Procedural vs substantive accountability in EMU governance: between payoffs and trade-offs

ABSTRACT

This article introduces a new normative framework for analysing accountability in the European Union’s Economic and Monetary Union (EMU). The framework is anchored in four normative ‘goods’ that accountability is supposed to ensure: openness, non-arbitrariness, effectiveness, and publicness. All of these can be achieved in a procedural or substantive way, depending on whether actors are held accountable for the quality of their decision-making processes or for the actual merit of their decisions. Transposed to EMU, this conceptualisation shows both the payoffs and trade-offs of prioritising procedural accountability. Using different examples across EMU governance, the article illustrates how current mechanisms of political, legal, and administrative accountability predominantly evaluate the procedures followed by EU institutions when performing their tasks. While such an approach can bring clarity, predictability, and autonomy for the actors involved, it distracts attention from the substantive assessment of EMU decisions. The article contributes to the EMU accountability literature by going beyond principal-agent expectations of democratic control rooted in the nation-state or legal debates about principles for accountable behaviour found in EU Treaties. The EMU, and the EU more broadly, need a different perspective on accountability focused on generally-applicable standards for holding power to account.

KEYWORDS:

• Accountability
• Economic and Monetary Union
• European Union
• EU law

Introduction

The Economic and Monetary Union (EMU) is a major achievement of European integration but also one of its major democratic accountability challenges. Few other policy areas of the European Union (EU) are as clearly about ‘who gets what, when, how’ (Lasswell, 1936), laying bare the necessity of controlling how power is exercised beyond the nation-state. Yet the academic and political debate about democratic accountability in EMU has reached a stalemate. On the one hand, it is widely acknowledged that EMU’s institutional architecture suffers from structural flaws determined by a multi-level system that blurs conventional accountability lines between those who hold political authority in representative democracy (citizens) and those who make decisions on their behalf (EU institutions) (Follesdal & Hix, 2006). In economic governance, the ‘democratic deficit’ emerges from citizens’ lack of control over the EU policy agenda through national or European elections. In national elections, voters can only hold their own governments accountable, while key decisions are taken collectively in the European Council, the Eurogroup, or the Economic and Financial Affairs Council (ECOFIN) (Brandsma et al., 2016, p. 625). In European elections, the performance of the European Parliament (EP) is not connected with how people vote – and even if it were – the EP has limited powers in economic governance (Crum, 2018; Rittberger, 2014) and a weak oversight role vis-à-vis the Commission (de la Parra, 2017). In monetary policy, the European Central Bank (ECB) is more independent from electoral competition than any other central bank in the world – with a mandate that can be altered only through a Treaty change requiring unanimous agreement of all EU Member States (De Haan, 1997, pp. 413–414).

On the other hand, scholars have emphasised the need to improve EMU democratic accountability since its inception (Verdun, 1998). Demands to make EMU more accountable follow the standard discourse on accountability in modern governance, whereby the concept is seen as having multiple positive connotations, holding (1) the promise of democracy (by ensuring the answerability of elected officials), (2) the promise of control (through mechanisms to oversee administrative action), (3) the promise of justice (through judicial and administrative review of decisions), and (4) the promise of performance (by incentivising public officials) (Dubnick, 2014, p. 29). Unfortunately, the institutional set-up of EMU reduces its potential to meet such promises. For example, in order to address the structural weaknesses of the EP in the EU political system, there is a proposal to institutionalise a subcommittee for Eurozone oversight (Chang & Hodson, 2019). It is doubtful, however, that such a committee would deliver on the ‘promise of democracy’ given the well-known disconnect between citizens and EP elections (Hix & Høyland, 2013, p. 184). In a similar vein, since the ECB’s mandate can only be altered through Treaty change, there is pressure for the institution to narrow its mandate, e.g., by excluding itself from financial assistance, or setting more specific supervisory objectives (Braun, 2017, pp. 6–7). The constraints on the ECB legal framework mean that the ‘promise of control’ present in accountability discourse is largely a voluntary exercise.

The mismatch between the structural flaws of EMU accountability and the incremental proposals put forth to reform it suggests that academic thinking about EMU accountability is at a stalemate. There is a gap between what is seen as necessary and what is feasible in an EMU governance framework marked by a hybrid constellation of intergovernmental and supranational actors. The problem is not that EU scholarship lacks ideas about how to reform EMU accountability, but that reform proposals cannot be easily implemented in the EU institutional structure. In this article, we identify the cause of the stalemate in the development of deductive and inductive approaches to EU accountability. Most deductive approaches apply national accountability standards to EMU and find shortcomings in their institutionalisation at the EU level – especially regarding the role of parliaments. In contrast, inductive approaches start from the Treaty framework and subsequently infer benchmarks of accountable behaviour. Both approaches inevitably fail: while the EU cannot meet national accountability standards, the principles set in the EU Treaties are too narrow and hence decoupled from generally-applicable accountability standards.

