Ever more soft law? A dataset to compare binding and non-binding EU law across policy areas and over time (2004

Abstract

What characterizes European Union soft law and what are its implications for the EU multilevel system? What is the proportion of hard and soft law in EU policy? Which types of soft law act are adopted in different policy sectors? This article introduces the conceptual and analytical framework that encompasses the EfSoLaw dataset and explains its methodology, advantages, and limitations. This dataset unites information on thousands of EU hard and soft law acts from seven different policy sectors, drawn from over fifteen years (2004–2019) and from various sources (EUR-Lex, DGs, agencies). We present implementation options of the dataset making it exploitable for other scholars and we propose hypotheses to explain the variation in the adoption of soft law in different policy sectors.

Keywords

Comparison database European Union law soft law

Introduction

Non-binding norms, generally termed soft law, are widely perceived to be on the rise (Guzman and Meyer, 2010). This is also the case in the European Union (EU) (Snyder, 1994; Stefan et al., 2019; Terpan, 2015). With ever more complex decision-making in Brussels, it has been stated that EU norms increasingly often take the form of soft law (Hartlapp and Hofmann, 2021). However, it is difficult to quantify how much soft law is being adopted in the EU. The Effects of Soft Law (EfSoLaw¹) dataset provides an answer to this question and offers an all-encompassing database to study how this affects EU policy making. The EfSoLaw dataset compares for the first time the number and characteristics of EU soft law acts across policy sectors and time. To illustrate the potential of the dataset, we present its structure and discuss two questions: What is the proportion of hard and soft law in specific policy sectors and over time? Which type of soft law acts are there, who are the authors, and what determines their adoption as regulatory tools at the EU level?

Many studies have either provided a clear definition or analyzed the nature of EU soft law decision making (Blutman, 2010; Guinard, 2013; Hillgenberg, 1999; Korkea-aho, 2015; Senden, 2004; Terpan, 2015). Others have explored specific types of soft law instrument such as the Open Method of Coordination (OMC) (e.g. Trubek and Trubek, 2005) or state aid guidelines (e.g. Cini, 2001). Zhelyazkova et al. (2015) quantify soft law acts, but limit their selection to acts produced by the European Commission, the European Council, or the EU member states. This raises the question of generalizability as EU soft law takes many more forms: opinions, codes of conduct, communications, notices, guidelines, and resolutions (including parliamentary resolutions). Yet, there is a dearth of empirical analyses that spans different EU soft law instruments as well as policy sectors. Providing a systematic descriptive account is therefore important to capture the provision and adaptation of non-binding norms within the EU more broadly. The dataset also offers a framework for developing existing analyses and arguments on EU soft law, understanding its scope, and deriving new hypotheses and theories on its role, as well as mechanisms through which it takes effect in the EU multilevel system. For example, descriptive features of soft law, such as the authorship of a norm or its enforcement mechanism, can further our understanding of EU governance. Engaging with these characteristics allows research on EU soft law to be connected more directly to EU policy-making and to link observable changes in EU governance to the dynamics of EU soft law.

The article introduces the conceptual and analytical framework of the EfSoLaw dataset, explains the underlying methodology, and discusses its advantages and limitations. This dataset complements previous databases created on EU hard law (e.g. Fjelstul, 2019; Schimmelfennig and Winzen, 2014) and qualitative comparisons of EU soft law across policy fields (e.g. Eliantonio et al., 2021), and it unites legal acts from various sources for the first time (EUR-Lex, Directorate Generals (DGs) and EU agencies archives). We manually coded the evolution of a wide range of EU soft law acts in seven policy sectors or sub-sectors of relevant legislation at the EU level: state aid policy, pharmaceutical regulation, food safety, the Common Foreign and Security Policy (CFSP), financial regulation, sustainable agriculture, and police and judicial cooperation. The data span over 15 years (2004–2019). This allows us to cover the more recent period not included in the study of legal instruments in the EU by Von Bogdandy et al. (2004) while at the same time taking the 2004 Eastern enlargement as a valuable starting point to analyze the changes in EU policy-making.

This article explains the methodology used for building the EfSoLaw dataset and includes the information that enables others to use it. We then explore the dataset confronting two different definitions of soft law, and investigate how much and what kind of soft law is produced at the EU level. Finally, we engage in an explanation of its patterns and differences between policy sectors and over time (further information in the form of EUR-Lex search criteria, a full codebook, descriptive statistics and coding scheme is available in the Online appendix).

Soft law in the EU multilevel system and the limitations of existing datasets

Although ‘soft law’ is not a clear category and therefore can have different meanings, the term is widely used in political science and legal literature (for an overview, see: Stefan et al., 2019). We present here two distinct strands of the literature, one focusing on the debates on EU soft law, based mainly on qualitative studies, the other investigating the characteristics and effects of EU law on law and policy through quantitative studies.

