All going well, it will be out in November 2019. We are now at the proofing stage.
Many theories in this 1000 words series describe multiple policymaking venues. They encourage us to give up on the idea of an all-knowing, all-powerful national central government. Instead, there are many venues in which to make authoritative choices, each contributing to what we call policy.
The word ‘multi-centric’ (coined by Professor Tanya Heikkila, with me and Dr Matt Wood) does not suggest that every venue is of equal importance or power. Rather, it prompts us not to miss something important by focusing too narrowly on one single (alleged) centre of authority.
To some extent, multi-centric policymaking results from choice. Many federal political systems have constitutions that divide power between executive, legislative, and judicial branches, or give some protection to subnational governments. Many others have become ‘quasi-federal’ more organically, by sharing responsibilities with supranational and subnational governments. In such cases, there is explicit choice to distribute power and share responsibility for making policy (albeit with some competition to assert power or shuffle-off responsibility).
However, for the most part, this series helps explain the necessity of multi-centric policymaking with reference to two concepts:
Both factors combine to provide major limits to single central government control. Elected policymakers deal with bounded rationality by prioritising some issues and, necessarily, delegating responsibility for the rest. Delegation may be inside or outside of central government.
1000 Words theories describing multi-centric government directly
Multi-level governance describes the sharing of power vertically, between many levels of government, and horizontally, between many governmental, quasi-non-governmental and non-governmental organisations. Many studies focus on the diffusion of power within specific areas like the European Union – highlighting choice – but the term ‘governance’ has a wider connection to the necessity of MLG.
For example, part of MLG’s origin story is previous work to help explain the pervasiveness of policy networks:
Polycentricity describes (a) ‘many decision centers’ with their own separate authority, (b) ‘operating under an overarching set of rules’, but with (c) a sense of ‘spontaneous order’ in which no single centre controls the rules or outcomes. Polycentric governance describes ‘policymaking centres with overlapping authority; they often work together to make decisions, but may also engage in competition or conflict’.
This work on polycentric governance comes primarily from the Institutional Analysis and Development (IAD) framework that helps compare the effectiveness of institutions designed to foster collective action. For example, Ostrom identifies the conditions under which non-governmental institutions can help manage ‘common pool resources’ effectively, while IAD-inspired studies of municipal governance examine how many ‘centres’ can cooperate as or more effectively than a single central government.
Complexity theory has a less clear origin story, but we can identify key elements of complex systems:
Systems contain many actors interacting with many other actors. They follow and reproduce rules, which help explain long periods of regular behaviour. Or, many actors and rules collide when they interact, producing the potential for many bursts of instability. In each case, the system is too large and unpredictable to be subject to central control.
1000 Words theories describing multi-centric government indirectly
Many other theories in this series describe multi-centric policymaking – or aspects of it – without using this term directly. Examples include:
Punctuated equilibrium theory suggests that (a) policymakers at the ‘centre’ of government could pay attention to, and influence, most issues, but (b) they can only focus on a small number and must ignore the rest. Very few issues reach the ‘macropolitical’ agenda. Multiple policymaking organisations process the rest out of the public spotlight.
Multiple streams analysis turns the notion of a policy cycle on its head, and emphasises serendipity over control. Policy does not change until three things come together at the right ‘window of opportunity’: attention to a problem rises, a feasible solution exists, and policymakers have the motive and opportunity to act. Modern MSA studies show that such windows exist at multiple levels of government.
The advocacy coalition framework describes the interaction between many policymakers and influencers. Coalitions contain actors from many levels and types of government, cooperating and competing within subsystems (see networks). They are surrounded by a wider context – over which no single actor has direct control – that provides the impetus for ‘shocks’ to each coalition.
In such accounts, the emphasis is on high levels of complexity, the potential for instability, and the lack of central control over policymaking and policy outcomes. The policy process is not well described with reference to a small group of policymakers at the heart of government.
The implications for strategy and accountability
Making Policy in a Complex World explores the implications of multi-centric policymaking for wider issues including:
This is a guest post by Chris Koski (left) and Sam Workman (right), discussing how to use insights from punctuated equilibrium theory to reform government policy making. The full paper has been submitted to the series for Policy and Politics called Practical Lessons from Policy Theories.
