Networks and governance

This chapter of the book moves away from the direct tools that can be switched on or off to consider the more complicated, indirect and often softer ones, which are harder to control and partly arise from the unintended actions of the state. For both within and beyond state bureaucracies there are sets of relationships, which in part can be manipulated by the existing tools or fostered directly, which has knock-on consequences for the effectiveness of government. The most powerful of these are the networks that link decisionmakers in government to other actors, both inside and outside the state. Inside the state are the bureaus of the central ministries and their agencies, which are replicated at other levels of government like regions and local authorities. Then there is a terrain of decision-making beyond the state in the forms of associations, interest groups, citizen groups and private bodies, such as companies or trusts. Whether inside or outside the state, networks may be thought of as regular relationships between decision-makers, sustained by mutual trust and knowledge, which tend to persist over time. Networks have properties that bestow advantages and costs on different actors, which in turn will affect outcomes outside government. Some of the properties are about the individuals in these networks; others are about the aggregate effects of the networks as the relationships add up to much more than the sum of their parts. This chapter seeks to explore the putative causal relationships that arise from networks. Linked to networks is governance, which is a broader concept. It describes

the wider framework and rules that shape how governments and core decision-makers operate. It forms a set of understandings and constraints on decision-makers. Governance can be a negative force because of the many obstacles to effective policy-making, but it can be used positively as states steer for better manoeuvre in a more complex world, allowing them to use practical knowledge acquired in these networks and to use the opportunities that can emerge in a more open environment. Governance is in part made up of networks, so many of the same arguments about reciprocity apply; but it is more than that as it indicates an overarching structure, which is about the management of the governing system itself. In spite of the indirect route of positive policy effects, governments around

the world have promoted stronger and better functioning networks and better

forms of governance, particularly in recent years. In the 1990s, governments and other public authorities sponsored partnerships and funding regimes that mixed private and public money in the name of getting public-sector actors out from their bureaucracies into long-term relationships with other power holders, whether in the private sector, or in the voluntary or third sectors. Latterly, there has been a greater attempt to connect different parts of the public sector together in initiatives, such as Joined Up Governance, which are supposed to embody the spirit of cooperation and to ensure public agencies work together to solve public problems. If the justification of those initiatives is correct, those on the ground should benefit. Whether the agencies actually do join up away from the rhetorical commitment is a critical point. Relationships on the ground may be poor because of long-standing rivalries, the difference in perspectives of agencies with particular tasks, or even the intense hatred that arises from past disputes and conflict of personalities. As ever with networks, there is much promise, but a large question remains over whether they deliver. A network is a regular link between one actor and another, usually as part

of a pattern of relationships among actors, which have a common interest. In public policy a network operates usually across organisations in a locale or in a sector of specialisation, but it can operate within organisations as well, such as between the directorates of a local authority, for example. The unit of analysis can either be the individual or the organisation, or even a faction or bureau within an organisation. In practice, networks in public policy are both individual and organisational, comprised of individuals whose relationships with each other comprise organisational roles. As a result it might be difficult to tie down exactly what a network is. Given the problems of definition, networks appear as a fuzzy concept and

encompass all kinds of human interaction. As a result, networks can appear too amorphous to have much bite when it comes to making a difference to policy outputs and outcomes. But the network idea can be useful in demarcating and understanding a set of relationships across organisations that are useful to achieve policy objectives and often vary in ways that are not necessarily correlated with other determinants of policy outcomes, such as wealth or organisational efficiency. What aspects of networks are potentially beneficial? One is the character

and quality of relationships between actors. This is the idea that a networked relationship can be strong or weak, with the idea that strong ties are better than weak ties (but this can work in different ways, as the chapter shows later on). Frequent ties can be effective means of information transfer, which can be efficient for all sorts of reasons, such as giving useful information over the implementation of public policy, for example details of potential problems or ideas about more effective means for delivering policies, or the necessary information to avoid the duplication of activities or the carrying out of inconsistent ones. The strength of contacts can be measured by frequency, but better is the sense of bonding and expectation of reciprocity

within networks, which may be related to frequency of contact, but is much more than that. The claim is that networks are associated with trust, which comes from the investment people and organisations put into networks and the experiences people have of each other as they interact over time. Small acts of cooperation can build up into the expectation that people will cooperate in the future. Trust can emerge from someone liking the person or organisation. Of course, it need not always work as expected, as problems emerge in partnerships, and lack of trust may beget lack of trust. A close network can lead to hatred and distrust, especially if relationships were good in the past but deteriorate because of misunderstandings and self-regarding actions that benefit one party but not the other. The other aspect of networks that can affect outcomes is the structure of

