How backsliding governments keep the European Union hospitable for autocracy: Evidence from intergovernmental negotiations

Abstract

The European Union (EU) is a democratic organization but faces severe cases of democratic backsliding. The literature deems the EU a hospitable environment for and reluctant to reign in backsliding. This study focuses on the tactics that backsliding governments employ to preserve this hospitable environment and the conditions under which they succeed. I argue that backsliding governments seek to repurpose the practice of accommodation that permeates EU decision-making for the protection of their backsliding projects. Doing so promises backsliders an escape from their precarious bargaining position in a democratic organization but comes with constraints. Backsliders must limit opposition carefully to a subset of EU competences, backsliding-inhibiting competences, that threaten their backsliding projects the most. Moreover, they can only rely on accommodation in the Council if the democratic member states perceive opposition as justified and remain insulated from political accountability by Europe’s parliaments. I present evidence based on quantitative and qualitative analyses of bargaining positions, processes, and outcomes in EU decision-making. The results have implications for understanding the EU’s autocratic predicament, the opportunities of backsliding governments, and the role of autocracies in regional and international organizations.

1 Introduction

The European Union (EU) is a democratic regional organization yet faces severe cases of democratic backsliding among its member states. Recent research highlights that EU membership helps fund democratic backsliding (Kelemen, 2020). Moreover, the member states remain deeply reluctant to interfere in core domestic affairs of a fellow government (Closa, 2019, 2021; Emmons & Pavone, 2021; Kelemen, 2022; Kelemen & Pavone, 2023; Oliver & Stefanelli, 2016). I focus on the tactics of backsliding governments. How do backsliders ensure that the EU remains a hospitable environment for them? What tactics do they employ? And what might be the conditions under which backsliding governments’ tactics succeed or fail in important negotiations?

I argue that backsliding governments repurpose the practice of accommodation, also known as the consensus norm, that permeates decision-making in the Council of the EU for the protection of their backsliding projects (Heisenberg, 2005; Lewis, 1998, 2005; Novak, 2013; Smeets, 2015; Thomson, 2011). The consensus norm promises backsliders an escape from their precarious bargaining position in a democratic organization, but it comes with constraints. First, backsliders, as other governments, must limit their opposition to select cases to retain political capital in the Council. Unlike other governments, they can be expected to focus opposition on a distinct sub-set of EU competences, backsliding-inhibiting competences, which threaten their backsliding projects. Whereas they oppose decisions such as the creation of an EU prosecutor to investigate the misuse of EU funds by governments, they support much market and regulatory legislation. Second, during bargaining processes, backsliders can only count on accommodation if the conditions sustaining the practice of accommodation are met. Democratic member states have to perceive backsliders’ opposition as justified and remain insulated from political accountability for accommodating backsliders. Herein – especially in European and national parliamentary accountability of the democratic member states – lies the main risk of bargaining failure for backsliding governments.

I present evidence based on a mixed-methods design. First, I analyze government positions in hundreds of decision-making processes, drawing on the Decision-Making in the European (DEU) dataset (Arregui & Perarnaud, 2022; Thomson et al., 2006, 2012). Second, I examine variation in bargaining processes and outcomes in case studies of the decisions that backsliding governments have opposed the most over the last decade. The two analyses reveal a pattern of selective opposition and variable bargaining success in line with the argument.

The analysis has implications for understanding the EU’s autocratic predicament, the opportunities of backsliding governments, and the role of autocracies in regional and international organizations generally. First, whereas much literature focuses on formal tools to constrain backsliding (e.g., Blauberger & van Hüllen, 2021; Closa, 2021; Oliver & Stefanelli, 2016; Sedelmeier, 2017), I stress informal practices that have only recently attracted attention (e.g., Emmons & Pavone, 2021; Holesch & Kyriazi, 2022; Kelemen, 2020). My institutional argument complements recent work on backsliding governments’ selective use of EU norms at the level of discourse (Emmons & Pavone, 2021). A key lesson is that the EU’s ability to reign in backsliding hinges on the willingness of the democratic governments to adapt their practices to the realities of a mixed-member organization. I find little evidence to suggest that governments are willing to adapt their practices accordingly but signs that parliamentary accountability has the potential to exert effective pressure in this direction.

Second, despite much attention to backsliders’ vetoes in salient cases, a lesson for them is that they obtain favorable results by abiding by procedural norms, such as the consensus norm, and employing them to evade substantive norms and decisions that threaten their backsliding projects. If they fail to do so, and there are signs that especially the Hungarian government might do so increasingly for domestic political reasons, they are likely to fare poorly in negotiations.

Third, beyond Europe, the literature increasingly examines the role of backsliders and autocracies in international organizations in terms of norm contestation (Debre, 2021; Flonk, 2021; Hafner-Burton et al., 2022; Hulvey, 2022; Meyerrose & Nooruddin, 2022). My findings underline the emerging view that these governments contest norms selectively and specifically to suit their regime goals. Furthermore, a lesson here that we also observe elsewhere is that backsliding governments and autocracies master formal and informal procedural norms, only to turn them against an organization’s substantive norms (e.g., Baturo, 2022; Lipps & Jacobs, 2022; Meyerrose & Nooruddin, 2022). The result is an intriguing mix of superficial or procedural socialization and norm contestation.

2 The EU as a hospitable environment for backsliding governments

Even though the EU is democratic in international comparison, the literature suggests that it is a hospitable environment for backsliding governments. There are two sides to this argument. On one hand, Kelemen (2020, pp. 490–494) highlights that backsliding governments benefit significantly from EU membership. EU funding has become a key source of GDP in Hungary and Poland. Moreover, the EU helps attract international investors and, by allowing skeptical citizens to work elsewhere, weakens domestic opposition and generates remittances. Finally, EU funding appears to benefit government supporters disproportionately, suggesting that the EU effectively funds domestic co-optation strategies. This analysis suggests that vocal criticism of the EU by backsliders serves domestic audiences (Schlipphak et al., 2023; Schlipphak & Treib, 2017) and does not indicate genuine skepticism of membership.

The other side of the argument that the EU proves hospitable for autocracy refers to the reluctance of the EU institutions and democratic member states to reign in backsliding. The EU has numerous levers: social pressure and dialogue (Emmons & Pavone, 2021; Oliver & Stefanelli, 2016; Sedelmeier, 2017), infringement procedures and judgements by the Court of Justice (Blauberger & Kelemen, 2017), the suspension of the voting rights under Article 7 of the treaties (Closa, 2021), and freezing funding under the EU’s rule of law conditionality and other procedures (Blauberger & van Hüllen, 2021). Yet, the EU has proven reluctant to use the more forceful of these tools and instead emphasized dialogue (Emmons & Pavone, 2021; Oliver & Stefanelli, 2016). This is despite most tools only requiring qualified majorities or, in the case of court proceedings, no voting. The literature agrees that the EU’s hesitation results significantly from a deeper reluctance of the member states to outvote and penalize another government or even to criticize domestic policies in core state domains (Closa, 2019, 2021; Emmons & Pavone, 2021; Kelemen, 2022; Kelemen & Pavone, 2023; Oliver & Stefanelli, 2016). Hence, the EU seems to be in a predicament with backsliding governments reaping gains and the democratic member states unwilling to reign them in (Kelemen, 2020).