Soft law in EU studies

Within the scope of EU studies, soft law is sometimes defined as lacking key qualities of ‘hard law’ (Stefan et al., 2019: 9). The European hard law has legal binding force and produces general and external effects. It is adopted according to specific EU procedures written in the Treaties (Senden, 2004). However, different definitions of what actually constitutes soft law make it difficult to draw broader conclusions. Moreover, a combination of soft and hard norms can be found in the same policy, and sometimes in the same act, leading to situations in which legal hybridity occurs (Korkea-aho, 2015; Trubek et al., 2006).

In the 1990s, Snyder (1994: 198) developed an understanding of soft law as ‘rules of conduct which, in principle, have no legally binding force but which nevertheless may have practical effects’. Senden (2004: 112) built on this approach and defined soft law as ‘rules of conduct that are laid down in instruments which have not been attributed legally binding force as such, but nevertheless may have certain (indirect) legal effects, and that are aimed at and may produce practical effects.’ With the emergence of new modes of governance, the analysis has focused in particular on the so-called OMC. Scholars have studied the co-occurrence of hard and soft law via the OMC in social policy (e.g. Trubek and Trubek, 2005) and have done so more recently with regard to the European Semester (e.g. Bekker, 2014). There is a legal theoretical debate as to whether the OMC is best captured by referring to the softness of its norms or their character as coordinating governance instruments (e.g. Craig and De Búrca, 2011). While the merit of these studies is to render soft EU instruments more visible, our article attempts to provide a more systematic assessment of a range of soft law instruments that might differ substantially from the OMC.

Today, studies of EU soft law often highlight the heterogeneity of the context and nature of EU soft law (Eliantonio et al., 2021). An instrument like the Cybersecurity Strategy in the CFSP relies on soft steering by summarizing policy goals and setting expectations for national industries and administrations. In financial regulation, the ESRB Recommendation (ESRB/2013/1) specifies the intermediate objectives and instruments of macro-prudential policy within the broader framework of macro-prudential supervision and therefore asks the addressees to comply or explain. Other instruments operate much more like traditional hard law. The environmental and energy state aid guidelines, for example, set precise obligations which are enforceable in the context of state aid control in the member states although stemming from soft law instruments. That is why in this article various definitions of soft law will be proposed. We discuss the results ensuing from these definitions of soft law to illustrate the plurality of points of view from which scholars can start to test their hypotheses through the EfSoLaw database.

Previous EU law datasets and their limitations for the study of soft law

Previous research has tested theories on EU law and EU law-making using quantitative data. Former quantitative data collections have focused almost exclusively on hard law, exploring the policy-making phase (König et al., 2006; Leuffen et al., 2012) or compliance with it (Börzel and Knoll, 2012; Toshkov, 2010), as well as reconstructing the links between EU hard law and decisions of the Court of Justice of the European Union (CJEU) (Fjelstul, 2019).

Efforts to measure the significance of soft law in the EU have been undertaken in relation to hard law (Von Bogdandy et al., 2004) or via sector specific analyses (Eliantonio and Stefan, 2021; Ripoll Servent and Trauner, 2014; Schelkle, 2007; Terpan and Saurugger, 2021). Von Bogdandy et al. (2004) suggested on the basis of CELEX data that EU law was in a process of inflation, both in terms of the number and type of legal acts. They estimated that around 11% of the legal acts produced by the EU have a ‘non-binding operating mode’ compatible with what we call soft law, of which 5% are recommendations and opinions (2004: 97). More recently, Hofmann (2021) using Eur-Lex estimated that 2.7% of the legislation in force consists of opinions and recommendations. Concurrently, a steady consensus has emerged in various studies that an increase in the use of soft law instruments is occurring (Hartlapp and Hofmann, 2021).

However, most of these studies have focused on the rationale for and the process of production and use of soft law, without analysing whether soft law has actually increased as a share of the EU law produced or not. An exception is the work of Zhelyazkova et al. (2015), which presents soft law produced by either the Council of the EU, the member states’ governments, and/or the European Commission drawing on data from Eur-Lex and the Official Journal.

These official sources provide systematic, easily available data. Yet, they are reduced to the most ‘law-like’ EU soft law instruments. Therefore, they risk producing a biased picture of the phenomenon, excluding, for example, most of the soft law produced by EU agencies, although it is an increasingly common kind of regulation in the EU (Chamon, 2016; Rocca and Eliantonio, 2019). Our dataset allows to fill this void in the literature, providing the empirical basis for a trend in EU policy-making that many EU law and policy scholars assume. Existing datasets on EU law can be divided into two types. One type of datasets relies at least partially on legal files being coded by humans (Pollack, 1994; Toshkov, 2010). The other type involves automated data collection using computer-based coding to derive information from various sources related to EU law and law-making, such as Eur-Lex, Pre-Lex, or other institutional sites (Fjelstul, 2019; Häge, 2011; Hurka et al., 2021; König et al., 2006; Ovádek, 2021). The advantages of automated data collection include the large amount of data involved, their potential to provide continuous updates, and the replicability for the data collection process. However, this type of data collection limits the inclusion of interpretative variables and any information that is not repetitive in form. Moreover, conducting an automated data collection risks reproducing the biases that exist in the initial data source (Häge, 2011: 458), which is relevant given the lack of reliability of Eur-Lex (Steunenberg and Toshkov, 2009).

