collective agreements. In positive scenarios, people
are social beings who make 'informed judgmentsabout uncertain benefits and costs' while

communicatingregularlywith others (Ostrom, 1990: 208). They share information, build

trust by becoming known as reliable and predictable, and come togetherto produce, moni-
tor, and enforce rules for the benefit of the group and the planet. They can be trusted to
cooperate to produce rules to solve their own problems rather than relying completely on
experts or governmentsto tell them what to do (Schlager, 2004: 168). These approaches
help put problemsin context and identify the conditions that encourage effective collec-
tive action.
This context seems remarkably complex, which makes the IAD difficult to understand
at first. While game theory laboratoryexperiments built on simple rules and relatively
-

small numbers of parameters -

produce relatively parsimonious analysis, field studies
embrace life's complexitiesto try to understand the key dimensions of each case study's
context. There are too many concepts, variables, global applications, and variations-by-
context to contain in a simple theory or model. Rather, the IAD offers a common language
to help organize research and encourage a family of theories to explain different parts of
the whole picture. As with any language, we learn it as we speak with people more fluent
than us. Unfortunately,it is also a highly technical language, and knowledge of its rules is
often as tacit, ill-communicated,and difficult to grasp as the working rules described by
the IAD (Heikkila and Andersson, 2018: 310). So if you see an IAD scholar with a pained
expression, saying 'it depends', 'it's complex', and 'there is no magic answer', you know
they really mean it. You should also give yourself a break if it takes a while to make sense.
 CollectiveAction Problems in Public Policy 129

Like Chapters 6 and 8, this chapter explores the consequences when policy problems
go beyond the boundaries of jurisdictions or political systems and become vulnerable to
complex decision-makingarrangements that may undermine central control and simple
institutional responses. The IAD makes frequent reference to 'complexity' and 'evolu-
tion'. It highlights the more 'spontaneous'and 'emergent' outcomes that arise when many
actors and institutions interact (Aligica and Tarko, 2012: 254; Cairney, 2013). Further,
EG emphasizes fragmentation when the overlap between jurisdictions, combined with
uncertainty and different ideas about how to solve problems, produces duplication or con-
tradictory policies. In that context, compare with chapters that explore the role of norm
enforcement (Chapter 11, p.192), policy learning, and coercive policy transfer (Chapter 12)
to address such problems.