An analysis of Lowi's policy network approach

Both the policy network approach and that employed by Lowi seek to provide general theories to explain how the policy-making process works. Traditionally, political science has focused on the role of formal institutions and procedures, such as political parties and the civil service. However, in the 1960s, several new approaches challenged this emphasis on the formal channels of participation and questioned the assumption that one could devise a general theory equally applicable to all policy areas. The policy network approach emerged which recognised a need to focus upon particular policy arenas because the relationships between actors will be different from one arena to another. Lowi's approach, by contrast, stressed the need to identify the issue at stake as this will determine the nature of the policy process; in short, the view that 'policy determines politics'.

Before comparing and evaluating these two approaches, I believe it is first necessary to note the fundamental problems any attempt to build a theory of the policy-making process will encounter. As Greenberg et al point out, 'public policy' is a notoriously difficult area to analyse. Firstly, the policy process takes place over a long time with many decision points and constantly changing values and therefore cannot easily be explained as a "simple unit". Moreover, the multiplicity of outcomes usually resulting from a policy proposal makes it difficult to categorise as does the frequently large number of participants with differing perceptions. Thus, the fact that public policy is "almost never a single, discrete, unitary phenomenon" (Greenberg et al, 1977, p1533) renders any attempt to formulate an overall theory of the way the policy process operates fraught with difficulty.

Bearing in mind these problems, the following analysis will begin by briefly outlining the policy network approach and the approach devised by Lowi. In doing this, I will try to draw attention to similarities and differences between the two. Next, I shall consider the many shortcomings of both approaches before reaching a conclusion on which approach is the most fruitful. I shall argue that, although undoubtedly in need of refinement and, in particular, sharper definition of its principal issue categories, Lowi's approach is potentially more useful if one is seeking to explain a general pattern in the policy-making process. The policy network approach suffers from serious weaknesses in that it is frequently vague and its central propositions are so wide and variable that, rather than a precise theory, it is often merely descriptive.

The policy network approach emerged in the 1960s as the focus shifted to the influence of pressure groups and informal actors in the policy process. Although there are various strands of thought and differing terminology, a central core of meaning exists within this approach. Rhodes & Marsh, for example, produced a typology in which 'policy network' is the generic term covering the many different types of network that exist from stable, highly-integrated policy communities with restricted membership to unstable, loosely-integrated issue networks with large numbers of participants (Rhodes & Marsh, 1992, pp182-188). Typically, Jordan and Richardson's argue that the true focus of decision-making is not the formal party or the Parliamentary arena but "the more hidden world of relations between government departments and interested pressure groups" (Greenaway et al, 1992, pp57-58).

Several general characteristics of the policy network approach can be identified. Firstly, it is believed that the policy-making process is segmented. Policies are formulated independently in each policy sector and the actors involved in the process are situated in a particular segment and interact with each other but not with those from other segments. As Richardson et al note, "each policy area develops into a semi-watertight compartment" ( 1982, p3) and the policy-making process is fragmented and power dispersed.

A policy network or community is usually centred around a government department and the actors involved will include both members of ...

This is a preview of the whole essay

A policy network or community is usually centred around a government department and the actors involved will include both members of that department and organised interests. Membership, however, is not the same for all networks as different types of network are said to possess their own characteristics, thus a policy community on defence will have highly restricted membership whilst that of education will be extensive. Similarly, the influence that various groups can expect to exert will vary according the particular nature of the network.

A high value is placed on consensus within a community as its autonomy from influences outside the segment minimises controversy, protects autonomy and leads to an ethic of mutual concessions or, as Marin & Mayntz, put it "antagonistic co-operation" (1991, p17). In most policy networks established rules for formulating policy exist and the process is rather routine. Moreover, change tends to be incremental rather than involving radical overhauls.

In contrast to the policy network approach which was developed largely as a consequence of the rise of pluralist thought, Lowi's model was developed mainly due to his dissatisfaction with the pluralist model which he regarded as having serious "logical and empirical weaknesses" (Lowi, 1963-64, pp681-2). Lowi's model is based on the two premises that the types of relationships amongst people are determined by their expectations and, in politics, expectations are determined by governmental outputs or policies. Thus, he argues that "a political relationship is determined by the type of policy at stake, so that for every type of policy there is likely to be a distinctive type of relationship" (Lowi, 1963-63, p688). This can be reduced to the simple assertion that policy determined politics.

