Preserving the old or building the new? Reputation-building through strategic talk and engagement with stakeholder inputs by the European Commission

ABSTRACT

How bureaucracies respond publicly to policy inputs received in the context of stakeholder participation in policymaking provides important insights into their reputation-building strategies. We examine the European Commission’s public engagement with stakeholders’ inputs. We argue this is driven by its attempts to simultaneously consolidate its core, well-established reputation as a responsible policymaker, and cultivate a newer reputation as a responsive to public demands institution. This helps navigating the challenges of having to find new sources of authority and legitimacy (embedded in participatory policymaking) while also maintaining its more established ones (embedded in evidence-based policymaking). We analyse a new dataset recording stakeholders’ inputs received as part of the EC’s legislative simplification programme. We find the Commission prioritised strengthening its core reputation over cultivating its emergent one. Evidence-based inputs were significantly more likely to receive an EC public response and inform its legislative simplification programme relative to opinion-based ones. Citizens’ inputs were significantly less likely to receive a response and inform the programme. Our findings diverge from research indicating bureaucracies prioritise defending their emerging reputations, highlight the Commission’s commitment to maintain its reputational and institutional uniqueness in the EU regulatory state, and reveal a reputation-building approach akin to that of EU agencies.

KEYWORDS:

• European Commission
• strategic communication
• stakeholder engagement
• reputation-building
• responsible and responsive governance

Introduction

Stakeholder engagement in public policymaking is a tool policymakers use to increase the legitimacy, democratic credentials and information quality of their decision-making. The establishment and institutionalisation of stakeholder engagement mechanisms marks an important shift from ‘insulation to responsiveness’ and engagement with the public and external audiences for executive bureaucracies and regulatory agencies (Braun & Busuioc, 2020). This offers them important opportunities for reputation and legitimacy building aimed at consolidating their power and authority (Busuioc & Lodge, 2017). Stakeholder engagement can help bureaucracies build and enhance their claim for input legitimacy by showing that their initiatives were informed by and reflect the demands of the different constituencies of affected stakeholders (Binderkrantz et al., 2021; Bunea, 2019; Busuioc & Jevnaker, 2022). It can also bolster their output legitimacy by showing that their decisions are embedded in expertise and policy relevant information facilitated by stakeholder participation (Arras & Braun, 2018).

Reputational approaches to the study of bureaucratic politics emphasise that a key mechanism through which stakeholder engagement helps bureaucracies build legitimacy involves allowing them to use the design and management of consultative fora and interactions with stakeholders as part of their institutional reputation-building and reputation-management strategies (Busuioc & Rimkutė, 2020b). Most of the research examines how the establishment of stakeholder engagement is used as a reputation-building instrument by EU regulatory agencies (e.g., Arras & Braun, 2018; Busuioc & Jevnaker, 2022) and US federal agencies (e.g., Moffitt, 2014; Potter, 2019). We note, however, that how policymakers handle and publicly respond to stakeholder policy inputs received through stakeholder engagement provides valuable insights into the type of institutional reputation these actors want to build with affected stakeholders and the general public. Nevertheless, the strategic use of public communication in the context of stakeholder participation in policymaking as a bureaucratic reputation-building device is currently under-researched, despite its relevance to the literature on stakeholder engagement and that on bureaucratic politics (see, however, Fink & Ruffing, 2020; Rimkutė, 2020b).

We address this gap and examine how the European Commission, a powerful executive bureaucracy and regulator, uses public communication and engagement with stakeholder policy inputs as a reputation-building strategy. We ask two questions: what explains the Commission’s decision to publicly engage with and answer some stakeholder inputs, but not others? And, relatedly, what explains the probability that its replies indicate that some stakeholder inputs informed its initiatives and programmes?

We answer by building on theories of regulatory governance and bureaucratic reputation. We argue that, similar to other modern bureaucracies, the Commission has incentives to use its public engagement with stakeholders’ inputs as ‘strategic talk’ and signalling device as part of its reputation-building strategies (Maor, 2020). This strategic communication is aimed at consolidating its long-established reputation for being a responsible policymaker (Koop et al., 2022, p. 46), while also making an attempt to cultivate a more recent, less established and somewhat more contested reputation as responsive to the public policymaker (Rauh, 2016; Reh et al., 2020). This reputation-building strategy helps the EC navigate the challenge of having to find new sources of authority and legitimacy (embedded in participatory policymaking) while maintaining its more established ones (embedded in expert-informed, evidence-based policymaking). The attempt to maintain its core reputation is consistent with a pattern of publicly engaging with inputs that makes the Commission more likely to prioritise their content and only respond to those that are evidence-based. The attempt to cultivate a new reputation as a responsive bureaucracy is consistent with a pattern whereby it prioritises the source of inputs and chooses to publicly engage with those coming from stakeholders representing the public interest broadly defined, and historically less present and less directly involved in supranational policymaking, such as citizens and non-business organisations. We argue the same logic underpins EC’s responses, indicating which inputs inform its initiatives and programmes.

Empirically, we examine how the Commission engaged with and responded to stakeholders’ inputs received through the ‘Lighten the Load’ online portal introduced in 2015 as part of the REFIT Platform, in the context of a broader Better Regulation reform. The Platform was established as an advisory body integral to the EC,¹ and was mandated to lead and inform the EC’s legislative simplification and modernisation programme (the REFIT programme). It invited citizens’ and organisational stakeholders’ inputs on legislative simplification through the portal. These inputs were considered by the Platform’s two Working Groups, which, based on their views and those coming from their own members, formulated a set of specific recommendations (opinions) on ‘how to make existing EU regulations more efficient and effective while reducing the administrative burden and without undermining policy objectives’ (EC, 2015).

We test our argument on a new dataset consisting of 446 stakeholder inputs received in the portal during 2015–2020. We examine first which inputs received an EC public reply, and, second, which received a response indicating how the input informed the work of the Platform and its adopted ‘opinions’, recommending ways to simplify legislation. We find strong support for the argument that the Commission’s communication was consistent with an attempt to consolidate its well-established reputation for being a responsible policymaker. Relative to opinion-based inputs, the evidence-based ones display a significantly higher probability of receiving a public response and informing the REFIT programme via the Platform’s opinions. We find little support for the Commission being more likely to publicly engage with and adopt inputs received from stakeholders representing the public interest and allowing it to build a reputation for being a responsive to the public policymaker. Contrary to expectations, citizens’ inputs were significantly less likely to receive a public response or one indicating they informed the Platform’s opinions. This indicates the Commission prioritised its core, long-standing reputation as an evidence-based policymaker while making less effort to cultivate its newer image as increasingly more responsive to the public.

Our study is relevant in several ways. First, it taps into a key issue of bureaucratic politics: how non-majoritarian institutions may use strategic communication in the context of stakeholder participation in policymaking as a reputation-building strategy (Schimmelfennig, 2020) to develop and maintain multifaceted reputations that support their legitimacy in times of increased public contestation of their decision-making, power and legitimacy (Bressanelli et al., 2020). Second, the study tackles a fundamental aspect of regulatory governance: when and how do executive and regulatory bureaucracies decide to cultivate new(er) and (significantly) less consolidated institutional reputations such as that of being institutions that are inclusive and responsive to stakeholder and public inputs (Busuioc & Rimkutė, 2020a; Koop & Lodge, 2020). And relatedly, how do bureaucracies manage established and emergent reputations in the context of stakeholder engagement mechanisms. This is important in light of a broader pattern of recent transformations describing regulatory states in general, and the EU in particular, which are marked by a shift from depoliticised and expertise-based forms of authority and legitimation of their decision-making towards more political, participatory and ‘socially-embedded’ forms of power and legitimacy (Braun & Busuioc, 2020, pp. 1605–1607). This creates important challenges for bureaucracies that must reconcile different principles and practices of policymaking that may at times contradict each other, while facing the difficult task of ensuring that by engaging in responsive governance they do not erode or jeopardise their institutional uniqueness which foundationally resided in responsible, depoliticised, evidence-based policymaking (Braun & Busuioc, 2020, p. 1607).