Yet, the vast array of shapes and forms that EU soft law documents take renders automatic coding impossible. An initial test on a limited number of Eur-Lex soft law documents showed incoherent automatic coding², and we wanted to create a useful dataset for those researchers asking questions about soft law in the EU multilevel system. The coding of variables relevant for implementation (e.g. the function of soft law, and its ways of enforcement) necessitated human coding given the variety of elements that need to be interpreted and the diversity of formats that soft law documents can have. For these reasons, we have opted for human-mediated coding leading to a higher reliability of the data.³

EfSoLaw: Introducing a new dataset

An operational definition of soft law for an inclusive database

Our definition of soft law borrows from how Abbott and Snidal (2013) understand soft law in international relations and international law as well as from legal scholarship (Baratta, 2014; Guzman and Meyer, 2010: 173–74). The most widely used definition limits soft law to non-binding instruments—for instance, recommendations, objectives, guidelines, or incentives—that resemble legal acts and/or produce legal effects (Snyder, 1994), and can be accompanied by some kind of soft enforcement (monitoring, peer review).

We follow this rationale and define soft law as non-binding norms. The focus on EU instruments rather than their constituting parts and single provisions facilitates the data collection. Unlike theoretically more refined definitions, such as the one based on a continuum proposed by Saurugger and Terpan (2021), this instrument-based definition does not require the different standards contained in an instrument to be studied to see whether there is an actual obligation.

This does not mean that our database avoids or simplifies the complexity of soft law. On the contrary, adopting an operational definition of soft law that includes a number of acts that are as large as possible has noticeable advantages, one of which is that it can serve different research purposes.

Constructing the EfSoLaw database

In the following paragraphs, we clarify how to generate the EfSoLaw database in detail, allowing other researchers to use the data knowing the conceptualization and operationalization processes that enabled it in the first place.

The database is composed of hard and soft law acts produced by the EU and its bodies and agencies in seven policy sectors. The policy sectors represent a sample that includes a diversity of EU competences, as well as different levels of technicality and styles of decision-making. The policy sectors vary based on two distinctions: the decision-making process (characterized as intergovernmental or supranational), and the age of a policy sector (long-standing EU policy making or a more recent development).

To make human-mediated coding feasible, we chose sub-sectors of various policy fields that would otherwise involve an unreasonable amount of acts to code manually: state aid from the competition policy; pharmaceutical regulation from public health and the internal market; food safety at the crossroads between consumers and health policy; financial regulation from the finance policy sector; sustainable agriculture from the common agricultural policy; police and judicial cooperation from the Area of Freedom, Security and Justice (AFSJ) policies. We added to these policy sectors one area of competence, the CFSP, which seems larger than the others but involves limited cooperation at the EU level. Thus, EfSoLaw includes seven policy sectors that are comparable but differ in terms of policy-making (the EU actors involved, the type of EU competences on a supranational to intergovernmental continuum, the time in which it became an integrated EU policy field).

Each of these sub-sectors shares many of the characteristics of the global policy fields to which they belong- e.g. food safety is a sector with an average degree of EU competences that contains many harmonized technical rules, which is typical in product regulation aimed at achieving public health goals. In contrast, the CFSP has fewer acts, yet shows a predominance of political rather than technical acts such as resolutions and conclusions, as would be expected in an area of intergovernmental policy that has only been introduced comparatively recently and to a partial degree among EU competences.

These policy sectors guided the collection of our data from the existing official EU database (Eur-Lex). We selected all EU acts (legal acts, preparatory documents) that were in force or published between 2004 and 2019, and included all of the EuroVoc thesaurus entries and directory codes relevant to our policy sectors (see the Online appendix). We then integrated data entries from the online archives of the DGs and EU agencies, adding instruments excluded from Eur-Lex, such as the pharmaceutical EudraLex law collection, as well as agency guidelines or reports, in order to complete our database.

For each act we cover up to 18 different variables. Some are ‘hard facts’ such as the date of entry into force and (where applicable) the expiry date, the main author and other actors involved in its elaboration (including agencies, DGs, member states, and committees), the policy areas covered and the nature of the addressees (internal or external) of the documents. Other codes required decisions that we made based on the existing literature. A good example is the functions that soft law can have. Scholars (Senden, 2004: 119; Zhelyazkova et al., 2015: 11) propose distinguishing preparatory and informative instruments (green papers, white papers, action programs and informative communications), interpretative and decisional instruments (the European Commission's communications and notices, certain guidelines, codes, and frameworks), and steering instruments (recommendations, opinions, and others that have emerged in community practice). Senden’s (2004) taxonomy allows a variety of instruments to be included that have grown in type and number since: Reports and impact assessments are seen as ‘informative communications’; resolutions and conclusions are seen as ‘steering’ instruments. Agency regulatory acts, in turn, could fall into one or more categories (‘steering’ or ‘interpretative instruments’) as this category bundles together acts produced by agencies that have different names, such as guidelines, codebooks, regulations, and which all have a regulatory purpose, but do not constitute hard law. We use a slightly different taxonomy that includes procedural, explanatory, and steering functions and we evaluate the content of the documents rather than relying on the instrument type alone (see the Online appendix).

