Why David sometimes defeats Goliath - The power of actors in disprivileged land-use policy networks

Since the 19th century, Danish policies regarding land use in the wider countryside have been strongly influenced by two competing policy networks. In the period 1866-1970, a strong ‘land reclamation network’ was privileged at the expense of a weak ‘nature protection network’. In contrast, during the last couple of decades, the situation has been turned on its head. Here, a strong nature protection network has been privileged at the expense of a weaker ‘land reclamation network’. In each of the above stable periods, however, a detailed case study of decision-making regarding a rejected project shows that sometimes the weaker network is able to defeat a stronger network in relation to concrete project decisions regarding the use of water bodies. Why is it that even very strong policy communities cannot expect to have all projects they support implemented? Adding the concepts of policy venue and policy image (Baumgartner & Jones 1993) to the rather static policy network approach provides an explanation – a theoretical framework which, for example, can be applied to decisions/non-decisions regarding nature restoration projects in those countries who during recent years have been trying to move beyond water body protection to restoration of water bodies to their natural state.


Introduction
Denmark ranks among the world's most cultivated countries. 63 pct. of the land area is registered as 'cultivated' and only 9 pct. as 'nature' (Hansen 2003: 36, 313).
Ironically, Denmark's long history of land reclamation is one reason why Denmark today is experiencing widespread nature restoration; e.g. with implementation of the Skjern River restoration project in the 1990s -one of Northern Europe's earliest and largest, nature restoration projects.
Denmark's uniform, cultivated landscape is partly the result of numerous land reclamation projects. Many of them implemented during the period between the Danish defeat in the second Schleswig War (1864) and the establishment of the Ministry of Environment (1971). Through a number of land reclamation projects, rivers were straightened, and lakes, moors and fjords drained with the aim of claiming fertile farmland. Similar comprehensive land reclamation policies are observable in e.g. the Netherlands, Germany, and the UK (Blackbourn 2006).
The Danish land-use policy regarding rivers, lakes, wetlands, etc. represented the scene of a struggle between two different policy networks. The prevailing land reclamation policy  was a result of the presence of a very strong land reclamation policy community, built after the defeat in the second Schleswig War with one overriding goal: to provide as much farmland as possible (Pedersen 2006(Pedersen , 2009). Actors in this policy community, of which the Ministry of Agriculture was an important member, were privileged throughout the period at the expense of actors in 3 another sectoral policy networkthe weak, disprivileged nature protection network built up towards the beginning of the 20th century, and with ministries responsible for nature protection among the actors. However, in spite of being the weaker of the two, the disprivileged network succeeded in fighting a small number of projectse.g. the Naeraa Beach land reclamation project in the 1940s, defeating here the privileged network.
Very strong policy networks (Marsh & Rhodes 1992) may succeed in living in isolation from attention from other networks, but often other policy networks are an important aspect of the context of a network -'the relationship between networks is clearly crucial' (Marsh 1998: 191)which is demonstrated e.g. in Smith's (1991) analysis of a clash between an agriculture network and a health network. When two, competing, sectoral policy networks are fighting over the same privilegese.g. preferred type of land useone would expect the stronger of the two to become privileged by the prevailing policy, and the weaker of the two to become disprivileged. In other words, the policy will normally favour the stronger of the two networks. It is argued, however, in this article that when concrete project decisions within a broader policy are under consideration, surprisingly, the weak are sometimes able to defeat the strong, either by changing the policy venue or by threatening to build up a new competing policy image. Apparently, the weak network can sometimes prove to be the more powerful of the two networks regarding specific project decisions, in sheer defiance of the fact that the general policy privileges the stronger of the two. The above argument is examined through detailed case studies of decisionmaking processes in two Danish land-use projects in the wider countrysideprojects which were both rejected. The Naeraa Beach case study is from the 'land reclamation 4 period'  and the Lake Soeborg case study is from the 'nature restoration period ' (1985-2006).
It is important to widen our knowledge of why some land-use projects are approved and others are not. It is argued that the proposed theoretical framework can be applied to land-use decisions/non-decisions in other countries (see e.g. Lowry 2003: 240ff)e.g. countries like the Netherlands, UK and Germany, where past land reclamation policy is rather similar to that in Denmark, and who are also trying today to move beyond protecting water bodies to restoring water bodies to their natural state.