We contribute to the literature on bureaucratic politics, regulatory governance and stakeholder participation in policymaking in several ways. Conceptually, we show how public communication and engagement with stakeholders’ inputs in the context of stakeholder participation mechanisms can be used as a reputation-building strategy. We show how this can help bureaucracies in general, and the EC in particular, build and project the image of responsive non-majoritarian institutions that are ‘seen to listen’ to external audiences (Koop et al., 2022, p. 55). This complements well and adds new insights to the literature on strategic communication as reputation-building, which predominantly examined how bureaucracies engage in communication through media and public statements to address public criticisms and defend their reputations (Bach et al., 2021; Boon et al., 2021; Gilad et al., 2015). Theoretically, we develop an argument illustrating how bureaucracies can use stakeholder participation and strategic public engagement with their inputs as reputation-building strategies to simultaneously consolidate core, well-established reputations and cultivate new(er) ones. We present an additional scenario about how bureaucracies can use stakeholder engagement and strategic communication as instruments aimed at building multifaceted reputations that emerged across time due to evolving political and policy circumstances marked by public contestation and the politicisation of regulatory and bureaucratic policymaking. Empirically, we explore this scenario in the context of a fundamental yet less researched issue in the EU regulatory state (legislative simplification) and examine theories of bureaucratic reputation in a context that goes beyond the study of supranational or national regulatory agencies which currently predominates in the literature. We analyse instead the EC, a powerful bureaucracy and regulator that is less examined from a bureaucratic reputation theory perspective and investigate the consequences of a shift from responsible to responsive regulatory policymaking regarding its reputation-seeking behaviour. We also address a recent call in the literature to systematically examine ‘the promise of bureaucratic reputation approaches for the EU regulatory state’ (Busuioc & Rimkutė, 2020b) and focus on one of its key institutions. We draw on a mix of qualitative and quantitative text analysis and examine a new, original dataset that allows unique insights into how direct, public communication with stakeholders can support bureaucratic reputation-building strategies.

Building a reputation for being a responsible and responsive bureaucracy and regulator

The literature on regulatory governance indicates that in the context of increased politicisation of regulation, public contestation and criticism of regulatory policymaking for being opaque, disconnected from the public and lacking direct accountability, regulatory bureaucracies shifted in time from a regulatory responsible mode towards a regulatory responsive one (Koop & Lodge, 2020, p. 1613). The former is consistent with the idea of predictability, expertise and professionalism that traditionally characterised bureaucracies that engaged in evidence-based policymaking, underpinned by expert knowledge, and decision-making isolated from politicisation and public interference (Koop & Lodge, 2020, p. 1615). This mode of governance is usually secluded and only indirectly accountable to the public, therefore less responsive to its demands (Moffitt, 2014, p. 225). It stands for a largely depoliticised form of policymaking that traditionally underpinned the work of independent regulators and the principles of technocratic policymaking. The latter, responsive mode is consistent with the presence of bureaucracies sensitive to and more interested in the public opinion and stakeholders’ public demands (Koop & Lodge, 2020, pp. 1616–1617). In this mode, bureaucracies are more open to inputs from citizens and interest groups (Rauh, 2016), and design procedures (e.g., stakeholder participation mechanisms) to ensure that their decisions are informed by and designed together with and in line with the public and stakeholders’ preferences (Schmidt, 2013). This, in turn, is consistent with a more responsive and accountable, although more prone to being politicised, mode of regulatory governance and decision-making (Braun & Busuioc, 2020). Whereas a responsible regulatory mode is usually associated with the idea of output- or performance-based legitimacy of policymaking, the responsive mode is usually associated with that of input, throughput and procedural legitimacy.

A key consequence of this shift is that bureaucracies now face the challenge of having to tap into multiple, and sometimes difficult to reconcile, sources of authority and legitimacy, to ensure their institutional survival, autonomy and power (Busuioc & Lodge, 2017). Some of these sources, such as the use of depoliticised, expert-informed policymaking and broad public consultations, which may increase politicisation, might come into conflict and create tensions within policymaking. Furthermore, while embracing the new opportunities and challenges of a responsive governance mode, bureaucracies cannot entirely drop the principles and practices of the responsible mode that constitutes their foundational source of legitimacy and authority (Braun & Busuioc, 2020, p. 1601). This creates a conundrum and an important challenge for modern regulatory governance: how to manage the need to tap into different sources of authority, legitimacy and power.

The literature on bureaucratic reputation indicates that reputation-building and reputation-management constitute a key approach bureaucracies use to address the challenge of having to harness different forms of authority and legitimacy (Maor & Sulitzeanu-Kenan, 2016) and adjust to changes in their institutional and political environment (Gilad et al., 2015). Reputation-management is underpinned by strategic considerations and allows for building different types of reputations that appeal to different audiences (Carpenter & Krause, 2012).² We contend that reputation-management, and especially the simultaneous fostering of different types of institutional reputation, represents a key strategy available to bureaucracies to maintain their legitimacy and power when facing the challenge of having to mitigate the logic and imperatives of operating as both a responsible and responsive policymaker in the context of modern regulatory states.

The responsible and responsive modes of governance are relevant for understanding the EC’s institutional and reputation-seeking behaviour, in light of its complex role within the EU regulatory state: that of an executive bureaucracy with extensive regulatory powers that was initially designed to engage in non-majoritarian, evidence-based and depoliticised policymaking (Reh et al., 2020, p. 425), and that of a ‘proto-executive’ with important formal and informal agenda-setting powers that has assumed an increasingly more active political role within the supranational polity in recent years (Reh et al., 2020, p. 420). Informed by this complex institutional role, the EC engages in ‘strategic legitimation’ and attempts to strategically build different reputations with different constituencies of interests and stakeholders across different issues and policy areas (Schimmelfennig, 2020). Research shows that while trying to maintain its longstanding reputation for being an expert-informed and evidence-based policymaker that proposes efficient and effective policy solutions (Rauh, 2016), the Commission made sustained efforts in the last two decades to build a reputation for being a responsive institution that takes into account and ‘listens to’ both the public opinion (Haverland et al., 2018) and stakeholders’ feedback (Meijers et al., 2019) when formulating policies and regulations. These more recent reputation-building efforts emerged in response to increased levels of bottom-up and top-down contestation and criticism targeting the EU political and institutional system in general (Bressanelli et al., 2020), and the EC in particular, as the symbol of non-transparent, unaccountable and remote from the European public supranational decision-making (Reh et al., 2020). They are part of a re-legitimation attempt aimed at ensuring the very survival of the supranational institutional order, of which the Commission is a ‘guardian’, and preserving the Commission’s institutional legitimacy, autonomy and power.

We argue that in the context of this shift from responsible to responsive governance that also characterises the EU polity, the EC has incentives to engage in a double reputation-building act through which it tries to harness both old and new sources of authority, legitimacy and power. Consequently, the manner in which the Commission decides to publicly engage with stakeholder inputs in the context of stakeholder engagement is driven by its decision to build a reputation as both a responsible and a responsive policymaker, for reasons and in a manner of approach we detail next.

Strategic talk and public engagement with stakeholder inputs as a signalling and reputation-building device

The literature on bureaucratic reputation indicates that bureaucracies use both stakeholder engagement and strategic communication as instruments to build multifaceted reputations that support their claim to different forms of legitimacy (Busuioc & Lodge, 2017). We argue that, in the context of stakeholder engagement mechanisms characterised by enhanced transparency and public visibility, the Commission has incentives to strategically use its public responses to signal to its external environment which inputs are prioritised, considered relevant and worthy of attention, and informative for its legislative initiatives and policy programmes. This, in turn, signals the type of reputation, and related legitimacy, the Commission wants to build by using strategic communication in the context of stakeholder participation in policymaking.