Similar qualitative coding decisions were made regarding the existence of links to other hard law or soft law acts, the type of relationship to the related hard law (whether it anticipates, follows or stands apart), the type of enforcement prescribed, the specification of the enforcer and the enforcement duties. A summary of these coding decisions can be found in the Online appendix.

In sum, the variables included in the EfSoLaw dataset enable a whole array of research questions related to EU soft law to be covered: from types of soft law, to questions about how it is produced and the actors in charge of it, its function and potentially, expected forms of compliance. A limitation of all quantitative data is that it does not reflect the substance or relevance of individual instruments (see Börzel, 2005). Therefore, conclusions about expansion or change of policy are difficult to draw directly. However, the database can be used in manifold ways to derive trends, patterns, and hypotheses on EU law production, preparing the ground or situating conclusive research in a larger context.

How much soft law is produced by the EU?

In this section, we show how trends can be derived from the dataset concerning the proportion of soft law produced over time. We use a broader as well as a narrower operative definition of soft law. One such definition is based on a loose array of instruments that have been classified as soft law in the existing literature (sample 1). The other is based on a more restrictive definition of the EU instruments that constitute soft law (sample 2).⁴

Inclusive versus restrictive definitions of soft law instruments

For the wider sample 1, we follow Senden’s (2004: 112) inclusive definition of soft law (cf. Introduction) and operationalize it along the non-binding form and the potential effects criteria. Instruments that do not include at least a minimal rule of conduct or incitation and are entirely informative or intra-institutional documents (e.g. staff working documents) are excluded. The resulting sample 1 consists of a body of 8131 acts, of which 3980 soft law acts. Reports, opinions, and agency regulatory acts are very strongly represented (Table 1).

Table 1.

Types of acts included in the EfSoLaw database (sample 1) and used in the EfSoLaw project (sample 2).

		Sample 1		Sample 2
	Type of act	N	%	N	%
Hard Law	Regulation	1005	12.4	1005	17.8
	Directive	200	2.4	200	3.5
	Legal instrument for implementation	1021	12.6	1020	18
	Decision (post-Lisbon)	810	10	554	9.8
	Decision (pre-Lisbon), common positions and joint actions	765	9.4	441	7.8
	International agreement	350	4.3
Soft law	Guideline	114	1.4	112	2
	Agency regulatory act	856	10.5	856	15.1
	Recommendation	107	1.3	104	1.8
	Communication	210	2.6	203	3.6
	Notice	80	1	65	1.2
	Conclusion	18	0.2	16	0.3
	Declaration	11	0.1	11	0.2
	Opinion	910	11.2	903	16
	Resolution	254	3.1	159	2.8
	Green/white paper	18	0.2
	Impact assessment	85	1
	Report	929	11.4
	Other	388	4.8
	Hard law subtotal	4151	51	3801	57
	Soft law subtotal	3980	49	2429	43
	Total	8131	100	5649	100

For sample 2, we adopt a more restrictive definition of soft law. Such a definition is particularly suited for studying the effects of EU soft law in the EU multilevel system. In this context and relying on set-theory (e.g. Beach and Pedersen, 2016), we argue that determining the type of formalization and the type of enforcement provides for a thorough conceptualization. They represent necessary features of hard law, which are also mirrored in soft law, but in a more modest form. Formalization may refer to the structure and the content of the document. Only where instruments have a law-like structure and where the goals and standards are formulated in a precise way they can be defined as law at all. This refers to the existence of a clear list of rules that the addressees are supposed to respect (although they are not legally bound). Enforcement refers to the prescription of legal and practical effects. Non-bindingness is constitutive of soft law. Yet, below the level of binding norms, there are various procedures and mechanisms that aim to ensure compliance with soft norms. Monitoring, review processes, and soft surveillance can be seen as forms of soft enforcement which, contrary to hard enforcement, do not include coercive means. Other attributes, such as author, addressee, or function, relate to the institutional context and implications of soft law instruments and can be added to specify subtypes (Beach and Pedersen, 2016: 106–9).

Based on these decisions, we arrive at sample 2 by following two steps. First, we eliminate from the database those acts that do not fall into our restrictive definition of soft law based on formalization and enforcement. Reports, impact assessments, and green/white papers lack the necessary degree of formalization in terms of their structure, obligations, and/or clear addressee and area of application.⁵ Secondly, we include only acts that directly or indirectly address EU member states. We therefore exclude international agreements and other arrangements that go beyond the multilevel EU system and also apply to actors outside the EU.⁶ For similar reasons, we exclude acts that do not involve general application, which only formulate obligations that are addressed to a single member state, to a private actor, or to European bodies (e.g. the decision to nominate a director for an agency).