Therefore, it is relevant to uncover why even very strong land-use policy communities cannot expect to have all projects they support implemented?

The power of policy networks
A characteristic of Danish policies regarding land use in the wider countrysidee.g. the land reclamation policy and the nature restoration policyis that, since the 1860s, interest groups, supported by government departments, have exercised a striking influence on the decisions involved, making it relevant to analyse the policy processes within a policy network framework (Marsh & Rhodes 1992). The approach is generally recognised for its descriptive qualities in periods with policy stability, thus making the approach applicable to the two stable periods described, in which the Naeraa Beach project and the Lake Soeborg project, respectively, were considered and surprisingly rejected.
The policy network approach addresses the policy consequences of different organisational arrangements and argues that different sectoral policy network types can be linked with differences in sectoral policiespolicy networks influence policy outcome (Marsh & Rhodes 1992;Daugbjerg 1998).
Marsh & Rhodes presented the idea of a network continuum based on the degree of integration exhibited by the individual policy network. At one end of the continuum we find the weak 'issue networks'. Issue networks are relatively open and they contain many actors who disagree on policy principles and procedures to approach policy problems. Furthermore, the pattern of interaction in the issue network is unstable and solely based on consultation among the actors. At the other end of the 5 continuum, the strong 'policy communities' are situated. Policy communities are tight, closed networks with a very limited number of actors. All actors in the network are in possession of valued resources (e.g. money, expertise, etc.) and there is resource interdependence among the actors. Actors in a policy community bargain mutually, and they have a consensus regarding policy principles and procedures to approach policy problems (Daugbjerg 1998: 44;Marsh & Rhodes 1992).
Both policy communities and issue networks operate under some structural constraints (Smith 1990: 34). However, the extent to which actors are privileged by established policy networks largely depends upon the character of the networks involved (Marsh & Rhodes 1992;Daugbjerg & Pedersen 2004: 223). Policy communities create power relations which privilege the actors organised in the network. A policy community will exercise structural power, and make rules, procedures and beliefs that support the interests of the actors in the policy community.
If a policy community is challenged by outsiderse.g. other policy networks (Marsh 1998: 191) it is normally rather easy for actors in the policy community to agree on a common position and counter the challenge (Daugbjerg & Pedersen 2004: 223).
In contrast, an issue network possesses no (or very little) structural power and actors in issue networks will therefore have to rely on possible, individual, non-structural power (resources) if they seek to make an impact on the policy in question (Smith 1990;Daugbjerg 1998). Therefore, if a policy area is characterised by competition between a strong and a weak policy network, it can be expected that the policy is privileging the strong network, and disprivileging the weak network.

Policy networks and variation within a general policy
Variability in concrete decisions within a general policy is an area of policy network research which is not thoroughly developed. However, Lowry (2003) has been focusing on this topic using Sabatier & Jenkins-Smith's Advocacy Coalition Approach (1993). Lowry analyses US restoration policies regarding dams, and observes that within general US river restoration policy, variability is detectable regarding decisions concerning implementation of individual river restoration projects. According to Lowry, two variables determine outcomes in specific US river 6 restoration proposals: i) the political receptivity to changes and ii) the physical complexity of potential changes (Lowry 2003: 8-27).
Lowry states that the degree of physical complexity depends on scale (spatial extent, number of political jurisdictions affected) and dimensionality (multidimensional vs. unidimensional). The political receptivity to change depends on whether or not the decision-making venue is tolerant of change; whether the costs of the status quo are high and readily apparent; and whether scientific information on potential benefits is widely embraced. When the anticipated physical complexity of changes is low and the anticipated political receptivity to changes is high, fundamental changes are possible (Lowry 2003: 8-27). However, it seems that Lowry's US framework is not directly applicable in the context of Danish land reclamation projects and nature restoration projects.
Concerning physical complexity, it is observable that several radical land reclamation projects and radical nature restoration projects with a large spatial extent have been implemented. For example, Denmark's largest river in terms of volume -Skjern Riverwas straightened in the 1960s to win farmland, and restored in the 1990s to attempt to recreate the natural conditions of the river valley (Pedersen 2006;2009). Furthermore, dimensionality of the project does not appear to be variable to any significant degree, although some variation is observable regarding the number of political jurisdictions affected. However, it is argued below that the difference in policy image between the political jurisdictions is more important than the number of political jurisdictions.
