Policy dismantling at EU level: Reaching the limits of ‘an ever-closer ecological union’?

Having explained their adoption, analysts are now trying to understand how EU environmental policies have subsequently evolved over time. In 2003, David Vogel famously speculated that having overtaken the US in the environmental race to the top, EU policies would also eventually succumb to policy gridlock, i.e. neither expanding nor dismantling. Empirical research has since confirmed that EU policy expansion is in decline, but less is known about why dismantling has also been very limited. This article breaks new ground by reconfiguring dismantling – a concept developed for national policy systems – to explain the various dismantling strategies deployed at EU level (1992 to 2016). It finds that the absence of significant dismantling is due both to the symbolic nature of early dismantling attempts and the failure of more recent attempts to build coalitions that overcome institutional obstacles to policy change in the EU.

Meanwhile, Vogel argued that weakened support for new policies in the US after 1990 also went hand in hand with political demands for what he termed policy roll-back-or what is now more commonly referred to as policy dismantling, that is, 'the cutting, diminution or removal of existing policy' (Jordan et al. 2013, p. 795). These demands were amplified in the subsequent W. Bush and Trump administrations (Layzer 2012;Klyza and Sousa 2013;Bomberg 2017). When Vogel completed his 2003 article, attempts to achieve dismantling were also running into political opposition, this time from pro-environment interests. The net result of limited expansion and limited dismantling was what he termed policy gridlock (Vogel 2003). Vogel concluded his article with a bold prediction: that having caught up with the US, the EU would continue to follow in its footsteps, gradually moving in a less environmentally ambitious direction. Crucially, he claimed that the state of policy gridlock in the US after 1990 would 'at some point occur in Europe' (Vogel 2003, p. 580). But has it? EU environmental legislation has grown from a few dozen laws in the 1970s to over 400 pieces (Delreux and Happaerts 2016). This growth has been at the heart of the EU's emergence as an environmental 'Regulatory State' (Majone 1999) and has attracted the attention of many policy scholars. Their work has revealed that after three decades of significant policy expansion, the Commission significantly reduced the number of new policy proposals after the late 2000s. Policy sectors such as the environment and health were particularly heavily affected (Kassim et al. 2017). This pattern of events has confirmed the first part of Vogel's prediction: that policy expansion would stall.
What about the other part: that policy dismantling would fail? The rapid rise up the EU's agenda of new discourses of better regulation and 'fitness checking' has certainly attracted the attention of analysts (Gravey and Jordan 2016;Steinebach and Knill 2017). But studies of environmental policy dismantling at the EU level have been much more limited. Dismantling has generally been studied at the national, rather than at the EU level. Over three decades ago, Hanf (1989), for example, analysed the deregulation of environmental policy in the Netherlands. More recently Jordan and Turnpenny (2012) and Bernauer and Knill (2012) have studied dismantling in the UK and Germany, respectively. Initial studies of the content of policies at the EU level have, however, suggested that significant dismantling is not occurring at the EU level. Various studies (e.g., Schaffrin et al. 2015;Gravey 2016;Steinebach and Knill 2017;Burns et al. 2019) have, for example, pointed out that in spite of multiple political demands for and the discourse of policy dismantling, EU environmental policies have not been dismantled across the board. In fact, in some cases the policies targeted for reform have become more not less stringent (Benson and Jordan 2014).
Hence, growing political demands for policy dismantling in the EU are not (yet) translating into significant policy change. What we do not really understand is why. In this article we make an original contribution by investigating how far the observed absence of significant policy dismantling can be explained by analysing the various dismantling strategies that have been deployed, that is, the 'mode[s], method[s] or plan [s] chosen to bring about a desired dismantling effect' (Bauer and Knill 2014, p. 37). We address the puzzling coexistence of growing dismantling discourses but limited dismantling of policy outputs by paying closer attention to the intervening process: namely the selection and implementation of dismantling strategies. We do so by analysing the selection and deployment of dismantling strategies at the EU level over an extended period of nearly 25 years . This period extends from the ratification of the Maastricht Treaty up to the 2016 UK referendum, prior to which the Cameron government listed policy dismantling as one of its four main demands in order to preserve the UK's membership (Cameron 2015).