The resulting sample 2 consists of 5649 acts, of which 43% are soft law (guidelines, recommendations, communications, notices, declarations, conclusions, resolutions, opinions, agency regulatory acts, compared to 49% in sample 1). It includes a high number of opinions and agency regulatory acts.

Importantly, the variety of soft law acts is much greater than that of hard law. This may come as no surprise, as hard law is clearly defined and thus limited in its features by the treaties (and this, despite the possibility of adopting atypical acts). Soft law, in turn, is an open category, allowing for ever more types of instrument and variations. Yet, this contrasts with the strong focus of the (political science) literature on one specific type of soft law, which results from the OMC, rather than considering the range and diversity of soft law instruments.

We now turn to a comparison of EU soft law and hard law in terms of their most important patterns of behavior. Where possible, we differentiate between policy sectors and dive into how they have developed over time.

The hard-soft divide

How does the distribution of soft law acts develop over time? At the aggregate level, we see a gradual increase in EU acts for both samples. Figure 1 displays the growing trends in the production of soft law acts from the lowest number of 86 total acts in 2004 to a peak of 219 acts produced in 2017 (sample 2). Then the number decreases (183 in 2019). A steeper increase in production can be seen in sample 1, with the lowest number of 117 soft law acts found in 2004 and a peak of 365 in 2017. The differential between the two samples can be explained through the increasing number of reports produced in recent years, as well as by the growing number of acts destined for actors outside of the EU in the framework of CFSP cooperation programs excluded from sample 2.

Figure 1.

Hard and soft law acts produced by the EU per year (sample 1 and 2).

The second way to get a glimpse of this evolution in the period 2004–2019 is to compare how the production of soft law fares in comparison to that of hard law. Figure 2 looks at the ratio of soft law to hard law for the same time period. Including all the acts in our dataset (sample 1), soft law shares increase from the lowest percentage of total acts (around 41% in 2004) to comprising the majority (52% in 2019), via a peak of 59% in 2012.

Figure 2.

Ratio of EU soft law produced per year in the EfSoLaw dataset in the EfSoLaw dataset (sample 1 and 2).

In sample 2, soft law is outnumbered by hard law for the same period. The soft law ratio plateaus around the 40% mark. A sudden peak above 50% (2011 and 2012) is linked to a slump in the production of hard law acts during the debt crisis period rather than a marked growth in soft law (cf. Figure 1). We can conclude from this that hard and soft law production in the EU has been growing over time at a similar rate while different definitions of soft law can produce very different results both in scale and direction in terms of the trends in soft law production.

It is noticeable that these findings match the results of an earlier study despite the fact that we employed different methodologies. Zhelyazkova et al. (2015: 12) identified a rise in the rate of soft law from 20% in 2004 to almost 40% in 2012 using a database that included all soft and hard acts produced by the European Commission, the European Council and the member states, while we find a 40% to 60% rise for the same period using a database which includes more types of soft law and all EU institutions as authors in our enlarged sample. Therefore, across different methodologies the rate of soft law produced by the EU was clearly on the rise in that period. Since there was a rise in hard law production following the debt crisis period, we can turn down arguments about a supposed drop in hard law production as advanced in the context of the OMCs (Dehousse, 2016).

EU soft law across policy sectors

Disaggregating the distribution of hard and soft law by policy sector suggests that the ratio of hard to soft law varies considerably (cf. Table 2 and Figure 3). Food safety and sustainable agriculture are policy sectors characterized by the significant preponderance of hard law acts, while other sectors have a preponderance of soft law acts. Within this group, pharma regulation and financial regulation have the highest share of soft law with three-quarters, while state aid and police and judicial cooperation are governed softly in (only) two-thirds of all acts. The category ‘other sectors’ includes acts that are primarily related to other policy sectors, but are relevant to the seven main sectors selected for our database, operating somewhat as a mean of the results found overall in the sample.

Figure 3.

Ratio of hard and soft law by policy sector in the EfSoLaw dataset (sample 1 and 2).

Table 2.

Total of hard and soft law by policy sector (sample 1 and 2).