We argue that the EC’s decision on whether and how to publicly respond to stakeholder inputs is informed by strategic considerations related to its desire to simultaneously strengthen and consolidate its long-standing reputation as a responsible, evidence-based policymaker, while trying to further cultivate a more recently developed, less established reputation as a responsive bureaucracy. Both reputations are important for its organisational survival, performance and status, and offer the Commission the chance to claim both input and output legitimacy, and to harness both old and new forms of institutional authority. The decision to engage in ‘public talk’ or remain ‘strategically silent’ (Maor et al., 2013) regarding stakeholder policy inputs thus constitutes an observable implication of its reputation-seeking behaviour.

We note first that the image of a responsible policymaker has always been at the core of the Commission’s institutional identity and organisational performance (Rauh, 2016). The Commission was originally designed as a non-majoritarian institution, drawing legitimacy from and embedding its policy initiatives in technocratic, evidence-based, expert-informed and largely depoliticised decision-making (Majone, 2002). While the recent literature speaks of an increasingly more ‘political’ Commission (Nugent & Rhinard, 2019), this institutional role clearly differentiates the Commission from its institutional counterparts (the Council and EP) and significantly contributed in time to its efforts to build and maintain a reputation as a responsible policymaker. Thus, this constitutes a foundational, core and vital reputation that justifies its existence and captures its institutional uniqueness in the EU polity (Busuioc, 2016). In its absence, the organisational survival of the EC would be seriously questioned. The cost of losing this reputation is high and poses a significant threat to the Commission’s institutional autonomy and status. This motivates the EC to further cultivate and consolidate it when inviting stakeholder inputs in supranational policymaking, despite this being a core, well-established reputation, which according to the literature on bureaucratic reputation should not necessitate further efforts to maintain it (Maor et al., 2013). Furthermore, the introduction of stakeholder participation mechanisms may themselves pose a threat to this type of reputation because public participation in policymaking may result in its increased politicisation (Bunea, 2020).³ This may be at odds with the image of a responsible policymaker and creates additional incentives for the Commission to try to preserve its established reputation as a responsible bureaucracy, even in those contexts consistent with the image of a responsive policymaker, such as stakeholder participation in policymaking (Rimkutė, 2020b).

In the context of stakeholder engagement, the Commission’s concern to maintain and consolidate its core reputation is likely consistent with a pattern of engagement that prioritises the substantive content of stakeholders’ inputs when deciding if and how to answer them publicly. We expect the Commission to discern between evidence-based and opinion-based inputs and to prioritise publicly answering the former over the latter. This form of ‘strategic talk’ allows the Commission to signal outside audiences that its decisions and policy initiatives are informed by evidence and expertise in line with the principles of evidence-based policymaking. Furthermore, to strengthen this signal, the Commission should be more likely to indicate in its responses that evidence-based inputs inform its initiatives. Thus, we expect that:

H1.1: The Commission is more likely to publicly engage with and answer stakeholder inputs that are evidence-based.

H1.2: The Commission’s responses are more likely to indicate that evidence-based inputs inform its initiatives and programmes.

Second, we note that the Commission has equally strong incentives to cultivate its newer, and, by comparison, weaker reputation that is more sensitive to public oversight and criticism, as responsive to public demands supranational bureaucracy (Rauh, 2016). Research shows that bureaucracies are particularly likely to engage in strategic public communication when receiving criticism from external audiences on issues that fall within their competences but for which their current reputation is weak or on issues on which their reputation is still evolving (Maor et al., 2013, p. 586). We build on this and further note that the Commission’s reputation as a responsive policymaker is a fairly recent development in its institutional history and constitutes a direct consequence of the increased politicisation and contestation of European integration and of the Commission’s role as its main institutional upholder (Reh et al., 2020). Building this reputation is an attempt to address such contestation and employ responsiveness as a re-legitimation strategy (Bressanelli et al., 2020, p. 337).

To cultivate this reputation in the context of stakeholder engagement mechanisms, the Commission has incentives to prioritise the source of inputs in an attempt to signal to external audiences which stakeholders represent relevant information sources and inform its initiatives. By implication, the type of stakeholder (citizens, business organisation, etc.) providing inputs represents an important feature that describes the source of comments. Publicly engaging with the inputs offered by only some stakeholders allows the Commission to signal which constituencies of interests it prioritises and responds to. Engaging publicly with inputs from citizens and organisations representing non-business interests supports well the Commission’s efforts to build a reputation as a regulator that is responsive to the public broadly defined. These stakeholders are more likely to represent the European public interest, and constitute a category of stakeholders usually less well represented in the participatory forums of supranational policymaking, despite the Commission’s sustained efforts to encourage their participation (Bunea, 2017). This provides the EC with extra incentives to publicly engage with their inputs to encourage their further participation in different engagement mechanisms set up as part of its extensive consultation regime. We derive two expectations consistent with a scenario in which the Commission tries to expand its reputational repertoire and cultivate the image of being a responsive policymaker:

H2.1: The Commission is more likely to publicly respond to inputs from citizens and non-business organisations relative to those from business organisations.

H2.2: The Commission’s responses are more likely to indicate that inputs from citizens and non-business organisations inform its initiatives and programmes relative to those from businesses.

Reputation-building in the REFIT programme

We argue that the ‘Lighten the Load’ portal set up by the EC to receive inputs about how to ‘make laws more effective and efficient’ as part of its REFIT programme offers an ideal context to deploy this multifaceted reputation-building strategy for reasons that we detail below. We also discuss the relation between the REFIT programme, the REFIT Platform and the portal.

First, we note that the EC’s Regulatory Fitness and Performance Programme (REFIT) was established as part of its Better Regulation agenda with the aim ‘to make EU laws simpler, more targeted and easier to comply with’.⁴ As part of this, the Commission established in 2015 the REFIT Platform, an entity that was an integral part of and served as an advisory body to the EC. The Platform was mandated to identify solutions for legislative modernisation and simplification. It consisted of two working groups: a ‘Stakeholder Group’, hosting 18 stakeholders⁵; and a ‘Government Group’, hosting national experts. The Platform solicited and considered inputs about regulatory simplification from citizens and organisations through the ‘Lighten the Load’ portal and from members of its Working Groups. These inputs were assessed by the two groups, and some were adopted as ‘Platform opinions’ through which the Platform made recommendations for concrete action on how to simplify legislation. Next, Commission DGs considered the recommendations and communicated follow-up actions through the EC Work Programmes and the REFIT Scoreboard.⁶ In line with the principles guiding its work (transparency, inclusiveness and responsiveness), the Platform, assisted by the EC Secretariat-General, managed its communication with stakeholders submitting inputs through the portal by employing the same online interface.

We focus empirically on this part of the Platform’s work that consisted of managing its (i.e., the EC’s) communication with stakeholders. We discern between two distinct aspects of this communication: whether the EC publicly responds to an input; and, the substantive content of this reply that identifies which inputs have informed the Platform’s adopted opinions and by extension the REFIT programme.

The Platform and portal facilitated stakeholders’ participation in retrospective regulatory review, a practice used to streamline legislation and usually associated with evidence-based approaches to regulatory governance, drawing extensively on expert-knowledge and cost–benefit analyses (DeMenno, 2017). By inviting public and direct stakeholder input into this, the EC introduced an important participatory and transparency dimension to a process that was previously the prerogative of more secluded, expert-based decision-making forums. Injecting public participation into elite regulatory governance marked a shift from responsible to responsive policymaking in this policy area and was consistent with other measures adopted as part of the EU Better Regulation agenda over recent years and especially during its last major reform in 2015–2016. This reform aimed to significantly develop and institutionalise the participatory dimension of supranational policymaking while strengthening its evidence-based credentials. It was a simultaneous attempt to enhance the EC’s input, throughput and output legitimacy and build a reputation for being both responsible (evidence-informed) and responsive to the public bureaucracy (Bunea, 2020).