Regarding the political receptivity to change, the argument here is that there are always interests/networks which are not receptive to policy change (unless they are paid to be receptive). Development of a real mutual understanding (which is not the result of power/influence being exercised) between very different interests, e.g. the mutual understanding Lowry (2003: 228-29) describes in connection with the Neuse River project, is exceptional. On the contrary, a privileged network in any given period will normally try forcing an implementation of all land-use projects favouring the network, while a disprivileged network will try to avoid implementation of the projects -still '... conflict, competition, leadership, and organization are the essence 7 of democratic politics' (Schattschneider 1975(Schattschneider (1960. Surprisingly, as the Danish cases show, it is the disprivileged network that sometimes is successful in this struggle. Therefore, analysing the Danish cases within an Advocacy Coalition Framework (ACF), as in Lowry's US study, would be problematic, because the ACF focuses on belief systems, policy-oriented learning and the triumph of better policy ideas over inferior ideas in a relatively open process (Smith 2000). In contradiction, the fight over Danish nature is centred around a raw power struggle, where actors are being excluded from being heard and where resources and resource interdependence seems to be very important, pointing to use of a policy network framework rather than the ACF. Furthermore, stable policy periods are analysed below, which also supports use of a policy network framework, because the ACF is developed to study policy change over the longer term (Smith 2000).
Here it is argued that variability within the local context might help the disprivileged of the two networks. First, with regard to a specific project it is possible to stop the project if there is the possibility of a venue change to a local/regional venue (or 'political jurisdiction'), with a policy image supporting the arguments of the disprivileged network. Policy venues are groups in a society that possess authority to make decisions concerning an issue (Baumgartner & Jones 1993: 31). During the 1980s and 1990s disprivileged actors, in general, used the possibility of venue change more frequently (Richardson 2000). On the contrary, a privileged network tries to avoid venue change. Second, another possibility for the disprivileged network is to threaten to build up a negative policy image in the media in connection with a given project. Policy image is 'how a policy is understood and discussed' (Baumgartner & Jones 1993: 25-26). Specialists/experts are able to explain issues in simplified, symbolic and favourable terms to non-specialists (the elite and the public) with only a passing interest in the area. Often, there are competing images connected with an issue. Inspired by Lowry's focus on costs and scientific information, it can be argued that if the project in question is very expensive compared to similar projects, or if scientific facts exist regarding severe negative effects of the project, the potential then exists for a negative policy image to be built up if the project is planned to be implemented regardless. A new negative policy image can, potentially, in the longer term, threaten to change the entire general policy that privileges the powerful network 8 (Baumgartner & Jones 1993;Richardson 2000;Pedersen 2006), and can make the privileged network refrain from implementing the project in question.
The concepts of 'policy venue' and 'policy image' are constituent parts of Baumgartner & Jones' (1993) punctuated equilibrium model, but the concepts might benefit the policy network approach too (Richardson 2000;Pedersen 2009). Models of how actors in groups/networks try to change the power relations through agendasetting processes have the potential of adding dynamic to the policy network approach, which has been criticised for its lack of explanatory force (e.g. Dowding 1995). Below, it is argued that even though a tight policy network constrains the policy agenda, some freedom to manoeuvre exists for the disprivileged network regarding concrete project decisions.

Method
One possible approach when analysing land reclamation and nature restoration is to characterise all involved actors as being part of one big (land-use) issue network.
However, e.g. Daugbjerg & Pedersen (2004: 222) show, when analysing policy networks, that an approach identifying two competing policy networks might be more dynamic (see also Smith 1991). Marsh (1998) admits that Marsh & Rhodes (1992) neglected the important fact that policy networks have a mutual impact on each other. Marsh (1998: 191) states: 'Individual policy networks don't exist in isolation; although tight networks may succeed in isolating the interactions in their network from the attention of other networks'. Policy networks potentially have the power to influence the character of other networks, and/or influence other networks' policy outcome. By identifying two networks below, it will be demonstrated that even tight policy networks may from time to time be influenced when challenged by apparently weaker policy network types. By accepting to lose a number of individual project decisions in the short term, tight policy networks may prevent their network disintegrating in the longer term.