	Sample 1			Sample 2
Policy sector	Hard law	Soft law	Total	Hard law	Soft law	Total
Other sectors	472	310	782	136	583	719
Sustainable agriculture	416	96	512	501	929	1430
Food safety	1439	337	1776	93	130	223
CFSP	790	337	1127	192	190	382
Police/judicial cooperation	208	344	552	158	88	246
State aid	95	185	280	1391	274	1665
Financial regulation	551	1760	2311	341	49	390
Pharma regulation	180	611	791	408	186	594
Total	4151	3980	8131	3220	2429	5649

These data show a pattern of variation in soft law production that is statistically significant.⁷ To explain this, we propose the hypothesis that the type of main author in charge of decision-making in a policy sector explains differences in hard and soft law adoption. Areas of exclusive competence or shared competence in which supranational decision-making prevails, such as agriculture, food safety, or state aid, should present higher relative shares of hard law production. Soft law production should be more relevant in sectors with limited and/or relatively new EU competences and a tradition of intergovernmental decision-making such as the CFSP and police and judicial cooperation. In these sectors, the European Council produces hard law and soft law and the European Parliament tries to contribute through soft law as legislative capacity is limited in terms of the treaties. Soft law production should be even more relevant in financial and pharma regulation as these are sectors in which actors other than the main three institutions of the EU have important roles, such as powerful agencies or the European Central Bank where financial regulation is concerned.

We test this hypothesis and look at the distribution of authors of soft law using our restricted definition in sample 2 (cf. Figure 4). As expected, the contingency table of lead authors of soft law by policy sector, on which the figure is based, shows strong statistical significance.⁸ State aid shows a strong presence of the European Commission in the production of soft law (83%). In sustainable agriculture, the CFSP, and police and judicial cooperation, the production of soft law is more evenly shared between the European Commission and the European Council, as well as the European Parliament and specialized committees or technical bodies in the field, such as agencies.

Figure 4.

Authors of soft law by policy sector in the EfSoLaw dataset (sample 2).

When comparing the two sectors with relatively fewer soft law acts (sustainable agriculture [49], the CFSP [88]) with those that have more (pharma regulation [581], financial regulation [929], food safety [274]), we notice that they differ in their relative maturity and centrality among EU policy competences. Sustainable agriculture is a relatively new (sub)field of policy, while the CFSP and police and judicial cooperation, despite formally being among the main policy objectives of the EU after Lisbon (Monar, 2012), are struggling to emerge from the lower output of regulation related to their intergovernmental form of policy-making (Trauner and Ripoll Servent, 2016), notably when compared to the other sectors mentioned. These sectors also differ in terms of the respective roles of the European Commission and the agencies in how they are regulated. In those sectors where agencies have a relevant regulatory function (Busuioc, 2013; Chamon, 2016), such as food safety, financial, and pharma regulation, we find a significant output of soft law (55%, 52%, and 85% of all soft law being produced by agencies in the three sectors, respectively).

Although there are manifold potential causal factors that explain the production of soft law (Cini, 2001; Peters, 2011), our results suggest that the maturity of the EU competences in the policy field as well as the form and distribution of regulatory competences among EU actors are both pertinent factors to explain the rate of adoption of soft law in the EU. More advanced models could be used to test these hypotheses more thoroughly. We hope that the European studies research community will use this dataset to develop inferential analyses that could contribute to the advancement of knowledge and theory in the field.

Conclusion

This article has introduced a new resource for scholars interested in the study of EU law and policy. The EfSoLaw dataset samples hard and soft law acts from seven different policy sectors covering the period of the Eastern enlargement onward (2004–2019). The original dataset contains more than 16,000 legal instruments and allows data to be selected according to 24 main variables, from the type of act to the authors, the actors involved, the policy sector, any links with other instruments, and the type of enforcement. To show the potential of the dataset to explore and study patterns of EU policy-making and its effects, we illustrate the evolution of soft law production over time and across policy sectors, adopting two different definitions of soft law, and we test, using an inferential analysis, certain hypotheses regarding the factors that determine the preference for certain types of soft law act in policy-making.

We find that soft law has seen an increase in production, and in terms of the sectors analyzed is now used as much in EU policy as hard law. This depends, however, on how stringent the definition of soft law is, as we showed diverging results based on two different definitions. Strong differences can be observed in the use of soft or hard instruments to regulate different policy sectors, mostly in connection with the EU lead actors in charge of policy-making, which depend on the level of EU competences and the relative maturity of the policy fields within EU policy-making.

Future work based on the dataset will be able to extend the analysis of causal explanations to encompass the preference for certain soft law acts over others, or the preference for soft law over hard law. At the very least, we will use the dataset as an operative tool to select cases for studying the effects of soft law on the EU multilevel policy system. Therefore, although it can immediately be exploited for quantitative studies of EU law, the EfSoLaw dataset is also a decisive resource for scholars using qualitative methods, as it allows potential case studies to be identified, and enables the construction of valuable comparative studies of the most similar or the most different soft law instruments within or between policy sectors.

With this database, we offer proof that soft law constitutes a substantial part of the legal output of the EU. Our results suggest that further attention should be paid to soft law among scholars who study the functioning of the EU and its legal system. Scholars can use the existing dataset or build on it, adding new policy sectors of interest, replicating the sampling procedures presented above and in the Online appendix.