This combination of evidence-based and participatory aspects makes the REFIT programme in general, and the EC communication on ‘Lighten the Load’ in particular, a relevant and substantively interesting testing ground for our theoretical argument given that the Commission had incentives to cultivate and project on its outside audiences its image as both a responsible and responsive actor when it comes to legislative simplification. On the one hand, the Commission had incentives to harness its old source of authority and legitimacy and maintain its reputation as a responsible, evidence-based policymaker to avoid the risk of being criticised for using its legislative simplification programme to push forward a de-regulation agenda in key areas such as consumer protection and public health (Bunea & Ibenskas, 2017). Conversely, the Commission had incentives to use this as an opportunity to cultivate its newer image as a responsive policymaker on a policy matter that was traditionally the realm of insulated, technocratic decision-making, and thus build capital for and avoid losing credibility regarding its commitment to expand the scope of responsive and inclusive policymaking at the supranational level.

Furthermore, relative to other EC stakeholder engagement mechanisms, the portal presented two notable features. First, the Commission exerted no control or agenda-setting power over stakeholders’ participation or the content and timing of their inputs. The Commission did not ask stakeholders to provide input on specific policy initiatives or legislative proposals, as is usually the case, for example, with public consultations or feedback opportunities. Second, the portal was one of the few instances where stakeholders’ individual inputs received EC answers that were publicly visible and fully accessible for scrutiny in real time, making it easy to match the reply to individual inputs. The Commission’s usual practice of responding to inputs in the context of other participatory mechanisms (e.g., public consultations) is to respond in an aggregate and indirect manner by referring to levels of aggregate support for different policy options gathered amongst different categories of stakeholders, and indicating which such options informed its decisions, thus circumventing the possibility of responding to individual stakeholder inputs. The portal allowed a unique level of scrutiny over stakeholders’ inputs and the Commission’s answers and created one of the few instances in which the EU consultative regime presents resemblances to a more interactive form of regulatory governance.

Finally, we note that, similar to other stakeholder engagement mechanisms, the Commission enjoyed full discretion in deciding which inputs receive a public answer. Our dataset shows the Commission answered publicly to approximately one-third of inputs. The substantive content of these answers varied: some indicated the inclusion of comments in the Platform’s recommendations/opinions, while others referred to procedural or competence aspects to explain why inputs did not inform the Platform’s opinions. This in turn presents a relevant variation describing the EC’s public communication in the context of stakeholder participation in supranational policymaking that we are able to examine empirically in our study.

Research design

We test our argument on a new dataset constructed in several stages. First, we collected all the information about stakeholder inputs, Commission replies, and stakeholder characteristics (actor type) available on the ‘Lighten the Load’ portal, covering the 11 November 2014 – 23 September 2020 period. The REFIT Platform mandate ended on 31 October 2019, being replaced by the ‘Fit for Future’ Platform in January 2021.⁷ We include 58 comments submitted beyond the formal end of the Platform’s mandate (October 2019) and September 2020 because the EC continued answering these inputs and we considered them relevant for our analysis.

We identified in total 574 comments: out of these, the substantive content of 117 was removed by the Commission due to violations of the feedback rules,⁸ and 11 were removed by us because they were duplicates or spam. We analysed 446 comments, out of which 177 (39.69 per cent) received public EC replies while 269 (60.31 per cent) did not. Most inputs were in English (293), but some (153) were in other EU languages. The EC answered in the language of the input: 112 replies in English and 65 in other languages. Before hand-coding the content of inputs and replies, the coders used Google translate to convert them into English (de Vries et al., 2018).

Stakeholders’ inputs and the EC’s replies consisted of short texts. The average number of words per comment is 176, while per reply it is 169.5. A careful inspection of inputs revealed their idiosyncratic content, which alongside their short length, informed our decision to code their content using qualitative text analysis instead of automated text analysis. Employing a deductive and inductive approach, we developed a coding scheme for the content analysis of inputs, and one for the EC replies. Our coding uses as a unit of analysis the entire input text and the entire reply text.

Dependent variables

We employ two dependent variables to capture two aspects of EC communication relevant for our study: whether the EC publicly replied, and whether this reply indicated the input informed the Platform’s recommendations for legislative simplification and featured amongst its adopted opinions and informing thus the REFIT programme.

The first variable is binary: 177 comments (39.7 per cent) received an EC reply, while 269 (60.3 per cent) did not. The second variable is binary and discerns between two broad categories of replies based on their substantive content: those indicating that the comment informed one of the Platform’s opinions/recommendations (Adoption) versus all other replies. This variable was constructed in two stages. First, we identified four types of substantive replies: Adoption (the reply indicates the comment was included in a Platform opinion); Policy substantive (the reply discusses a specific policy matter); Competence (the reply indicates the comment points to a policy issue that falls outside the EU’s competence or the REFIT remit); and Procedural (the reply refers to procedural aspects; e.g., it indicates the comment refers to legislation under formulation or consultation, or more recent than the past two years). Table A1 in the Appendix exemplifies all reply categories. Their distribution was as follows: 16 (9 per cent) Adoption; 31 (17.5 per cent) Policy substantive; 77 (43.5 per cent) Competence; and 50 (28.2 per cent) Procedural. Second, we created a dichotomous variable distinguishing between replies indicating Adoption (1) and all other replies (0). We use this in our regression analysis examining the probability an input informs the Platform’s opinions.

Explanatory variables

Stakeholder inputs

Following a careful reading of stakeholder inputs, we coded their content to discern between evidence-based and opinion-based comments. The key difference is the extent to which the former relies on numbers, statistics, paper citations or a detailed, reasoned explanation for the conveyed message, while the latter lacks all these elements and employs statements of personal opinions, preferences or recommendations in the absence of the aforementioned elements. Table A2 in the Appendix exemplifies our coding. A binary variable discerns between (1) evidence-based (45 per cent) and (0) opinion-based (55 per cent) inputs.

Stakeholder type

For each stakeholder we identified their actor type based on the portal. A categorical variable distinguishes between 67 business stakeholders (15 per cent); 31 non-business stakeholders (7 per cent, including academic/research institutions, consumer organisations, NGOs, public authorities, trade unions); and 306 citizens (68 per cent). Forty-two contributions were anonymous (9 per cent). Based on their substantive content, we coded them as inputs submitted by citizens since they were written in a manner similar to other citizen inputs in our dataset.

Control variables

We control for several factors that may shape the Commission’s propensity to publicly answer stakeholders’ inputs and include them in the REFIT Platform’s opinions.

First, we control for an important input characteristic, namely whether it indicated specific recommendations for regulatory simplification or not. Our content analysis revealed that stakeholders expressed a variety of views regarding regulatory simplification. Some comments indicated specific simplification recommendations, while others suggested the maintenance of the regulatory status quo, or asked for increasing the amount of EU legislation and regulation across different issues and policy areas. Given the goal of the REFIT programme, we control for the intuition that the Commission was more likely to publicly answer and adopt inputs referring to the reduction of regulation/legislation, while disregarding the others because they were substantively less relevant. Thus, we coded a categorical variable labelled Comment demand type discerning between inputs asking for: a reduction of legislation/regulation (23.1 per cent); an increase in regulation (50.7 per cent); the harmonisation of regulation (14.6 per cent); or made a neutral statement regarding simplification (11.7 per cent). Table A3 in Appendix exemplifies these categories.