The governments in the Council of the EU are not the only relevant actors. The European Commission commands important tools, including the ability to stop funding from the EU budget through administrative procedures (Closa, 2019; Scheppele & Morijn, 2023). The Court of Justice can issue judgements and financial penalties against backsliding governments (Blauberger & Kelemen, 2017). However, the reluctance of national governments to act against backsliding appears to have had a detrimental effect beyond the European Council and the Council. The Commission has used its tools and supplied the Court with legal cases only reluctantly, not least due to deference to, and anticipation of resistance by, national governments (Closa, 2019; Kelemen & Pavone, 2023). The intergovernmental politics on which I focus thus constitute a critical obstacle for EU action against democratic backsliding.

In contrast, the European Parliament has demanded EU action against democratic backsliding more consistently than the Commission and Council, especially after the center-right group severed ties with the Hungarian governing party after the 2019 elections. Moreover, some national parliaments have taken up the issue intermittently, albeit with governing parties often rejecting strong resolutions against backsliding (Chiru & Wunsch, 2023; Granat, 2023; Meijers & van der Veer, 2019). Yet, whether parliamentary accountability influences intergovernmental negotiations around democratic backsliding has remained unclear.

While existing work highlights the benefits of membership for backsliding governments and the hesitation of the EU, less attention has been paid to the tactics of the backsliding governments themselves. Specifically, given the benefits at stake, backsliders cannot take the EU’s hesitation for granted. They are likely to employ tactics designed to preserve the EU’s reluctance to use its tools. Understanding the tactics of backsliders is closely related to appreciating their limits, and thus ways out of the EU’s autocratic predicament. Clearly, such limits must exist. In December 2022, in a (partial) reversal of previous behavior, the Council for the first time voted to suspend funding Hungary was to receive from the EU budget, and the Commission refused to pay additional cohesion and Covid recovery funds. Did Hungarian tactics reach a limit? To answer a question such as this, we need to understand what backsliders’ tactics, and the conditions for their success and failure, are.

Some recent work illustrates key features of the tactics of backsliding governments. For example, Kelemen (2020, pp. 487–489) and Holesch and Kyriazi (2022) show that backsliding governments garner support by voting with conservative parties in the European Parliament and form coalitions with each other to build voting power in the Council. At the level of discourse, Emmons and Pavone (2021) show that backsliders selectively reference the EU’s own norms, claiming that measures against backsliding would put the EU’s own values at risk, jeopardize intergovernmental cooperation, and violate the EU’s own laws. They do so while obfuscating or downplaying domestic institutional changes, thus making it easier for other governments, who value the Council’s informal norms and are cautious with respect to EU interference in domestic affairs, to opt for inaction. This tactical selectivity, combined with democratic governments’ reluctance to rethink their practices, might also help understand the EU’s interaction with backsliding governments in intergovernmental decision-making.

3 Backsliding governments and the Council’s practice of accommodation

I argue that the core tactic of backsliding governments lies in re-purposing the “consensus norm” or practice of accommodation that characterizes decision-making in the European Council and Council of the EU for the protection of their backsliding projects (Lewis, 1998).¹ Backsliding governments play by this norm just as other member states, but employ it specifically to avert threats to their backsliding projects. Yet, advantageous as this tactic is, it requires that backsliding governments comply with the nuanced conditions sustaining the Council’s practices. This is a limit that backsliders control only partly and that can lead to them losing in key decisions.

There is agreement that consensual practices characterize Council decision-making (Heisenberg, 2005; Lewis, 1998, 2005; Novak, 2013; Smeets, 2015; Thomson, 2011). The Council rarely votes against a government. Only 2.5% of the 29,000 votes cast in the Council over the last decade were ‘No’ votes (Pircher & Farjam, 2021). The overwhelming majority of instances of opposition were isolated (two or fewer opponents) (Novak et al., 2021, p. 487). For backsliding governments, this constitutes an opportunity. They might be able to wield a de facto veto to safeguard their backsliding projects.

Yet, the literature also points to significant limits of this de facto veto. The practice of accommodation in the Council requires that backsliders give and take. Under a norm of diffuse reciprocity, everyone is expected to make concessions (Lewis, 2005). Indeed, the average member state position, not the status quo country, best predicts bargaining outcomes (Thomson, 2011). Moreover, while outvoting others is to be avoided, so is saying “No” frequently (Smeets, 2015). Governments must conserve political capital, use objections sparsely, offer a path forward, and give justifications perceived as valid (Smeets, 2015; Tallberg, 2008, p. 695). In fact, governments often remain silent in a decision (Novak, 2013) or agree to being outvoted to signal having put up a fight to domestic audiences (Hagemann et al., 2019), rather than expend political capital on blocking progress (Smeets, 2015). The consensus norm, in other words, entails demanding conditions and refers to concessions of the minority as much as willingness to avoid a vote by the majority.

It is important to clarify that the Council’s practice of accommodation does not imply the absence of “hard bargaining” tactics by backsliding and other governments. Indeed, examples of tactics such as issue-linkage and veto threats are easy to find (Aksoy, 2012; Kirpsza, 2023). Early work tended to juxtapose these tactics and the consensus norm (Lewis, 1998), but according to recent literature, the normative constraint is not about the choice of tactics as such but, as in the case of the de facto veto, about how they are employed (Novak et al., 2021; Smeets, 2015; Tallberg, 2008). The countries that, in select cases, seek to benefit from a de facto veto under the practice of accommodation also use other bargaining tactics, but lose political capital if doing so is deemed unjustified or excessive. They are expected to make concessions beyond what their bargaining power might warrant. If the conditions for the operation of the consensus norm are met, countries might expect to benefit from their de facto veto in parallel to other bargaining tactics, even if these prove fragile.

For backsliding governments, the opportunities (the de facto veto) and demanding conditions mean that they must strike a delicate balance. Mimicking the behavior expected of all governments, they must be selective in their opposition, cultivate political capital and maintain a veneer of constructive conduct to benefit from the consensus norm in the decisions that matter to them the most. In practice, this means first and foremost that they can oppose decisions only selectively and must also frequently be seen to support the compromises that form around the average member state position. Second, this implies that backsliding governments might lose the protection of the consensus norm if they fail, due to tactical mistakes or factors beyond their control, to observe the conditions underlying the practice of accommodation. The first point refers to the position-taking and opposition behavior of backsliders, the second to bargaining processes and outcomes. I discuss both in turn.