To break the stalemate, we propose a new deductive framework for studying accountability more suitable to EMU and the EU setting. Drawing on liberal and republican political theory, legal scholarship, and the public administration literature, we develop a normative conceptualisation of accountability that seeks to answer a basic question: ‘what is accountability good for?’ Accordingly, we identify four normative ‘goods’ of accountability: openness, non-arbitrariness, effectiveness, and publicness. Different accountability forums can examine an actor’s conduct in line with the four goods: parliaments through legislative oversight, courts through judicial review, ombudsmen and court of auditors through administrative review. Furthermore, we show that accountability forums can address the normative goods in two ways: one centred on the processes through which actors take decisions (procedural accountability), the other focused on the merit of the decisions themselves (substantive accountability). We argue that there are both payoffs and trade-offs in choosing one alternative over the other. While procedural accountability brings clarity and predictability for the actors involved, it detracts from the underlying goals of the normative goods (by drawing public attention from the policies EU officials pursue to the procedures by which they do so). In contrast, substantive accountability is more costly to achieve, but has the merit of maintaining the concept’s normative ethos.

We begin by explaining the ‘stalemate’ in the EMU accountability literature, showing the need for a deductive, normative perspective that is applicable to the field without being specific to it. The second section introduces the four normative goods of accountability and describes the possibilities to enforce them in a procedural or substantive way. The third section applies this conceptualisation to EMU by showing how political, legal, and administrative forums deliver the normative goods of accountability both procedurally and substantively. The fourth section highlights both the dominance and the limits of procedural accountability. Given the political salience of EMU, we conclude that procedural mechanisms are overused in its current governance structure.

The stalemate of accountability research on EMU: between deduction and induction

To organise the EMU accountability literature, we propose a distinction between deductive and inductive approaches. The classification illustrates the normative reasoning behind accountability assessments: how do scholars judge whether an actor has acted accountably in EMU? Do they start from general standards of accountability and seek to apply them to specific institutions like the ECB? Or alternatively, do they first examine a given institutional setting, e.g., the legal framework of the ECB, and then derive accountability standards applicable thereof? The distinction between deductive and inductive methods is well-known in scientific inquiry. From Aristotle to Francis Bacon to William Whewell, philosophers have discussed two directions of the scientific method: either beginning with general and fundamental principles that are then applied to specific cases (deduction), or starting with the specificities of what is observed and then developing general principles (induction) (Andersen & Hepburn, 2016). In accountability research, the distinction is useful because it transcends regular disciplinary boundaries between political science and law (Dubnick, 2014, p. 33). We can identify deductive and inductive studies which examine all the classic institutional mechanisms of accountability, whether political, legal, or administrative.

In the EMU accountability literature, deductive approaches revolve around two general standards of accountability: ensuring democratic control and preventing abuses of power (Bovens, 2007, p. 463). In respect to political accountability, the two goals are visible in the role assigned to parliaments in economic and monetary affairs respectively. In economic governance, the importance of parliaments for democratic control is grounded in understanding the concept as the counterpart to delegation in the ubiquitous principal-agent model. The logic is straightforward: if ‘A is obliged to act in some way on behalf of B’, then ‘B is empowered […] to sanction or reward A for her activities or performance in this capacity’ (Fearon, 1999, p. 55). Although the EU chain of delegation is complex (Art 10 TEU), parliaments are seen as crucial in closing the gap between citizens as the ultimate principals of EU economic decisions and various executive agents (e.g., the European Council, the Council, or the Commission). Accordingly, scholars argue that democratic control in economic governance can be improved by empowering the EP (Fasone, 2014; Fromage, 2018; Rittberger, 2014) and national parliaments (Hallerberg et al., 2018; Maatsch, 2017) in both decision-making and legislative oversight. Despite having different analytical foci, these studies share the normative assumption that parliamentary involvement in economic governance can help Member States – and hence the EU – ‘meet their legitimacy obligations to their own publics’ (Lord, 2017, p. 676).

