(a) Coding national convergence

The main aim of this paper is to explain variations in national legal convergence of EU consumer sales law. This is based on information from a Convergence Index which also covers variations of European convergence (see sub-section (b) below). The set of components of this index is summarised in Table 1. The study underlying this index explains the choice and coding of these components in detail,Footnote ¹⁸ and a sample of the coding explanations is available online.Footnote ¹⁹ In the following, we also provide an overview of the considerations behind the coding of the different components, as well as the choice of countries and directives.

Table 1.

National dimension of Convergence Index

The first component, national transposition technique, entails that if a national legislator chooses to transpose European legislation through a code or satellite law, the rules are likely to reflect a harmonised European identity. However, if European consumer rules are accommodated in a civil code, they will very likely have to fit into the existing national private law framework and thus the converging effect might be weaker. For example, if a national legislator must amend a civil code to accommodate European legislation, it will adapt the latter to the terminology used by established codification, and thus the European identity of the rules will be diluted.

The second component, reception of selected novel concepts, stands for the identity and interpretation of selected substantive provisions that are new to a national legal order. As European terminology lies at the core of the harmonisation debate, novel concepts are defined here as European legal concepts giving rise to legal rights or obligations, independent of any similar rights or obligations that might exist at national level. An example is the way some EU directives include the right of consumers to withdraw from a contract in a ‘cooling-off period’: here, while Member States may achieve the same effect through other legal concepts, such as termination of the contract, this leads to less convergence than the precise use of the novel European concept.Footnote ²²

Similarly, the third component, reception of selected open-ended norms, focuses on the way in which national legal systems accommodate blanket clauses. Open-ended norms are behavioural and moral standards that parties are held to respect in a contractual relationship, depending on their specific circumstances. This factor is partially connected to terminology that comes with a separate content from whatever equivalents might exist in the national legal order, which should ideally be treated independently. A good illustration of this problem is the concept of good faith.Footnote ²³ The Unfair Contract Terms Directive uses this concept in its general test, and the danger of not acknowledging the European roots of good faith in the context of unfair terms leads to the application of the national equivalent. The ideal interpretation of good faith in a European context is shown by Justice Mann in Foxtons, when citing the Preamble of the Directive for a better understanding of its definition: ‘That recital brings in the concept of good faith. This is an autonomous Community expression and has been elaborated in English authority as will in due course appear’.Footnote ²⁴

The fourth component, reference to European law in end legislation, deals with the way a national legal instrument can signal its European source in order to guide practitioners into applying it as such, and not solely as a national set of rules. Being aware that specific consumer sales rules stem from the European legislator facilitates the interpretation as a source of European law (which is required for EU hard law, such as directives), instead of mistaking this identity for a national framework.Footnote ²⁵

Timely transposition, the fifth national component, is relevant since any delay hinders effective harmonisation. Failure to implement a directive by a specified date leads to national courts having to interpret national law in the light of the purpose of the directive, as mandated by direct effect. However, sometimes it might prove difficult for national courts to do just that. Although untimely transposition may also be the subject of infringement procedures, often such applications become obsolete because Member States have, in the meantime, enacted transposing legislation. In order to avoid the bias of overlooking these cases, this component focuses on the initial failure of a timely transposition.

Sixth, the component on infringement procedures covers Member State behaviour when assessing willingness to engage in harmonisation efforts. Infringement procedures may be taken against a Member State pursuant to Art 258 TFEU for the failure to fulfil an obligation under the Treaties, be it regarding the untimely adoption of implementing measures, or issues regarding the conformity and correct application of these measures. However, this last factor only takes into account resilience to harmonisation, manifested in the incorrect transposition of European law. In order to avoid any overlap with timely transposition, this component therefore only considers those infringement procedures resulting in CJEU decisions on incorrect transposition.