The major analytical problem for Lowi then is to identify possible types of outputs or policies. He does this be defining categories of policy in terms of their impact, or expected impact, on society and, by this method, arrives at three, although a fourth is added later, functional categories; namely, distributive, regulatory and redistributive.

The distributive category deals with government decisions which, in the short-term, can be made without regard to limited resources; that is, decisions in which the costs of conferring benefits on a certain group are so widely dispersed that they appear to be negligible. Examples of distibutive decisions would be the awarding of certain government contracts such as road-building or defence orders. The politics operating in this arena is characterised by stability and a low level of controversy as the disaggregation of issues individualises and therefore minimises conflict. The regulatory arena involves a direct choice of who is to be indulged and deprived. Decisions cannot be disaggregated to the individual level because the regulatory rules must be generally applied and therefore are usually only disaggregable down to the sector level and so there is likely to be greater conflict between groups and greater instability. The final category employed by Lowi is that of redistributive issues. These issues affect broad categories of individuals in either a positive or negative way. Interests tend to be ranged along class lines in a highly stabilised political structure.

I shall now evaluate the two competing approaches, beginning with the policy network approach. As already argued, one of the major weaknesses of this theory is its vagueness. For example, Rhodes & Marsh identify the characteristics of what they believe to be the two extreme types of policy network, namely the policy community and the issue network. They themselves acknowledge that this only sets the "outer limits of the analysis" and that there will be an infinite number of intermediary cases ( Rhodes & Marsh, 1992, p188). Consequently, the model is so vague and undetailed that it can be used to explain all situations and, as such, is unfalsifiable and of little theoretical value. It is perhaps better therefore to regard the policy network approach as a tool, useful for guiding empirical studies towards the informal relationships involved, rather than a model for explaining such relationships.

Another failing of the policy network approach is the underdevelopment of a comparative framework. The approach tends to be descriptive rather than comparative and, as it was developed principally in Britain, no comparative cross-national element has been incorporated. Furthermore, the emphasis that the model places on consensus and incremental change can be questioned. The approach's inherent assumption that governments prefer consensus to control was seriously challenged with the election of the more ideological, conflictual Thatcher government in 1979 and, as Greenaway et al point out citing the example of local authorities relationship with central government, the policy network approach completely underestimates the extent to which central government can control or bypass outside interests should it wish to (1992, p61). Moreover, the model has no explanation for major change which is unarguably occurring, instead it points to continuity and only minor change. The influence of the external environment is also neglected,. despite the fact that change is often generated by external factors such as currency stability.

Finally, it can be argued that a major defect of the policy network approach lies in its assertion that the policy- making process is divided into autonomous segments. Hogwood, for example, argues that the approach overemphasises the degree to which a policy is contained in one particular area (cited by Greenaway et al, 1992, p61). Segments may well be interconnected in some cases, the issue of budgetary politics in which spending departments and the treasury are effectively in competition and thus integrated being a good example. Furthermore, the approach ignore the existence of a hierarchy between segments and the fact that a decision made by one department may easily affect a 'less important' department. Similarly, the approach gives no general guidelines as to the size of the segments involved or the degree of specialisation of groups and officials. The stress on the sectorisation of policy-making then by this approach would appear to be problematic given the political realities.

I shall now consider the major weaknesses of Lowi's approach. Firstly, there is the problem of 'cut-off'. In order to test a policy under Lowi's model, one must be able to place it into a category and then see whether the pattern of politics corresponds with that predicted. However, the pattern surrounding an issue is likely to be continually changing and therefore the question of whether Lowi's predictions hold true is likely to depend upon where the point in time that one chooses to test.