In monetary affairs, deductive studies focus on ‘preventing the development of concentrations of power’ and ensuring an appropriate system of checks and balances of the ECB (cf. Bovens, 2007, p. 462). The emphasis here is different because the ECB is a non-majoritarian institution whose need for independence from electoral competition has been one of the cornerstones of the EMU since its creation (Art 130 TFEU). The EP is thus often cited as an ‘accountability forum’ and not the principal of the ECB in monetary policy and banking supervision; consequently, the ECB ‘owes’ the EP transparency and justification of decisions but not obedience or even political responsiveness (Amtenbrink & van Duin, 2009; Collignon & Diessner, 2016; Maricut-Akbik, 2020; Markakis, 2020). Conversely, the importance of preventing abuses of power is much stronger in the area of legal accountability, i.e., judicial review of ECB decisions by national and EU courts (Dawson et al., 2019, pp. 88–91; Goldoni, 2017). In legal accountability, deductive and inductive approaches are sometimes intertwined, as courts apply the general standard of ‘curtailing the abuse of executive power’ in reference to existing regulations (Bovens, 2007, p. 466). A combination of deductive and inductive approaches can be found for instance in the work of Markakis, who examines the accountability of the ECB in relation to the price stability objective prescribed in Article 127(1) TFEU (Markakis, 2020).

Conversely, most studies on legal accountability are ‘purely’ inductive, as scholars use EMU’s legal architecture as the starting point for evaluating accountability. A clear example is offered by case-law analyses of ECB or European Stability Mechanism instruments. In this category, scholars do not apply overarching accountability standards in order to evaluate judicial decisions, but selectively employ the Treaty framework in order to identify specific features relevant for the case under investigation. For instance, the independence of the ECB is connected to different headings and degrees of judicial review, e.g., the duty to state reasons in Gauweiler.¹ In this context, scholars offer contrasting interpretations of the stringency with which the Court of Justice of the European Union (CJEU) should uphold different Treaty principles. The ECB’s introduction of the Outright Monetary Transactions (2012) programme, for example, divided lawyers on potential violations of the Treaty’s ‘no-bailout’ clause (Art 125 TFEU) and the extent to which national and EU courts should intervene to hold the ECB accountable for acting ultra vires (Goldmann, 2014; Zilioli, 2016).

Another variant of inductive studies examines the normative peculiarities of EMU governance. In a recent article, Steinbach argued that the ‘normative choice for accountability’ in EMU need not be democratic in the principal-agent sense of empowering parliaments; rather, EU economic governance has created an accountability regime of its own (Steinbach, 2019, pp. 1357–1358). This regime is institutionalised in the EU Treaties, which highlight the Union’s commitment to create ‘an open market economy with free competition’ (Articles 119(1–2), 120, and 127(1) TFEU). The emphasis on the ‘free market’ is seen as a constitutional norm which ‘implies the absence of state intervention in the market-based price determination process’ (Steinbach, 2019, p. 1359). Accordingly, states and private actors in EMU are/should be primarily accountable to markets rather than political actors operating in a flawed democratic system. The issue is whether accountability to the market is normatively justifiable in a democratic system; after all, the belief that the market will hold actors accountable ‘just the right amount’ is rooted in ordoliberal assumptions of political economy that have been seriously challenged since the crisis (Ryner, 2015).

The point of this review is not merely to demonstrate the typology between deductive and inductive approaches but to show its problems. First, deductive studies drawing on principal-agent theory are stuck in a vision of accountability designed for the nation-state that is simply unattainable in EMU. National parliaments and the EP are important accountability forums, but they cannot be expected to deliver in the same way as legislatures in national democracies. In EU politics, the delegation chain from elected representatives to executive actors is either short-circuited (in the case of the EP) or too tortuous to function properly (for national parliaments). In contrast, deductive studies focused on preventing ‘abuses of power’ offer a more compelling standard of accountability in the EU political system and should be developed further – as we propose below.

Second, inductive studies display different problems, namely the replacement of general accountability standards with situation-specific benchmarks assessing the performance of an actor in a given setting. In EMU, the EU Treaties constitutionalise certain principles (like the ‘no-bailout clause’ or the ‘free-market orientation’) that remain open to political contestation and ordinary decision-making domestically (Grimm, 2015). One of the difficulties of this constitutionalisation is that EU institutions can be held accountable for the extent to which their decisions comply with Treaty principles but not for the principles themselves. Political accountability, however, may concern the overall principles governing EMU (a level of contestation that inductive studies – as they are oriented by these principles themselves – cannot capture). Normative standards of accountability should be broader than policy-specific benchmarks for holding actors accountable in a particular context. We introduce such an approach in the next section in relation to the EU setting.