Many people assume that the main problem faced by governments is an information deficit. However, the opposite is true. A surfeit of information exists and institutions have a hard time managing it. At the same time, all the information that exists in defining problems may be insufficient. Institutions need to develop a capacity to seek out better quality information too.
Institutions, from the national government, to state legislatures, to city councils – try to solve the information processing dilemma by delegating authority to smaller subgroups. Delegation increases the information processing capacity of governments by involving more actors to attend to narrower issues.
The delegation of authority is ultimately a delegation of attention. It solves the ‘flow’ problem, but also introduces new ‘filters’. The preferences, interests, and modes of information search all influence the process. Even narrowly focused smaller organizations face limitations in their capacity to search and are subject to similar forces as the governments which created them – filters for the deluge of information and capacity limitations for information seeking.
Organizational design predisposes institutions to filter information for ideas that support status quo problem definitions – that is, definitions that existed at the time of delegation – and to seek out information based on these status quo understandings. As a result, despite a desire to expand attention and information processing to adapt to changes in problem characteristics, most institutions look for information that supports their identity. Institutional problem definitions stay the same even as the problems change.
Governments eventually face trade-offs between the gains made from delegating decision-making to smaller subgroups and the losses associated with coordinating the information generated by those subgroups.
Governments get stuck in the same ruts as when the delegation process started: status quo bias that doesn’t adjust with change problem conditions. There is a sense among citizens and academics that governments make bad decisions in part because they respond to problems of today with the policies of 10 years ago. Government solutions look like hammers in search of nails when they ought to look more like contractors or even urban planners.
When institutions become stultified in their problem definitions, policymakers and citizens often misdiagnose the problem as entirely a coordination problem. The logic here is that a small group of actors have captured policymaking and are using such capture for their own gain. This understanding may be true, or may not, but it leads to the “centralization as savior” fallacy. The idea here is that organizations with broader latitude will be better able to receive a wider variety of information from a broader range of sources.
There are two problems with this strategy. First, centralization might guarantee an outcome, but at the expense of an honest problems search and, likely, at the expense of what we might call policy stability. Second, centralization may offer the opportunity for a broader array of information to bear on policy decisions, but, in practice will rely on even narrower information filters given the number of issues to which the newly centralized policymaking forum must attend.
The alternative, more delegation, has significant coordination challenges as we find bottlenecks of attention when multiple subsystems bear on decision-points. Also, simply delegating authority can predispose subsystems to a particular solution, which we want to avoid.
Our solutions do solve fundamental problems of information processing in terms of sorting and seeking information – such problems are fundamental to humans and human-created organizations. However, while governments may be predisposed to prioritize decisions over information, we are optimistic that our recommendations can facilitate better informed policy in the future.
This week, we continue with the idea of two stories of British politics. In one, the Westminster model-style story, the moral is that the centralisation of power produces clear lines of accountability: you know who is in charge and, therefore, the heroes or villains. In another, the complex government story, the world seems too messy and power too diffuse to know all the main characters.
Although some aspects of these stories are specific to the UK, they relate to some ‘universal’ questions and concepts that we can use to identify the limits to centralised power. Put simply, some rather unrealistic requirements for the Westminster story include:
If life were that simple, I wouldn’t be asking you to read the following blog posts (underlined) which complicate the hell out of our neat story:
You don’t know what policy is, and it is not only made by a small number of actors at the heart of government.
We don’t really know what government policy is. In fact, we don’t even know how to define ‘public policy’ that well. Instead, a definition like ‘the sum total of government action, from signals of intent to the final outcomes’ raises more issues than it settles: policy is remarkably difficult to identify and measure; it is made by many actors inside, outside, and sort of inside/outside government; the boundary between the people influencing and making policy is unclear; and, the study of policy is often about the things governments don’t do.
Actors don’t possess comprehensive knowledge about the problems and solutions they describe
It’s fairly obvious than no-one possesses all possible information about policy problems and the likely effects of proposed solutions. It’s not obvious what happens next. Classic discussions identified a tendency to produce ‘good enough’ decisions based on limited knowledge and cognitive ability, or to seek other measures of ‘good’ policy such as their ability to command widespread consensus (and no radical movement away from such policy settlements). Modern discussions offer us a wealth of discussions of the implications of ‘bounded rationality’, but three insights stand out:
Policymakers cannot turn policy intent into policy outcomes in a straightforward way
The classic way to describe straightforward policymaking is with reference to a policy cycle and its stages. This image of a cycle was cooked up by marketing companies trying to sell hula hoops to policymakers and interest groups in the 1960s. It is not an accurate description of policymaking (but spirographs are harder to sell).
