Europeanization Mechanisms and Process Tracing: A Template
for Empirical Research*
KYRIAKOS MOUMOUTZIS1
and SOTIRIOS ZARTALOUDIS2
1
King’s College London2
University of Birmingham
Abstract
While the literature on Europeanization has exhibited considerable awareness of the methodological
challenge of establishing causal relations between non-binding EU stimuli and domestic
change, less work has been done on how this challenge might be met. This article contributes to
the literature’s attempts to meet this challenge by: 1) reformulating four explanations of Europeanization
based on four distinct causal mechanisms (instrumental learning, social learning, naming
and shaming and peer pressure); 2) specifying their observable implications for three intervening
steps between EU stimuli and change in national policy (the definition of the policy problem, the
alternative courses of action considered and the manner in which they were assessed); 3) defending
process tracing against critiques of its usefulness for research on Europeanization; and 4) providing
practical guidelines on how process tracing can be used to test these four explanations empirically,
using examples from employment policy, where non-binding EU stimuli feature most
prominently.
Keywords: Europeanization; mechanisms; process tracing; causality; employment policy
Introduction
It has become commonplace to state that establishing causality in the literature on Europeanization
is a challenge, especially for researchers who attempt to establish causal relations
between non-binding European Union (EU) stimuli and change in national policy
(cf. Featherstone, 2003; Haverland, 2006; Radaelli, 2012). While a sizeable part of the literature
has exhibited considerable awareness of this challenge, less work has been done
on how it might be met. In the small part of the literature that has attempted to address
this issue, it has become equally commonplace to argue that process tracing might be used
to establish the causal significance of the EU (Moumoutzis, 2011; Radaelli, 2006;
Zartaloudis, 2013). Process tracing is a within-case1
qualitative method that ‘attempts
to identify the intervening … causal mechanism between an independent variable …
and the outcome of the dependent variable’ (George and Bennett, 2005, p. 206). Its attempt
to identify causal mechanisms is typically considered process tracing’s
distinguishing feature (Beach and Pedersen, 2013, p. 1).
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main
Street, Malden, MA 02148, USA
* Acknowledgments: We would like to thank Kevin Featherstone, Claudio Radaelli, Anand Menon, participants in a
UACES Annual Conference panel in Leeds in September 2013 and participants in a departmental research seminar at
KCL in November 2013 for their useful comments on earlier drafts of this article. Funding from the Bodossaki
Foundation is gratefully acknowledged. Any errors remain the authors’ responsibility.
1
As has been pointed out (Bennett and Elman, 2006, p. 462), many small-n studies have been presented as comparisons,
while in fact they use process tracing to make within-case causal inferences.
JCMS 2016 Volume 54. Number 2. pp. 337–352 DOI: 10.1111/jcms.12294
While the use of this method in the literature on Europeanization has been widespread,
it has not been particularly advanced. It has been consistent with what has been
termed ‘detailed narratives’ that make limited use of theory (George and Bennett, 2005,
pp. 210–11). Simply put, researchers frequently merely mention process tracing and
then present a historical narrative of relevant EU- and domestic-level developments.
Regardless of the amount of detail that they include, these narratives often do not
succeed in establishing the causal significance of the EU.
This is regrettable, given the progress that the literature on within-case qualitative
methods has made since the mid-1990s (see Bennett and Elman, 2006 and Mahoney,
2007; for process tracing in particular see Bennett and Checkel, 2015). This shortcoming
is not inherent in process tracing, as this method is consistent with more advanced analyses
that make extensive use of hypotheses and variables. Thus, the challenge for researchers
is to transform these narratives into ‘analytical explanations’ (George and
Bennett, 2005, pp. 211–12). The intervening steps between cause and effect, on which
process tracing focuses, should be ‘theoretically predicted’ (Checkel, 2006, p. 636). Additionally,
an ‘uninterrupted’ causal path between cause and effect is required and ‘it is
not sufficient that a hypothesis be consistent with a statistically significant number of intervening
steps’ (George and Bennett, 2005, p. 207).2
The causal path might be
interrupted if relevant theories make no predictions about a certain intervening step or
an indeterminacy problem might arise if they make similar predictions (George and
Bennett, 2005). As has been pointed out (Rohlfing, 2012, pp. 173–74), process tracing
per se cannot eliminate the potential for such indeterminacy. Consequently, it is necessary
to hypothesize observable implications for the intervening steps between cause and effect
that are as distinctive as possible.
The purpose of this article is to contribute to the literature’s attempts to meet the challenge
of establishing the causal significance of the EU, especially in policy areas where
EU stimuli are not binding.3
It does so by reformulating four explanations of Europeanization
based on four distinct causal mechanisms: 1) instrumental learning; 2) social
learning; 3) naming and shaming; and 4) peer pressure. Such mechanism-based explanations
‘detail the cogs and wheels of the causal process through which the outcome to be
explained was brought about’ (Hedström and Ylikoski, 2010, p. 50). The analysis here
also specifies the observable implications of the aforementioned mechanism-based explanations
for three intervening steps between cause and effect: 1) the definition of the policy
problem; 2) the alternative courses of action considered; and 3) the manner in which the
2
For example, while evidence that non-binding EU stimuli have caused changes in the definition of policy problems has
been reported, there is not as much evidence that these changes have been followed by changes in the alternative courses
of action that national policy makers considered or changes in the way in which these alternatives were assessed before
the decision to change policy was made. Consequently, the evidence that has been reported has been insufficient to establish
an uninterrupted causal path between cause and effect (cf. Zeitlin, 2005). For these intervening steps, see below.