Four normative goods of accountability beyond the nation-state

The objective to conceptualise accountability beyond the nation-state is typically found in global politics (Grant & Keohane, 2005). Reviewing this literature, Michael Goodhart concluded that all approaches to global accountability share a similar understanding of the term, namely the ‘question of making those who wield power answerable to the appropriate people’ (Goodhart, 2011, p. 45). In his view, this definition is based on a Westphalian notion of the state that is unworkable in world politics. In the EU, the absence of a European demos and the resilience of national demoi (Nicolaïdis, 2013) implies that accountability cannot be organised around the ‘appropriate’ forum because this is simply not feasible in a large-scale political unit where citizens have few opportunities to influence governing decisions (Dahl, 1999). In this context, Goodhart proposes to shift our thinking about accountability ‘from who is entitled to hold power to account to the reasons why accountability is justified in democratic theory’ (2011, p. 51). He is interested in the nexus between democracy and human rights norms that ‘constrain the exercise of power and enable meaningful political agency’ in the global arena (2011, p. 52).

We agree with Goodhart that accountability beyond the nation-state must begin with an understanding of the normative goods to which accountability is aimed. However, we believe that focusing on human rights reduces accountability to the role of an instrument necessary to achieve other democratic objectives rather than giving it credit as a democratic goal in itself. We subsequently identify four normative goods of accountability based on a survey of liberal and republican political theory and the broader public administration literature. Our approach is deductive, focused on general standards of accountability that can be applied to specific cases. By their nature, these goods are not intended to preclude disagreement about their content but rather to structure and allow distinctions between different accountability claims to be identified. While other authors have sketched similar frameworks (e.g., Bovens, 2007, pp. 462–464 or Dubnick, 2014, p. 29), our approach is different because it excludes the dimension of political responsiveness to the appropriate forum as inapplicable beyond the nation-state. We outline the four goods below.

The first good is openness. Liberal thinkers from Bentham onwards have long argued that public confidence in official action will increase if public policy is conducted under the public gaze (‘publicity’) (Bentham, 1999). The openness of public policy has thus been linked to a number of public goods such as the avoidance of corruption (Kolstad & Wiig, 2009), the improvement of public knowledge, and the republican demand that free citizens should enjoy ‘non-domination’ through the ability to question and contest official action (Heidelberg, 2017). We might want accountability because we see it as a device to ensure that public action is open, transparent, and contestable.

The second good is non-arbitrariness. There is a deep tradition tying accountability with principle-agent insights of how (political) principles can control (administrative) agents to whom they have delegated powers (Fearon, 1999). This is a narrower instance of a broader accountability good, namely that those who wield public power should do so in a limited manner and that they should exercise coercion only to the degree necessary to achieve their goals (Bovens, 2007, p. 463). In political theory, constraining arbitrariness is a key pillar in civic republicanism (Pettit, 1997, p. 55). Non-arbitrariness is also linked to more general limits on public action such as human rights or due process guarantees that seek to regulate the relationship between the individual and the state (Allan, 2013, p. 77). Accountability – by making officials answer for conduct – provides a means by which arbitrary distinctions or applications of power can be identified, and later remedied.

The third good which accountability seeks to render concerns effectiveness. While openness and non-arbitrariness seem highly normative values, accountability may be sought for more utilitarian reasons, namely that accountable officials are more likely to deliver high-quality services. From this perspective, accountability holds the promise of performance (Dubnick, 2014, p. 29). By making an official answer for their conduct, and by offering the possibility to correct potential errors, accountability is a mechanism to improve government effectiveness (Bovens, 2007, pp. 463–464). Here, the premise is that the need to justify and even correct conduct will likely improve, and encourage reflection upon, the design of policy-making or implementation.

The final good is publicness or the idea that official action should be oriented towards the common good – and therefore justified by public or universal reasons (Waldron, 2016). This involves demonstrating both that officials were not personally enriched and that their decisions are fairly balanced, taking into account different societal interests and perspectives. Once again, accountability can ensure the publicness of official action in this sense – when parliamentarians scrutinise government agencies, or courts conduct judicial review, a key demand is that actors show how their activities forwarded the national or collective interest (Oliver, 1991, p. 28). Accountability is thus a device to advance the normative good of public policy grounded in the public interest.

Having established the normative goods of accountability, the question is how they can be delivered in practice. Our proposal is to distinguish between procedural and substantive ways of providing the four accountability goods. To put it simply, actors are procedurally accountable if they can demonstrate that the processes or steps they followed in performing their tasks were open, limited (non-arbitrary), effective, and/or public. In contrast, actors are substantively accountable if they can demonstrate that the decisions themselves or the outcomes to which they led were open, limited, effective and public. We further explain the distinction below.