Section two briefly reviews what has been learnt-both empirically and theoretically-about dismantling since Vogel's article. It sets out how we will analyse different dismantling strategies and, more specifically, how we will differentiate between them empirically. We claim that the explicit and visible attempts to dismantle noted above align with two distinct strategies developed by : active dismantling and symbolic dismantling. We investigate whether the absence of notable dismantling is best explained by the deployment of active dismantling strategies that fail (i.e., an active decision to dismantle that nevertheless yields no policy dismantling) or the use of symbolic dismantling strategies (i.e., dismantling in name only), by identifying and testing three possible explanations.
Drawing on Pierson (1996) and , a first explanation would be that decision-makers failed to assemble a sufficiently large winning coalition on what and how to dismantle-in effect foundering on the various hurdles in the EU's hyper consensual system. A second possible explanation is that although politicians managed to agree what and how to dismantle, they subsequently faced opposition from interest groups and voters (what Vogel claimed had happened in the US after 1990). A third possible explanation is that the commitment to dismantle was only ever symbolic. Symbolic dismantling refers to a situation in which explicit commitments to dismantle are deliberately not followed through by decision-makers, being limited to symbolic activities such as 'relabeling policies [and] … commissioning consultations/evaluation reports' (Bauer and Knill 2012, p. 46).
Section three presents our empirical findings, encompassing the selection and deployment of dismantling strategies at the EU level. Finally, section four discusses our findings in relation to Vogel's predictions about the long-term relationship between expansion and dismantling and arrives at some conclusions with regard to how far EU environmental policy has, like the US, succumbed to a form of policy gridlock. It also reflects on the implications of our findings for future research on policy dismantling, EU integration and environmental policy leadership.

| STUDYING POLICY DISMANTLING: THE NEGLECTED ROLE OF REFRAMING AND VENUE-SHOPPING
Since Vogel's article on US environmental policy gridlock was published in 2003, other analysts have tried to add more fine-grained empirical detail. For example, Klyza and Sousa (2010) argued that the US environmental policies adopted in the 1960s, 1970s and 1980s created a supportive policy regime-a 'Green State'-that opposed dismantling in Congress. Thus, pro-environmental actors had 'simply to play defense to protect old commitments, typically an advantageous position in the American political system ' (2010, p. 445). This broadly equates to our Explanation 1. It also echoed both Vogel's findings and (incidentally) Pierson's earlier arguments (1996) about how policy feedback sustains the welfare state (Jordan and Matt 2014).
Vogel concluded by pointing to the way in which dismantling was increasingly being blocked by Congress, but more recent studies have underlined the need to look 'beyond legislative gridlock' (Klyza and Sousa 2013) and examine attempts both to strengthen and weaken policies across multiple venues-that is, in the courts, in different states, etc. Korte and Joergens (2012) have usefully shown how a particular venue-the executive branch-can be used to weaken policies in the absence of congressional support, through budget cuts and political appointments. Hence, when, as in the US, governance is multi-levelled, analysts should be alive to the possibility that dismantling advocates (and opponents) may choose to switch their strategies to venues that are more (or less) conducive to the building of coalitions of support.
Finally, in exploring the agenda-setting stages of the policy process, Layzer (2012) emphasized that advocates of dismantling must also win the battle for attention just as advocates of policy expansion must, for example by emphasizing particular frames or ideas. She argued that after the 1980s, the US environmental movement lost the battle of ideas with conservatives, who reframed the issue into one of how to boost economic growth by dismantling existing policies and cutting 'red tape'. In the similarly multi-levelled governance context of the EU, frames and venues are potentially even more important than in the US; they are a means to 'gain … attention' and 'build … credibility' (Princen 2011, p. 928)-necessary conditions that have to be satisfied to achieve a winning coalition in favour of policy change, including dismantling.