Supplemental Material

sj-docx-1-eup-10.1177_14651165221111985 - Supplemental material for Ever more soft law? A dataset to compare binding and non-binding EU law across policy areas and over time (2004–2019)

Supplemental material, sj-docx-1-eup-10.1177_14651165221111985 for Ever more soft law? A dataset to compare binding and non-binding EU law across policy areas and over time (2004–2019) by Bartolomeo Cappellina, Anne Ausfelder, Adam Eick, Romain Mespoulet, Miriam Hartlapp, Sabine Saurugger and Fabien Terpan in European Union Politics

Supplemental Material

sj-do-2-eup-10.1177_14651165221111985 - Supplemental material for Ever more soft law? A dataset to compare binding and non-binding EU law across policy areas and over time (2004–2019)

Supplemental material, sj-do-2-eup-10.1177_14651165221111985 for Ever more soft law? A dataset to compare binding and non-binding EU law across policy areas and over time (2004–2019) by Bartolomeo Cappellina, Anne Ausfelder, Adam Eick, Romain Mespoulet, Miriam Hartlapp, Sabine Saurugger and Fabien Terpan in European Union Politics

Supplemental Material

sj-dta-3-eup-10.1177_14651165221111985 - Supplemental material for Ever more soft law? A dataset to compare binding and non-binding EU law across policy areas and over time (2004–2019)

Supplemental material, sj-dta-3-eup-10.1177_14651165221111985 for Ever more soft law? A dataset to compare binding and non-binding EU law across policy areas and over time (2004–2019) by Bartolomeo Cappellina, Anne Ausfelder, Adam Eick, Romain Mespoulet, Miriam Hartlapp, Sabine Saurugger and Fabien Terpan in European Union Politics

Supplemental Material

sj-xlsx-4-eup-10.1177_14651165221111985 - Supplemental material for Ever more soft law? A dataset to compare binding and non-binding EU law across policy areas and over time (2004–2019)

Supplemental material, sj-xlsx-4-eup-10.1177_14651165221111985 for Ever more soft law? A dataset to compare binding and non-binding EU law across policy areas and over time (2004–2019) by Bartolomeo Cappellina, Anne Ausfelder, Adam Eick, Romain Mespoulet, Miriam Hartlapp, Sabine Saurugger and Fabien Terpan in European Union Politics

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The EfSoLaw project has been funded by a Franco-German grant provided by the Agence nationale de la recherche (ANR) and the Deutsche Forschungsgemeinschaft (DFG) (ANR-18-FRAL–0011).

ORCID iD

Bartolomeo Cappellina

Supplemental material

Supplemental material for this article is available online.

Notes

References

Abbott

Snidal

(2013) Law, legalization, and politics: an agenda for the next generation of IL/IR scholars. In: Interdisciplinary Perspectives on International Law and International Relations: The State of the Art. Cambridge: Cambridge University Press, pp.33–57.

Baratta

(2014) Complexity of EU law in the domestic implementing process. The Theory and Practice of Legislation 2(3): 293–308.

Beach

Pedersen

(2016) Causal Case Study Methods: Foundations and Guidelines for Comparing, Matching, and Tracing. Ann Arbor: University of Michigan Press.

Bekker

(2014) EU coordination of welfare states after the crisis: further interconnecting soft and hard law. International Review of Public Administration 19(3): 296–307.

Blutman

(2010) In the trap of a legal metaphor: international soft law. International and Comparative Law Quarterly 59: 605–624.

Börzel

(2005) Mind the gap! European integration between level and scope. Journal of European Public Policy 12: 217–236.

Börzel TA and Knoll M (2012) Quantifying non-compliance in the EU: A database on EU infringement. Berlin Working Paper on European Integration 15: 1-24. http://www.infringement-db.jmce.org/Index.html

Busuioc

(2013) European Agencies: Law and Practices of Accountability. Oxford: Oxford University Press.

Chamon

(2016) EU Agencies: Legal and Political Limits to the Transformation of the EU Administration. Oxford: Oxford University Press.

10.

Cini

(2001) The soft law approach: commission rule-making in the EU’s state aid regime. Journal of European Public Policy 8(2): 192–207.

11.

Craig

De Búrca

(2011) The Evolution of EU Law. Oxford: Oxford University Press.

12.

Dehousse

(2016) Has the European Union moved towards soft governance? Comparative European Politics 14(1): 20–35.

13.

Eliantonio

Korkea-aho

Stefan

(eds) (2021) EU Soft Law in Member States: Theoretical Findings and Empirical Evidence. Bloomsbury: Hart Publishing.

14.

Eliantonio

Ştefan

(2021) The elusive legitimacy of EU soft law: an analysis of consultation and participation in the process of adopting COVID-19 soft law in the EU. European Journal of Risk Regulation 12(1): 159–175.

15.

Fjelstul

(2019) The evolution of European Union law: a new data set on the acquis communautaire. European Union Politics 20(4): 670–691.

16.

Guinard

(2013) A propos du soft law en droit de l’Union Européenne : analyses théorique et pratique. Revue de L’Union Européenne 12(573): 609–617.

