ANALYSING POLICY NETWORKS

Abstract

This article argues that an ideational variant of David Marsh and Martin Smith's (2000) ‘dialectical’ model of policy networks adds significant value to their analysis by taking certain kinds of ideas, which Berman (1998) labels ‘programmatic beliefs’, seriously as independent, causal factors. It further contends that Marsh and Smith's approach moves beyond Marsh and Rod Rhodes's (1992) earlier structural model of interest group intermediation by focusing attention on both structure and agency and defends it against the rational choice critique advanced by Keith Dowding (2001). It observes, however, that Marsh and Smith's approach does not pay sufficient attention to the role of ideas in explaining policy-making, which should not be treated as mere rhetorical devices (Marsh & Smith, 2000), nor situated only at the macro, ideological level of paradigm shifts (Hay, 2004a). It argues instead for an approach to analysing policy-making that treats programmatic beliefs as independent variables, policy networks as intermediate variables, and policy outcomes as dependent variables. The article demonstrates the potential utility of such an approach by briefly examining David Toke and Marsh's (2003) analysis of policy change on the issue of GM crops in the UK, and suggests that in order to answer the question of why as well as how a policy was developed, it is often necessary to examine the programmatic beliefs motivating members of a policy network.