Providing accountability goods: procedural vs substantive ways

The simplest way of understanding the distinction between procedural and substantive accountability is through the categories of public law (Harvey, 2017). Judges often differentiate between reviewing parliamentary acts on procedural or substantive grounds (O’Cinneide, 2013, p. 392). When conducting procedural review, a judge will enquire into the robustness of the process through which a parliamentary act was adopted (Lenaerts, 2012). When conducting substantive review, what is important is not the process of adopting an act but its substantive provisions per se and their likely impact. To give an example, if a Court is enquiring whether a statute setting out minimum requirements for religious schools infringes freedom of religion, it might assess the infringement either procedurally (did Parliament consider the impact of the bill on religious minorities, or incorporate their views, when adopting it?) or substantively (is the statute likely to infringe religious freedom or is it neutral vis-à-vis different systems of belief?).

If we transport this distinction to the world of accountability, procedural accountability suggests an accountability relation oriented around the process by which a particular decision was rendered. If we are holding an actor to account procedurally, we are calling them to account for the procedural steps they undertook in forming or executing a policy decision. If we are holding an actor to account substantively, by contrast, we are calling them to account for the substantive worth of the policy decision itself. Thus, a parliamentary committee examining the implementation of the bill above might, procedurally, ask how often religious schools had been inspected or how parents of religious minorities had been consulted. Parliamentarians might also seek, however, to hold the inspectorate substantively accountable – did the implementation of the statute achieve the goals it originally sought, or why did the inspectorate choose to prioritise inspecting one set of schools or aspect of curricula over another? In the latter case, what is at issue is not the form of decision-making but its substance i.e., did the actor make substantively worthwhile, just, or efficient decisions? The official is thus being held accountable against a substantive benchmark: to justify the worth of their action.

From a conceptual perspective, process is irrelevant in this case – the inspectorate could be judged by parliamentarians to have implemented the statute justly or effectively even in circumstances where the procedure by which they had done so was inadequate just as the inspectorate could demonstrate a transparent and inclusive procedure yet still be seen by parliamentarians as substantively failing to adequately explain the correctness or efficacy of their decisions. In simple terms, the two ‘forms’ of accountability carry different lenses through which to understand whether accountability has been adequately rendered. The next section applies this conceptualisation to EMU.

Procedural and substantive accountability in EMU

The four normative goods of accountability – in both their procedural and substantive variants – can be identified across EMU’s governance architecture. The examples below focus on different EU institutions and accountability relationships in order to illustrate general trends, namely the predominance of procedural approaches since the euro crisis.

Let us start with the first good – openness. Transparency has been for decades a key concern of accountability research (Hood, 2010). An official can satisfy the demand for openness procedurally by providing the public with regular information and documents. Accountability forums (such as parliaments, auditors, or courts) may then investigate these procedures and demand their reform if found wanting. Many accountability requests in EU economic governance are of this nature. To take an example from banking supervision, the largest number of questions asked by Members of the European Parliament (MEPs) to the ECB are requests for information, seeking to address information asymmetries (Maricut-Akbik, 2020, p. 9). Keeping in mind that ‘transparency is a necessary but insufficient condition for accountability’ (Curtin, 2017, p. 43), it would make sense for MEPs to first ask for information before acting on it substantively. But when political accountability does not move beyond transparency requests, the value of openness remains procedural.

Openness can also be achieved substantively. The test of substantive openness is the de facto transparency of official action. To be substantively open, an official must not simply demonstrate transparent procedures, but provide information in a sufficiently relevant and timely manner that it is likely to be used by accountability forums (Kitrosser, 2015, p. 16). This establishes a corresponding obligation on accountability forums, namely that they utilise their information rights to understand decisions, and to make clear to the public the substantive choices, including achievements and errors, of public actors. To return to the example above, the key difficulty in the parliamentary accountability of the ECB is not only the availability of information but its volume and complexity (Maricut-Akbik, 2020, p. 12). MEPs simply do not have the expertise to identify the most salient questions that would allow them to substantively challenge ECB decisions. Substantive openness often requires additional resources, raising difficult questions (elaborated later) about who should bear its costs.