4 Positions of backsliding governments

A crucial condition for backsliders to benefit from the Council’s practice of accommodation is that they can only oppose select EU decisions. This is true for all member governments, as evident in the rare “No” votes noted earlier. What distinguishes backsliding governments and reveals that they seek to use the consensus norm to protect their backsliding projects is, I suggest, that they adopt opposing positions on a distinct set of EU decisions, while supporting most other decisions. I refer to this set as the EU’s backsliding-inhibiting competences.

Backsliding-inhibiting competences are areas in which EU decisions could constrain backsliding projects. For many areas of EU activity, this is not likely. Indeed, as noted earlier, the EU’s market, regulatory, and subsidy regimes benefit backsliding governments at least as much as other member states (Kelemen, 2020). In these areas, backsliders have an opportunity to display supportive and constructive behavior and gain the political capital needed to sustain opposition in other domains.

To identify more problematic areas, I assume that backsliding, as autocratization generally, relies on repression, co-optation, and legitimation strategies (Gerschewski, 2013). The EU could constrain these strategies in various ways. Generally, it is highly active in “core state powers” such as justice and home affairs (Genschel & Jachtenfuchs, 2014; Herschinger et al., 2011; Lavenex & Wagner, 2007). It seeks to monitor spending in the member states, especially in light of concerns about corruption (e.g., Vachudova, 2009). Controversial border and migration policies also fall within the EU’s remit (Lavenex, 2006; Schimmelfennig, 2021). In these domains tensions between EU decisions and backsliding projects are likely.

The inclusion of not only repression and co-optation but also legitimation strategies, and EU policies that might counter them, warrants further justification. Even in consolidated autocracies, legitimation, if successfully pursued, complements repression and cooptation by eliciting voluntary public acquiescence. It is as crucial for the stability of autocracies as the other two dimensions (Gerschewski, 2013). For governments between autocracy and democracy, including backsliders, whose power remains vulnerable and for whom electoral loss and anti-government mobilization remain real (if increasingly unlikely) risks, legitimacy is even more important. Indeed, this is particularly so since their authority over key means of repression remains incomplete (Bermeo, 2016; Levitsky & Way, 2002). In the EU context, several studies stress the opportunistic and instrumental use of illiberal policies and discourses by backsliding governments for the purpose of obtaining public support for their rule (Scheppele, 2019; Vachudova, 2020). While there is some disagreement on the extent to which legitimating policies facilitate autocratization or vice versa, there is little dispute that the EU’s backsliding governments rely strongly on their purported “social purpose” for domestic mobilization (Bohle et al., 2023, p. 8; Scheppele, 2019; Vachudova, 2020).

Table 1 lists EU competences that could be considered backsliding-inhibiting. Consider first the backsliding strategy of repression. Repression relies on the government’s discretionary control of the bureaucracy, police, and judicial system to undermine political opponents (Bermeo, 2016; Gerschewski, 2013). EU competences over these institutional arenas or related citizen rights are thus sensitive from the perspective of backsliding governments, as evident in controversial attempts of the EU to prevent the restructuring of the Hungarian and Polish judiciary (e.g., Kelemen & Blauberger, 2017).

Table 1 Backsliding-inhibiting competences

As co-optation entails financing loyalists (Kelemen, 2020), the EU could interfere by enforcing strict rules regarding the distribution of funds. The EU could tie funding to the rule of law, monitor spending, and challenge questionable public procurement (Blauberger & Kelemen, 2017; Blauberger & van Hüllen, 2021; Priebus, 2022a). The EU has also created, and Hungary and Poland have refused to join, a European Public Prosecutor’s Office, which strengthens the EU’s ability to investigate crimes against the budget and could compromise the Hungarian and Polish governments’ co-optation strategies (Kelemen, 2020, p. 490).

The EU could also constrain backsliders’ legitimation strategies. These strategies, which have a strongly conservative, nationalist, and populist orientation (Bogaards, 2018; Bohle et al., 2023; Buzogány, 2017; Scheppele, 2019; Vachudova, 2020), have included policies such as restricting material related to gender and sexual orientations in schools,² prohibiting recognition and change of gender identity, and restricting adoption.³ The Hungarian and Polish governments have also adopted restrictive policies on refugees.⁴ EU policies that foster (the redistribution of) immigration, loosen border controls, protect minority rights, strengthen equality more broadly, or prevent the politicization of education are thus problematic for backsliding governments.

Note that any specification of backsliding-inhibiting competences is imperfect as the EU interprets its authority creatively. For example, the EU has referenced anti-discrimination in the workplace and various other competences (e.g., audiovisual services, e-commerce, the freedom to provide services, competition policy) to target backsliding. However, these less conspicuous areas are unlikely to trigger opposition systematically. The implications for ongoing backsliding are too uncertain at the time of the decision. Backsliding governments are unlikely to invest political capital into opposition, which they could instead expend on more critical decisions.

For backsliding governments, opposition to decisions in these areas is necessary to protect their domestic goals. Yet, the other part of the argument – backsliders’ support for decisions in other areas – is equally important. It helps backsliding governments to build the political capital needed to benefit from the Council’s practice of accommodation in select cases. Their selective profile of opposition also distinguishes backsliding from Eurosceptic governments. The latter might oppose decisions in backsliding-inhibiting competences, which often relate to core state powers, but should not oppose primarily or only these decisions due to a nuanced tactic to balance political capital and the protection of domestic backsliding.

5 Backsliding governments and intergovernmental bargaining

My second expectation focuses not on the input side (the positions and opposition behavior of backsliding governments), but on the process and outcomes of intergovernmental bargaining. Being selective in their opposition is only one condition that backsliders must observe to rely successfully on the Council’s practice of accommodation. I highlight further conditions that backsliding governments must observe. If these conditions are not met, backsliders are left with conventional tactics such as vetoes, their autocratic coalition, or issue-linkage. These tactics leave them vulnerable in an organization that is overwhelmingly democratic and experienced in circumventing bargaining blockades.

A first condition is that backsliding governments must uphold the appearance of constructive and legitimate conduct even if they invoke their de facto veto. Recall that this veto is not a formal tool but embedded in the give-and-take practices of the Council (Lewis, 1998; Smeets, 2015). Governments must explain why their opposition is justified under, rather than a violation of, the practice of accommodation. Tallberg (2008), for example, highlights that high level government officials, including heads of state and government, put strong emphasis on the need to justify opposition in ways perceived as valid. Similar findings have been reported based on interviews at lower levels of the European Council and Council hierarchy including the level of permanent representatives and working groups (Heisenberg, 2005; Lewis, 1998; Novak et al., 2021; Smeets, 2015).