Instead, for decades we have tried to explain the ‘gap’ between the high expectations of policymakers and the actual – often dispiriting- outcomes, or wonder if policymakers really have such high expectations for success in the first place (or if they prefer to focus on how to present any of their actions as successful). This was a key topic before the rise of ‘multi-level governance’ and the often-deliberate separation of central government action and expected outcomes.
The upshot: in Westminster systems do you really know who is in charge and who to blame?
These factors combine to generate a sense of complex government in which it is difficult to identify policy, link it to the ‘rational’ processes associated with a small number of powerful actors at the heart of government, and trace a direct line from their choices to outcomes.
Of course, we should not go too far to argue that governments don’t make a difference. Indeed, many ministers have demonstrated the amount of damage (or good) you can do in government. Still, let’s not assume that the policy process in the UK is anything like the story we tell about Westminster.
In the seminar, I’ll ask you reflect on these limits and what they tell us about the ‘Westminster model’. We’ll start by me asking you to summarise the main points of this post. Then, we’ll get into some examples in British politics.
Try to think of some relevant examples of what happens when, for example, minsters seem to make quick and emotional (rather than ‘evidence based’) decisions: what happens next? Some obvious examples – based on our discussions so far – include the Iraq War and the ‘troubled families’ agenda, but please bring some examples that interest you.
In group work, I’ll invite you to answer these questions:
I’ll also ask you to identify at least one blatant lie in this blog post.
I am at a workshop today on policy ‘bubbles’, or (real and perceived) disproportionate policy responses. For Moshe Maor, a bubble describes an over-reaction to a problem, and a negative policy bubble describes under-reaction.
For Maor, this focus on bubbles is one way into our increasing focus on the role of emotion in policymaking: we pay disproportionate attention to problems, and try to solve some but not others, based on the ways in which we engage emotionally with information.
This focus on psychology is, I think, gaining a lot of traction in political science now, and I think it is crucial to explaining, for example, processes associated with ‘evidence-based policymaking’.
In taking this agenda forward, there remain some outstanding issues:
How much of the psychology literature is already reflected in policy studies? For example, see the social construction of target populations (emotion-driven treatment of social groups), ACF (on loss aversion and the devil shift), and the NPF (telling stories to exploit cognitive biases).
What insights remain untapped from key fields such as organisational psychology? I’ll say more about this in a forthcoming post.
How can we study the psychology of policymaking? Most policy theory begins with some reference to bounded rationality, including PET and the identification of disproportionate information processing (policymakers pay disproportionate attention to some issues and ignore the rest). It is largely deductive then empirical: we make some logical steps about the implications of bounded rationality, then study the process in that light.
Similarly, I think most studies of emotion/ policymaking take insights from psychology (e.g. people value losses more than gains, or they make moral judgements then seek evidence to justify them) and then apply them indirectly to policymaking (asking, for example, what is the effect of prospect theory on the behaviour of coalitions).
Can we do more, by studying more directly the actions of policymakers rather than merely interpreting their actions? The problem, of course, is that few policymakers may be keen on engaging in the types of study (e.g. experiments with control groups) that psychologists have used to establish things like fluency effects.
How does policymaker psychology fit into broader explanations of policymaking? The psychology of policymakers is one part of the story. The other is the system or environment in which they operate. So, we have some choices to make about future studies. Some might ‘zoom in’ to focus on emotionally-driven policymaking in key actors, perhaps at the centre of government.
Others may ‘zoom out’. The latter may involve ascribing the same basic thought processes to a large number of actors, examining that process at a relatively abstract level. This is the necessary consequence of trying to account for the effects of a very large number of actors, and to take into account the role of a policymaking environment, only some of which is in the control of policymakers.
Can we really demonstrate disproportionate policy action? The idea of a proportionate policy response interests me, because I think it is always in the eye of the beholder. We make moral and other personal evaluative statements when we describe a proportionate solution in relation to the size of the problem.
