Going full circle: the need for procedural perspectives on EU responsiveness

ABSTRACT Research on policy responsiveness of the European Union has adopted the systemic model from national contexts. This focusses on the static congruence or the dynamic adaptation of aggregated policy output and similarly aggregated public opinion. Approaches in this vein provide relevant insights and uncover a surprising degree of EU responsiveness. Yet, this debate contribution argues that they only insufficiently capture indirect accountability chains and the emerging challenges of public EU politicization and mediatization. We establish the need for procedural perspectives that addresses how the different EU institutions perceive and digest public opinion and subsequently influence it through communication. To further this research agenda, we sketch the contours of a procedural model by highlighting possible variation at the input, throughput and output stages. Going full circle, we suggest, allows us to better understand the responses to public opinion and their wider implications for the societal acceptance of the unfinished supranational polity.


Introduction
In contemporary European Union politics, understanding whether and how supranational policy aligns with the preferences of the wider citizenry seems to be more important than ever. The recent rise of populist challenger parties throughout Europe combines fundamental opposition to supranational authority with calls for more direct democracy. The electoral success of this particular mobilization strategy suggests that many EU citizens believe that EU policies do not respond to their preferences. But where and how does responsiveness to citizen preferences actually figure in the complex EU policy-making machinery?
We argue that this question cannot be tackled sufficiently by simply transferring what the introduction to this collection has called the systemic model of responsiveness to EU policy-making. This model concentrates on electoral accountability as the key mechanism establishing responsiveness and empirically assesses it along the static congruence or the dynamic adaptation of aggregated public opinion polls and similarly aggregated policy output. Based on the assumption that vote choices are driven by largely correct retrospective assessments of policy output, political actors caring about re-election have incentives to pander to public opinion in an anticipatory manner when formulating policy. In this view, public opinion has famously been conceptualized as a 'thermostat' that consistently corrects direction and output of government policy (Hakhverdian 2012;Wlezien 1995). Correlational analyses following this standard model demonstrate a rather high level of aggregated responsiveness in national democracies and highlight moderating effects of public issue salience (Burstein 2010), the ideological leeway of governing actors (Jensen and Mortensen 2014), or the clarity of electoral accountability chains in different institutional variants of representative democracy (Wlezien and Soroka 2012).
This model provides highly valuable insights on the functioning of substantive democracy at the national level. We also have to note that its applications to the EU have uncovered surprisingly strong empirical links between the support for political integration in aggregated public opinion and the aggregated amount of supranational policy output (Bølstad 2015;Franklin and Wlezien 1997;Toshkov 2011).
Yet, like the other contributions in this debate section, we have doubts that mere comparisons of aggregated input and output indicators sufficiently illuminate the causal mechanisms linking the preferences of the wider citizenry to substantive policy choices in the EU. Like Meijers et al. (2019) we are not convinced that this highly aggregated perspective fits the multiplicity of actors and 'their' audiences involved in EU policy-making. And similar to Steunenberg (2019), we also do not consider the adoption of EU legislation as the final point in time at which responsiveness should be assessed.
However, these contributions remain fundamentally wedded to the concept of responsiveness as a product. This product, they argue from different perspectives, should then be related to independent variables to establish causality. Our contribution, in contrast, suggests moving beyond comparative statics by focusing on the process of responding rather than the product of responsiveness as the main phenomenon to be explained. We come to advocate this approach by highlighting three more specific context conditions of contemporary EU policy-making that challenge the systemic model of responsiveness research in section two: (1) an active policymaking role of non-majoritarian institutions; (2) the differentiated politicization of European integration; and (3) the mediatization of politics more generally. These challenges, we argue, warrant a procedural perspective on EU responsiveness. We sketch the contours of such an alternative model in section three. Scrutinizing and comparing how different EU institutions perceive, digest, and influence public opinion promises deeper insights on how responsive supranational policy-making actually is and which adjusting mechanisms it offers for accommodating the current public scepticism.