Two indicators of this idea lend themselves for cross-case comparisons. The first indicator is whether a government’s objections are shared by other governments. The demand to avoid a vote is seen as more legitimate and less costly in political capital if shared by several governments (Smeets, 2015). Second, the de facto veto is likely to be questioned if it prevents an urgent decision. This could be the case due to formal deadlines that can only be missed at significant costs, as in the budget process (Heisenberg, 2005, pp. 168–169). Pressure to resolve a crisis, which could otherwise impose costs on many member states also creates urgency. Both factors reinforce each other. Under time pressure, opposition from other governments will be rare due to being difficult to reconcile with the practice of accommodation, thus leaving backsliding governments isolated.

In addition to the need to appear constructive and legitimate, a second condition is that the Council remains insulated from political pressure. As Lewis (2005, p. 947) observes, a crucial feature of Council decision-making “is a high degree of insulation from the normal currents of domestic constituent pressure” and the commensurate “capacity to reshape domestic constraints” by obfuscating the concessions and compromises that the practice of accommodation requires. Obviously, insulation is never perfect, but the key point is the insulation of negotiators from political accountability for making concessions to backsliding governments. Kelemen (2020) considers the lack of accountability for the Council’s accommodation of backsliding a key ingredient of the EU’s predicament. Unsurprisingly, the practice of accommodation can falter if the risk of accountability rises with, for example, pressure from the European Parliament (Häge, 2011) or enhanced transparency (e.g., Cross, 2013a; Stasavage, 2004). However, the most important indicator that negotiators have to fear political accountability is if national parliaments, to whom they are directly bound via chains of delegation and accountability (Bergman, 2000; Winzen, 2022), oppose concessions to backsliding governments.

These conditions suggest when backsliding governments can hope to benefit from the Council’s practice of accommodation, but also help clarify the limits of their tactics. If backsliding governments find themselves isolated in urgent decisions in which other member states face national parliamentary pressure, they cannot rely on accommodation in the Council. In these cases, as noted earlier, they might still employ conventional tactics, even at the risk of losing political capital, but these tactics leave them in a weak position. For one, the Lisbon Treaty abolished almost all unanimity voting. Moreover, the main backsliding governments of Poland and Hungary cannot form a blocking coalition even with their closest allies from the Visegrád group (Czech Republic, Hungary, Poland, and Slovakia). Issue-linkage requires a contemporaneous negotiation with a veto opportunity, and it can be broken if the other member states can exclude backsliders from a decision through legislative differentiation (Duttle et al., 2017; Kroll & Leuffen, 2015). Issue-linkage can also be used by democratic member states to put pressure on backsliders. Given these constraints, I expect that backsliding governments are likely to lose in decisions in backsliding-inhibiting competences if the conditions for the operation of the consensus norm are not met.

6 First analysis: Backsliding governments’ positions

This study presents two analyses: one of backsliders’ positions and another of bargaining processes and outcomes. I begin with the data and results of the first analysis.

I rely on a narrow and a broad measure of backsliding. There is a controversial debate about which countries qualify as instances backsliding. Employing a broad and a narrow measure allows, instead of taking sides in this debate, to inspect results at the strict and lenient end of the range of plausible standards. The narrow measure treats Hungary (since 2010) and Poland (since 2015) as cases. Even sceptics of backsliding regard these as clear cases (Cianetti & Hanley, 2021). These cases have also dominated the debate (e.g., Blauberger & van Hüllen, 2021; Emmons & Pavone, 2021; Kelemen, 2020; Scheppele, 2022). The broad approach applies the standard of the Varieties of Democracy (V-Dem) project (Lührmann & Lindberg, 2019): a decline of 0.1 on the 0–1 polyarchy index in a connected period. Three EU members meet this standard: Hungary (since 2010), Poland (since 2016), and Slovenia (since 2012). As this broad measure should be designed to be inclusive, I add three countries that come close: Croatia (2013–2019, 0.099 decline), Czech Republic (since 2009, 0.091), and Greece (since 2013, 0.086). These countries face issues that could qualify as backsliding (e.g., Bugarič, 2015; Haggard & Kaufman, 2021; Soteropulos, 2018; Vachudova, 2020), but could also be conceptualized differently (Cianetti & Hanley, 2021). In the Czech Republic and Slovenia, backsliding ended with opposition victories in the 2021 and 2022 elections. Countries are coded as backsliding only from the onset and for the period in which backsliding is observed.

Ideally, we could measure precisely when backsliders invoked their de facto veto. Yet, due to the informality of doing so, this is not feasible. Instead, I rely on data on whether backsliders adopted Eurosceptic or pro-EU positions on a decision in the Council. A Eurosceptic position demands weaker, vaguer, or more (nationally) discretionary rules, leaves more authority to national governments or to the Council (rather than the Commission), limits the EU’s budget and spending, reduces the authority of EU institutions, and reduces the scope and actors bound by EU rules – in short, a position that reflects the kind of opposition behavior, backsliding governments would be expected to display in backsliding-inhibiting competences. Note that such a measure of positions does not tell us whether a government subsequently makes voluntary concessions to join a compromise or insists on its opposition, but it provides an indication of the issues a government tends to oppose, and thus of the extent and patterns of selective opposition.

I measure Eurosceptic and pro-EU government positions using the Decision-making in the European Union (DEU) dataset (Arregui & Perarnaud, 2022; Thomson et al., 2006, 2012).⁵ Based on 494 interviews with EU policymakers, the DEU dataset records initial government positions on 363 controversial issues arising in 141 legislative proposals introduced between 1996 and 2018. For each issue, each government’s stance is recorded on a 0-100 scale, with the most extreme positions providing the ends of the scale. I manually identified the most Eurosceptic of the observed positions for each issue in the DEU data.⁶ In principle, governments’ pro-EU or Eurosceptic positions can then be measured as the distance from the Eurosceptic position (higher values thus mean pro-EU). My coding correlates highly (r = 0.73) with Wratil’s (2018) who had coded a similar variable for a subset of the DEU data. For all discrepancies, I double-checked the validity of my coding.

In practice, government positions in the DEU data—and thus also Eurosceptic and pro-EU positions—follow a strongly bimodal distribution. Two-thirds of the observations are on either extreme (pro-EU or Eurosceptic). 80% are either on the 0–30 end or the 70–100 end of the 0-100 range. Given this distribution, I employ a binary measure that splits the data at 50 and indicates whether governments adopt a pro-EU (1) or Eurosceptic (0) position.⁷

I argued that backsliding governments oppose decisions in backsliding-inhibiting competences. To identify these competences, I classified DEU issues following the structure of Table 1. Related to repression strategies, the DEU data contains decisions related to citizens’ rights, to the organization of the public administration, and to courts and the justice system. In relation to co-optation, the data encompasses decisions on the spending of EU funds, especially the distribution of funds, transparency and monitoring of spending, and sanctions related to spending. Finally, as to legitimation, there are crucial recent decisions on border control and migration. Appendix 1 shows all legislative proposals in which at least one issue was coded as backsliding-inhibiting and the replication materials include a complete overview of my coding of the DEU data.