The two cases -Naeraa Beach and Lake Soeborgare selected due to their extreme or unique character (Yin 1994: 39). They are considered extreme because the output of the decision-making processes is surprising in both cases; the weaker networks defeat the stronger, privileged networks, despite the fact that the stronger networks 9 were able to implement a considerable number of other projects. This is interesting because the variability displayed here regarding decisions concerning land-use projects cannot be explained by changes in e.g. the broader context (socioeconomy, public opinion, government, etc.). Furthermore, both cases concern rather large-scale projects; a fact that made the privileged networks particularly interested in implementing these projects, making the policy output even more surprising. Through analysis of the two cases we might learn more about those rare cases, where weak networks beat strong networks. Finally, the two cases represent different periods (land reclamation 1866-1970; nature restoration 1985-2006); periods which are separated by radical policy changes, and periods where different networks can be seen to be privileged.
Needless to say, a multiple case design would have permitted more general conclusions. However, Peter John (2001: 148) observes that many policy network studies lack 'the feel of the policy process'. To get the feel of the policy process, it is necessary to dig deep down into the involved networks to trace the policy dynamics, thus imposing a limit on the number of cases able to be analysed in a journal article.
The case studies have primarily been performed through analyses of hitherto unpublished documentsarchive files on the two cases from the involved authorities having decision-making power. Regarding the land reclamation case, the material is primarily based on the very detailed meeting minutes from the Land Reclamation boards administrative procedure for the Naeraa Beach case, while the nature restoration case is primarily based on archive files on the Lake Soeborg case from the Danish Forest and Nature Agency. Regarding both cases, hundreds of policy documents have been read. This material is supplemented by an interview with a former Minister of Environment and Energy, Mr. Svend Auken. Both cases are also analysed with materials reported in Pedersen (2006) newspaper articles, parliamentary proceedings, other archive files, reports, etc. A good case study will use as many sources as possible (Yin 1994: 80).

The two competing policy networks in the land reclamation phase (1866-1970)
Two land use networks have been competing over the use of Danish rural territory: a land reclamation network vs. a nature protection network. For a long period 10 agricultural purposes were highly dominant, but, just as in other European countries, agricultural policies today are challenged by other sector policies e.g. rural policies and environmental policies (see Lowe 2006).
Following the Danish defeat in the Second Schleswig War (1864), where Southern Jutland was lost to Prussia and Austria, a private foundation called the Danish Heath Society was established by a group of influential men (primarily landowners). The purpose of the foundation being to 'reclaim' lost territory (farmland) through domestic land reclamation projects and, needless to say, as a way of earning an income for those involved in the projects. In the following decades, remaining heathland areas were cleared and ploughed by the Heath Society. Early, the Heath Society turned its attention to other tasks as well: dewatering meadowlands, draining lakes and moors, straightening rivers, etc. A strong parallel can be drawn to Maass' (1951)  The nature protection network was a very weak issue network. 13 It can be questioned whether fishery interests actually were part of the nature protection network or whether they formed their own network. However, placing fishery interests in the nature protection network can be defended based on the view that all actors in the nature protection network were connected by means of their interest in stopping the land reclamation projects by protecting the natural area which remained. It is obvious that due to it being a very weak network with almost no structural power and infrequent contact, in general, actors of the network often acted on their own, without support of fellow network actors. The most important lesson, however, to learn from the case study of Naeraa Beach below is that weak actors do sometimes have the possibility to defeat actors organised in very strong policy communities.

The Naeraa Beach land reclamation project
Naeraa Beach is a 685 hectare bay on the coast of the Danish island of Funen.
Following the enactment of the act on land reclamation (1940), it was soon proposed to reclaim Naeraa Beach. In February 1941, the Land Reclamation Board travelled to Naeraa Beach to meet with three local landowners who were interested in the project and representatives for local fishermen who were against the project.
The archive files of the Land Reclamation Board show that the theme of the meeting with the landowners was the expected benefits from such a project. There was a potential problem: even though the landowners affected would receive state grants to cover 67 pct. of the project's costs, there was a risk that the landowners' costs would exceed the landowners' benefits from the project. But, at the meeting, the landowners affected assured that they were willing to support the project despite this fact.
However, the Land Reclamation Board was permitted only to approve land reclamation projects where it was ensured that benefits corresponded to at least 33 pct. of project costs (Pedersen 2006: 291-92;LR Board). Therefore, the project was in fact illegitimate.