Three challenges of the standard model in contemporary EU policy-making To start with, the almost exclusive focus on electoral accountability as the key responsiveness mechanism seems ill-suited to capture the influence of nonmajoritarian institutions in contemporary EU policy-making. Of course, virtually all institutionalized forms of representative democracy deliberately attempt to isolate some portion of governance functions from public opinion. Implementing agencies, central banks, and constitutional courts are usually not supposed to be responsive to the vagaries of public opinion (even though they are sometimes, e.g., Hall 2014;McGuire and Stimson 2004). They are rather expected to act on stable, impartial rules and high levels of sectoral expertise. In contrast to most national political systems, however, such non-majoritarian institutions have much more pronounced policy-making roles in the EU polity. This holds with regard to tertiary law-making by agencies and regulatory networks (Rittberger and Wonka 2011). Yet, it also holds with regard to secondary law-making that has been the prime focus of most EU responsiveness research thus far. Non-majoritarian institutionsmost notably the European Commission and the European Court of Justiceperform key tasks in the policy cycle such as the initiation, drafting, and review of binding EU law.
For our argument, two things are especially noteworthy. First, the empowerment of these non-majoritarian actors has been an intentional choice of institutional design: delegating policy powers to originally insulated organizations was meant to overcome credible commitment problems of national governments facing varying short-term political pressures at home (e.g., Majone 2000;Moravcsik 1998). Second, even though these institutions have often acted as policy entrepreneurs and also gained high public visibility, they are not linked to the European electorates through direct, institutionalized accountability chains thus far. 1 Against this relevance of non-majoritarian institutions, explaining responsiveness only along electoral accountability risks to miss important elements of EU policy-making. This does not dismiss research on the responsiveness of EU institutions with more direct accountability chains such as the Council (Schneider 2019;Wratil 2018a) or the European Parliament (Williams and Spoon 2015). Nor does it suggest that the European Commission, the European Court of Justice, or other non-majoritarian agencies are not responsive at all (to the contrary, see below). But it does suggest that a more complete understanding of EU responsiveness needs to explicitly cover the particular incentives of non-majoritarian institutions as well. For these institutions, possible clashes with their original mandates, organizational setups and policy legacies, as well as the risk to alienate their traditional principals when giving in to short-term political pressures emerging from public opinion has to be factored in as well.
A second challenge that hampers the straightforward application of the systemic model to EU policy-making emerges from the politicization of European integration (De Wilde 2011). At first sight, and in line with the standard model's emphasis on public salience as a moderating variable, the increasing risk that European decisions become subject of highly visible and controversial public debates should create incentives for more responsiveness at the European level. Indeed, initial research shows that both majoritarian (Schneider 2019; Wratil 2018a) but also non-majoritarian institutions adapt policy choices in the light of intensifying public debates about European integration (Blauberger et al. 2018;Rauh 2018;Van der Veer and Haverland 2018). But at the same time, the public politicization of European integration also increases the complexity of political conflict against which possibly responsive European policy has to be formulated.
Three things are particularly noteworthy for our argument. First, EU politicization and the salience of specific policy issues vary not only over time but also across issue areas and, most notably, across member states ( The systemic model's focus on one-dimensional policy choices (often: more or less spending) falls short of capturing such political complexities. This has already been noted in research focussing on responsiveness in national contexts (most recently: Rasmussen et al. 2019). But we argue that this shortcoming is particularly pronounced in EU policy-making. Even if supranational policy-makers have incentives to respond to public opinion, they will have to make choices as to which parts of the European public they want to focus on. And even if they can identify a policy that satisfies the relevant parts of the European public, they will also have to evaluate whether this very public also prefers to have this choice taken at the supranational level at all. Rather than assuming a consistent public opinion that functions as a simple thermostat, a convincing account of responsiveness in EU policymaking should thus be able to factor in the various trade-offs that policy-makers face against the background of differentiated politicization of European integration.
A third and final challenge to the input-output comparisons the systemic model relies upon is the accelerating mediatization of contemporary politics. That a crucial part of the interaction between policy-makers and the affected societies occurs through indirect encounters in mediated public spheres is in itself nothing new (Gamson and Wolfsfeld 1993;Koopmans 2004). Indeed, media coverage of policy output has been identified as an important link in the 'thermostat' model of responsiveness (Soroka and Wlezien 2019;Williams and Schoonvelde 2018). But we also know from news value theory that such mass-mediated discourses introduce distortions, biases, and adverse incentives. It amplifies conflict, personalization, and negativity in political communication (Galtung and Ruge 1965;Manin 1997: 218-34).