In summary, the US literature suggests that venue-shopping and/or reframing are likely to be central to the pursuit of policy dismantling in the EU. Crucially, the vagaries of agenda setting explain why 'retrenchment opportunities wax and wane while institutional arrangements change slowly or not at all' (Sheingate 2000, p. 336). For EU analysts, the US literature also underscores the need to think about how pro-and anti-dismantling coalitions are built and sustained over time, including via the selection and deployment of different dismantling strategies, as these will be central to our Explanation 1.
In their study of national policy, Bauer and Knill (2014) suggested that there are, in principle, four ideal-types of dismantling strategy (see Table 1). These are based on two questions: first, whether the 'political decision to dismantle is actively and consciously taken or not'; and second, to what extent 'political actors wish to hide or reveal their dismantling activities ' (2014, p. 38). This creates two cleavages: between active and passive dismantling (i.e., whether politicians actively pursue dismantling or let it happen during their term of office without engaging with it); and between open and hidden dismantling (i.e., whether politicians try to avoid blame or claim credit).
As this article focuses on explaining the puzzling coexistence of dismantling talk and very little dismantling in practice (Gravey and Jordan 2016), we are particularly interested in the distinction between active dismantling and symbolic dismantling. Bauer and Knill (2014) suggest that these two are ideal-types: symbolic dismantling is evidenced by the commissioning of reports and/or the announcement that cuts will be made in future; by contrast, active dismantling is evidenced by cuts to the number of or the stringency of policies. But how can we distinguish between them empirically? And more specifically, how can we distinguish between failed active dismantling and symbolic dismantling? We do so by conceptualizing the strategies as two opposite points on a continuum of dismantling types, ranging from no or very limited initiative to pursue dismantling to a number of concrete policy initiatives (successful or not) to cut, remove or weaken policy outputs (see Table 2).
As one moves from the 'symbolic' to the 'active' end of this continuum, the empirical indicator is the growing number of deliberate policy initiatives, such as attempts to reframe dismantling in a new way or change the venue in T A B L E 1 Four ideal-types of dismantling strategy which it is pursued in order to increase support for (or decrease opposition to) it. This relates to what Princen (2011, p. 933) has described as linking policy initiatives with established overall values that are held to be central to the EU's 'purpose and identity' or with 'stated policy priorities and commitments' to garner sufficient support. Hence, a member state strategy of commissioning a report on EU 'red tape' would be considered symbolic. But if the member state subsequently followed up on that by building a political coalition to turn 'tackling red tape' into an EU-wide effort it would move towards the more active end of our continuum.
The rest of this article analyses the environmental policy dismantling strategies that have been employed at EU level from 1992 to 2016. How dismantling is framed by political actors and in which venues it is pursued are explored using two types of historical data. First, documents from the Commission, Parliament and member states, as well as reports from think tanks and civil society organizations were collected and coded using NVIVO. These data were used to identify the relationship between particular political actors, the targets of policy dismantling and the associated frames and venues. Second, we undertook 17 elite interviews with actors in the EU institutions (in the Commission, member states and Parliament) and in wider civil society (e.g., environmental NGOs, business lobbies, etc.).

| ENVIRONMENTAL POLICY DISMANTLING: A LONG-TERM PERSPECTIVE
On 2 June 1992, the Danish people voted down the Maastricht Treaty-the first time that voters had openly voted against deeper European integration. The EU Presidency at the time (July-December 1992) was held by the UK. It was tasked with bringing Denmark back into the fold. The UK eagerly seized this opportunity to advocate the dismantling of a range of (mostly environmental) policies which were causing it implementation difficulties at home (Jordan 2003). It did so by reframing dismantling as being consistent with subsidiarity, a principle that had long been discussed at EU level but was only formally integrated into the EU Treaties, alongside proportionality, by the Maastricht Treaty (Golub 1996). The UK calculated that such reframing would be a more successful dismantling strategy if it was perceived to achieve 'stated policy priorities and commitments' (Princen 2011, p and to the most problematic parts of the existing acquis.