For example, in tobacco policy, a well-established argument in public health is that a proportionate policy response to the health effects of smoking and passive smoking (a) has been 20-30 years behind the evidence in ‘leading countries’, and (b) has yet to happen in ‘laggard’ countries. The counterargument is that the identification of a problem does not necessitate the favoured public health solution (comprehensive tobacco control, towards the ‘endgame’ of zero smoking) because it involves major limits to personal liberties and choice.
Is emotion-driven policymaking necessarily a bad thing?
[excerpt from my 2014 PSA paper ] This is partly the focus of Alter and Oppenheimer (2008) when they argue that policymakers spend disproportionate amounts of money on risks with which they are familiar, at the expense of spending money on things with more negative effects, producing a ‘dramatic misallocation of funds’. They draw on Sunstein (2002), who suggests that emotional bases for attention to environmental problems from the 1970s prompted many regulations to be disproportionate to the risk involved. Further, Slovic’s work suggest that people’s feelings towards risk may even be influenced by the way in which it is described, for example as a percentage versus a 1 in X probability (Slovic, P. 2010: xxii).
Haidt (2001: 815) argues that a focus on psychology can be used to improve policymaking: the identification of the ‘intuitive basis of moral judgment’ can be used to help policymakers ‘avoid mistakes’ or allow people to develop ‘programs’ or an ‘environment’ to ‘improve the quality of moral judgment and behavior’. Similarly, Alter and Oppenheimer (2009: 232) worry about medical and legal judgements swayed by fluid diagnoses and stories.
These studies compare with arguments focusing on the positive role of emotions of decision-making, either individually (see Constantinescu, 2012, drawing on Frank, 1988 and Elster, 2000 on the decisions of judges) or as part of social groups, with emotional responses providing useful information in the form of social cues (Van Kleef et al, 2010).
Policy theory does not shy away from these issues. For example, Schneider and Ingram (2014) argue that the outcomes of social construction are often dysfunctional and not based on a well-reasoned, goal-oriented strategy: ‘Studies have shown that rules, tools, rationales and implementation structures inspired by social constructions send dysfunctional messages and poor choices may hamper the effectiveness of policy’. However, part of the value of policy theory is to show that policy results from the interaction of large numbers of people and institutions. So, the poor actions of one policymaker would not be the issue; we need to know more about the cumulative effect of individual emotional decision making in collective decision-making – not only in discrete organisations, but also networks and systems.
And finally: if it is a bad thing, should we do something about it?
Our choice is to find it interesting then go home (this might appeal to the academics) or try to limit the damage/ maximise the benefits of policymaker psychology to policy and society (this might appeal to practitioners). There is no obvious way to do something, though, is there?
You should get the impression from 1000 words that most policy changes are small or not radically different from the past: Lindblom identifies incrementalism; punctuated equilibrium highlights a huge number of small changes and small number of huge changes; the ACF compares routine learning by a dominant coalition to a ‘shock’ which prompts new subsystem and policy dynamics; and multiple streams identifies the conditions (rarely met) for major change.
Yet, I just gave you the impression that we don’t know how to define policy. If we can’t define it well, how can we measure it well enough to come to this conclusion so consistently?
Why is the measurement of policy change important?
We miss a lot if we equate policy with statements rather than outputs/ outcomes. We also miss a lot if we equate policy change with the most visible outputs such as legislation. I list 16 different policy instruments, although they tend to be grouped into smaller categories: focusing on regulation (including legislation) and resources (money and staffing) to accentuate the power at policymaker’s disposal; or regulatory/ distributive/ redistributive to suggest that some policy measures are more difficult to ‘sell’ than others.
We also give a limited picture if we equate change with outputs rather than outcomes, since a key insight from policy studies is that there is generally a gap between policymaker expectations and the actual result.
What are the key issues in measurement?
So, as in defining policy change, we need to make choices about what counts as policy in this instance to measure how much it has changed. For example, I have (a) written on one output as a key exemplar of policy change – legislation to ban smoking in public places for Scotland, England/ Wales, the UK, and (almost) EU – to show that a government is signalling major changes to come, but also (b) situated that policy instrument within a much broader discussion – of many tobacco policies in the UK and across the globe – to examine the extent to which it is already consistent with a well-established direction of travel.
To make such choices we need to consider:
How do we solve the problem?