The second good – non-arbitrariness – also carries procedural and substantive elements. Procedurally, public actors are commonly bound by statutes which specify their substantive mission, meaning that officials must explain why particular decisions are necessary to fulfil their mandate (Harlow, 2006). One function of judicial review is often then to check whether this is the case. In the ‘suitability’ test of proportionality review, courts play an important role in limiting arbitrariness by verifying that there is a justified connection between the measures officials adopt in forwarding their mandate and that mandate itself (Harbo, 2010). Similarly, public institutions may adopt procedures to ‘mainstream’ rights-based limitations into their policy-making i.e., to conduct impact assessments by which officials may demonstrate that human rights have been taken into account (McCrudden, 2005). Here, arbitrariness is limited procedurally in the sense that public actors bind themselves via process-based limits on their action. For example, under the 3rd Greek bailout (European Commission, 2015), the Juncker Commission prepared social impact assessments on the implications of financial assistance programmes on social rights as part of the commitment to improve EMU’s social dimension.

Substantively, non-arbitrariness carries a higher bar. As with openness, the important element is whether public action was de facto limited and non-arbitrary (in that room for policy manoeuvre was meaningfully constrained). Non-arbitrariness also concerns whether a given policy arbitrarily discriminates against a given group in society, or infringes an individual’s rights. Accordingly, the existence of a procedure to mainstream human rights considerations within an institution (or to give reasons to a verifying court on why a measure was necessary) would not fulfil the non-arbitrariness requirement if the outcome of such mainstreaming (or reason-giving) violated a core right. To return to the example of financial assistance, while the Commission indeed conducted a social impact assessment for the 3rd Greek bailout, scholars criticised the adequacy of this assessment and whether it actually generated any meaningful changes in the implementation of Greece’s stability programme (Copeland, 2020, pp. 144–145). The key test for substantive accountability as non-arbitrariness is thus whether policy choices in EMU plausibly aimed for and achieved non-arbitrary results.

The third accountability good, effectiveness, would seem to be inherently substantive in nature. Nevertheless, the ‘explosion’ of auditing in the 1990s as a means to control government (Power, 2003) suggests that effectiveness can also be assessed in a limited procedural way. In 2016, the European Court of Auditors (ECA) evaluated the operational performance of the Single Supervisory Mechanism and the role of the ECB thereof (European Court of Auditors, 2016). The evaluation report included a section on the ‘difficulty in obtaining audit evidence’, as the ECA officially complained that the ECB provided the auditing team ‘very little’ of the information required (European Court of Auditors, 2016, p. 20). The ECB was given the chance to respond to the report, emphasising its different understanding of ‘sufficient information’ for auditing purposes. Specifically, the ECB claimed that ‘all audit evidence covered by the Court’s mandate to audit the “operational efficiency of the management of the ECB” had been provided’ and that any exception concerned documents that were not related to the SSM’s operational efficiency (European Court of Auditors, 2016, p. 123).

The problem here concerns the legal limits imposed on auditing the ECB’s substantive effectiveness by the ECA. According to EU primary law, the ECA has ‘full power to examine all books and accounts of the ECB’ but only in respect to examining the ‘operational efficiency of the management of the ECB’ (Article 27, Protocol no. 4 TFEU). The conflict between the two institutions regarding the SSM report reflects the different interpretations of ‘operational efficiency’ and the substantive need for information to make such an assessment. In the absence of said information, the ECA’s report was reduced to evaluating procedural aspects such as staffing, e.g., whether on-site inspections of banks should be run by ECB staff as opposed to representatives of national supervisors (European Court of Auditors, 2016, p. 11). Despite these legal limitations, the ECA clearly understood the value of substantive effectiveness, as the report recommended that the ECB develops a public and formal performance framework to assess the effectiveness of its supervisory activities (European Court of Auditors, 2016, p. 12).

Lastly, the fourth accountability good of publicness can also be rendered procedurally or substantively. In both cases, publicness requires that public officials demonstrate the orientation of their conduct towards the common good. In the EU, this may take on a specific meaning, namely the duty of EU actors to demonstrate that their policies take the interests of the EU as a whole into account (and not just selected industries or states). A common mechanism to achieve procedural accountability as publicness concerns establishing procedures for public participation in government action (Stewart, 2014). An accountability forum may therefore assess whether a robust process of consultation existed when adopting policies. In the EMU context, the leaders of the EU’s main institutions committed, via the 2015 ‘Five Presidents Report’, to greater involvement of social partners in fiscal coordination via the European Semester, with some commentators arguing that this led to a gradual ‘socialisation’ of the Semester’s priorities (Zeitlin & Vanhercke, 2018). In this case, accountability as ‘publicness’ is pursued by procedurally involving relevant actors in policy-making.