3
We refer to these EU stimuli as ‘non-binding’ in the sense that they are not justiciable. We prefer to refrain from using the
term ‘soft’ EU law because of the problems related to the distinction between ‘hard’ and ‘soft’ law that the literature has
identified (see Trubek and Trubek, 2005). Additionally, the terms that we use here have clearer implications for the causal
mechanisms through which Europeanization might produce policy change. As the Commission cannot bring charges of
non-compliance with non-justiciable EU stimuli before the European Court of Justice (ECJ), the ECJ cannot enforce compliance
with such stimuli by imposing sanctions. Finally, it should be noted that the analysis presented here also applies to
policy areas where binding EU stimuli are available, yet unlikely to be adopted due to divergent national preferences (see
Radaelli, 2008, p. 239).
Kyriakos Moumoutzis and Sotirios Zartaloudis338
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd
latter were assessed. The analysis of these observable implications renders these explanations
amenable to an advanced application of process tracing.
The first part of the article presents the aforementioned mechanism-based explanations.
These explanations depart from conventional analyses of the causal mechanisms
on which they are based. While the literature often conflates various mechanisms into
one, this article argues that it is problematic to equate ‘naming and shaming’ with ‘peer
pressure’ and ‘social learning’ with ‘socialization’. The systematic distinction between
different causal mechanisms allows us to formulate explanations of Europeanization that
are theoretically more coherent and easier to test empirically.
The second part of the article defends process tracing against critiques of its usefulness
for research on Europeanization. It is argued that those who have advocated random selection
on the independent variable as the most effective way of establishing causality
have disregarded the fact that cases of Europeanization are cases of equifinality, to which
the use of process tracing is better suited. It is also argued that the generalizability of findings
depends on case selection rather than the use of process tracing.
The third part of this article demonstrates how process tracing might be applied empirically
using examples from employment policy, where the use of non-binding EU stimuli
has been extensive (see Heidenreich and Bischoff, 2008; Van Rie and Marx, 2012;
Zartaloudis, 2013). While this part of the article does not present findings of an empirical
case study, it does provide detailed practical guidelines on the use of process tracing.
I Four Explanations of Europeanization and their Observable Implications
Several mechanisms of Europeanization in policy areas with non-binding EU stimuli have
been identified in the literature (cf. Trubek and Trubek, 2005). Nevertheless, none has
been specified with a degree of precision that would render them empirically testable
through process tracing. As has been recently pointed out (Bennett and Checkel, 2015,
p. 30), before employing process tracing it is necessary to specify as precisely as possible
what one should observe within a case in order to empirically substantiate each hypothesized
explanation. This section serves this purpose by presenting four explanations of
Europeanization based on four distinct causal mechanisms (instrumental learning, social
learning, naming and shaming and peer pressure) and specifying their observable impli-
cations.4
Mechanisms are conceptualized here as ‘entities’ and the ‘activities’ in which
they engage and which produce particular types of change (Hedström, 2005, p. 25;
Machamer et al., 2000, p. 3). As has often been the case (Hedström, 2005, p. 2), the ‘entities’
and ‘activities’ on which the analysis focuses are actors and their actions. In each of
the four explanations presented here, the analysis of different actions exhibits a ‘productive
continuity’ (Machamer et al., 2000, p. 3) between the same cause (non-binding EU
stimuli) and the same effect (change in national policy).5
As has been pointed out (Hall, 2008, p. 27), it is particularly important to derive ‘predictions
that are consistent with one theory but inconsistent with its principal rivals’. In
4
It should be noted that further operationalization of the explanations presented here is likely to be necessary once a specific
case has been selected.
5
For a framework that considers non-binding EU stimuli’s effects other than change in national policy see Zeitlin et al.
(2014).
Europeanization mechanisms and process tracing 339
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd
this vein, the analysis here increases the level of ‘theoretical uniqueness’ of each explanation’s
observable implications (the extent to which they can be derived from only one of
these four explanations in principle) and the extent to which they are mutually exclusive.6
At least one of the observable implications of each explanation is both theoretically
unique and incompatible with the observable implications of the other explanations for
the same intervening step between cause and effect. Finally, this section argues that discourse
(in the sense of a process that serves to disseminate policy ideas) is more useful for
explaining the Europeanization of domestic politics rather than national policies.