the network. This is the idea that networks have different shapes, such as having a central point through which all or much information may flow, or may be divided into different sub-groups or cliques, or be weakly connected with large numbers of isolated points. It may be the case the networks with central points may overcome coordination problems because of the ability of one or a few actors to be the link points, or it may be the case the networks have efficient links between the different elements, so information transmits efficiently across the network. There are distinct if complementary ways to understand networks. One

account is informal, where the network idea is a metaphor for understanding relationships among policy actors, and is often used in the policy network school of political analysis, concerned with describing the character of relationships within policy sectors, such as education or health (Marsh and Rhodes, 1992). Here the network can be seen as relatively open or closed, with consequent impacts on policy-making, but there is not much discussion of the structure of the network other than these features, one leading to closed decisions controlled by a small elite and producer groups, another more open, with many groups and open access to the media. There is a discussion of the power relationships in this example but with little discussion of the consequences for the quality of policy-making and the outcomes that are produced. The other way of thinking about networks is through the mapping of the

linkages by precise measurement of each contact or relationship, such as whether there is a contact or not, or noting the frequency or quality of contacts. It is possible to produce statistics that measure the structure of the network by counting the number of contacts, how many there are in relation to the total possible ties (their density), the ease of getting from one part of the network to the other (betweenness) and whether certain actors are at central points and have power because of that. This is sometimes known as formal network analysis or social network analysis (see Scott, 2000 for a review), which is a mixture of mathematics and empirical sociology. The mathematics is about the theoretical foundations and measures using this kind of data, what is called graph theory; the sociology is about most of the applications,

such as to friendships, family or employment relationships. The study of political and policy networks is very much an extension to the large literature rather than at its core. Most of the politics and policy applications are largely descriptive or concerned with the political outputs of the networks. The large sociological literature does examine the impact of networks on policy outcomes in health and employment, based on the idea networks are facilitative, helping provide support and useful information, in dense helping networks or through weak ties, which are useful for accessing information in other less proximate networks. This chapter is not directly concerned with these kinds of networks, but they do show the importance of networks and how measuring them in a formal way can reveal how the structure of the network can impact on policy outcomes. The question is whether the robust results from social networks in the wider society at large generalise to policy networks or whether governing networks are different in character. Network accounts of politics and policy have been criticised on many

grounds. This chapter has already alluded to the difficulties of defining networks. It is often not clear what they measure, as the term network is amorphous in involving contact, friendship, professional allegiance, information sharing or jointly working together, either one of several or even all of these. Not only is there a problem of definition, but measurement presents difficulties. Many of the standard ways of measuring networks, such as a questionnaire recording contacts between elites or looking at the crossmembership of governing boards, may measure a particular kind of contact, say for meetings or for a particular event, but may not pick up the other aspects of the network that may be as or even more important. These measures may pick up something that is unstable and changes over time rather than the long-term relationship inherent in the network concept. Thus researchers know policy networks change rapidly, which might suggest these relationships are purely functional, needed to get business done rather than being long term and sustaining in time. Then there are some well-known measurement problems with networks, which mean instruments based on respondents recalling their contacts tend to exaggerate the ties, sometimes by a factor of about two, when compared to observational data (Barnard et al., 1980, 1985). As a result, it is not possible to rely on the measures used in most forms of network research. If this is the case, then most of the studies purporting to claim an impact of networks are based on shaky foundations. The bigger problem with studies of networks is their account of causation.

The policy network literature claims networks have a causal impact on outcomes, such as power within networks and the kinds of policies that emerge (Marsh and Rhodes, 1992). This kind of approach has come in for a heavy attack that the networks do not have such an impact (Dowding, 1995). The argument is network structure may reflect power relationships so may not have causal properties at all. This would be consistent with the measurement problems alluded to above – that networks are so changeable is evidence they

emerge because of the importance elites and organisational structures place upon them rather than the impact of networks. Dowding is less critical of formal network analysis, which he considers to

provide the potential for understanding the causal effect of networks: ‘If the properties of networks are to be clearly identified as causally efficacious then only this type of technique will demonstrate this’ (1995: 158). But from working through some examples of the use of this technique, he concludes they either make little progress or tend to assume causal relationships, which are tautological. This results from the same problem – the network is the result of a power relationship determined outside the network rather than having causal properties itself. Overall the contribution of formal network analysis is at best modest, which may be part of the weakness of the method as well as the limited impact of networks themselves. If networks are limited in impact, it may not make much sense governments investing in them. The rest of the chapter will examine whether this statement is true. As with the other chapters, such as with institutions, there are two elements

to understanding how the tool works – one is to understand the causal relationship between the process and the outcomes; the other is the creation of the tool itself, which may involve using other tools in forming stronger networks, which may be difficult to achieve. So with networks the chapter reviews the evidence that links networks to outcomes, how to introduce better networks between organisations and whether changing networks has the desired effect.