In the study underlying this section, the values of these six national convergence components were coded for five directives (Doorstep Selling, Unfair Contract Terms, Distance Selling, Consumer Sales and Unfair Commercial Practices)Footnote ²⁶ and seven Member States (Belgium, France, Germany, Ireland, the Netherlands, Romania and the UK).Footnote ²⁷ Several reasons account for this selection. First, it focuses on key consumer sales directives from the consumer acquis that address substantive issues of sales law and apply a general level of consumer protection, while other directives either deal with specific procedural aspects or industriesFootnote ²⁸ or were too recent for the Index to be relevant.Footnote ²⁹ Second, as regards the countries, the selection includes both the civil and common law countries, as well as newer and older Member States.Footnote ³⁰ In addition, the targeted analysis of a limited number of units is deliberate in this paper given the use of fsQCA as an extension of a case-based method of social science research.Footnote ³¹

The resulting aggregate values on national convergence are displayed in Table 2. We use these aggregates, rather than the individual values, as our main point of interest because they provide good variation across the 35 observations. It is also in line with research that expresses a preference for composites or ‘bundles’ of variables in order to capture substitutes and complements of individual variables,Footnote ³² as well as the more general literature on index construction which accepts the usefulness of composite indicators.Footnote ³³

Table 2.

National convergence across five Directives and seven Member States (maximum 6)

(b) Coding European convergence

This paper explores the determinants of these aggregate values of national convergence. An initial point to consider is the role of the EU in stimulating convergence. This EU dimension of the convergence debate can draw on an extensive literature in general European private lawFootnote ³⁴ and further afield.Footnote ³⁵ Specifically, the following has to consider that the five directives may have different predispositions to stimulate national convergence. Here, we use the European dimension of the Convergence Index. It consists of the components of Table 3. Again, the study underlying this index explains the choice and coding of these components in detail;Footnote ³⁶ thus, here, we just provide a summary.

Table 3.

European dimension of Convergence Index

The first component in the European set, law versus policy, reflects the nature of the European instrument: mandatory or voluntary. It can be argued that if an instrument is mandatory (eg it is a directive), the national transposition efforts reflect top-down harmonisation, and that is more far-reaching than informal coordination policies. Even if the latter may prove successful, legal norms theoretically have a higher converging effect than political norms, because of – among other reasons – their enforceable nature. Moreover, it must be kept in mind that not only this component, but all the components below are assumed to be indicators of legal convergence, and not (mere) policy convergence.Footnote ³⁸

The second component, type of European instrument, focuses on the types of legislative acts. In the field of consumer sales law, directives are the main legislative acts (eg Unfair Contract Terms, Unfair Commercial Practices, etc). In more recent years, the proposal for a Common European Sales Law (CESL) has paved the way to regulation-based harmonisation; however, the CESL has been withdrawn by the Commission, which in turn presented three new directives on related matters.Footnote ³⁹ The main difference between CESL and the existing (and proposed) directives is that the CESL had been shaped into a regulation, which would have been applicable as such (even as an optional instrument), whereas directives still need to be transposed in national legislation.

The third component, nature of policy, looks into the degree of harmonisation. While in earlier days, consumer contracts were governed by minimum harmonisation, the trend in more recent times has been that of maximum harmonisation. This shift affects the way in which national jurisdictions accommodate European law. Moreover, it can be said that by not allowing Member States to derogate from common standards, national regimes become more similar.Footnote ⁴⁰

The fourth component, reference to self-regulation, assesses any self-regulation references in a directive, as well as the extent to which it is promoted by that directive. Private law actors complement state-made law with their own standards (eg trustmarks, codes of conduct, etc). Acknowledging the activity of the private sector may lead to the creation of public-private structures that can check this activity and make sure it does not fall short of the European standards it is supposed to fulfil.Footnote ⁴¹

The fifth component, general clause, deals with standard tests imposed by the law. A general clause is a provision in the text of the directive that uses a statutory test to determine whether the professional in a B2C transaction has fulfilled a specific legal standard. The general test normally includes several generic conditions, designed in a way that can include different factual patterns applicable to consumer contracts.Footnote ⁴²

The sixth component, black list, reflects the listing of specific factual situations which must be interpreted in the same manner throughout all Member States when it comes to specific consumer transactions. It can be argued that if the European legislator sets out examples of practices which national courts are required to follow, then it is more likely that legal systems converge more, since such a list has a unifying role in relation to practice. An example of this is the black list found in the Unfair Commercial Practices Directive. Not the same can be said about all such lists: while black lists have the unifying effect mentioned before, grey lists only include examples of practices that must still be tested against circumstantial facts.Footnote ⁴³

The seventh component, CJEU case law, reflects the usefulness of patterns that can be found in the volume of case law decided on by the ECJ/CJEU.Footnote ⁴⁴ The assumption here is that an unclear directive will lead to more cases as national courts struggle with the interpretation and application of European law. While the subsequent European case law may then foster European convergence, it does not change the fact that the original directive was drafted in a way necessitating judicial clarifications.Footnote ⁴⁵ This view is supported by the nature of the CJEU's activity: in recent years the Court has issued orders instead of judgments for various cases coming specifically from Slovenia, the Czech Republic, Hungary and Romania. Procedurally, the Court may choose to do so for questions on matters that have already been decided. Based on the fact that national courts can consult this case law, and still take the liberty to ask questions through the preliminary reference procedure, it is therefore argued that the provisions needing interpretation are so complex that national courts still do not know how to relate to them.