A downside is that DEU is limited to legislation that reached some degree of controversy among governments but is not geared towards selecting particularly controversial cases (Arregui & Perarnaud, 2022). There has been extensive debate about the (un-) representativeness of the DEU data due to the selection of somewhat controversial cases (Arregui & Perarnaud, 2022; Thomson, 2011). For my purposes, the exclusion of much technical legislation, to which neither backsliders nor other countries tend to object, seems reasonable. A more serious concern is that the DEU data, reflecting the moderate to low controversy around most EU activity, does not ensure the inclusion of the most controversial decisions. Indeed, as the list in Appendix 1 confirms, it omits some of the EU’s major advances in backsliding-inhibiting competences, such as the rule of law conditionality regulation. In other words, it includes issues that backsliders should oppose in principle, but that they might not oppose as strongly as some cases that are not covered, and that they might eventually concede to preserve political capital. This is important to keep in mind as it should work against finding evidence in line with the argument.

For example, I coded the first issue in Proposal for a Directive on fight against fraud to the Union´s financial interests by means of criminal law (2012/0193/COD) as a backsliding-inhibiting competences. The issue, according to the DEU codebook (see issue 338), was: “What are the preferences of stakeholders regarding the inclusion of the VAT fraud into the scope of competencies of the European Public Prosecutor Office?” The positions were:

• 0: No inclusion of the VAT fraud into the scope of competences of the EPPO (current status quo)

• 70: Inclusion of VAT fraud with two conditions: cross-border fraud, and valued more than 10 million euros.

• 100: Inclusion of VAT fraud without conditions

This issue is backsliding-inhibiting because fraud-prevention capacities of the EU could threaten backsliders’ co-optation strategies. Yet, it is less threatening for backsliding governments such as Hungary than the fundamental choice, not included in the DEU, of whether the EU should have a public prosecutor in the first place.

A final concern is that the DEU data contains few cases for the period of democratic backsliding. The dataset was collected in three waves covering legislation proposed in 1996–2000, 2003–2008, and 2012–2018. For the present analysis, the first wave provides 145 legislative issues, the second 134, and the third, which was smaller, 32. However, this still leaves many issue-government observations for the analysis. Moreover, most, albeit not all, decisions in backsliding-inhibiting competences in the period of backsliding (especially after the onset of backsliding in Poland in 2015) fall into the area of border policies and migration and are thus related to backsliders’ legitimation strategies. I have argued above that these strategies should be included in the analysis, but it is important to note that the analysis relies strongly on the corresponding cases. The inclusion of the first wave is debatable but, in an analysis over time, can help understand government behavior before the onset of backsliding. Empirically, including the first wave makes no difference (Appendix 2).

6.1 Results

I analyze the data with multilevel logistic regression models. The dependent variable is the measure of Eurosceptic (0) or pro-EU (1) positions and the independent variables include the measure of whether a country is currently backsliding (Backsliding), whether the decision under negotiation falls under backsliding-inhibiting competences (Inhibiting competence), the interaction between these two variables (Backsliding*inhibiting), and additional variables. Specifically, measures of the EU support of government parties and whether a decision falls into the sensitive justice and home affairs domain should help ascertain that the results are due to backsliding and backsliding-inhibiting competences, rather than the right-wing and Eurosceptic orientation of the governments and the sensitivity of policy areas in question.

I estimate several models with different approaches to the nested structure of the data and the alternative conceptions of backsliding. In Fig. 1, Models 1 and 2 (M1, M2) include random intercepts for the country and the proposal under negotiations. M3 includes only observations from countries that experience backsliding during the analysis period (i.e., they have far fewer observations) and fixed effects controlling for time-invariant country effects. The models in panels (a) and (b) are identical except that panel b) adopts the narrow conception in which backsliding only occurs in Hungary and Poland. All models were estimated in a Bayesian framework with loose priors.

Fig. 1

Backsliding and opposing positions in EU decision-making. Note: Observations for the logistic regression models in both panels (cases/countries/proposals): M1: 5547/28/127. M2: 4827/28/118. M3a: 780/6/114. M3b: 264/2/64. The bars are 90% (dark blue) and 95% (light blue) Highest Posterior Density Intervals (HPDI or credible intervals). For the detailed regression results, see Appendix 2

The results show that backsliding governments and other governments do not differ in their pro- or anti-EU positions outside of backsliding-inhibiting competences (see results for Backsliding). This is consistent with the claim that backsliders behave similar to other member states in most decisions. In backsliding-inhibiting competences (see results for Inhibiting competence), all governments adopt cautious positions, which is what the literature suggests (e.g., Genschel & Jachtenfuchs, 2014). Intriguingly, however, in model 3, which includes only countries that backslide at some point, there is no negative relationship. Keeping the interaction in mind (see below), this means that these countries did not oppose EU decisions in backsliding-inhibiting competences before the onset of backsliding, when these decisions did not yet threaten autocratic ambitions. The results so far remain nearly unchanged across the broad and narrow definitions of backsliding.

I turn to the other side of the interaction effect, that is, the question of whether backsliding governments adopt more Eurosceptic positions in backsliding-inhibiting competences (see results for Backsliding*inhibiting). In the random effects models (M1 and M2 of panel a), the evidence is consistent with this claim in that the 90% credible interval excludes any positive parameter estimates and the median estimate is clearly negative. In the fixed effects model and in the models based on the narrow conception of backsliding, in which less information to estimate the interaction is available, the credibility intervals are wider, even though the median point estimate is similar. Given the limitations of the data, some uncertainty thus remains, but the overall picture suggests at least tentatively that backsliding governments indeed adopt a distinct profile of selective opposition to EU decisions focused on backsliding-inhibiting competences.

Figure 2 illustrates that backsliding in combination with backsliding-inhibiting competences reduces the median probability of a government adopting a pro-EU position by 10% points. There is no difference in the probability of pro-EU positions outside of backsliding-inhibiting competences. In the Council, in which few governments consistently find themselves on the opposing side (Cross, 2013b; Kleine et al., 2022; Thomson et al., 2006, 2012), this is an important relationship – certainly if we recall that the DEU data excludes key decisions to which backsliders might be most opposed.