The UK's preferred dismantling strategy was to draw up lists of policies that the EU should agree to dismantle.
However, the key venue for dismantling would be the Commission (as it held a monopoly over new EU legislative proposals), not the Council. Therefore, the UK Presidency invited the Commission President Jacques Delors to draw up an EU-wide hit-list (Independent 1992). However, as a venue, the Commission was not especially conducive to dismantling. The Commission, particularly under Delors, was emphatically pro-integration. Unsurprisingly, the Commission's list contained 20 proposals for new legislation-among them one (on zoos) that was only indirectly environmental-that could be removed, not items of the existing acquis. However, the Commission made somewhat vague and symbolic promises to update rather than dismantle existing legislation. Dissatisfied with Delors' list, the UK Presidency produced its own list (Financial Times 1992). Leaked to the press shortly before the Edinburgh summit, it listed 37 directives and regulations (plus 34 proposals for new policy) for dismantling. Among the environmental items on that list, the UK suggested that one directive be completely repealed (shellfish waters) and five others weakened (relating to wildlife habitats, hazardous waste, bathing, drinking and ground-water, respectively).
In effect, the lack of political agreement at the EU level (our Explanation 1) spawned two parallel processes, each focused on a particular venue. First, a Commission-led process centred on the production of annual reports on the implementation of the subsidiarity and proportionality principles (European Commission 1993). Second, the UK, France and, eventually, Germany engaged in more listing activities. The UK and France produced a combined hit-list in 1993, targeting a mixture of new proposals and existing policies (covering habitats, zoo animals and environmental impact assessment) (Reuters 1993). Germany followed up with its own hit-list (with a sole environmental target: the proposal for a zoos directive) which received little support. However, Germany profoundly changed its tactics during In addition to targeting specific policies, the Six Presidencies also attempted to reframe the better regulation agenda onto the narrower issue of administrative burden reduction (Six Presidencies 2004), an approach originally developed by the Netherlands and subsequently adopted by other member states, such as the UK and Germany (Wegrich 2009  Once again, policy dismantling was being framed as pro-EU, a way to 'make sure the overall project is not going to be fundamentally threatened by a relatively negative discussion that we now see in a number of member states' (Interview: European Commission, November 2014). But this time, pushing for dismantling was perceived as being specifically linked to 'the UK … referendum and the opt-outs' (Interview: European Commission, November 2014).
Hence, Cameron's renegotiation letter requested the Commission to continue its work of 'scaling back unnecessary legislation', asking specifically for 'a target to cut the total burden on business' (Cameron 2015, p. 3 (1996) and Bauer and Knill (2014), we have argued that there are potentially three explanations for this failure to dismantle: (1) a failure to assemble a sufficiently large winning coalition on what and how to dismantle; (2) a failure to implement an agreement to dismantle because of opposition from interest groups and/or voters; or (3) symbolic dismantling where dismantling was never seriously intended by politicians. We tested these explanations by examining the various dismantling strategies deployed by various actors at the EU level. We did so by taking Bauer and Knill's (2014) distinction between active and symbolic dismantling strategies and reconceptualizing them as representing two ends of a continuum of types. We argued that analysts can determine whether a dismantling strategy is indeed 'active' or simply 'symbolic' by investigating the number and type of policy initiatives (such as attempts to reframe dismantling or change venue).
We found notable changes in the dismantling strategies employed over time. The strategies deployed in the 1990s tended towards the 'symbolic' end of the continuum (i.e., Explanation 3); they could be described as either paying lip-service to dismantling (such as the yearly reports from the Commission) or as short-lived and unable to summon majority support (e.g., hit-lists from a handful of member states). In the 2000s and 2010s, the dismantling strategies moved away from the 'symbolic' to the more 'active' end of our continuum. They were sustained over lon- the 'American policy tools of cost-benefit analyses and regulatory impact assessment' have had 'little impact on the actual regulatory policies adopted by the EU' (Vogel 2013, p. 321).