Possibly the most fundamental problem encountered by this approach though is that of mixed policies, making it virtually impossible to place them with any degree of certainty into one of the three categories. Many policies may include elements of redistribution and regulation: James Wilson for example cites the case of urban renewal programmes which might equally regulate the use of land, redistribute the housing supply and distribute benefits to certain contractors. However, even though the allocation of a policy to the correct issue type is central to the predictive ability of his theory, Lowi gives, as Greenberg et al note, "little guidance for determining how a policy is to be classified in any but the simplest cases" (1977, ppp1534-5). As Lowi is not explicit in explaining how to determine the correct issue type, it is virtually impossible to empirically test the hypotheses he presents.

The only answer Lowi has to this problem is to argue that an a policy should be classified according to people's expectations of its impact upon past experience. However, this only works when expectations are unanimous which is unlikely, and gives no guidance as to classifying innovative policies as there is no experience upon which to base expectations (Greenberg et al, 1977, p1535).

Another problem faced by Lowi's approach is that it is based on subjective not objective perceptions of a policy. It is quite possible that the same policy may be viewed quite differently by two different societies. It could be argued that the institutional and cultural nature of a society determines perceptions which then determine policy categorisation and the politics involved. If this is accepted then Lowi's model loses one of its main attractions, its supposed objectivity.

Lowi's approach has been criticised on several other counts. For example, Alexander Smith argues that participation and conflict is widespread on symbolic and emotive issues such as national identity and not merely on redistributive issues. Lowi also appears to turn his back on simpler, more concrete description and explanation of lower levels of the policy process and, in so doing, perhaps ignores the peculiarities of particular issues. No obvious distinction between various phases of the policy process or between urgent and non-urgent issues is incorporated into Lowi's model. However, I would regard these minor problems which do not in themselves pose a fundamental threat to the approach.

Both the policy network approach and that employed by Lowi suffer from serious weaknesses and neither, in their present forms, provide a convincing explanation for the workings of the policy process. The policy network approach is so blurred and vague that it covers all possible scenarios and therefore ends up being descriptive rather than explanatory. Although it provides a useful set of criteria by which the policy-making process should be examined, namely membership, integration, resources and power, it fails to predict how these criteria will manifest themselves in particular instances. I would suggest therefore that the policy network approach, although failing to establish an explanatory model, can be used fruitfully as a guide to carrying out empirical case-studies. In contrast, Lowi has attempted to develop a general model of the policy-making process and, by identifying three categories of issue, each related to a distinctive type of politics, has incorporated substantial predictive and explanatory elements. However, it is virtually impossible to falsify or validate Lowi's theory for, as Greenberg et al argue, "As it stands, Lowi's theory is not testable because the basic concepts are not operationalisable....researchers must make a number of guesses and assumptions that create a situation where they can no longer be sure just whose theory they are testing" (1977, p1536). As such, Lowi's theory tends to have more admirers than followers. Although a worthy attempt to create an explanatory rather than a merely descriptive model, the weakness of definition and guidance in using the central categories means that, in its current state, the theory cannot be judged in practice. However, despite its shortcomings, I believe that Lowi's model is the best explanatory theory of the policy-making process that has been developed so far.

Bibliography

J. Greenaway, S. Smith & J. Street, Deciding Factors in British Politics, 1992, Routledge

G. D. Greenberg, J. A. Miller, L. B. Mohr & B. C. Vladeck, "Developing Public Policy Theory, Perspectives from Empirical Research", American Political Science Review No. 71, 1977

A. Grant Jordan, "Iron Triangles, Woolly Corporatism, or Elastic Nets", Journal of Public Policy Vol. 1 Part 1, 1981

T. J. Lowi, "American Business, Public Policy, Case Studies and Political Theory", World Politics Vol. 16, 1963-64

P. Kenis & V. Schneider, "Policy Networks and Policy Analysis", in R. Marin & R. Mayntz eds., Policy Networks, 1991, Westview Press

R. A. W. Rhodes & D. Marsh, "New Directions in the Study of Policy Networks", European Journal of Political Research Vol. 21 Nos. 1-2, 1992

J. Richardson, G. Gustafsson & G. Jordan, "The Concept of Policy Style", in J. Richardson ed., Policy Styles in Western Europe, 1982, George Allen & Unwin