Second, we control for the argument that bureaucracies are more likely to pay attention to and publicly engage with negative public comments/evaluations (Maor et al., 2013). This constitutes a classic reputation management technique. We construct a variable capturing the tone of the stakeholder comment and control for the possibility that inputs with a more negative tone are more likely to be answered. We construct this variable at comment level by using sentiment analysis conducted with the help of the ‘syuzhet’ R package and the NRC dictionary available in this package (Mohammad & Turney, 2013). We computed the overall negative tone by subtracting the number of words with positive sentiment from the number of words with negative sentiment. Higher scores indicate that the comment contained mostly words with a negative connotation. This measure ranges from most positive words (−25) to most negative words (12) per comment (mean −3.75; sd. 4.77).

Third, we computed a measure of text readability for each comment. This allows controlling for the impact of the comments’ textual complexity (readability) on the EC’s decision to respond. More complex (and less easy to read) texts should convey more complex information and be more aligned with the standards of evidence-based/responsible policymaking. This in turn makes comments with higher textual complexity more likely to receive an answer and inform the Platform’s opinions. We use the Dale-Chall (1995) measure of text readability, which is based on the length of sentences and includes a measure of the proportion of ‘difficult words’ that do not appear on a list of common words. A lower score indicates that the text is less readable and more complex. The textual complexity of inputs ranges from −50.9 to 47.6 (average 15.9; sd. 13.2).

Lastly, we control for the policy topic to which the comment refers to. This accounts for the intuition that, since the level of European integration and supranational regulation varies across policy areas, the Commission might be more likely to engage with and adopt comments referring to policy issues and areas well-integrated and characterised by high levels of supranational legislation. As we do not have any theoretical assumptions signalling which topics might emerge from stakeholder inputs, and to overcome the challenge of creating a categorisation scheme and sorting documents into categories, we use quantitative text analysis. As several comments mention more than one topic, we used a document-level mixed-membership topic model (LDA). These models increase the accuracy of the topic models as compared to single-membership models (Grimmer & Stewart, 2013). We used the R ‘stm’ package to estimate the topics and calculate the expected topic proportions for each comment. We decided an 8-topic model suits our data well, after statistical testing based on a trade-off between semantic coherence and exclusivity (Roberts et al., 2014) and evaluating whether the topics made substantive sense. Table A5 in Appendix shows the most representative words for each topic. A thorough reading of the most representative documents for each topic was necessary to fully interpret and systematise the topics. We labelled the topics as follows: Consumer protection/product regulation (average topic proportion: 0.09); Environment/energy (average topic proportion: 0.18); Social policy (average topic proportion: 0.09); Institutional affairs (average topic proportion: 0.10); REACH/chemical regulation (average topic proportion: 0.08); Single market (average topic proportion: 0.17); Transport (average topic proportion: 0.11); Laws and human rights (average topic proportion: 0.15).

As the measurement for each topic is the expected proportion of the comment that corresponds to each of the eight topics, they together necessarily equal 1. This creates an issue of multicollinearity. We thus decided to exclude one topic from the analysis (REACH/chemical regulation) based on it being the least prominent.

Analyses

We tested our hypotheses in two stages with the help of logistic regression models. First, we examined the probability that the Commission answers a comment publicly (H1.1 and H2.1). Table 1 presents the results of the logistic regression analyses for all inputs and EC replies.

Table 1. Explaining the probability of an input receiving a public EC reply.

	Model 1	Model 2
Evidence-based comment	1.638*** (0.218)	2.055*** (0.280)
Stakeholder type (Ref. category: Business)
Citizens	−0.572* (0.298)	−0.699** (0.330)
Non-business	−0.310 (0.462)	−0.392 (0.484)
Controls
Comment demand type (Ref. category: Reduce legislation)
Neutral		0.794** (0.402)
Increase		−0.216 (0.289)
Harmonise		−0.388 (0.365)
Negative tone		0.043* (0.025)
Comment text complexity		0.007 (0.010)
Policy topic
Consumer protection/product regulation		1.028 (0.765)
Environment/energy		0.263 (0.767)
Social policy		1.333 (0.818)
Institutional affairs		1.473* (0.796)
Single market		0.759 (0.750)
Transport		1.308 (0.797)
Laws and human rights		0.477 (0.819)
Intercept	−0.757** (0.308)	−1.533** (0.660)
Log-Likelihood	−260.838	−250.334
AIC	529.676	532.668
N	446	446

Note: Logistic regressions. Standard errors in parentheses. *p < 0.1; **p < 0.05; ***p < 0.01.

Models 1 and 2 support hypothesis 1.1 and show that evidence-based inputs are significantly more likely to receive a response than opinion-based ones. However, contrary to hypothesis 2.1, models 1 and 2 indicate citizens are significantly less likely than businesses to receive a reply. Non-business organisations are also less likely to receive a reply, but the difference is not statistically significant.

Table 2 presents the predicted probabilities of receiving an EC reply computed for our two explanatory variables, across their categories. Across stakeholder types, the predicted probability of an evidence-based comment to receive a reply is significantly higher than that of an opinion-based comment. The predicted probability of an evidence-based comment to receive a reply when the comment comes from a business actor is 0.47 higher than that of an opinion-based comment. For citizens, this probability is 0.46 higher, whereas for non-business organisations it is 0.47 higher. This is consistent with our argument that the Commission wants to consolidate its reputation for being a responsible regulator by displaying a systematic higher propensity to publicly answer inputs aligned with the standards of evidence-based policymaking.

Table 2. Predicted probabilities of an EC reply in model 2 (95% CIs in parentheses).

Comment	Stakeholder
	Business	Citizen	Non-business
Evidence-based	0.78 (0.65; 0.89)	0.64 (0.50; 0.77)	0.70 (0.50; 0.89)
Opinion-based	0.31 (0.17; 0.48)	0.18 (0.10; 0.28)	0.23 (0.09; 0.45)

Contrary to expectations, for both evidence-based and opinion-based inputs, citizens and non-business actors are less likely to receive an EC reply relative to business. The predicted probability that a citizen’s evidence-based comment receives a reply is 0.64; for business this probability is 0.78. For opinion-based comments, the probability of a reply is 0.18 for citizens and 0.31 for businesses. Comments from non-business organisations are less likely (by 0.08) to receive a response relative to businesses regarding evidence-based and opinion-based comments, although these differences are not statistically significant.

Most control variables do not have a significant impact. The notable exception is the negative comment tone indicating that indeed negative comments are more likely to receive a reply. The change in the predicted probability of receiving a reply, when the negative tone variable changes from its minimum to its maximum, marks an increase of 0.29. Also, inputs providing neutral comments are significantly more likely to receive a public response compared to comments suggesting a reduction of legislation/regulation. This might be because many comments in this category usually asked a question or made a comment about the portal, which the EC might have felt compelled to answer in order to legitimise and optimise its functioning.

In the second stage, we tested hypotheses 1.2 and 2.2 (positing that evidence-based inputs and those coming from citizens and non-business actors are more likely to be consider for inclusion in the Platform’s opinions), by analysing the sub-sample of inputs that received an EC reply. We proceed in two stages. First, in Table 3, we cross-tabulated the categories of our explanatory variables with all EC reply categories we identified.

Table 3. Distribution of EC reply categories per comment and stakeholder type.