Coding the five directives of consumer law, we find that they have different predispositions to stimulate national convergence. Aggregates of the seven components of Table 3 lead to the score of ‘1’ for the Unfair Contract Terms Directive, ‘3’ for the Doorstep Selling, Distance Selling and Consumer and Consumer Sales Directives, and ‘4’ for the Unfair Commercial Practices Directive.Footnote ⁴⁶ We also considered using other conditions, for example, the length of the directives and their transposition periods.Footnote ⁴⁷ Yet, as this information would be identical for all of the countries under consideration, there is no benefit in including more conditions on the respective directives. It would potentially be interesting to code whether some of the Member States opposed a directive in the Council. However, as also noted in the literature,Footnote ⁴⁸ such information has only been made available recently; in particular, it is impossible to gather such information for the 1980s and 1990s:Footnote ⁴⁹ thus, it could not be considered in the analysis in this paper.

(c) Possible further determinants for national convergence

In addition to the European dimension of the convergence index, this paper examines which other factors influence legal convergence in consumer sales law. This can draw on a rich literature from legal research on private law harmonisation, as well as other fields and topics of European research.Footnote ⁵⁰ For present purposes, first, we identified conditions that aim to complement the national factors of the Convergence Index, given that such national characteristics of the legal system can help to clarify the role of the national legislator in legal convergence. Secondly, we considered the political dimension of the transposition debate with the aim of testing the Convergence Index on conditions that deal with national politics.

Thus, based on our reading of the literature, we identify six further conditions that can be divided into two groups of three conditions dealing with ‘country characteristics’ on the one hand and ‘national politics’ on the other. The selection of the conditions also considered that, for all of them, we expect only a unidirectional causal relationship for national-level convergence; thus, we do not face issues of endogeneity which have been a source of concern for some of the previous empirical research on the relationship between law and society (including political and economic factors).Footnote ⁵¹

The summary of the corresponding six conditions in Table 4 indicates that we have used different moments in time for different conditions. This is due to the fact that we expect some conditions to be relevant for the substantive orientation of the directive (eg whether a government is left- or right-wing) and therefore consider the moment of the national enactment. In other instances, our expectation is that the entire period from EU adoption to national enactment has shaped the domestic process of implementation (eg public support for the EU).

Table 4.

Possible determinants for national convergence

The three conditions on ‘country characteristics’ consider, first, whether there is no existing legislation in the matter of the respective directive.Footnote ⁵⁴ The rationale is that existing legislation creates path-dependence and therefore, potentially, less legal convergence for the Member State in question. Secondly, we include a condition on the quality of the legal system. The rationale is that it can be expected that in countries with good protection of rights in general, there is also the motivation to provide high levels of consumer protection. Similar conditions have been used in related studies; for example, Börzel and colleagues consider bureaucratic quality and rule of law.Footnote ⁵⁵ Specifically, our choice for the data on ‘Legal system and property rights’ by the Fraser Institute is due to the fact that it provides information for the 1980s and 1990s,Footnote ⁵⁶ while most of the other datasets (eg the various corruption indicators) only start in the late 1990s. The third condition, support for the EU, is also commonly used in studies that evaluate the non-compliance with EU law.Footnote ⁵⁷ It may be relevant here since the law-making institutions of a country can reflect the view of the population as regards either more complete or more limited legal convergence.