Instrumental and Social Learning
Policy learning has been the most popular explanation of policy change as the outcome of
Europeanization (Borrás and Radaelli, 2010; Kröger, 2009; Zeitlin, 2005). Policy learning
is understood here as ‘relatively enduring changes in thought or behavioural intention that
result from experience and/or new information concerned with the attainment or revision
of policy objectives’ (Heclo, 1974, p. 306).7
While various types of learning have been
discussed in the literature, we argue that the distinction between ‘instrumental’ and ‘social
learning’ is more useful for empirical research because it produces observable implications
that allow researchers to distinguish between them empirically. As was mentioned
above, the extent to which causal arguments make distinctive predictions about intervening
steps between cause and effect should be a guiding principle in the process of refining
causal mechanisms.
In this vein, during EU-level interactions, national policy-makers acquire new information
drawn from the experience of other EU Member States, assess it and, if they calculate
that practices originally pursued elsewhere might achieve their policy goals more effectively,
voluntarily incorporate them into their policy. This type of learning has been referred
to as ‘instrumental’, ‘single-loop’ or ‘adaptive’ learning and it results in a change
in policy means (Argyris and Shön, 1978, pp. 20–26; May, 1992, p. 335; Visser, 2005,
p. 180). In contrast, ‘social’, ‘double-loop’ or ‘reflexive’ learning results in a change in
policy ends, as the information that national policy-makers acquire and assess is not
merely related to policy instruments, but also policy objectives.8
The causal significance
of the EU in this case lies in the fact that it is the channel through which new information
6
For the term ‘theoretical uniqueness’ see Rohlfing (2014, p. 614), who has refined Van Evera’s (1997, p. 31) conception
of ‘uniqueness’. It should be noted here that a complete discussion of observable implications would also require an assessment
of their ‘empirical uniqueness’ (the extent to which the observable implications of one or more than one of these four
explanations can be tested in a specific case) and their ‘certainty’ (the probability that they will be confirmed). As Rohlfing
(2014, pp. 613–14) has pointed out, this can only be determined in relation to a particular case and therefore it is not possible
within the confines of this article.
7
See also Sabatier (1987, p. 654). Hall has defined ’social learning’ as ’a deliberate attempt to adjust the goals or techniques
of policy in response to past experience or new information’ (1993, p. 278). In this sense, ‘learning is indicated when policy
changes as the result of such a process’. As Bennett and Howlett (1992, p. 285) have pointed out, this assumption is fairly
common amongst explanations of policy change based on learning. Surely, however, it is conceivable that certain policy
makers might learn and intend or even attempt to pursue policy change, yet they might fail to convince or impose the lessons
they have learnt on their colleagues. Furthermore, if these policy makers have complete information at the early stages
of the policy-making process, they are likely to anticipate this outcome and therefore they may not attempt to pursue policy
change. Consequently, Hall’s definition is too restrictive as it excludes cases of learning, where the latter is not followed by
(attempts to pursue) policy change.
8
The outcomes of instrumental learning are consistent with what Hall (1993) refers to as first- and second-order change,
while the outcomes of social learning are consistent with third-order change.
Kyriakos Moumoutzis and Sotirios Zartaloudis340
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd
becomes available to national policy-makers, which in turn modifies national policymakers’
conception of the policy problem (social learning) or the range of alternatives
available to them (instrumental learning).
Instrumental learning is a common element of rationalist explanations of policy change
(Dobbin et al., 2007, p. 463; James and Lodge, 2003, p. 181; Risse et al., 2001, p. 12).
According to the latter, (national) policy-makers use ‘at some later point’ the information
that they have acquired during (EU-level) interactions ‘to alter strategies, but not preferences,
which are given’ (Checkel, 2001, p. 561). Such explanations predict that national
policy-makers calculate the costs and benefits of EU-recommended courses of action in
terms of policy objectives, not the objective of re-election. In the case of social learning,
these cost–benefit calculations are made in terms of policy objectives, which have previously
been redefined upon reflection on information relating to the nature of the policy
problem.
Naming and Shaming
‘Naming and shaming’ and ‘peer pressure’ have also featured prominently in the literature.
The precise relationship between these two mechanisms remains unclear. The literature
uses them practically interchangeably to refer to the negative implications for
national policy-makers whose country is under-performing with regard to indicators
jointly agreed upon at the EU level (see for example Hodson, 2004; Zeitlin, 2005). National
policy-makers, it is argued, pursue policy change in response to non-binding EU
stimuli in order to avoid the ‘informal sanction’ of naming and shaming or the pressure
of their peers (Heidenreich and Bischoff, 2008, p. 504; Trubek and Trubek, 2005, p. 357).
This article departs from conventional analyses of naming and shaming and peer pressure
and argues that they constitute distinct causal mechanisms. Naming and shaming can
form the basis of a rationalist explanation of policy change, whereas peer pressure is only
consistent with a constructivist approach. The key distinguishing feature of naming and
shaming is the publicity that its effective functioning requires.9
In this sense, evaluations
of national performance that are made public are expected to embarrass national governments
who are found to have performed badly and result in reputational losses, as they are
available for opposition parties, interest groups and the media to use in their discourse
(Büchs, 2008; De Ruiter, 2013; Hodson, 2004; Zeitlin, 2005).10
Such reputational losses
9
It appears as though the conflation of these two mechanisms originated in Hodson’s analysis, which remains the most detailed.