Substantively, publicness concerns not just the existence but the effects of participation on decisions. While the participation of civil society actors in the European Semester may be an end in itself, linking this participation to a debate about the Semester’s policy priorities illustrates the weakness of procedural accountability. The real question is whether civil society involvement in the Semester matters or is simply another abstract commitment boldly stated in policy documents only to be safely disregarded when substantive decisions about EU fiscal policy are made (Dawson, 2018).

With all four categories, the guiding distinction between procedural and substantive accountability is between process and merit. For the former, the connecting point between actor and forum are the steps taken to make public action accountable; for the latter, the basis for interaction is the merit of official action vis-à-vis alternatives. The ability of different institutions – from courts to parliaments – to deliver one form of accountability or other may highly vary, leaving a crucial question: why institutionalise an accountability regime focused on one form of accountability rather than the other? This is the subject of the next section.

Procedural accountability – and its limits

There are good reasons why institutions may favour procedural accountability. One reason concerns the clarity and predictability of the standards used to orient accountability relations. Under the substantive reading, a potentially broad set of standards are at play, with the actor under a potentially limitless justificatory burden. To take one of the four categories discussed above, requirements of publicness in public policy are complex and subject to significant disagreement (Waldron, 1999). This diversity in interpretation applies to both actors and forums. Assuming that officials may be accountable – in an EU setting – to multiple accountability forums, finding an appropriate balance to satisfy these interests can be an overwhelming task. In the EMU context, the ECB’s bond buying programme was of such volume and complexity to have potentially limitless consequences on a wide variety of societal interests – a concern central to the recent ultra vires ruling of the German Constitutional Court.² How could the ECB therefore adequately demonstrate that this policy was non-arbitrary, effective, and oriented towards the common good?

A second reason (or ‘payoff’) regarding procedural forms concerns the unique independence of key actors operating in the EMU context. Both the ECB and other central banks have high operational autonomy in order to guarantee the credibility of their price stabilisation functions. By providing high discretion over how to fulfil their mandate, such institutions benefit from a procedural form of accountability that restricts substantive political pressure in order to protect the autonomy and credibility of their decisions (for an elaboration of this discussion, see the Online Appendix).

Under procedural accountability, the primary duty of the actor is to follow an established process (under the assumption that, if followed, this process will lead to positive outcomes). The forum must then verify that the correct process has been implemented. As a result, a significant burden is lifted and externalised: for the actor, the burden of calculating and adjusting conduct according to its concrete effects; for the forum, the burden of policing how the actor utilised the discretion it was given. In social psychology, experimental research has demonstrated that procedural (or process) accountability is linked to situations when actors lack knowledge about the specific outcomes they are expected to achieve; consequently, they shift their focus to the quality of their decision-making process (Schillemans, 2016, p. 1412). This suggests that procedural accountability can easily become decoupled from substantive results.

This potential ‘replacement effect’ of procedures for substance is precisely the first major limitation of procedural accountability. In regard to actors, Jane Mansbridge famously discussed the interplay of material and moral incentives in the process of delegation (Mansbridge, 2014). The introduction of material incentives may destroy moral incentives by making what was once a duty (to act in the right way) a material question (is this action in my interest?). A similar risk applies in relation to procedural accountability. Procedural accountability assumes that if actors follow correct procedures, they will orient their activities towards correct substantive outcomes. The risk is that the opposite applies – that actors begin to care only about procedures at the expense of the substantive goods those procedures are designed to secure (and without reflecting on their adequacy).

In EMU, the replacement effect of procedural for substantive accountability is certainly present. ECB accountability in particular is grounded in procedural devices such as transparency, which are seen as better suited to the institution’s operational independence (Dawson et al., 2019, p. 81). Nevertheless, transparency obligations tend to limit contestation of the functional mission of the ECB, as accountability debates become hijacked by discussions over the ECB’s secrecy regime (Curtin, 2017). Consequently, the focus on transparency limits the substantive contestation of the actual merit and distributive implications of ECB decisions. While accountability forums like the EP are thus permitted and even encouraged to ask questions of ECB officials, their ability to contest and seek to influence the direction of monetary and supervisory policy is minimal (Collignon & Diessner, 2016; Maricut-Akbik, 2020). Accountability in substance seems needed but also excluded a priori by EMU’s institutional structure.