Further conditions related to ‘country characteristics’ were considered in our preliminary assessment but eventually not included in the final analysis,Footnote ⁵⁸ such as the divide between civil and common law countries or differences based on cultural variables (eg the Hofstede dataFootnote ⁵⁹). The two common law countries of our study tend to show slightly higher scores for national convergence.Footnote ⁶⁰ In terms of culture, it may be hypothesised that aspects such as individualism or uncertainty avoidanceFootnote ⁶¹ could reflect a country's approach to consumer protection. However, the problem is that these conditions would not vary across time. Thus, the validity of any results based on the current dataset would be too slim.Footnote ⁶²

The group of conditions dealing with ‘national politics’ also consists of three potential conditions. First, the condition on the effectiveness of the government examines whether a powerful government can more easily enact laws, leading to full convergence – or, in other words, whether the need to make compromises leads to less convergence. This condition is related to conditions used in the literature, for instance, on the number of veto players and constraints of government branches.Footnote ⁶³ In this paper we consider the political diversity of the government, the voting power of the parties of government in parliament and the specific case of the French ‘cohabitation’. Secondly, as far as national politics is concerned, it can be hypothesised that left-wing governments are more willing to support consumer protection and therefore also more willing to implement measures leading to full national convergence (noting that four out of five directives used in this paper are based on the concept of minimum harmonisation). It may also be argued that this factor can capture that right-wing governments can be nationalistic and Eurosceptic. Thus, in this regard, this condition supplements the coding of the EU support of the general population. Thirdly, if the implementation period is interrupted, convergence may be incomplete or delayed. Such a condition is used in related studies.Footnote ⁶⁴ It also fills a gap left by the condition ‘effective government’, which only considers the situation at the moment of the national enactment; thus, the information about the interruption of the implementation period can be relevant as it captures the potential influence of the previous government.

(a) FsQCA in a nutshell

This paper applies the method of a fsQCA in order to analyse the determinants for national convergence of EU consumer sales law. FsQCA has become a popular method across many academic disciplines. For example, in political science and European studies, it has been applied to the transposition of EU legislationFootnote ⁶⁵ and the ‘gold-plating’ of European directives.Footnote ⁶⁶ There are also examples from legal scholarshipFootnote ⁶⁷ and management studiesFootnote ⁶⁸ which have used fsQCA in order to understand legal and regulatory differences between countries.

FsQCA differs from regression analysis, because it does not require a large number of observations and a small number of explanatory variables, ie it can work with a small number of observations and a relatively large number of explanatory factors (in the fsQCA terminology: ‘conditions’). More specifically, it aims to ‘facilitate a dialogue between theory and evidence’;Footnote ⁶⁹ thus, it also asks researchers to use qualitative skills and knowledge in research design and evaluation. In the present case, fsQCA in combination with configurational narratives can indicate variants of causal relationships for the specific countries and directives under investigation (as the following will show). This also means that this approach can achieve a relatively high level of internal validity (though with a lower level of external validity than studies of ‘large n’ regression analysis).

The use of fsQCA is an extension of ‘Qualitative Comparative Analysis’ (QCA). Charles Ragin, the founding father of QCA, summarises QCA as follows:Footnote ⁷⁰ it is an ‘analytic technique that uses Boolean algebra to implement principles of comparison used by scholars engaged in the qualitative study of macro social phenomena’. Its aim is that, ‘by formalizing the logic of qualitative analysis, QCA makes it possible to bring the logic and empirical intensity of qualitative approaches to studies that embrace more than a handful of cases – research situations that normally call for the use of variable-oriented, quantitative methods’. The ultimate goal of this exercise is to show ‘the different combinations of conditions that produce a specific outcome’.

Conventional ‘crisp-set’ QCA codes the conditions in a binary way (‘0 and ‘1). The ‘fuzzy-set’ in fsQCA means that the analysis allows intermediate numbers that range from ‘0’ to ‘1’. Thus, this approach opens this line of research to types of data as collected in the current paper as all but one of the conditions are of a non-binary nature.Footnote ⁷¹ We therefore normalised the data of all possible conditions to the scale of ‘0’ to ‘1’, ie the lowest value was set as ‘0’, the highest as ‘1’ and the others were scaled accordingly in a continuous way, rounded to two decimal places.Footnote ⁷² Subsequently, the dataset was imported into the main software for fsQCA.Footnote ⁷³

(b) Methodological choices

The main outcome analysed in the following is the level of national convergence.Footnote ⁷⁴ In addition, it is possible to set the outcome as ‘negated’.Footnote ⁷⁵ This means that, in the present case, we also examined what factors account for particularly low levels of convergence. The possible causal conditions are the seven conditions discussed in the previous section.