Strikingly, Hodson drew on Olson’s analysis of social pressure, which has made explicit that two distinct ‘kind(s) of
social pressure may occasionally be operative’ (Olson, 1971, p. 63, fn 18). One type of social pressure is generated ‘through
person-to-person friendships’, the other ‘through mass media’ (Olson, 1971, p. 63, fn 18). We believe that our distinction
between an explanation based on naming and shaming, which occurs in public, and an explanation based on peer pressure,
which occurs in private, is more consistent with Olson’s analysis.
10
Empirical evidence shows that the formulation of country-specific recommendations that follows the evaluation of national
performance is ‘a politically negotiated process … between the Commission and the Member States’ (Copeland
and Ter Haar, 2013, p. 31). The Member States try to influence the wording of the Commission’s recommendations or
to have (parts of) recommendations removed and occasionally they succeed (Copeland and Ter Haar, 2013, p. 32; Mailand,
2008, pp. 356, 358–9). It does not follow, however, that ‘recommendations are therefore quasi-extensions of the Member
States acting in areas they themselves identify as a priority’ (Copeland and Ter Haar, 2013, p. 32). If the Member States
could exercise such control over the recommendations, they would be fairly responsive to them. The evidence, however,
shows that responsiveness to recommendations is variable, yet overall rather limited (see Copeland and Ter Haar, 2013,
pp. 30–1). Consequently, while the Member States regularly try to and occasionally succeed in influencing the Commission’s
recommendations, the latter remain beyond their full control and are more than a mere reflection of national priorities.
Europeanization mechanisms and process tracing 341
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd
may in turn result in electoral losses. This causal argument is based on the premise that
the electorate is both informed about these evaluations and willing to electorally punish
a government under which their country has performed badly.11
While the ‘cost/benefit
choice mechanism’ (Checkel, 2001, p. 556) is the same as in the explanation based on instrumental
learning, the type of cost–benefit calculations is different. Consequently, researchers
can empirically test this mechanism by investigating the type of cost–benefit
calculations that national policy-makers made before pursuing policy change. Establishing
that the EU was causally significant via naming and shaming requires evidence that
national policy-makers calculated the electoral costs of the continuation of the policy previously
pursued.12
Peer Pressure
In contrast to naming and shaming, peer pressure does not require publicity, but instead
the ‘political determination of (EU) member-states to use (it) … against recalcitrant countries’
(Radaelli, 2000, p. 34). National policy-makers must be willing to highlight discrepancies
between the behavioural norms of their community (the EU) and the behaviour of
their peers, and they must be willing to exert pressure on their peers to change their behaviour.
Furthermore, national policy-makers ‘must consider a peer rebuke … to be
costly’ (Hodson, 2004, p. 240). It remains unclear, however, why they might consider
such pressure costly and yield to it. It has been suggested that socialization ‘would certainly
enhance the impact of peer pressure on domestic policy choices’ (Radaelli,
2003b, p. 54, fn 20). In fact, a recent review of the relevant literature concluded that ‘socialization
for governmental actors’ has been the main effect of non-justiciable EU stimuli
on welfare states, even though ‘they are the ones that are least willing to use it to reflect
upon how to alter problem and/or solution framing in the process of ongoing welfare state
reform’ (De la Porte and Pochet, 2012, p. 343). In this case, the term ‘socialization’ has
been used rather lightly. If national policy-makers are not even willing to consider alternative
definitions of policy problems or policy solutions presented during EU-level interactions,
they have hardly been socialized. Socialization is ‘a process of inducting actors
into the norms and rules of a given community’ (Checkel, 2005, p. 804). The nature of
this process is highly transformative. According to constructivist approaches, with which
this mechanism is consistent (Checkel, 2001, p. 557), once these norms and rules have
been internalized, actors are driven by a logic of appropriateness, namely by considerations
of what constitutes normal, right or good behaviour within the context of a given
community (March and Olsen, 2004). In the case of the EU, socialized national policy-
11
Although Zeitlin (2005) does not specify the reasons why a national government would prefer to avoid such ‘embarrassment’,
Hodson (2004, p. 239) and Büchs (2008, p. 24) refer to naming and shaming’s negative implications for ‘a government’s
credibility in the eyes of the electorate’ and ‘election turnout’ respectively. We argue that these implications are
unlikely to provide sufficient incentive for policy change because they fall short of electoral punishment. The scope condition
for the effective operation of this particular mechanism should be more restrictive: voters should be willing to electorally
punish their government for bad performance.
12
Naming and shaming does not make information available only to domestic actors, but also to financial markets (see
Hodson, 2004, p. 239). In cases where this information has implications for an EU Member State’s creditworthiness, it
might be the response of the financial markets that causes change in national policy. Such a causal relation will also be
reflected in national policy-makers’ cost-benefit calculations that precede policy change, as it will increase the financial cost
of the policy previously pursued.