EC reply	Adoption	Competence	Procedural	Policy substantive
Comment
Opinion-based	0 (0.0%)	29 (53.7%)	19 (35.2%)	6 (11.1%)
Evidence-based	16 (13.0%)	50 (43.7%)	32 (26.0%)	25 (20.3%)
Stakeholder
Business	8 (20.0%)	13 (32.5%)	15 (37.5%)	4 (10.0%)
Citizens	6 (5.0%)	56 (46.7%)	33 (27.5%)	25 (20.8%)
Non-business	2 (11.8%)	10 (58.8%)	3 (17.6%)	2 (11.8%)

The descriptive statistics show that no opinion-based comments received a reply indicating they were considered for the Platform’s opinions. However, 13 per cent of evidence-based inputs did receive such a reply, suggesting that this category was more likely to be considered relevant for the Platform’s work relative to the opinion-based ones, in line with hypothesis 1.2. Nevertheless, only a small proportion of evidence-based inputs received this reply: the majority received EC replies referring to competence, procedural or substantive policy aspects, which does not support hypothesis 1.2. Table 3 shows that only 5 per cent of citizens’ comments and 11.8 per cent of non-business actors’ comments were considered relevant/informative for the Platform’s opinions, while as many as 20 per cent of businesses’ comments received such a reply. This contradicts hypothesis 2.2.

Second, we used logistic regression to provide a formal test of hypotheses 1.2 and 2.2. We employed a dependent variable discerning between inputs that received an ‘adoption’ reply (1) vs. those receiving any other type of reply (0). Since the number of observations with ‘adoption’ replies is low (16), and as there are no cases in which the evidence-based comment covariate takes the value 0 and no ‘adoption’ occurs, we used the penalised maximum likelihood estimation. Models 3 and 4 in Table 4 report the estimates of models with and without control variables.

Table 4. Explaining the probability of receiving an EC reply indicating comment consideration for the REFIT platform’s opinions.

	Model 3	Model 4
Evidence-based comment	2.534* (1.429)	1.437 (1.123)
Stakeholder type (Ref. category: Business)
Citizen	−1.197** (0.566)	0.136 (0.683)
Non-business org.	−0.351 (0.814)	1.127 (0.981)
Controls
Comment demand type (Ref. category: Reduce legislation)
Neutral		−2.325 (1.569)
Increase		−2.018*** (0.746)
Harmonise		−1.358* (0.822)
Negative tone		−0.085 (0.059)
Comment text complexity		−0.033 (0.021)
Policy topic
Consumer protection/product regulation		−0.262 (1.337)
Environment/energy		−0.567 (1.560)
Social policy		−1.647 (2.364)
Institutional affairs		−2.219 (1.691)
Single market		0.476 (1.600)
Transport		−1.457 (1.752)
Laws and human rights		0.532 (1.780)
Intercept	−3.709** (1.458)	−1.923 (1.570)
Log-Likelihood	−45.871	−37.920
AIC	99.743	107.841
N	177	177

Note: Penalised maximum likelihood logistic regressions. Standard errors in parentheses. *p < 0.1; **p < 0.05; ***p < 0.01.

Model 3 indicates that evidence-based inputs were significantly more likely to receive an ‘adoption’ reply: for these the predicted probability of receiving this reply increases by 0.21 for businesses, 0.08 for citizens and 0.16 for non-business actors. When control variables are included, the effect loses statistical significance, although it remains substantial in size. Given the relatively small size of the sample and high number of control variables, this is not surprising. Together, these findings provide mixed evidence regarding hypothesis 1.2.

Model 3 shows that business actors’ inputs are more likely to receive an ‘adoption’ reply relative to citizens and non-business actors, although the difference is statistically significant only for citizens. This contradicts hypothesis 2.2. These differences disappear when all control variables are included (model 4). Here, the coefficients of stakeholder type are positive and substantial in size for non-businesses but close to zero for citizens. We conclude that there is no evidence supporting hypothesis 2.2 and the claim the Commission made systematic efforts to build a reputation for being responsive to the general public by being more likely to engage with and provide public answers to inputs from citizens and non-business actors.

Overall, our analyses show that the EC prioritised the substantive content of stakeholder inputs and focused less on their source when deciding when and how to publicly answer them, in line with evidence-based policymaking.

Conclusion

We examined the strategic use of public communication in the context of stakeholder engagement mechanisms as a reputation-building strategy used by bureaucracies to cultivate multifaceted institutional reputations. We focused on how the EC engaged with and publicly answered stakeholders’ policy inputs as a reputation-building strategy. Similar to other bureaucracies, in the face of increased public contestation and criticism, the Commission underwent a shift from a responsible to a more responsive mode of (regulatory) governance in the recent period. This, in turn, created a need to tap into different sources of institutional authority and legitimacy, and build a multifaceted reputation for being simultaneously an expert-informed (evidence-based) and responsive to the public’s demands policymaker. We argued this double reputation-building act helped the Commission to address the challenge of having to tap into different and sometimes difficult to reconcile sources of power and legitimacy, and shaped its public communication and engagement with stakeholders’ inputs. We tested our argument in the context of stakeholder participation in the EC’s legislative simplification programme, a telling example of the EC’s efforts to shift towards a more responsive mode of governance in an area that was traditionally the realm of insulated, expert-informed decision-making.

Our findings revealed a pattern in the EC’s engagement with stakeholders’ inputs consistent with a reputation-building behaviour that used strategic public talk mainly to consolidate its foundational, long-established, core reputation for being an evidence-based policymaker, and less to cultivate its more recent, less established image as a responsive to the public institution. This contrasts the research on bureaucratic reputation arguing that bureaucracies have fewer incentives to consolidate their core, well-established reputations through strategic talk, while having stronger incentives to use public communication to cultivate their newer and less established reputations when facing reputational threats and public scrutiny (Maor et al., 2013). In this respect, our findings are consistent with previous research on EU agencies’ reputation-seeking behaviour, and suggest the EC reputation-building efforts resemble in many ways that of its regulatory agencies, which use public (media) communication (Busuioc & Rimkutė, 2020a; Müller & Braun, 2021; Rimkutė, 2020a) and/or stakeholder engagement (Busuioc & Jevnaker, 2022; Rimkutė, 2020b) to consolidate their technical (evidence-based) and performative reputations, and emphasise their output legitimacy in line with the founding principles of the EU regulatory state (Busuioc & Rimkutė, 2020b). Similar to the EU regulatory agencies and the European Central Bank, the EC attempts to diversify its institutional reputation and public image across time and policy contexts, but in doing so it does not want to compromise its reputational and institutional uniqueness in the EU polity, i.e., that of being an evidence-based, expert-informed, responsible policymaker. This, in turn, reiterates the importance of expert-informed, evidence-based policymaking for the EC as a source of authority, legitimacy and power, and indicates the long-lasting effects of foundational institutional reputations for how bureaucratic actors in general, and the EC in particular, develop their reputation-seeking behaviour across time (Busuioc & Rimkutė, 2020a). Furthermore, in line with previous research on the EC’s stakeholder engagement mechanisms (Binderkrantz et al., 2021; Bunea, 2019), our findings re-emphasise that the Commission’s use of this reputation-building instrument may vary across time and circumstances and may be deployed in different ways according to policy goals, contexts, and its (re-)legitimation needs and objectives (Bressanelli et al., 2020). Last but not least, our findings also corroborate the most recent research on the Commission’s public communication through press releases: examining the Commission’s communication style across thirty-five years and almost 45,000 press releases, Rauh (2021) found that when communicating to the European public, the European executive adopted an ‘extremely technocratic style of communication’ characterised by the use of ‘complex language’ and ‘specialised jargon’. This communication style is consistent with the principles and practice of evidence-based policymaking and the image of a responsible policymaker that our findings also point to.