The problem is that we can produce very different accounts of policy change from the same pool of evidence, by accentuating some measures and ignoring others, or putting more faith in some data more than others (e.g. during interviews).
Sometimes, my preferred solution is to compare more than one narrative of policy change. Another is simply to ‘show your work’.
Take home message for students: ‘show your work’ means explaining your logical process and step-by-step choices. Don’t just write that it is difficult to define policy and measure change. Instead, explain how you assess policy change in one important way, why you chose this way, and shine a light on the payoffs to your approach. Read up on how other scholars do it, to learn good practice and how to make your results comparable to theirs. Indeed, part of the benefit of using an established theory, to guide our analysis, is that we can engage in research systematically as a group.
Evolutionary theory is prevalent in policymaking studies and it can be useful if we overcome some initial barriers. First, ‘evolution’ comes with a lot of baggage when we move from a discussion of animals to people. We can blame ‘social-Darwinism’ for the racist/ sexist idea that some people are more evolved than others.
Second, the word ‘evolution’ is used frequently in daily life, and academic studies, without a clear sense of its meaning. When it is used loosely in everyday language, it refers to a long term, gradual process of change. However, evolution can also refer to quick, dramatic change; the idea of ‘punctuated equilibrium’ is that long spells of stability and gradual change are interrupted by relatively short but profound bursts of instability. When we get into the details of studies, there are other sources of potential confusion about, for example, the nature of evolution (does it refer to advancement as well as change?) and the nature of ‘selection’ (do species simply respond blindly to their environments or help create them?).
This sort of confusion can be found in the study of public policy where evolution can refer to a wide range of things, including:
The most prominent theories of politics and policymaking draw on references to evolution in different ways. For example:
Multiple Streams Analysis (Kingdon). Although policymaker attention may lurch from one problem to another, problems will not be addressed until policy solutions have evolved sufficiently within a policy community and policymakers have the motive and opportunity to adopt them. ‘Evolution’ and the ‘policy primeval soup’ describe the slow progress of an idea towards acceptability within the policy community.
Punctuated Equilibrium Theory (Baumgartner and Jones). ‘Incremental’ policy change in most cases is accompanied by ‘seismic’ change in a small number of cases – an outcome consistent with ‘power laws’ found in the natural and social worlds. Kingdon’s picture of slow progress producing partial mutations is replaced by Baumgartner and Jones’ fast, disruptive, pure mutation.
Complexity theory. People, institutions and their environments are interacting constantly to produce rather unpredictable outcomes (or outcomes that may ‘emerge’ locally, in the absence of central control). This might be broken down into three steps:
In other words, different ‘worlds’ are in constant collision, producing new ways of thinking and behaviour that ‘emerge’ from these interactions. They are then passed down through the generations, but in an imperfect way, allowing new forms of thinking and behaviour to emerge.
To describe these processes as ‘evolutionary’, we should use the language of evolution – variation, selection and retention – to describe and explain outcomes. The idea in the natural world is that certain beings (including humans) want to do at least two things: (1) pass on their genes; (2) cooperate with others to secure resources and share them out to their kith and kin. In the political world, the equivalent is passing on ‘memes’ (as described in the 70s by Richard Dawkins) – the ideas (beliefs, ways of thinking) that we use to understand the world and act within it:
The distinctive aspect of applying evolutionary theory to policymaking relates to the idea of passing on memes through the generations. In nature, we think of passing on genes through the generations as a process that takes hundreds, thousands or millions of years. Passing on memes through the ‘policy generations’ is more like the study of fruit flies (months), viruses or bacteria (days or weeks). Ways of thinking, and emerging behaviour, change constantly as people interact with each other, articulating different beliefs and rules and producing new forms of thinking, rules and behaviour. Big jumps in ways of thinking may be associated with generational shifts, but that can take place, for example, as one generation of scientists retires (as described by Kuhn) or, more quickly still, one generation of experts is replaced (within government circles) by another (as described by Hall).
I have discussed in other ‘1000 words’ posts what happens when theories, derived from cases studies of US politics, are applied to other countries and cases. ‘Evolutionary theory’ is more difficult to track, because it is a body of disparate work, loosely related to work in natural science, applied in a non-coordinated way. The same can be said for studies of complexity theory.