Fig. 2

The probability that backsliding governments adopt pro-EU positions. Note: The predicted probabilities were generated based on 16,000 draws from the posterior distribution of Model 1

I present further discussion and results in Appendix 2. These suggest that the results are not due to the Euroscepticism of backsliding governments, a time trend, or a correlation between backsliding-inhibiting competences and the policy area of justice and home affairs. Appendix 2 also includes a placebo test in which I assign backsliding to other groups of countries, which raises confidence in the findings, and results that exclude observations from the first wave of DEU data. Finally, I vary the broad operationalization of backsliding by coding fewer or additional countries as backsliding – specifically as Greece, Bulgaria, Romania, and Malta. This again leaves the results largely unchanged.

Summing up, I argued that backsliders oppose EU decisions selectively in backsliding-inhibiting competences while behaving as other countries in other areas, to retain the protection of the Council’s practice of accommodation. This first analysis offers three pieces of evidence in line with this argument. First, backsliding governments do not generally adopt opposing positions in EU decision-making. They do not differ from other countries in most cases. Second, we saw in the fixed effects models that the (later) backsliders did not even oppose EU decisions in backsliding-inhibiting competences before the onset of backsliding, even though these decisions tend to be sensitive for all governments. Third, the evidence suggests that backsliders adopt Eurosceptic positions if backsliding-inhibiting competences are at stake. It is necessary to acknowledge some uncertainty around this finding in the analyses with fewer observations and narrower definitions of backsliding, but the overall picture and median parameter estimates are consistent with the argument. As noted, the DEU data, while the best-available data for this study, does not include some of the most problematic decisions for backsliders and, thus, poses a rather strict test for the argument.

7 Second analysis: Bargaining processes and outcomes

The second analysis examines the claim that backsliding governments’ ability to benefit from the Council’s practice of accommodation hinges on conditions: not being isolated, the EU not facing time pressure, and the insulation of Council members from political accountability, especially from national parliaments. For this analysis, I examine EU decisions that backsliders clearly opposed and sought, with varying success, to prevent.

Table 2 Cases of strong Hungarian and Polish opposition to EU decisions

I focus on cases that the main backsliders, Hungary and Poland, opposed strongly, with opposition considered strong if it featured in specialized news. I selected cases if the headline of an article on the country pages for Hungary and Poland of the Euractiv.com website indicated opposition to an EU decision.⁸ This website publishes specialized news on EU decision-making and is likely to cover most significant controversy in the Council. For these cases, I collected further information cited below. Table 2 shows all cases, the bargaining constellation, the result, and whether Hungary and Poland obtained their demands.

I excluded Council conclusions in external relations, which are not legally binding. The analysis includes the formal decision on sanctions against Russia, however. Moreover, I combined certain cases. First, in energy and environment policy, decisions related to the implementation of the EU’s climate agenda are often highly intertwined. The Table includes a key conflict (case 4) in the reform of the EU’s Energy Union that pitted backsliding governments and some other member states against the majority. This case can be seen to illustrate environmental and energy policy conflicts that have appeared in several parts of the EU’s climate agenda. Second, I discuss an emergency decision on refugee reallocation in 2015 in the context of the reform of the Dublin regulation (case 3) as these decisions are substantively linked. Third, the Table does not list decisions – specifically, the 2021–2027 EU budget, the EU’s legislative implementation of a global corporate tax deal and Hungary’s opposition to financial aid for Ukraine – in which backsliders’ opposition was widely seen to serve predominantly as leverage for other decisions, but I discuss these issue-linkages in the context of the relevant decision (cases 8 and 10).

I briefly note key cross-case patterns. First, strong opposition is rare with only 10 cases in over a decade of backsliding. This is less than 1% of the EU’s 1050 legislative decisions from 2010 to 2019 (3 years short of the period here) (Pircher & Farjam, 2021, p. 478). Even if the selection based on Euractiv.com missed half of the cases or more, the conclusion that opposition is rare would not change. It is in line with the claim that backsliders oppose EU decisions highly selectively.

Second, also in line with selective opposition, Hungarian and Polish opposition focuses on backsliding-inhibiting competences. The backsliders opposed the European Public Prosecutor, reforms of the border and migration regime, and decisions to penalize rule of law violations. The reappointment of Donald Tusk as European Council president (case 1) empowered a leading opposition politician. The oil embargo against Russia (case 9) fits given Hungary’s alleged role as Russia’s “trojan horse” and gains from cheap oil (Orenstein & Kelemen, 2017). The only exception is the backsliders’ opposition to energy union and certain facets of the EU’s climate agenda more broadly (case 4), which is plausibly explained by the backsliders’ reliance on regulated energy prices and coal power.

Third, the backsliding governments frequently negotiate successfully. Even when they lacked a veto or blocking coalition, they often reached their goals. And in cases that the backsliding governments lost, such as the EU’s adoption of the rule of law regulation (case 8) and the suspension of funding for Hungary (case 10), they secured at least some concessions. The only full loss is the appointment of Donald Tusk, in which the backsliders were isolated and the EU under time pressure, created by the end of the previous term, so that a vote was taken.

Subsequently, I present case studies. Regarding the question of how the backsliders keep the EU hospitable for their domestic goals, two groups of cases promise most insight. First, cases in which the backsliders won, despite lacking a veto or blocking minority, rule out voting power as an explanation. They promise insights as to other sources of success, including possibly the Council’s practice of accommodation. I examine the migration reform cases (3 & 7). I do not include the suspension of voting rights (6), which has been covered in detail in the literature (Closa, 2021, p. 7; Priebus, 2022b). Second, the cases in which the backsliders lost (partly) might reveal the limits of their tactics, including potentially their reliance on accommodation, and whether isolation, time pressure, and political accountability mattered. I examine the rule of law regulation (8) and suspension of funding (10) cases. Finally, Appendix 3 discusses case 4. As a rare case of strong Hungarian and Polish opposition to a decision outside of the area of backsliding-inhibiting competences, this case helps explore whether these governments make concessions to build political capital, as expected under the consensus norm.

7.1 The Dublin regulation and the Pact on Migration and Asylum (cases 3 and 7)

The Dublin regulation determines which countries are responsible for asylum-seekers arriving in the EU. The status quo is that the state of first arrival bears this responsibility. Reform debates were reinforced by increased arrivals in the context of the Syrian civil war in summer 2015. The Southern member states demanded that responsibilities and costs be allocated more evenly across member states. Hungary, Poland as well as Austria, the Czech Republic, and Slovakia strongly opposed reallocation, the key element of the reform.⁹ After three years of negotiations, without shift in the positions of the opponents, the Commission gave up on the idea of reallocation and withdrew the proposal.¹⁰ The Commission revived the idea in September 2020 in the Pact on Migration and Asylum, but, in a replay of the previous negotiations, the idea of mandatory reallocation stalled and became a voluntary declaration.¹¹

Formal voting rules do not explain this outcome well. The decision could have been adopted by majority voting. The opponents did not have a blocking minority. This point is underlined by the fact that the EU had adopted a one-time, emergency reallocation mechanism in September 2015, under acute pressure from Greece and Italy, to react to the arrivals at the external border, outvoting Hungary and other countries.¹² Notably, this decision was taken before the Polish government changed and backsliding started – and the Polish government voted for reallocation. Even with Poland added to the list of opponents, a blocking minority would not have been reached, however.