Kyriakos Moumoutzis and Sotirios Zartaloudis342
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd
makers under peer pressure become convinced13
that it is appropriate for them to comply
with EU recommendations that would enable their country to achieve EU goals. Under
such conditions, national policy-makers should be expected to yield to the pressure of
their peers.14
Consequently, establishing that the EU was causally significant via peer
pressure requires evidence that national policy-makers did not make cost–benefit calculations,
but attempted to identify the appropriate course of action for a member of their
community (the EU).
The precise relationship between socialization and social learning has thus far
remained unclear, presumably because both explanations predict change in policy goals.
In certain formulations, socialization and social learning do not appear to constitute separate
mechanisms.15
This article argues that their usefulness will be maximized if they
remain distinct. Learning by definition requires the provision and assessment of new information.
Initially, socialization will indeed entail learning, as new members of a community
will first be required to learn what the community’s norms and rules are.
Subsequently, however, socialization will require normative arguments, which will demonstrate
discrepancies between national policy and EU rules and which will not necessarily
be based on new information.
Discourse
Discourse has also been considered a mechanism through which Europeanization produces
domestic change. As will be shown below, while there are analyses of discourse
which are useful for research on Europeanization, these analyses tell us more about the
Europeanization of domestic politics rather than national policies. It has been argued
(Trubek and Trubek, 2005, p. 92) that ‘discursive transformations’ may explain policy
change, as national policy-makers that make use of concepts introduced at the EU level
‘tend to shift policy orientation’. Others have implicitly suggested fairly specific causal
and temporal sequences. In these analyses, discourse is understood as ‘both the policy
ideas that speak to the soundness and appropriateness of policy programmes and the interactive
processes of policy formulation and communication that serve to generate and
disseminate those policy ideas’ (Schmidt and Radaelli, 2004, p. 193). Moreover, discourse
has been identified both as a dependent variable that Europeanization might affect
and as an intervening variable that affects the likelihood of the Europeanization of national
policies (see Radaelli, 2003a, pp. 36, 49). In this sense, discourse as a set of ideas
is initially Europeanized, and subsequently discourse as a process that serves to disseminate
these ideas facilitates the Europeanization of national policies. This article argues
13
Checkel (2001) uses the term ‘social learning’ to refer to this mechanism that entails ‘persuasion’ and ‘convincing someone
through argument’. As was mentioned above, we follow May (1992) and retain the use of the term ‘social learning’ for
cases, where actors reflect on information relating to the policy problem itself.
14
There might be a set of cases in which national policy-makers should be expected to yield to the pressure of their peers
even in the absence of socialization. It is plausible to argue that this might be the case in Member States, where a national
‘compliance culture’ prevails and compliance per se takes precedence over other goals. According to the worlds-ofcompliance
typology (see Falkner et al., 2007, p. 405), this is the case in Denmark, Finland and Sweden, which belong
to the ‘world of law observance’. Nevertheless, it should also be noted that the empirical record of this approach is at best
mixed. See among others Thomson (2009, p. 14), who used ‘exactly the same data’ as Falkner et al. (2007), but reached
‘starkly different conclusions’.
15
Radaelli (2008, p. 240), for example, has argued that Europeanization may entail ‘learning by socialisation’, which ‘make
(s) policy makers more aware of their interdependence and can inspire more commitment towards EU-level goals’.
Europeanization mechanisms and process tracing 343
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd
that the first part of this sequence cannot be analytically separated from learning. If discourse
as a set of ideas is related to policy instruments, the Europeanization of discourse
will be the result of instrumental learning. If it is related to policy objectives, it will be the
result of social learning.
The second part of this sequence appears more promising. Information acquired during
EU-level interactions might be used as a political resource to construct a discourse – in the
sense of a process that serves to disseminate policy ideas – that might serve various purposes:
to present a policy problem as a legitimate addition to the agenda or prioritize a
problem that is already on the agenda, to serve to check domestic pressure for a policy
solution by providing one in a time-efficient fashion or to serve to legitimate a policy solution
that has already been selected (Bennett, 1991; Graziano et al., 2011; LópezSantana,
2006). On the one hand, ‘co-ordinative discourse’16
(the use of ideas by policy
actors in their attempts to co-ordinate agreement on policies) results in change in the preferences
of (some of) the actors that participate in the policy-making process. On the other,
‘communicative discourse’ (the use of ideas by political actors in their attempts to communicate
to the public the policy programme agreed upon in the context of co-ordinative
discourse) results in change in mass public opinion. As Europeanized ideas become
shared by a larger number of domestic actors and/or the public, reform capacity is enhanced,
thus rendering policy change more likely (Schmidt, 2002, p. 900). In such cases
of usage of EU resources in domestic discourse, no change in the definition of the policy
problem or the alternative courses of action considered should be observable. In fact, this
process is more accurately conceptualized as the Europeanization of domestic politics
rather than national policies.