In contrast to regression analysis, fsQCA requires a more active position of the researcher in the analysis,Footnote ⁷⁶ as will be shown in the following. First, the fsQCA program presents the result in a ‘truth table’, which shows the different combinations of the causal conditions with the corresponding consistency thresholds.Footnote ⁷⁷ The default consistency threshold is 0.8, and for a small number of observations and few outcomes with more than one case (as here), it is recommended to choose a threshold of ‘1’.Footnote ⁷⁸ We have no reason to deviate from this recommendation in the present case, but also checked alternative specifications and the results remain unchanged.

Secondly, the next decision to take is whether researchers expect the causal conditions to be ‘present’, ‘absent’, or either ‘present or absent’. To be precise, their information helps the software to distinguish ‘easy’ from ‘difficult’ counterfactuals during logical minimisation. In the dataset of this paper, the explanatory factors were defined in a way that we expect a positive effect for a high level of national convergence (and vice versa for a low level of national convergence). Sometimes the fsQCA program also presents the researcher with a ‘prime implication chart’ indicating model ambiguities. In the present case, any such ambiguities can also be resolved by way of expressing a preference for a positive effect for a high level of national convergence (and vice versa for a low level of national convergence).

Thirdly, the fsQCA results are then presented as ‘complex’, ‘parsimonious’ and ‘intermediate’ solutions. These three solutions indicate a trade-off between coverage and consistency (ie breadth and accuracy of the solutions).Footnote ⁷⁹ The complex solution has the highest consistency but the lowest coverage, while it is the opposite for the parsimonious solution. Some of the literature, discussing fsQCA results, presents all three solutions,Footnote ⁸⁰ but more commonly researchers are selective in the way they report the solutions – ie whether to focus on the complex, parsimonious or intermediate solution, or combinations of those.Footnote ⁸¹ In the present case, we discuss the parsimonious and intermediate solutions, as the complex solution did not reduce the data to an extent that it would facilitate interpretation of the configurational outcomes.

With these methodological explanations and choices, we can now turn to the results.

(c) Result and discussion for high national convergence

Table 5 presents the results for the main question of interest of this paper: what determines high levels of national convergence for the observations of this study? The solution coverage and solution consistency of the intermediate solution have high numbers (above 0.8). As one would expect, the parsimonious solution has lower overall consistency but higher overall coverage. The main focus of the following analysis will be the intermediate solution with some references to the parsimonious one.

Table 5.

Determinants for high national convergence (‘intermediate solution’ with results also in ‘parsimonious solution’ in large signs)Footnote ⁸²

Overall Solution Consistency 0.84 (intermediate), 0.71 (parsimonious)

Overall Solution Coverage 0.93 (intermediate), 0.98 (parsimonious)

Different from the output of a regression analysis, Table 5 does not contain different models but different pathways. Thus, fsQCA considers that different combinations can lead to the same outcome of interest (in the present case: high convergence).Footnote ⁸³ We also established that none of the conditions alone is necessary for the outcome.Footnote ⁸⁴ In the comparison of the different pathways, it can be seen that all of the paths have high consistency and that there is a trade-off with the raw coverage of a path as it decreases from the first to the last pathway.

In an ideal world, an interpretation of fsQCA results would be able to show precisely how each of the pathways accounts for a plausible configuration of conditions. In the present case, we can provide some explanations of the configurations; however, as a caveat, we also need to note that it is not possible to present perfect narratives that would rationalise all of the conditions and pathways (nor can we exclude that, for particular cases, idiosyncratic conditions may not also play a roleFootnote ⁸⁵).

To start with, it is noticeable that EU convergence is often relevant but only in combination with other factors, which confirms the non-empirical finding by Baghi, emphasising how the harmonisation debate must be equally viewed in the light of the social and political aspects of the legal systems which are desired to be converged.Footnote ⁸⁶ On the basis of the factors analysed in this study, it can be said that the favourable attitude of an EU directive towards convergence is mediated through the domestic level. In particular, it can be seen that in two of the three pathways EU convergence is relevant in combination with ‘no existing legislation’. This configuration is plausible, namely that lack of prior domestic legislation followed by a favourable EU directive (ie a directive with a high score for European convergence) leads to a high level of national convergence.

For the two other groups of conditions, the general picture is that ‘national politics’ features more often than ‘country characteristics’, notably also with two of the conditions in the parsimonious solution. The role of national politics is rarely explored in consumer protection literature:Footnote ⁸⁷ thus, this is an important finding, pointing to the need to ‘bring in politics’ to the debate about convergence and consumer sales law. The fact that country characteristics are less relevant can be interpreted in a progressive way, namely that countries are not somehow impaired by circumstances such as performing poorly in indicators of ‘good law’ or low public support for the EU when it comes to the capacity to implement full legal convergence.