Instead, ample evidence indicates that the backsliders benefitted from the consensus orientation of the Council. In a detailed reconstruction, Smeets and Beach (2022, p. 1421) highlight that the European Council determined early on that the Dublin reform was to be decided by consensus, and affirmed this commitment throughout the negotiations, rendering agreement on the controversial idea of mandatory reallocation impossible.¹³ The backsliding governments benefitted from having several allies – not enough to form a blocking minority but too many to let the proponents consider voting by majority.

One might object that the September 2015 majority decision is evidence against the role of the consensus norm. However, it was taken under an acute sense of urgency which can lead to deviations from consensus (Heisenberg, 2005, pp. 168–169). This pressure faded as arrivals declined. Moreover, Smeets and Beach (2022, p. 1421) note that the decision was unusual in that it was taken at the ministerial level, at the initiative of the Council presidency and the Commission, against the demand by the heads of government to achieve consensus. It was thus an exceptional process that illustrates the existence of a majority coalition for the reform, but that does not invalidate the point that governments typically seek consensus.

7.2 The rule of law regulation (case 8)

On 16 December 2020, the EU adopted regulation 2020/2092 on a general conditionality regime to protect the EU budget. The conditionality regulation allows the EU to suspend, by qualified majority, payments from the budget to member states that breach the rule of law. The backsliding governments of Hungary and Poland vehemently opposed the adoption of this regulation but failed to block it. However, they secured a mixed result by watering down a broad proposal to a narrow regulation in which the EU must demonstrate precisely that rule of law violations “seriously risk affecting the sound financial management of the Union budget or the protection of the financial interests of the Union in a sufficiently direct way” (Article 4). Showing this direct link raises the standard of evidence and creates uncertainty as to violations that affect EU funds indirectly. Why did the backsliding governments fail to prevent the conditionality regulation and how did they obtain concessions?

As to the concessions, issue-linkage played a role. The backsliding governments did not have a veto or blocking minority. In fact, both countries were outvoted by a qualified majority at the ambassador level on 16 November 2020, and thus came under intense pressure to prevent legal adoption at the ministerial level.¹⁴ They benefitted from the fact that the EU was simultaneously negotiating a 7-year multiannual financial framework (MFF) and its Covid recovery fund. Both required unanimous agreement. The backsliders announced that they would veto both decisions and that each of them would only lift the veto if the other also did so.¹⁵ Attempting to justify their veto, they claimed to be counteracting a vague and divisive regulation and empowerment of the EU to meddle in core state competences.¹⁶ Via issue-linkage, the backsliding governments thus turned a majority decision in which they had been outnumbered and outvoted into a unanimous decision, in which they had a strong bargaining position.

One might now have expected accommodation efforts leading, together with the issue-linkage, to a favorable outcome for the backsliding governments. Yet, the issue-linkage quickly faltered and there is little evidence of accommodation and strong signs that time pressure explains this. Indeed, the EU was under intense time pressure. It had to adopt the decision by the end of the year to face the pandemic with updated spending ceilings and a recovery fund. In line with the expectation that its willingness to accommodate declines under time pressure, the EU moved swiftly to weaken the backsliders’ bargaining stance. The Commission announced that it would propose alternative solutions for the Covid recovery fund within a month relying on the enhanced cooperation procedure or a system of national guarantees, leaving Hungary and Poland without a veto and excluded from the funds.¹⁷ This solution would not have solved the problem of the budget ceilings, however, exposing other member states to billions of Euros of cuts in cohesion funding. The narrow regulation that was ultimately adopted was a compromise giving the backsliders access to funds in exchange for unblocking the negotiations and accepting the narrowed down conditionality regulation.

7.3 The decision to suspend funding for Hungary (case 10)

On 15 December 2022, the Council adopted decision 2022/2506 to suspend ca. 6.3bn Euro of funding from Hungary (15% of its allocation from the 2021-27 multiannual budget) under the new conditionality regulation. This was the first Council decision to impose significant costs on a backsliding government.¹⁸ Yet, the Commission had proposed a higher amount and critics had demanded to suspend all funding (Scheppele et al., 2022), so that the outcome could have been worse for the backsliding governments.

As before, this case also features issue-linkage and illustrates its limits. When the Commission proposal to suspend funding came into view, Hungary blocked an agreement on a global corporate minimum tax that the EU sought to implement.¹⁹ It also announced a veto against aid for Ukraine.²⁰ However, both issue-linkages proved fragile. With respect to the tax deal the EU was not under immediate pressure and developed the idea to adopt the deal under its enhanced cooperation procedure, excluding Hungary.²¹ Regarding Ukraine aid, the EU not only threatened but in fact adopted the legislation by specifying two financing options (via the EU budget or member state budgets) rather than only the EU budget. This circumvented the unanimity requirement and left the financing mechanism, but not the aid, for Hungary to choose, with the risk of alienating its autocratic partner and Ukraine ally, Poland.²²

As in the previous case, Hungary could ultimately not benefit from the Council’s practice of accommodation. However, there is clear evidence that this was not the result of unwillingness of the other governments. The Council tried to accommodate Hungary even though, having broken the issue linkage, they could have outvoted it. First, the Council, including the French and German governments, requested a new assessment of the rule of law in Hungary from the Commission.²³ Considering that it gave the Commission only days, although the original assessment had stressed that reforms would have to be assessed in practice and over an extended period (European Commission, 2022, p. 33/para. 122), the Council’s request cannot plausibly be interpreted as genuine interest in an update. It rather signaled unwillingness to pass sanctions and the ambition to pressure the Commission to backtrack. When this failed, member states agreed on sanctions, for reasons discussed below, but only after reducing the penalty by 15% (from 7.5bn to 6.3bn Euro) and highlighting the “significance of remedial measures that have been satisfactorily implemented by Hungary.”²⁴ The Council could have voted and the basis of the Council decision was the updated Commission assessment, which Commissioner Hahn summarized as “Hungary has [not] done enough to enable the Commission to adapt or even retract our proposal.”²⁵ The reduced penalty was thus a concession not warranted by practical policy change or the voting constellation.