What do our findings tell us about the more general phenomenon of policy dismantling?  hypothesized that dismantling advocates would pursue open 'credit claiming' strategies instead of more hidden 'blame-avoidance' strategies to dismantle green rules. Although the dismantling strategies we detected were open, actors often framed their arguments to avoid blame, stressing that dismantling was not deregulation, or arguing that dismantling would strengthen European integration. In that respect, environmental policy dismantling at the EU level is an example of the ongoing politicization of the EU level, which has 'affected all EU actors' actions and interactions, making for a "new" dynamics of EU governance that is more political in every way' (Schmidt 2018(Schmidt , p. 1548. Our findings show that the politicization of policies at the EU level can also be used by some to defend the EU's polity. This defence can be linked both to Explanation 1-some member states struggling to build a sufficiently large coalition because their peers view dismantling as anti-EU-but also, gradually, under Explanation 2, where civil society becomes more involved in shaping the implementation of dismantling 'on the ground'. The Nature Alert campaign in 2016 prevented the Commission from dismantling the nature directives; time will tell whether this campaign was an anomaly or a harbinger of things to come.
Our central focus in this article has been on the strategies used by actors to achieve or to block dismantling.
More research is needed into the preferences driving policy dismantling-and on how opposition to the EU interacts with policy preferences. For example, are the greener member states favouring local over EU-level action, lest the latter succumbs to dismantling? The Commission's policy preferences certainly warrant further attention. Since the Danish 'no' vote, a profound culture change appears to have taken place inside that organization. The dominant frame in EU policy-making now appears to be that of 'self-restraint', but it has taken a long time to take root (it was first espoused by Sir Leon Brittan back in 1992 (Agence Europe 1992c)). The US experience of dismantling provides some possible explanations for this change. Layzer has argued that US Conservatives had built 'a compelling antiregulatory storyline to counter the environmentalist narrative' (Layzer 2012 p. 4 to openly supporting more active dismantling. This finding is significant because it goes against standard assumptions that the Commission is hard-wired to oppose dismantling as it is a threat to its raison d'être-advancing European integration (Löfstedt 2007). It is also not the case that (see Steinebach and Knill 2017) the Commission has simply engaged in 'dismantling by default', that is, allowing a de facto reduction of environmental standards by failing to update them to changing environmental conditions (Bauer and Knill 2014). It is also different from the argument that the environmental acquis has reached such a mature stage that the Commission's main task is to revise and update existing policies (Burns et al. 2019). Rather, our research reveals something new-the Commission is actively engaging in a sustained campaign to limit new and dismantle existing regulations as part of an explicit pro-European integration strategy. It extends to all areas of its work, not just DG Environment (Kassim et al. 2017).
The changes in the Commission may change policy dynamics at the EU level. In the past, dismantling was advocated by states but opposed by the Commission and the Parliament (Explanation 1). However, if the Commission henceforth moves from opposing to supporting dismantling, then the remaining opponents will be the greener member states and the Parliament-a much-diminished coalition. It is also worth remembering that as well as making policy dismantling more likely at the EU level, the Commission's complete control over the legislative agenda also means that its change of attitude may also make the risk of gridlock much more likely-and give it a particularly European flavour. Thus, unlike the US, gridlock would not come from opposing forces during the legislative stage (Explanation 1)-but through the removal of policy proposals at the pre-legislative stage, making it difficult for green states and the European Parliament to prevent it: … if the Commission wants to withdraw a proposal, it withdraws a proposal and that's it. Parliament can show they are unhappy about it but they can't do anything. (Interview: Environment NGO, Finally, the growing support for dismantling within the European Commission raises questions not only about the future of the acquis but about the very nature of the EU itself. Should the EU even be considered a regulatory state if the Commission actively dismantles the regulatory acquis, perhaps heralding the start of a new era of European disintegration by stealth?