Several implications follow from our findings. First, similar to other supranational institutional actors, the EC’s reputation-building behaviour seems context-dependent.⁹ Our findings are consistent with previous research indicating that when it comes to reputation and legitimacy-building strategies, the EC is adaptable and aptly adjusts its efforts to the political and policy context in which it operates and needs to build or defend its power and legitimacy (Bunea, 2020; Koop et al., 2022; Van der Veer, 2021). Therefore, any attempt to develop a theoretical model and/or empirical analysis of its reputation-building behaviour requires an in-depth understanding of the policy and political context in which the Commission operates and deploys different strategies for re-legitimation and reputation-building. Second, our findings suggest that, similar to other bureaucratic actors, the Commission may face important trade-offs when adding an extensive participatory layer to its time-tested, evidence-based policymaking in its attempt to address public criticisms and build the image of being a responsive to the public policymaker. When stakeholder participation results in inputs of varying informational quality and extent to which they meet the standards of evidence-based policymaking, the Commission may have to prioritise defending its image of a responsible regulator and sacrifice that as a responsive one because the reputational cost of being perceived as an institution making decisions based on low-quality information and policy evidence is too high and seriously threatens its reputational and institutional uniqueness. This points to an important conundrum bureaucracies may face when introducing stakeholder engagement into their decision-making mechanisms (Braun & Busuioc, 2020) and the potential informational costs and reputational challenges of such a move (Busuioc & Rimkutė, 2020b). While in the short term this choice helps preserve a time-honoured, desired, institutional image, in the long-term, it might bring reputational costs and criticisms by providing a relevant example of how the Commission is unresponsive to the general public’s inputs on relevant issues for the EU regulatory state. Relatedly, our findings show how even in the context of instruments that may help build a reputation for responsiveness (i.e., stakeholder engagement mechanisms), bureaucracies may choose to cultivate their image of evidence-based decision-makers. This suggests the need to assess empirically the extent to which participatory instruments and procedures adopted by bureaucratic actors actually support responsive policymaking or serve as mere symbolic efforts to create the image of responsive policymaking by for example inviting citizens’ inputs into policymaking but then disregarding them because they may not match the information needs and standards of evidence-based policymaking.

Our study developed a general argument about the EC’s reputation-seeking behaviour through public communication in the context of stakeholder engagement in policymaking, but tested it empirically in the context of only one policy stage (i.e., policy evaluation) and policy issue (i.e., legislative simplification and modernisation). This is a limitation that future research could fruitfully address by expanding the empirical remit of analysis and testing our argument in the context of other policy stages in which the Commission receives policy inputs from stakeholders and the general public (i.e., agenda-setting and policy formulation) and across different policy issues and policy domains. This would require overcoming the challenge of identifying EC’s responses to individual stakeholder inputs in the context of public consultations and feedback mechanisms highlighted earlier in our study, and would provide important comparative insights into how the EC articulates its reputation-seeking behaviour and uses public engagement with stakeholders’ inputs to this end in different policymaking contexts.

Acknowledgements

The authors would like to thank Christel Koop, Eva Ruffing, Raimondas Ibenskas, colleagues attending the departmental and PhD-seminars at the Department of Comparative Politics, University of Bergen, and the reviewers for their very helpful and insightful comments on this manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This article is part of a project that has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (ERC StG 2018 CONSULTATIONEFFECTS, grant agreement no. 804288). H2020 European Research Council.

Notes on contributors

Adriana Bunea

Adriana Bunea is professor at the University of Bergen, Department of Comparative Politics.

Idunn Nørbech

Idunn Nørbech is PhD candidate at the University of Bergen, Department of Comparative Politics.

Notes

1 The EC Secretariat-General provided the Platform’s secretarial services. The EC Vice-President Timmermans chaired the Platform’s plenary meetings. The chair of the EC Regulatory Scrutiny Board chaired its two Working Groups. The Platform was part and parcel of the EC. We thus refer to the Commission in our theoretical argument and to the Platform and its online portal in the case selection and research design.

2 Following Carpenter and Krause (2012, p. 26), we define reputation as ‘a set of beliefs about an organization’s capacities, intentions, history and mission that are embedded in a network of multiple audiences’. Bureaucracies make intentional efforts to shape these beliefs by pursuing different reputation-building strategies with the aim to project a certain institutional image onto external audiences and maintain/enhance their institutional autonomy and power. The literature indicates some reputations are established early in time, during the foundational moments, while others emerge later in their organisational life (Busuioc & Rimkutė, 2020b). This is the case for the EC’s reputation for being a responsible policymaker (foundational reputation) vs. a responsive policymaker (recent reputation).

3 See Müller and Braun (2021, p. 679) for a similar argument about ECB’s fears of politicisation of its core competences and how this shapes its public communication and reputation-seeking behaviour.

4 https://ec.europa.eu/info/law/law-making-process/evaluating-and-improving-existing-laws/refit-making-eu-law-simpler-less-costly-and-future-proof_en (Accessed: 22 March 2022).

5 Nine civil society organisations, six business organisations, three social partners, one representative of the Committee of Regions and one of the Economic and Social Committee.

6 EC, Detailed information on the follow-up by the Commission to REFIT Platform opinions: 2018 annual burden survey, Publications Office, 2019, https://data.europa.eu/doi/10.2792/64888.

7 https://ec.europa.eu/info/law/better-regulation/have-your-say-simplify (Accessed: 26 January 2021)

8 https://ec.europa.eu/info/law/better-regulation/rules-feedback-and-suggestions_en (Accessed: 13 June 2021).

9 See the 2020 Journal of European Public Policy 27(3) special issue on ‘EU Actors Under Pressure: Politicisation and Depoliticisation as Strategic Responses’.

Appendix

Table A1. Examples of EC reply categories.

Type of reply	EC reply text
Adoption	‘On behalf of the REFIT Platform, we would like to inform you that on 7 June 2017 the REFIT Platform has adopted an opinion on your suggestions regarding “identity and travel documents” (REFIT Platform ref. XIII.3.a). The adopted Opinion has also been published on the REFIT Platform’s website. The follow-up to this adopted opinion will be made in the context of the preparation of the CWP2018 and published on our website thereafter. With many thanks again for your interest in these issues’
Procedural	‘On behalf of the REFIT Platform, thank you for submitting your suggestion on the efficiency, effectiveness or burden reduction of EU regulation or its implementation by Member States. The Platform has concluded that it does not intend to prepare an opinion on the issues raised. This is because the issues are covered by an EU legislative measure which only entered into application less than two years ago and so there is too short a period of experience of the application of the measure for sound conclusions on its performance to be drawn. Your suggestion has been forwarded to the Commission service responsible for the relevant policy area which has produced an explanation of the current context, which we attach for your information. With thanks again for your suggestion which will be retained for future reference’
Competence	‘On behalf of the REFIT Platform, thank you for submitting your suggestion on the efficiency, effectiveness or burden reduction of EU regulation or its implementation by Member States. The Platform has concluded that it does not intend to prepare an opinion on the issues raised. This decision is based on the fact that this submission is a competence of the Member States, and falls beyond the remits of the REFIT Platform. We would also like to provide you with a reference to the Circular Economy Strategy and Plan: http://ec.europa.eu/environment/circular-economy/index_en.htm’
Policy substantive	‘Commercial rates of interest, commissions and other charges are based on many factors. One of the main factors for the rate of interest is the risk that an individual loan might pose to the bank. For instance, a company with a low risk profile will most likely receive an offer with a low rate, while the offer for a company with a higher risk profile will most likely include a higher rate. These market forces are something that the European Union cannot, and should not, influence. The Commission can however implement policies to help reduce asymmetric information exchange between banks and companies, to ensure that there is a clearer link between the cost of the loan and the risk it really entails. The Commission also supports diversity in the sources of financing for companies for instance by supporting credit unions so that companies can be less dependent on bank financing’.

Table A2. Examples of evidence-based and opinion-based stakeholder comments.

Type of comment	Comment text
Evidence-based	‘Please find in the attachment suggestions and activities recognized at working-seminars (2016) for businesses and NGO’s, that could be implemented at the EU level. They cover 7 policy areas: collaborative/ sharing economy, intellectual property framework, public procurement, mutual recognition, prevention of non-compliant products in the EU market, cloud computing and free flow of data, reducing VAT related burdens and obstacles when selling across borders’.
Opinion-based	‘The gun ban that EU is pushing ahead is illegal. Primarily Germany and France, but also other southern EU countries have messed up the whole idea of EU with their sloppy immigration politics and are now trying to make smaller and northern countries suffer because of their stupid & criminal decisions made under influence of blind humanism’.

