The aim of this study is twofold. First, it sets out to structure the debate on policy networks and show that the network concept is useful in the study of public policy. Second, it studies the differences in agri-environmental policies and agricultural policy reforms in Sweden, Denmark and the European Community. The main thesis is that network analysis can help to understand why policy makers make different policy choices.

The main finding of the study is that fundamental policy changes in an old, wellestablished sector is most likely to occur when the existing sectoral policy network is not cohesive, when the organizational structure of the state centralizes authority, and when parliamentary support of the group subject to policy change is limited. In situations where the sectoral network is cohesive, where the state disperses and where the group facing pressure for change has a high degree of support in parliament, policy change is likely to be modest or only marginal.

Before confronting the empirical findings of the different policy reforms with the policy network concepts, Daugbjerg claims that it is necessary to give a critical overview of the literature on policy networks. This makes sense, since the literature on policy networks is indeed very heterogeneous and there is hardly any consensus on how to apply the network approach to the study of policy making. Daugbjerg's critical review of the literature on policy networks is unfortunately very much more centered around networks of policy network authors (the group interaction approach, personal interaction approach, formal network approach, structural approach) than on analytical concepts. This can be a useful approach, since not only the traditions but also the backgrounds in the traditions of policy network scholars can be alluded to. I believe, however, that in the case of this study it would probably have been more useful to distinguish the literature as follows. First, by those who see policy networks as a (new) form of governance (compared, for example, to market and hierarchy) and thus define and conceptualize the phenomenon as well as argue that it constitutes a significant form of coordination in modern societies. Second, by those who ask questions about the conditions for the development and differences in structures and processes of policy networks. And third, by those who concentrate on the consequences of policy networks. If Daugbjerg had used this conceptual framework to analyze the policy network literature, he could not only have avoided the presentation of 'paper war' (Bruno Latour), but could also have avoided misinterpretations such as presenting the 'formal network approach' as a specific 'approach' instead of seeing it as a necessary instrument to empirically analyze the characteristics of networks. Moreover, the author's own approach, which is the study of the consequences of policy networks, would have stood out much more clearly. For example, it would have become apparent that this study is in the first place designed as a contribution to public policy analysis rather than as a

contribution to a theory of policy networks. The reason being that Daugbjerg's study in the first place tries to explain policy making and not policy networks. 'Policy network theories' explain variations in the development, structure and process of policy networks in the same way as organization theories for organizations and market theories for markets. Consequently, what one would have expected in the theoretical part of this study is a chapter comparing theories of policy making and an argument why, from a theoretical point of view, policy networks can be expected to explain policy making better than other approaches.

Chapter 3 ('Moving Beyond the Meso-level Analysis') sets out to specify a number of complementary factors explaining policy making. It is argued that parliament as well as state structure influence policy making and policy reforms. Rather than presenting these merely as alternative explanations to the policy network explanation, states that '[t]hey have an impact on the creation of meso-level policy networks and on the choice of sectoral or sub-sectoral policies.' Including the role of macro-variables in any study of policy making does, of course, make sense theoretically. What remains unclear is why these, and not other macro-variables have been chosen and what the relationship is between these macro-level and meso-level variables. The first deficit has to do, again, with the fact that a theoretical chapter on policy making is missing in the book. The issue about the relationship between macro and meso-variables will be addressed in the final part of this review.

In Chapter 4 ('Classifying Policy Reforms and New Policies'), Daugbjerg discusses and ways to conceptualize his dependent variable. This is an important chapter, since it reviews the policy literature to find criteria by which the three case studies can be classified.

The strength of the book is Chapter 5 ('Danish and Swedish Nitrate Policies') and Chapter 6 ('Agricultural Policy Reforms in the European Community and Sweden'). Chapter 5 describes Danish and Swedish nitrate policies and analyzes whether these policies differ fundamentally from the nature of established agricultural Chapter 6 addresses the question of whether agricultural policy reforms in the EC and Sweden were fundamental policy changes or merely adjustments existing policies. These chapters contain a lot of interesting information, at least as judged by somebody who is not an expert in the field of agricultural policy. The first conclusion of the chapters is that the Swedish policy is a high cost policy and the Danish policy is a low cost policy. This means that Swedish farmers, in contrast to Danish fanners, had to accept that the polluter pays principle was applied more strictly. The second conclusion is that the agricultural policy reform in the EC was much more moderate than it was in Sweden. These conclusions seem convincing, judged by the evidence presented in the two chapters.

Chapter 7 ('Cohesion of Agricultural Policy Networks') and Chapter 8 ('Farmers' Structural Power in Parliament and State Structures') present the values of the explanatory factors of the theoretical model for the three case studies. Here, it is first illustrated that both the Danish and EC agricultural policy networks are more cohesive than the Swedish network. Daugbjerg defines cohesion on the basis of network membership, the degree of integration and institutionalization of the networks. Second, it is demonstrated that the structural power of farmers in the Swedish parliament has decreased considerably since the early 1930s and that, in contrast, Danish farmers' structural parliamentary power has eroded only slightly. As far as differences in organizational structures of the state, it is demonstrated that Sweden is more centralized than the EC, but that it is not possible to show any significant difference in the degree of centralization when Denmark and Sweden are compared.

On the basis of all this information, Daugbjerg concludes that network analysis can help to explain why agri-environmental policy reforms differ. According to my reading, this conclusion should, however, for a couple of reasons be qualified. One reason being that the network approach has been applied only minimally. Instead of analyzing only the role of cohesion, many other network characteristics could have been taken into consideration (see, for example, Provan & Milward 1995 for such a comprehensive approach). The other reason is that Daugbjerg himself qualifies his main finding (i.e., that policy network analysis explains policies) by introducing two other explanatory variables. The commonplace statement that networks cannot be seen in isolation from the macro-context in which they are embedded does not really help here. Analytically speaking, linking macro and meso in an analysis can be done in a number of ways: first, macro-variables are introduced in order to explain the specific structure and development of a policy network across countries. In this case, they would explain policy networks and not policy outcomes which are not the issue in this research. Second, macro-variables are introduced to show that they explain less, compared to network variables (consider also the possibility that one function of networks could be to hinder the impact of macro-effects). Third, macro-variables are introduced to show that policy outcomes are the result of the complex interdependency the meso and macro-levels (for such a promising approach, see Döhler & Manow 1997). Finally, one could, as Daugbjerg does, consider that policy outcomes are best explained by aggregating the effects of macro and meso-variables. Given this approach and given Daugbjerg's statement that '[fjuture research should also attempt to establish the relative explanatory power of each of these variables' (p. 190), his finding, i.e. that policy networks play a central role in his explanation, should be qualified. The 'Defence of Policy Network Analysis' which is formulated in Chapter 9, the final chapter, does conceptually make sense, given the argument that policies could theoretically be explained by network variables. However, the empirical case on which the argument is based seems to me less convincing.

Despite some redundancy and occasional mistakes (e.g. '. . . Denmark applies a high cost nitrate policy . . .' (p. 186)), the book is, generally speaking, clearly written and well structured. And although the study does not fulfil the full promise of demonstrating the explanatory strength of policy networks, it is an interesting comparative study about agri-policy reforms and an interesting discussion about the policy network literature.

Department of Political Science and Public Administration

Free University, Amsterdam

REFERENCES

Döhler, M. & Manow, P. 1997. Strukturbildung von Politikfeldem. Opladen: Leske &
Budrich.

Provan, K. G. & Milward, H. B. 1995. 'A Preliminary Theory oflnterorganizational Network
Effectiveness,' Administrative Science Quarterly 2, 1-33.