17.

Guzman

Meyer

(2010) International soft law. Journal of Legal Analysis 2(1): 171–225.

18.

Hartlapp

Hofmann

(2021) The use of EU soft law by national courts and bureaucrats: how relation to hard law and policy maturity matter. West European Politics 44(1): 134–154.

19.

Häge

(2011) The European Union policy-making dataset. European Union Politics 12(3): 455–477.

20.

Hillgenberg

(1999) A fresh look at soft law. European Journal of International Law 10(3): 499–515.

21.

Hofmann

(2021) Types of EU soft law and their national impact. In: Eliantonio

Korkea-aho

Stefan

(eds) EU Soft Law in the Member States: Theoretical Findings and Empirical Evidence. Oxford: Hart Publishing, pp.39–56.

22.

Hurka

Haag

Kaplaner

(2021) Policy complexity in the European Union, 1993-today: introducing the EUPLEX dataset. Journal of European Public Policy: 1–16. DOI: 10.1080/13501763.2021.1938174.

23.

König

Luetgert

Dannwolf

(2006) Quantifying European legislative research: using CELEX and PreLex in EU legislative studies. European Union Politics 7(4): 553–574.

24.

Leuffen

Rittberger

Schimmelfennig

(2012) Differentiated Integration: Explaining Variation in the European Union. London: Macmillan International Higher Education.

25.

Monar J (2012) Justice and Home Affairs. Journal of Common Market Studies 50: 116–131.

26.

Ovádek

(2021) Facilitating access to data on European Union laws. Political Research Exchange 3(1): e1870150.

27.

Peters

(2011) Soft law as a new mode of governance. The Dynamics of Change in EU Governance 21: 423–424.

28.

Pollack

(1994) Creeping competence: the expanding agenda of the European Community. Journal of Public Policy 14, 95–145.

29.

Ripoll Servent

Trauner

(2014) Do supranational EU institutions make a difference? EU asylum law before and after ‘communitarization’. Journal of European Public Policy 21(8): 1142–1162.

30.

Rocca P and Eliantonio M (2019) European Union Soft Law by Agencies: an Analysis of the Legitimacy of their Procedural Framework. SoLar Working Paper 1–40.

31.

Saurugger

Terpan

(2021) Normative transformations in the European Union: on hardening and softening law. West European Politics 44(1): 1–20.

32.

Schelkle W (2007) EU fiscal governance: Hard law in the shadow of soft law? Columbia Journal of European Law 13(3): 705–731.

33.

Schimmelfennig

Winzen

(2014) Instrumental and constitutional differentiation in the European Union. JCMS: Journal of Common Market Studies 52(2): 354–370.

34.

Senden

(2004) Soft Law in European Community Law. Oxford: Hart.

35.

Snyder

(1994) Soft law and institutional practice in the European Community. In: Martin

(ed.) The Construction of Europe: Essays in Honour of Emile Noël. Dordrecht: Kluwer Academic Publishers, pp.197–225.

36.

Stefan

Avbelj

Eliantonio

, et al. (2019) EU Soft law in the EU legal order: A literature review. SSRN Electronic Journal. DOI: 10.2139/ssrn.3346629.

37.

Steunenberg

Toshkov

(2009) Comparing transposition in the 27 member states of the EU: the impact of discretion and legal fit. Journal of European Public Policy 16(7): 951–970.

38.

Terpan

(2015) Soft law in the European Union-the changing nature of EU law. European Law Journal 21(1): 68–96.

39.

Terpan F and Saurugger S (2021) Soft and hard law in times of crisis: Budget monitoring, migration and cybersecurity. West European Politics 44(1): 21–48.

40.

Toshkov

(2010) Taking stock: a review of quantitative studies of transposition and implementation of EU law. Working paper. Institute of European Integration Research, Maastricht.

41.

Trauner

Ripoll Servent

(2016) The communitarization of the area of freedom, security and justice: why institutional change does not translate into policy change. JCMS: Journal of Common Market Studies 54(6): 1417–1432.

42.

Trubek

Cottrell

Nance

(2006) ‘Soft law,’ ‘hard law,’ and European integration: toward a theory of hybridity. In: Scott

de Búrca

(eds) New Governance and Constitutionalism in Europe and the US. Bloomsbury: Hart Publishing, pp.65–107.

43.

Trubek

(2005) Hard and soft law in the construction of social Europe: the role of the open method of co-ordination. European Law Journal 11(3): 343–364.

44.

Von Bogdandy

Arndt

Bast

(2004) Legal instruments in European Union law and their reform: a systematic approach on an empirical basis. Yearbook of European Law 23(1): 91–136.

45.

Zhelyazkova

Börzel

Schimmelfennig

, et al. (2015) Beyond uniform integration? Researching the effects of enlargement on the EU’s legal system. MAXCAP Working Paper Series 8: 1–44.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

3.78 MB

0.00 MB

0.10 MB

7.33 MB