After the support was given, representatives for 14 local professional fishermen's organisations entered the meeting room. At the meeting, the fishermen protested against the project, claiming that Naeraa Beach was the region's most important 14 breeding ground for fish. The board members, however, explained to the fishermen representatives that it was the board's duty to increase employment and agricultural production, and weigh these benefits against possible costs, e.g. the loss of a fishery (Pedersen 2006: 292;LR Board). These arguments seem exaggerated, because, as mentioned above, expected benefits from the project were very small. Furthermore, the employment effect of land reclamation projects was, in general, insignificant and, unofficially, employment effects were not prioritised by the land reclamation board (see above). But, these facts were not mentioned to the fishermen. In spite of this, however, the protests on the part of the fishery continued, supported by the Directorate of Fisheries.
A couple of further meetings were held between the Land Reclamation Board and fishery interest groups (Pedersen 2006: 292-301; LR Board)the same arguments being repeated again and again. In addition, the Land Reclamation Board appealed to the fishermen's patriotic feelings, claiming that projects like the Naeraa Beach project were necessary due to the difficult situation the country was in, Denmark being occupied by Germany at the time. When this argument had no effect, the Land Reclamation Board proposed a 'horse trade': the fishermen should abandon protests against the Naeraa Beach project and in return be allowed to 'win' their protests against a planned Odense Fjord project. But the appeals did not soften the opponents; the fishermen refused to give up their protest, the sediments at Naeraa Beach representing a better breeding ground than those at Odense Fjord. However, these sediments would create good farmland as well, as far as the Land Reclamation Board were concerned. When appeals and horse trades did not work, the Land Reclamation Board warned the fisheries against a possible public reaction against the industry: ' ... it will be very difficult to explain to the Danish people that due to respect for the fishery we cannot reclaim a single fjord. We are interested in having as many people as possible When arriving at Naeraa Beach, they were rowed across the bay and in the clear water they noted that, in fact, there was a considerable amount of fry, which was also confirmed by some of the local residents. Most decisive for the opinion of the Permanent Under-Secretary, however, was the statement of two local lockkeepers he met on the trip, who confirmed the extraordinary numbers of fish at Naeraa Beach.
The Permanent Under-Secretary admitted, when reporting to the Land Reclamation Board, that he did not believe the initial reports from fishery experts, but that the two 16 lockkeepers represented agricultural interests and therefore represented a more trustworthy source (LR Board).
The Permanent Under-Secretary reported to the minister that the project was 'risky', because the benefits were not particularly high and an entire local sector was in danger of being destroyed. He concluded that the board would 'lose the Danish people's sympathy for land reclamation if the board drives too hard'. Other members of the Land Reclamation Board did not, however, agree. I.e. a board member who was also an MP and a farmer himself stated that it was difficult to assess the Naeraa Beach project, 'because the fishermen are protesting in a way that makes it difficult to believe them; therefore, an objective measure needs to be used' (LR Board). In July 1941, the Minister of Agriculture rejected the Naeraa Beach project. Odense Fjord (475 hectares), however, was 'reclaimed' and drained in the 1940s.
In conclusion, fishery interests from the disprivileged nature protection network were able to bring an end to a Naeraa Beach land reclamation project on their own, despite the fact that the strong policy communitythe land reclamation networkwas interested in implementing the project. A cocktail of scientific factsit was beyond doubt that Naeraa Beach was of high importance to the local fisheryand question marks surrounding the agricultural benefits of the project were, potentially, damaging to the land reclamation policy in this case, because fishery interests posed a threat of unrest that would evolve if the project was given the rubber stamp. The threat of a negative policy image connected to the general land reclamation policy of the time caused the land reclamation network to abstain from implementing the specific project in question. Surprisingly, the result described above occurred despite that fishery interests in this case were not supported by other actors in the very weak nature protection network. It is therefore even more impressive that it was possible to break the strong structural power of the land reclamation network. This supports Dowding's (1995: 146) critique of the policy network approach. One of Dowding's points is that the nature of politics is more about the actors' resources (in the Naeraa Beach case the fishery interest's knowledge/information) than their networks. However, even though the land reclamation network lost this case, thousands of other cases were won by the land 17 reclamation network; often supported by land reclamation network rhetoric that these projects, in general, were very important for employment, despite that figures show they were not (Pedersen 2006;2009). Apparently, support is found for both Marsh & Rhodes (1992) and Dowding (1995); policy communities influence policies, but sometimes/on rare occasions actors not supported by a strong policy community can conquer the policy community just by relying on actor resources. Nevertheless, in the Lake Soeborg nature restoration project, the weaker land reclamation network succeeded in the battle against its implementation.