II In Defence of Process Tracing for Research on Europeanization
Process tracing is not without its critics in the literature on Europeanization. This
section draws on the broader literature on process tracing in order to address these
critiques and defend the usefulness of process tracing for establishing causality in research
on Europeanization. According to Haverland (2006, pp. 137–8), process tracing
is unlikely to allow researchers to establish the causal significance of the EU,
regardless of whether the observable implications of alternative explanations that
do not attribute causal significance to the EU are also process-traced. In this vein,
the problem lies in developments, such as globalization, which occurred in parallel with
European integration, and produce similar effects (Haverland, 2006, p. 137). In contrast
to most of the literature on Europeanization, Haverland (2006, pp. 135–6, 139) has
suggested that establishing the causal significance of the EU requires a comparison
between EU Member States and non-members in order to introduce variation in the
independent variable.
This critique disregards crucial differences between qualitative and quantitative research.
More precisely, Haverland has suggested research strategies that are better suited
to quantitative research for a literature that is predominantly qualitative. Indeed, random
selection on the independent variable is one of the distinguishing features of quantitative
16
For the distinction between ‘co-ordinative’ and ‘communicative’ discourse see Schmidt and Radaelli (2004).
Kyriakos Moumoutzis and Sotirios Zartaloudis344
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd
research (Mahoney and Goertz, 2006, pp. 239–41). Qualitative researchers, however,
typically select positive cases on the dependent variable (Mahoney and Goertz, 2006,
pp. 239–41).
Furthermore, Haverland’s critique disregards the notion of equifinality – the idea that
there are multiple causal paths to the same outcome – which has been identified as an integral
part of the understanding of causality in qualitative research (Mahoney and Goertz,
2006, pp. 236–7).17
It should be made explicit that distinguishing the causal significance
of the EU from that of other factors does not require researchers to establish that the substantive
content of EU stimuli is unique. Indeed, as the use of non-binding stimuli is often
intended to spread best practice, the EU by definition promotes policies which are already
being pursued by certain national governments. Similarly, national governments in powerful
bargaining positions might determine the outcome of EU policy-making on specific
policy issues and impose national policy as EU policy, in which case the EU will again
promote policies that are already being pursued by certain national governments. Finally,
the EU might make policy recommendations that are similar to those made by international
organizations, such as the Organisation for Economic Cooperation and Development
or the International Monetary Fund, which might also result in domestic change
independently of the EU.
It follows from the above that the outcomes of Europeanization may also be produced
by other factors either at the domestic or the international level, and therefore cases of
Europeanization are cases of equifinality. If the substantive content of EU stimuli and
the outcomes of Europeanization were unique, one would merely need to observe the outcomes
that are unique to this process in order to safely conclude that this process has occurred.
The existence of alternative sources of policy ideas and practices suggests that
research questions related to Europeanization are indeed puzzling and worth exploring.
The fact that cases of Europeanization are cases of equifinality suggests that ‘the statement
that the EU causes a particular outcome’ does not ‘(imply) that if the EU were to
absent (sic) that particular outcome would not occur’ (Haverland, 2006, p. 135), precisely
because other developments might also produce it. Similarly, equifinality renders evidence
of similar domestic changes in both EU Member States and non-members insufficient
to establish that the EU is not causally significant for changes in its Member States.
Domestic change in non-members may indeed be the effect of causes other than the EU,
such as globalization or a newly elected government’s reform agenda. Such evidence,
however, does not suffice to establish that similar domestic changes have been produced
through the same causal path in all cases.
Given that cases of Europeanization are cases of equifinality, within-case methods,
which establish causality by determining the causal mechanism through which an outcome
has been produced (Bennett and Elman, 2006, p. 459), are particularly important
for empirical research on Europeanization. As has been pointed out, given that such
methods do not establish causality through comparison, they are ‘not susceptible to selection
bias’ (Bennett and Elman, 2006, p. 461). Thus, the literature’s emphasis on process
tracing is justified because the latter is precisely such a method that makes it possible to
17
Strikingly, Haverland’s discussion of globalization implicitly acknowledges that cases of Europeanization are cases of
equifinality, yet his definition of causality does not.
Europeanization mechanisms and process tracing 345
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd
establish whether the EU has been causally significant.18
It also makes it possible to determine
the mechanism through which Europeanization has produced policy change, as
the intervening steps between cause and effect, on which process tracing focuses, provide
‘diagnostic pieces of evidence’ that can (dis)confirm alternative explanations (Bennett,
2010).
It has also been argued – ironically, by process tracers themselves (Checkel,
2006, p. 367)19
– that the findings of research based on within-case qualitative
methods such as process tracing are problematic because they are not generalizable.
On the one hand, the scope for generalizability is indeed typically narrower in qualitative
research, because its complex conception of causality directs it towards different
research goals (Mahoney and Goertz, 2006, p. 237–8). On the other, this critique
should not be overstated. Generalizability does not depend on the method per se, but
on a different element of research design: case selection. For example, if process
tracing produced no evidence of Europeanization in a case where all or at least most
of the conditions that render Europeanization likely were present – that is, in a most
likely critical case of Europeanization – it would be safe to conclude that causal arguments
regarding Europeanization have been substantially weakened and should be
revised or abandoned. In other words, methodologically sound case selection ensures
the generalizability of findings produced through process tracing, not process tracing
per se.