Second, procedural accountability is limiting from the perspective of accountability forums. One challenge relates to information asymmetries. In principal-agent theory, agents are expected to ‘shirk’ their obligations before principals either by hiding information before they are appointed (adverse selection) or by hiding their behaviour while on the job (moral hazard) (Moe, 1984, pp. 754–755). Procedural accountability extends these problems, as explained by Heidelberg:

Because the accountor is exposed to the rules and subject to them, it is not unusual for the accountor to have a better understanding of how the rule system works than the accountee, in which case, the rule system is as much an instrument for the accountor as the accountee. (Heidelberg, 2017, p. 1386)

The discussion about exceptions to the Commission’s application of the Stability and Growth Pact (SGP) are a clear example of an actor exploiting such a rule-based system. The Commission’s past record of sanctioning some Member States but not others for budgetary transgressions has raised important accountability questions of arbitrariness and equal treatment (Chalmers, 2012, p. 684). MEPs have regularly accused the Commission of unfair treatment of Member States, e.g., for the failure to take action against France’s or Germany’s fiscal breaches in 2014–2015 (de la Parra, 2017, p. 114). The problem is that the relevant legislation – the Two and Six-Pack – gives the Commission ample discretion to calculate budget deficits by deciding which expenditures, fluctuations, and one-off investments are taken into account and which are excluded (Dawson, 2019, p. 66). This has two important implications. First, the Commission benefits from information asymmetries as it is the only institution with the expertise to understand and apply the intricate web of legislative and non-legislative rules that constitute the European Semester. Second, the Commission can make strategic use of these rules in order to explain away decisions not to sanction Member States, as illustrated by the French and German decisions. Overall, procedural accountability seems to favour the actor at the expense of the forum.

Conclusions

In this article, we introduced a new normative framework for analysing accountability in EMU governance. The framework has been determined deductively by surveying the relevant literature in political theory, law, and public administration in order to identify four normative goods of accountability that are general rather than EU-specific. The originality of the approach comes from changing the analytical optics in accountability research from the appropriate forum to whom actors should answer (as emphasised by deductive studies) to the question ‘what should actors be accountable for?’ This is particularly applicable at the EU level, where the distance between the ultimate democratic principal (citizens) and their supposed agent (EU institutions) remains great. We argued that we can break the stalemate of EMU accountability research by examining the extent to which national and EU institutions provide four normative goods: openness, non-arbitrariness, effectiveness, and publicness. All can be achieved in a procedural or substantive way, but the latter imposes higher standards of behaviour for actors (who must defend the merit of decisions rather than simply the process through which decisions were reached). Nowadays, the EMU seems dominated by procedural accountability, which produces specific pathologies. Most problematically, there are dangers of a ‘replacement effect’ of procedures for substance and a consistently weak position for accountability forums vis-à-vis actors with expert knowledge.

Nevertheless, we do not conclude that procedural accountability is of limited value universally. Its advantages of clarity and predictability make it an important avenue in providing accountability in EMU. However, without a substantive component, EMU accountability is likely to remain focused on processes of decision-making, which are insufficient for evaluating decisions in a policy field that is increasingly distributive in nature. Conversely, substantive accountability carries trade-offs in terms of its complexity and costs. The limited knowledge of forums seeking substantive accountability (such as judges and parliamentarians), and the need to allow economic discretion, remain constraining factors. Moreover, accountability forums need expertise in many fields to balance information asymmetries, implying high costs. For these reasons, the choice for procedural or substantive accountability should in our view be political and revised at regular intervals. The political consensus that constitutionalised the EMU at Maastricht is increasingly contested today. Consequently, EMU’s political architects have to decide again if the costs of substantive accountability are worth the payoffs, or alternatively, if they are willing to accept the trade-offs that procedural accountability entails.

Acknowledgements

This research was supported by the European Research Council under the European Union’s Horizon 2020 research and innovation programme (grant agreement no. 716923). We are grateful for helpful comments on earlier versions of the manuscript to the three JEPP reviewers as well as to the participants in a workshop organised at the European University Institute in November 2019.

Disclosure statement

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

Additional information

Funding

This research was supported by the H2020 European Research Council under the European Union’s Horizon 2020 research and innovation programme (grant agreement no. 716923).

Notes on contributors

Mark Dawson

Mark Dawson is Professor of European Law and Governance at the Hertie School in Berlin, Germany.

Adina Maricut-Akbik

Adina Maricut-Akbik is Postdoctoral Researcher at the Hertie School in Berlin, Germany.

Notes

1 Case C-62/14, Peter Gauweiler et al. v Deutscher Bundestag, EU:C:2015:400.

2 German Bundesverfassungsgericht, Cases 2 BvR 859/15, 2 BvR 980/16, 2 BvR 2006/15, 2 BvR 1651/15, Judgment of 5 May 2020, https://www.bundesverfassungsgericht.de/SharedDocs/Entscheidungen/EN/2020/05/rs20200505_2bvr085915en.html.