Table A3. Examples of stakeholder comment demand type.

Category demand	Comment text
Reduce	‘Radiocentre’s concerns on behalf of the UK commercial radio industry relate to Article 4 of Consumer Credit Directive 2008/48/EC (CCD) which requires standard information to be provided in any credit advertising indicating an interest rate or cost of credit. This affects advertising for financial services, retail and motors, typically adding 12–15 s of airtime to a radio advert. Our concerns include: Cost: we estimate that complying with Article 4 of the CCD costs commercial radio stations and advertisers around €95 million a year in the UK alone. Failure to inform and protect consumers: independent research commissioned by Radiocentre shows low levels of recall of key figures – only 4% of listeners recalled the total amount payable Access to market: advertisers have to pay for additional airtime because of the CCD – a burden, which few advertisers, let alone SMEs, can afford Platform neutrality: radio is the only medium where this information has to be read out in full so the CCD has a disproportionate impact on the radio sector Inconsistency with existing EU consumer law: the Unfair Commercial Practices Directive, which pre-dates the CCD, acknowledges that consideration should be given to “limitations of space or time” of the medium used to communicate No impact assessment: no impact assessment was conducted by the Commission prior to the adoption of the Directive – a key component of Better Regulation. Suggestion: We fully support the principles of the CCD, however consumers ignore information which is too complex or difficult to remember and there is a lot of evidence that simpler information with fewer figures is much more effective at landing critical messages at the advertising stage. Radiocentre have invested (and continue to invest) in independent research exploring the effect of different treatments and this has already highlighted that limiting the information can lead to a three-fold increase in recall of critical figures. We would therefore like to be able to work with Members of the REFIT platform, advertisers, regulators and consumer groups to deliver a solution for the benefit of consumers and businesses’.
Increase	‘Dear Comission, Is there any laws in the planning stage that will restrict the use of packaging to a minimum necessary and fine overpackaging? This kind of law is necessary now and will be even more necessary for the future. It will help to educate the market to a more responsable use of resourced and will reduce general trash impact in Europe. Me as a consumer, I feel forced by the industry and market to buy products which are completely over packed, so i am forced to produce an amount of garbage which in my point of view is insane, considering the state of the world. With many products there is no option, as all brands use similair (& unnecessary) packaging. How can one, as a citizen help to forward this kind of law? Thanks very much for your attention and response!’
Neutral	‘I understand that there is to be an on-line consultation regarding REFIT and the Habitats and Birds Directives which is due to take place in April 2015 – so far I have been unable to find it. The on-line consultation for the banning of driftnets passed un-noticed by most organisations and the NFFO would regard a similar exercise for REFIT as unacceptable unless the on-line questionnaire is widely publicised’.
Harmonize	‘EU Driving Licences. It was decided in 2013 I believe that these permits should be uniform across the Member States. The EU did not go far enough as each licence must be changed or renewed in each Member State when exercising freedom of movement. It would be better for all citizens if there was one universal EU driving licence permitted to be used in each State’.

Table A4. Policy topics identified with the help of structural topic models (STMs).

Name of topic	Min	Mean	Max
T1: Consumer protection/product regulation	0.001	0.094	0.986
T2: Environment/energy	0.003	0.181	0.934
T3: Social policy	0.0002	0.091	0.985
T4: Institutional affairs	0.001	0.103	0.947
T5: REACH/chemical regulation	0.0004	0.083	0.984
T6: Single market	0.0006	0.172	0.953
T7: Transportation	0.001	0.119	0.947
T8: Laws and human rights	0.0008	0.157	0.973

Table A5. Most representative words exclusive to each topic (FREX).

Topic	Top words
T1:	vat, return, advertis, insur, consum, incom, sale
T2:	applic, habitat, water, energi, natur, wildlif, climat
T3:	medicin, italian, famili, educ, women, itali, abus
T4:	attach, smoke, search, recognit, engin, petit, manag
T5:	see, label, substanc, hazard, product, clp, classif
T6:	citizen, countri, languag, immigr, translat, movement, sweden
T7:	region, communiti, speed, account, aid, data, secur
T8:	vote, photo, peopl, websit, social, politician, retir

Table A6. Logistic regression explaining the probability of receiving a public EC reply – ENGLISH TEXTS ONLY (robustness check for Table 1 in the Article).

	Model 1	Model 2
Evidence-based comment	1.311*** (0.272)	1.787*** (0.352)
Actor type (ref. category: Business)
Citizens	−0.805** (0.349)	−0.995** (0.405)
Non-business	−0.668 (0.522)	−0.681 (0.552)
Controls
Comment demand type (ref. category: Reduce legislation)
Neutral		0.804 (0.501)
Increase		−0.308 (0.367)
Harmonize		−0.289 (0.443)
Negative tone		0.062* (0.033)
Comment text complexity		0.0001 (0.013)
Policy topic
Consumer protection/product regulation		1.809* (0.995)
Environment/energy		0.347 (0.913)
Social policy		2.226** (1.006)
Institutional affairs		1.847* (1.062)
Single market		1.435 (0.908)
Transport		1.974* (1.066)
Laws and human rights		0.594 (1.003)
Intercept	−0.458 (0.362)	−1.385* (0.813)
Log likelihood	−174.670	−162.951
AIC	357.341	357.902
N	294	294

Note: Logistic regression models. Standard errors in parentheses for regression coefficients. *p < 0.1; **p < 0.05; ***p < 0.01.

Table A7. Distribution of EC reply categories per comment type and stakeholder type – ENGLISH TEXTS ONLY (robustness check for Table 3 in the article).

EC reply	Adoption	Competence	Procedural	Policy substantive
Comment
Opinion-based	0 (0.0%)	23 (59.0%)	12 (31.0%)	4 (10.0%)
Evidence-based	13 (18.1%)	20 (27.8%)	23 (31.9%)	16 (22.2%)
Stakeholder
Business	8 (25.8%)	8 (25.8%)	11 (35.5%)	4 (12.9%)
Citizens	3 (4.0%)	31 (45.0%)	21 (30.0%)	14 (20.0%)
Non-business	2 (18.2%)	4 (36.4%)	3 (27.3%)	2 (18.2%)

Table A8. Explaining the probability of getting an EC reply indicating comment consideration for the REFIT programme (adoption) – ENGLISH TEXTS ONLY (robustness check for Table 4 in main manuscript).

	Model 3	Model 4
Evidence-based comment	2.444* (1.437)	1.438 (1.220)
Stakeholder type (ref. category: Business)
Citizen	−1.413** (0.690)	0.163 (0.940)
Non-business org.	0.089 (0.887)	1.799 (1.161)
Controls
Comment demand type (ref. category: Reduce legislation)
Neutral		−2.870* (0.939)
Increase		−2.246*** (0.939)
Harmonize		−1.118 (0.893)
Negative tone		−0.119 (0.076)
Comment text complexity		−0.022 (0.032)
Policy topic
Consumer protection/product regulation		1.221 (1.991)
Environment/energy		0.781 (2.238)
Social policy		−0.205 (2.840)
Institutional affairs		−0.085 (2.348)
Single market		1.331 (2.125)
Transport		0.318 (2.462)
Laws and human rights		1.373 (2.503)
Intercept	−3.325** (1.471)	−3.128 (2.025)
Log-Likelihood	−31.931	−24.734
AIC	71.862	81.468
N	111	111

Note: Penalised maximum likelihood logistic regressions. Standards errors in parentheses. *p < 0.1; **p < 0.05; ***p < 0.01.