The Lake Soeborg nature restoration project
Lake Soeborg was a 500 hectare lake north of Copenhagen which since the late 18th century had experienced several land reclamation projects. As a result of the last project in 1960, there was no longer any water remaining in the lake.
The Lake Soeborg restoration plan was first shelved in 1986 due to lack of resources in the Danish Forest and Nature Agency under the Ministry of Environment (DFN Agency; Pedersen 2006), but in connection with the local authority election campaigns in 1993, local Social Democratic politicians proposed a full-size nature restoration project at the 'lake'. One of the local Social Democrats had very good 19 relations with actors who were part of the nature restoration network, because he was chairman for two local angling organisations; he was furthermore a member of the Danish Outdoor Council, which supported the proposal. All other actors in the nature protection network supported the proposal too: the Danish Society for Nature Conservation, the Heath Society (who had swapped networks), the Danish Ornithological Society, and the Danish Forest and Nature Agency, all expecting a stunning nature area (Pedersen 2006: 316-324).
The Danish Forest and Nature Agency decided to file for protection of the area, which was one of two possible nature restoration instruments. Another possibility was to start buying land at the 'lake' and in the surrounding area and then, through purchase and exchange with the relevant landowners, acquire a state-owned project areaexpropriation was a possibility too, when using this latter instrument. Local attention was massive, and local farmers and affected landowners were not positive about the creation of a large lake on farmland -claiming e.g. that 'farmland nature' is of scenic value in itself, that a restoration project would result in a mosquito nuisance, and that the farmland to be lost was of high quality. Many critical voices were heard in the local newspaper and, on election day, the representation of the political parties supportive of the restoration project (e.g. the Social Democratic Party) was reduced in both the municipality and the county where Lake Soeborg is situated. Instead, the Liberal Party, generally supportive of farmers' viewpoints, gained power. This represented an important development, as winning the case for protection demanded that either the municipality or the county supported the proposal in the Nature Conservation Board. And as a result of the power structure neither did (Pedersen 2006: 316-324;DFN Agency).
Representatives from the Danish Forest and Nature Agency were rather bitter, claiming that the arguments of the old land reclamation network were based on feelings and not facts (DFN Agency). Again, the mutual understanding between actors organised in two competing networks is shown not to be very great.
The Minister of the Environment, a Social Democrat, regretted the decision.
However, he did have the option of using the other possible nature restoration instrument of state ownership, but desisted from using it. Today, the minister explains 20 that the case was too controversial due to the massive local resistance (interview Svend Auken 2004). Apparently, the case was so controversial that he feared damaging the positive policy image associated with nature restoration policy in general.
However, in 2001, another attempt was made by the nature protection network to bring about restoration of Lake Soeborg. A private foundation -Aage V. Jensen's Foundation, another actor in the nature protection networkwas interested in buying one of four large farms at the 'lake', with the purpose of implementing a small nature restoration project in this area. As part of a standard procedure, a local land-use commission was supposed to approve, that the designation for agricultural use of the area was abolished. These commissions were regional commissions consisting of three representatives: two representatives from agricultural organisations and one lawyer, each representative holding veto power. Surprisingly, as far as the Danish public were concerned, one of the agricultural representatives, a local pig farmer, voted 'no' to abandoning the agricultural designation due, in his opinion, to the area being excellent farmland; the two other representatives voted 'yes'. The use of the veto power presented a surpriseit was the first time in ten years that a land-use commission had exercised a veto, and it was probably the first time, the commission had stood in the way of a nature restoration project.
The local pig farmer was a member of the same regional agricultural organisation as the landowners at Lake Soeborg who did not support the project. Representatives from the nature protection network and the county regretted the decision, which received extensive media attention. The decision, however, was not legally wrong so, once again, Lake Soeborg remained dry (Pedersen 2006: 324-325).
However, there is an epilogue to the events. In connection with the discontinuation of county-level administration in Denmark from January 2007, the Minister of Environment, Mrs. Connie Hedegaard, changed the composition of the regional nature conservation boards. County representatives were replaced with representatives appointed by the Minister of Environment, thereby diminishing local/regional power in these boards. If protection of the Lake Soeborg area is filed for in future, the chance 21 that the outcome will be positive, in terms of the nature protection network's viewpoint, is now greater.