III How to Process Trace the Observable Implications of the Four Explanations of
Europeanization
Despite the usefulness of process tracing for establishing causality in research on
Europeanization, little work has been done on the precise intervening steps on which
it should focus. Following Moumoutzis (2011), we argue that process tracing should
focus on the following dimensions of the policy-making process: 1) the definition of
the policy problem; 2) the alternative courses of action considered; and 3) the manner
in which the latter were assessed. These intervening steps have been selected because
they are the ones for which the observable implications of the four
explanations presented above are most easily distinguishable.20
As was mentioned
above, the preceding analysis increased the level of ‘theoretical uniqueness’ of at
least one of the observable implications of each explanation for these intervening
steps and the extent to which they are mutually exclusive, thus facilitating empirical
testing.
18
This is not to suggest that comparison is not useful, but only that Haverland argues that it can resolve a problem (selection
bias) from which process tracing does not suffer.
19
It should be noted that Checkel appears to have modified his position on this issue in his more recent work – see Bennett
and Checkel, 2015: 13–14, 29.
20
While we only have enough space here to cover explanations that attribute causal significance to the EU, researchers
should also identify the observable implications of explanations that do not attribute causal significance to the EU for the
same intervening steps. On selecting sources, generating data relevant to these intervening steps from them and drawing
inferences from this data see among others Bennett and Checkel (2015, pp. 27–9, 31–2) and Rohlfing (2012, especially
chapters 7 and 8).
Kyriakos Moumoutzis and Sotirios Zartaloudis346
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd
As a first step, empirical research should establish how national policy-makers defined
the policy problem that they intended to address.21
Evidence of this dimension of the
policy-making process has diagnostic value, as different explanations have different observable
implications for this dimension. Explanations based on peer pressure and social
learning predict changes in national policy-makers’ definitions of the policy problem and
their policy objectives, while explanations based on instrumental learning and naming
and shaming do not.
Consider employment policy. Both the Commission and the Council have identified
problems which national employment policies should address, such as low female employment
rates or labour market segmentation (Commission, 2010, 2013; Council of
the European Union, 2010). It can be hypothesized that if the EU can effectively induce
change in its Member States’ employment policies, national policy-makers will begin to
identify low female employment rates and labour market segmentation as problems that
their policy should address. While explanations based on instrumental learning or naming
and shaming cannot explain such changes because they predict no variation in the definition
of the policy problem, the explanation based on social learning predicts that national
policy-makers will begin to consider low female employment rates and labour market
segmentation as employment policy problems after collecting relevant information during
EU-level interactions. Alternatively, the explanation based on peer pressure predicts that
during EU-level interactions, national policy-makers will become convinced that continuing
to disregard low female employment rates and labour market segmentation as employment
policy problems would undermine the EU’s ability to achieve its goals. If empirical
research shows no such changes in the definition of the policy problem, their absence
should be considered disconfirming evidence for both the explanation based on social
learning and the explanation based on peer pressure.
Subsequently, researchers should identify the various alternative courses of action that
national policy-makers considered. While evidence of this particular dimension of the
policy-making process does not allow researchers to distinguish between alternative
mechanisms of Europeanization, it has diagnostic value for the distinction between explanations
that attribute causal significance to the EU and explanations that attribute causal
significance to other variables. To continue with the example of employment policy,
the Commission and the Council have specified courses of action that might address
the problems of low female employment rates and labour market segmentation. These include
a particular conception of ‘flexicurity’, which requires changes in employment protection
legislation that address discrepancies in employment protection for different types
of contracts, lifelong learning policies that upgrade the skills of parents who return to
work after taking care of family dependents and changes in social security systems that
provide affordable access to childcare facilities (Commission, 2010, 2013; Council of
the European Union, 2010). It can be hypothesized that if the EU can effectively induce
change in its Member States’ employment policies, national policy-makers will begin to
consider these EU-recommended courses of action as alternatives to national policy. If
21
It should be noted here that there might be cases where one or more policy-makers made calculations different from those
of the majority of their colleagues. Empirical testing of the explanations presented above requires empirical evidence of the
calculations that dominated the policy-making process in each intervening step between cause and effect. See also footnote
7 on this point.
Europeanization mechanisms and process tracing 347
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd
empirical research shows that national policy-makers continued to disregard both the policy
problems and the courses of action identified at the EU level, their disregard should be
considered disconfirming evidence for all four explanations presented here.