For our argument, two things are especially noteworthy here. First, this mediatization is accelarating. While traditional television, for example, was prone to unite the citizenry into a single public that was exposed to similar news around the same time (Wigbold 1979), contemporary digital media outlets and the proliferation of political information through social media channels shorten the news cycle and create niche publics with selective information exposure (Garrett 2009). Second, focussing on immediate communication and short-term newsworthiness challenges the traditional modes of policy-making in the EU which emphasize compromise solutions, long-term planning, and strong reliance on expertise.
This accelerating mediatization renders it questionable that low-frequency polls are good indicators for the information that policy-makers have when making their choices. 2 Rather, mediatization should make it more difficult to figure out what 'the public' as a single body actually wants. The salience of specific policy issues becomes much more volatile so that mediatization speeds up the feedback loops between policy output and public opinion formation. Thus mediatization creates new, often rather pragmatic challenges for policy-makers aspiring to be responsive. In conjunction with public EU politicization, furthermore, mediatization requires active communication and justification efforts on part of the policy-makers themselves. This challenges the traditionally more technocratic ways of decision-making in the EU. However, mediatization might also present an opportunity by enabling more direct communication with citizens. The institutions that are remote from electoral politics can more easily defend their policy choices themselves and in real time (Gilad et al. 2015), thus circumventing otherwise long and indirect accountability chains. We see, for example, that governance institutions beyond the nation state professionalize their public communication efforts when being faced with public politicization (Ecker-Ehrhardt 2018). Mediatization thus creates pragmatic hurdles in identifying public opinion, but it also incentivizes and facilitates what the Zhelyazkova, Bølstad, and Meijers (2019) as well as Meijers et al. (2019) discuss as 'rhetorical' or 'signalling' responsiveness.
To sum up, we identify three context conditions of contemporary EU policy-making that challenge key assumptions and approaches in the systemic model of responsiveness. First, the relevance of non-majoritarian institutions problematizes electoral accountability as the key theoretical mechanism. Second, the politicization of European integration problematizes the assumption of one-dimensional policy-making and the identification of stable majorities in the European citizenry. And third, the accelerating mediatization problematizes the identification of a single public opinion and highlights the role of autonomous communicative responsiveness by individual EU institutions.

A procedural perspective on responsiveness in EU policymaking
Against these three challenges, we propose to complement comparisons of aggregated public opinion and aggregated policy output with research that takes a decidedly procedural perspective on how public opinion and specific policy choices interact. More systematic knowledge on how different EU institutions perceive, digest, and inform public opinion is needed. Such a procedural perspective promises to identify the break-points and the possible adjusting mechanisms of responsive or non-responsive policymaking in the EU. More complete causal accounts of how public opinion is digested by and within individual organizations would also tackle the question of whether responsiveness is 'actively' created or just reflects a spurious correlation driven by cueing efforts of policy-making elites (cf. Zhelyazkova, Bølstad, and Meijers 2019). With the above sketched challenges of non-majoritarian institutions, public EU politicization, and mediatization in mind, Figure  1 sketches the contours of such a procedural model of EU responsiveness. This model is inspired by theories of organizational mediatization and reputation management developed primarily in the context of national bureaucracies (e.g., Maor and Sulitzeanu-Kenan 2016;Schillemans 2012;Schulz 2004;Thorbjornsrud et al. 2014). Rather than assuming that policymakers and citizens have perfect information about public opinion, policy output, and policy effects, respectively, we propose to analyse responsiveness through the lens of behaviour within the policy-making organization of interest.