Considering the typical cases, ie those with more than 0.5 membership, or even more than 0.7 membership,Footnote ⁸⁸ the analysis of the pathways can be explained as follows. The first pathway includes the largest number of cases, notably four of the respective Romanian, Dutch, Belgian and Irish cases. The two conditions of a favourable attitude of the respective EU directive and public support for the EU complement each other well. It may then also be plausible that the main countries of this pathway are four of the relatively less populous Member States as here there may be greater willingness to respond to such conditions than, say, in France, Germany and the UK. All four cases with more than 0.7 membership included in the first pathway are linked to the Unfair Commercial Practices Directive, the only maximum harmonisation directive in the study.

By contrast, the cases of both the second and the third pathway are predominantly from Germany and the UK. Both share the condition of an effective government, which can also be explained in terms of trust in public institutions. In the second pathway, this is combined with the condition of a good legal system and in the third pathway it is combined with the conditions of EU convergence and no existing legislation. It can therefore be said that both countries show more complex configurations when it comes to the determinants for high national convergence, with the second pathway being somehow more generic and the third pathway linked to the gap a new directive may be able to fill.

The fourth pathway applies to all of the Romanian cases. This is likely to be due to the fact that Romania transposed all directives at around the same time, given that all these instruments pre-date Romania's accession to the EU in 2007. In substance, the conditions can be seen as typical for one of the EU accession countries: there is no experience with EU legislation, no existing legislation on matters covered by EU directives, and public support for the EU is high because of the low level of trust in national institutions, and high level of trust in supranational institutions, combined with no disruption of the law making process by new elections.

The conditions of the fifth and sixth pathways are the most complex ones (and they do not show any cases with more than 0.7 membership), with the former being mainly about EU convergence and country characteristics and the latter being mainly about national politics. As the sixth pathway only includes one case, it is more interesting to consider the fifth pathway more closely. Here, it is noteworthy that four of the five cases concern the Distance Selling Directive. The relevant conditions show that there has been a gap in the respective Member States filled by a favourable EU Directive and supported by left wing politicians in government, presumably since the latter are keen to restrict the power of large multi-national corporations engaged in distance selling.

(d) Result and discussion for low national convergence

The fsQCA results discussed in the following reverse the question, asking which pathways lead to particularly low levels of national convergence. Here too we established that none of the conditions alone is necessary for the outcome.Footnote ⁸⁹

In Table 6, EU convergence is not relevant in any of the pathways. This is an interesting finding, as it shows that it is mainly domestic factors, not the EU approach, that lead to low levels of legal convergence. The condition most relevant – and which is also present in the parsimonious solution – is that there is an interruption by elections (to be precise, not ‘no interruption by elections’). This is plausible, as such an interruption may negatively affect both the timeliness and the substance of an implementation of EU law.

Table 6.

Determinants for low national convergence (‘intermediate solution’ with results also in ‘parsimonious solution’ in large signs)Footnote ⁹⁰

Overall Solution Consistency 0.63 (intermediate), 0.50 (parsimonious)

Overall Solution Coverage 0.70 (intermediate), 0.75 (parsimonious)

Considering the typical cases, ie those with more than 0.5 or even 0.7 membership,Footnote ⁹¹ the analysis of the three pathways can be explained as follows: in the first pathway are three cases from the Netherlands, two from Belgium and two from Ireland. These are countries with often complex coalition governments; thus, it is plausible that here we have situations where both the lack of an effective government and interruption by election lead to low levels of national convergence.

The second pathway refers to situations where, in addition to the interruption by elections, low support for the EU has had a negative impact on national convergence. This includes, perhaps unsurprisingly, two cases from the UK – which are also the ones that even have 0.7 membership – but also three from France, two from Germany and one from Belgium. Those latter cases are primarily about the more recent directives; thus, they may reflect the rise of Eurosceptic populism across the EU, notably in France with the Front National, which also impacts on the mainstream political parties. By contrast, the third pathway mainly concerns the convergence as regards the two older directives; thus, this may reflect that the ‘interference’ of EU law into existing prior legislation was felt to be more severe than in later years.