Furthermore, the composition of local land-use commissions has also been changed.
The Lake Soeborg veto resulted in comprehensive media attention and complaints from charitable foundationsthe support of (minimum) one of the three nonagricultural representatives is required.
In conclusion, actors from the weak land reclamation network were able to stop the Lake Soeborg Project, despite the fact that the project was supported by the most powerful network of the twothe nature protection network. First, the project was stopped when the land reclamation network built up a negative policy image locally around the project and coupled the policy image to the regional policy venuethe Nature Conservation Board (in practice, the county and the municipality). At the county and municipal level, the Liberal Party, generally the party most supportive of farmer viewpoints, was in power. The last attempt was stopped when the land reclamation network used its veto power in a policy venue (the land-use commission) a formal opportunity which was extremely rarely used in practice. The case shows that sometimes it is possible for the weaker network to exploit local/regional policy venues to make decisions favour the weaker network. However, the case also shows that the stronger network, the nature protection network, has actually, in general, been able to change the organisation of structural power subsequent to their defeats by changing two policy venues, thereby increasing the possibility of a Lake Soeborg nature restoration project, and other projects in future.

22
At first sight, Danish policies regarding land use in the wider countryside are characterised by classic policy outcomes. Actors in strong policy networks are privileged by the policies at the expense of actors organised in weak, disprivileged policy networks. In the land reclamation phase (1866-1970) a very strong land reclamation network was privileged when areas of nature were converted to farmland, and at the same time a nature protection network was disprivileged. In the nature restoration phase  the situation was turned upside down. Here, the nature protection network was the stronger policy network and was privileged by the new policy reverting 'reclaimed' land to nature areas, while the land reclamation network was disprivileged.
However, the case studies of the Naeraa Beach land reclamation project rejected in the 1940s and the Lake Soeborg nature restoration project rejected in the 1990s and 2000s display surprising results: sometimes it is possible for David to defeat Goliath regarding concrete project decisions within a broader general policy. At Naeraa Beach, the weak nature protection network defeated a strong land reclamation network. At Lake Soeborg, a weak land reclamation network defeated a strong nature protection network. The above examples represent unexpected outcomes, because the two 'Goliaths' were able to obtain approval for a number of other projects in their time. This is interesting because the variability displayed here regarding decisions concerning land-use projects cannot be explained by changes in e.g. the broader context (socioeconomy, public opinion, colour of government, etc.).
Baumgartner & Jones' concepts of 'policy image' and 'policy venue' add dynamism to the static policy network approach (see also Richardson 2000). Even a very strong policy community cannot expect to achieve implementation of all desired projects.
The general policy is not necessarily under threat, but it can be possible for actors in disprivileged networks to stop concrete projects either by using local/regional policy venues with a negative policy image or by threatening to build up a negative policy image around the general policy in the public arena. Actually, it might be a wise longterm strategy for the most powerful network to lose some cases. By accepting to lose some 'dangerous', concrete project decisions in the short termwhich, if implemented, would entail a potential threat to the general policy imagetight policy networks may prevent their network from disintegrating in the longer term. And as the 23 Lake Soeborg case shows: after being defeated regarding a concrete project, the privileged network can attempt to improve the structural power balance to avoid similar single case defeats in future.
Furthermore, the case studies indicate, in contrast to the US findings of Lowry (2003), that sometimes it is difficult to develop a mutual understanding between interests organised in two competing networks. The interaction between the land reclamation network and the nature protection network can be characterised as a power struggle more than development of a mutual understanding between opposing interestsactors in competing networks are not inclined to trust expert information from fierce opponents. In addition, the analysis shows that it is methodologically beneficial, as recommended by Daugbjerg & Pedersen (2004), to identify two competing networks instead of one large 'issue network', when analysing environmental policy networks.
In general, land-use politics are characterised by numerous struggles between opposing interests. The findings challenge the traditional expectation that strongest interests always are the winners. Supposedly, the theoretical framework can be applied to e.g. decisions/non-decisions regarding restoration of water bodies in the countriese.g. Australia, Costa Rica and the member states of the European Union (see Lowry 2003: 240ff;Blackbourn 2006) which, at the turn of the millennium, are trying to move beyond protection of water bodies to restoration of water bodies to their natural state.