Regarding responsiveness at the input stage of EU institutions (cf. Meijers et al. 2019), we propose to not only ask whether policy makers take positions in response to public opinion but how they do so. To which extent and for which reasons does the organization in question dedicate resources and staff to monitor polls, media coverage, and online debates considered relevant to its specific policy competences? What sources do officials employ, which elements of 'public opinion' do they consider important, and how systematic is the production and distribution of this knowledge within the organization? These questions call for systematic, comparative case studies of organizational resources and behaviours across individual EU policy-making institutions and time. Organizational charts and strategic development plans as well as questionnaires and interviews among EU institutional staff could shed light on this. Such research would illuminate what supranational policy makers actually believe to know about public preferences and would shed light on the incentives of non-majoritarian institutions, in particular.
Regarding throughput, we propose to investigate the extent to which this knowledge and dedicated communication staff are leveraged during policymaking. At which stage do public relation considerations enter the policy-formulation process? How much importance is ascribed to public relations relative to other organizational goals and mandates? How influential are public relations officials within the organization? These questions call for research that focusses on individual policy choices as the unit of analysis and process tracing analyses. Comparing the anticipation of public reactions across policy-drafting processes in the Commission (e.g., Hartlapp et al. 2014: Chapter 9) or Council negotiations (Wratil and Hobolt 2019), for example, should illuminate how policy-makers deal with the complexity of EU politicization and value it relative to more functional concerns that have traditionally dominated EU policy-making.
Finally, regarding output, we propose to investigate the extent to which the adoptionand also the implementation (cf. Steunenberg 2019)is flanked by pro-active and professionalized communication efforts to claim credit for responsive policy (cf. Meijers et al. 2019). Which media channels and audiences are addressed? How intense are communication efforts and campaigns? How is public communication of policy output timed? Which self-legitimation standards are employed in justifying or preparing a given policy choice? These questions suggest mapping the supply of public, policy-specific communication especially through content analysis methods across EU institutions and time, revealing how the link between public opinion and policy contents is constructed (see e.g., De Bruycker 2017). This would provide insights as to how far EU institutions are willing and able to deal with the challenges of differentiated politicization and mediatization.
Empirical and systematic research along these lines would allow us to trace how public opinion 'enters' the different organizations involved in EU policymaking, how it 'travels' within these organizations and, finally, how it affects not only the policies but also the communication that 'leave' the organization. Rather than understanding responsiveness as a repeated one-shot game in which policy is adapted to a single, consistent public opinion that exists prior to decision-making, this model describes an adaptive circle in which governance organizations interact (with relevant parts of) the European citizenry through mediatized discourses (cf. Lee 2014).
One might object that allowing for purely communicative action blurs the concept of policy responsiveness. Of course, our point is not to suggest that a high level of procedural responsiveness would compensate for a lacking congruence of citizen preferences and EU policy. Yet, given that public opinion on the EU is influenced by elite cueing to a significant degree (e.g., Steenbergen et al. 2007), given that knowledge about the EU varies strongly among European citizens (e.g., Hobolt 2007;Karp et al. 2003), and given that public salience of issues exceeding the domestic domain is inherently volatile (Oppermann and Viehrig 2011), we consider this procedural and communicative broadening of research on responsiveness as a much needed asset. In particular, the approach we sketch here promises to explain the conundrum that significant portions of the citizenry still hardly perceive EU policymaking as responsive despite the large congruence between aggregated opinion polls and aggregated EU policy output that the systemic models have uncovered. Notes 1. A notable, yet currently contested exception is the attempt to link the Commission presidency to the outcome of European elections through the 'Spitzenkandidaten' process (Hobolt 2014). 2. More generally, it also questions in how far polls initiated by policy-makers provide unbiased measures of public opinion. Especially the Eurobarometer polls contracted by the European Commission have recently been rather interpreted as strategic tools to influence the public discourse Höpner and Jurczyk 2015).
European Union Politics' (Lorentz Center Leiden, January 29 -1 February 2018) for sharing their insights and debating the current state of research with us. We also thank two anonymous reviewers that have helped to clarify our arguments.

Disclosure statement
No potential conflict of interest was reported by the authors.

Notes on contributors
Pieter de Wilde is Associate Professor in European Studies at the Norwegian University of Science and Technology (NTNU).
Christian Rauh is a Senior Researcher in the Global Governance department of the WZB Berlin Social Science Center (www.christian-rauh.eu).