The governments in the Council thus intended to accommodate Hungary, but still suspended significant funding. Several pieces of evidence suggest that governments and the European Commission would have faced political accountability if they had decided against the suspension. The European Parliament adopted, with cross-party support, resolutions labelling Hungary an electoral autocracy, demanding action from the Commission and Council, and rejecting a positive assessment of Hungarian reforms (European Parliament, 2022a, b). Earlier in the year, several parliamentarians had also raised, but not acted on, the possibility to censure the European Commission for failing to use the EU’s tools against democratic backsliding.²⁶ In addition, the Dutch, French, and German parliaments demanded that their governments support the suspension of funding (Commission des Affaires Européennes, 2022; Deutscher Bundestag, 2022; Ministerie van Financiën, 2022). In particular, the German government was legally obliged to follow the parliamentary vote and justify deviations in writing. The EU’s parliaments thus raised accountability to a much higher level than before.

Against this background, the seemingly contradictory behavior of the Council and the French and German governments (i.e., first trying to avoid a vote by having the Commission revise the assessment but then voting for suspended funding) makes sense. The first action sought to accommodate Hungary and avoid political accountability. When reconciling both goals became impossible following the Commission’s refusal to rewrite its assessment, key governments prioritized avoiding a conflict with their national parliaments over accommodating Hungary, as the literature on the limits of the Council’s practice of accommodation suggests.

7.4 Lessons from the analysis of bargaining processes and outcomes

Overall, the second analysis shows that backsliders rarely oppose EU decisions but often obtain their demands if they do. The case studies underline the view that backsliding governments benefit from the Council’s practice of accommodation (Closa, 2021, p. 20; Emmons & Pavone, 2021; Kelemen, 2022; Oliver & Stefanelli, 2016). This works particularly well if other member states share their opposition, as in the migration reform cases. Yet, even under high pressure, as in the case of the suspension of funding to Hungary, backsliders obtained concessions beyond what policy substance and bargaining power warranted. Yet, the case studies also reveal limits of the backsliders’ tactic. Confronted with isolated backsliders, urgency, and pressure from the EU’s parliaments, democratic governments did not accommodate backsliding beyond limited concessions. Conventional bargaining tactics alone left backsliders in a precarious position. Even seemingly robust issue linkages faltered, sometimes in a matter of days.

8 Conclusion

I began by asking how backsliding governments preserve the hospitability of the EU for their backsliding projects (Kelemen, 2020). Existing literature highlights the reluctance of the EU to use its tools to reign in backsliding (e.g., Closa, 2021; Emmons & Pavone, 2021; Kelemen, 2022; Oliver & Stefanelli, 2016), but what tactics do backsliders employ to maintain this reluctance? And what are the limits of these tactics? I have argued that, as their core tactic, backsliding governments seek to benefit from the Council’s practice of accommodation or consensus norm and tailor their behavior to the demands of this norm. I presented evidence consistent with the argument that backsliders selectively oppose EU decisions in backsliding-inhibiting competences but not in other domains. Moreover, a series of case studies showed that backsliders’ ability to rely on accommodation in the Council declines if the necessary conditions are not met. If the Council faces time pressure, political accountability, and isolated backsliders, it outvotes or circumvents backsliders’ opposition.

The analysis has implications for understanding the EU’s autocratic predicament. Recent debates have focused on the EU’s tools to reign in backsliding (e.g., Blauberger & van Hüllen, 2021; Holesch & Kyriazi, 2022; Oliver & Stefanelli, 2016; Sedelmeier, 2017). Complementing recent work on the discursive strategies of backsliding governments (Emmons & Pavone, 2021), my analysis suggests that the EU’s informal norms and practices are at least as important. Backsliding governments have fared well in decision-making when they could rely on the Council’s practice of accommodation and, knowing this, behave accordingly. When backsliders could not rely on accommodation, they lost key negotiations. The ability of the EU to reign in backsliding might thus depend on its willingness to adapt organizational norms and practices to the realities of being a mixed-membership organization. My analysis offers little reason to expect that national governments could drive such change but indicates that the EU’s parliaments could do so by demanding accountability for the accommodation of backsliding governments.

The analysis also holds a key lesson for backsliding governments, namely that they obtain results not by disrupting the organization but playing the norm-abiding member. Importantly, this does not pertain to substantive norms of democracy and the rule of law but to procedural norms – such as the Council’s practice of accommodation – that grease the wheels of decision-making. The EU’s backsliders have understood this, remained selective in their opposition, and managed to secure the benefits of membership for over a decade. Yet, after two mixed outcomes (see above) and under domestic pressure, there are signs that the backsliding governments – Hungary in particular – might change tactics with veto threats against Finish and Swedish NATO accession, sanctions against Russia, and a general intensification of aggressive rhetoric and actions (Bohle et al., 2023). Doing so might cost political capital, galvanize parliamentary attention in other countries, create situations in which the EU faces intense time pressure to pass a decision and, more broadly, violate the conditions under which a government can expect accommodation in the Council. Under these conditions, my analysis indicates, backsliders have so far fared poorly in EU negotiations.

A full analysis of why backsliding governments would change tactics goes beyond my scope. However, part of the answer might be found in the observation that the backsliding governments’ illiberal projects have become increasingly entrenched and, by 2020, shifted towards a phase of radicalization and assertiveness (Bohle et al., 2023, pp. 10–11). The domestic entrenchment of illiberalism and the promotion of a radical ideology to legitimate backsliding might reduce the capacity of backsliding governments for nuanced tactical behavior in the Council.

Finally, do these lessons matter for understanding the role of backsliding and autocratic governments in regional and international organizations beyond the EU? This debate has been framed in terms of norm contestation. It has highlighted that autocracies are selective and specific in their opposition. They oppose some aspects of good governance, certain human rights norms, or particular forms of internet governance, for example (Debre, 2021; Flonk, 2021; Hafner-Burton et al., 2022; Hulvey, 2022; Meyerrose & Nooruddin, 2022). The selective opposition of backsliding governments to EU decisions observed here underlines this link between domestic regime goals and specific patterns of international opposition.

Furthermore, the lesson that governments might work the procedural norms of an organization expertly but use their procedural expertise to avoid and undermine an organization’s substantive norms is relevant for all organizations. We increasingly observe that backsliders and autocracies employ rather than reject international organizations, and their specific procedural opportunities, to protect their regimes and backsliding projects (e.g., Baturo, 2022; Lipps & Jacobs, 2022; Meyerrose & Nooruddin, 2022). The problem is probably greatest in democratic organizations confronted by democratic backsliding. These organizations have procedural norms made for and by the interaction of democracies, which might not be well-suited for conflicts between democratic and autocratic regimes. In contrast, mixed-member organizations should be more familiar with such conflicts. In the United Nations, for example, China’s initiatives to change internet governance have resulted in polarization rather than accommodation between democracies and autocracies (Hulvey, 2022). Nevertheless, even mixed organizations aim to teach autocracies the formal and informal norms of decision-making and have generally seen this as a contribution to socializing these countries into a broadly liberal and democratic international order (Kelley, 2004; Lake et al., 2021). They have to face the possibility that autocracies learn the procedural norms willingly, albeit only to contest substantive norms effectively.