As a final step, empirical research should establish whether national policy-makers calculated
the costs and benefits of the alternatives that they considered and, if so, the type of
costs and benefits that they calculated. Evidence of this dimension of the policy-making
process also has diagnostic value. The explanation based on naming and shaming predicts
that national policy-makers calculate the costs and benefits of EU-recommended courses
of action for their re-election, while the explanations based on learning predict that they
calculate their costs and benefits for policy objectives. The explanation based on peer pressure
predicts that national policy-makers do not make cost–benefit calculations and that
they take EU-recommended courses of action because they become convinced that it is
the right thing to do for officials of an EU Member State. If empirical research shows that
policy change was the result of cost–benefit calculations (regardless of type), these calculations
should be considered disconfirming evidence for the explanation based on peer
pressure. If empirical research shows that national policy-makers selected an EUrecommended
course of action, such as improving access to childcare services because
they calculated that it would facilitate the participation of women in the labour market or
introducing more open-ended contracts for labour market outsiders because it would reduce
labour market segmentation, these calculations should be considered disconfirming
evidence for the explanation based on naming and shaming. In contrast, if empirical
research shows that national policy-makers selected such EU-recommended courses of
action because they calculated that if they had not, they would have endangered their reelection,
these calculations should be considered disconfirming evidence for the explanations
based on learning.22
Conclusions
In contrast to frequently under-specified analyses of mechanisms of Europeanization in
the literature, this article has made the explanatory logic underlying each mechanism explicit.
The guiding principle in this process of refining these causal mechanisms has been
the need for distinctive observable implications for the intervening steps between EU
stimuli and change in national policy, on which process tracing focuses. The result of this
process of refinement is that none of the four explanations presented here are observationally
equivalent. They hypothesize causal pathways that differ from one another with regard
to at least one intervening step between EU stimuli and domestic change. In
particular, the explanation based on social learning and the explanation based on peer
pressure predict change in the definition of the policy problem, while the explanation
based on instrumental learning and the explanation based on naming and shaming do
22
It should be noted here that EU Member States that are subject to the Excessive Deficit Procedure are required to submit
an economic partnership programme which describes the necessary structural reforms in a manner that develops their national
reform programmes and is consistent with Europe 2020 and the guidelines for the employment policies of EU Member
States (see Regulation (EU) No 473/2013). It is therefore conceivable that sanctions might be imposed on euro-area
member states that fail to comply with decisions, which are made within the context of an excessive deficit procedure
and are related to employment policy. If change in national employment policy is the effect of such sanctions, this will
be reflected in national policy-makers’ cost–benefit calculations that precede policy change, as such sanctions will increase
the financial cost of the policy previously pursued.
Kyriakos Moumoutzis and Sotirios Zartaloudis348
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd
not. The explanation based on social learning predicts that change in the definition of the
policy problem is the result of new information, whereas the explanation based on peer
pressure predicts that it is the result of normative arguments. The explanation based on
instrumental learning predicts that national policy-makers make cost–benefit calculations
in terms of policy objectives, while the explanation based on naming and shaming predicts
that they make cost–benefit calculations in terms of the objective of re-election. In
contrast, the explanation based on peer pressure predicts that national policy-makers do
not make cost–benefit calculations, but attempt to identify the appropriate course of action
for a member of their community (the EU) instead.
This article has provided detailed guidelines on how process tracing can be used to test
the four explanations presented here empirically, using examples from employment policy,
where the use of non-binding EU stimuli has been widespread. Process tracing is well
suited to establishing causality in cases of equifinality and the use of process tracing demonstrated
here is uncommonly advanced. As both the intervening steps between cause and
effect and the observable implications of the relevant mechanisms for these intervening
steps have now been specified, future research can move from ‘historical narratives’ to
‘analytical explanations’ of policy change as the outcome of Europeanization. The analysis
presented here can serve as a template for research on the Europeanization of policy
areas other than employment, especially where EU stimuli are not binding (e.g. foreign
policy or social inclusion).
Correspondence:
Sotirios Zartaloudis
Department of Political Science and International Studies (POLSIS),
University of Birmingham, B15 2TT, UK
email: s.zartaloudis@bham.ac.uk
References
Argyris, C. and Shön, D.A. (1978) Organizational Learning: A Theory of Action Perspective
(Reading, MA: Addison-Wesley).
Beach, D. and Pedersen, R.B. (2013) Process-Tracing Methods: Foundations and Guidelines
(Ann Arbor: University of Michigan Press).
Bennett, C.J. (1991) ‘How States Utilize Foreign Evidence’. Journal of Public Policy, Vol. 11,
No. 1, pp. 31–54.
Bennett, A. (2010) ‘Process Tracing and Causal Inference’. In Brady H.E. and Collier D. (eds) Rethinking
Social Inquiry: Diverse Tools, Shared Standards (Lanham: Rowman and Littlefield
Publishers).
Bennett, A. and Checkel, J.T. (2015) ‘Process Tracing: From Philosophical Routes to Best Practice’.
In Bennett, A. and Checkel, J.T. (eds) Process Tracing: From Metaphor to Analytic
Tool (Cambridge: Cambridge University Press).
Bennett, A. and Elman, C. (2006) ‘Complex Causal Relations and Case Study Methods: The Example
of Path Dependence’. Political Analysis, Vol. 14, No. 3, pp. 250–67.
Bennett, C.J. and Howlett, M. (1992) ‘The Lessons of Learning: Reconciling Theories of Policy
Learning and Policy Change’. Policy Sciences, Vol. 25, No. 3, pp. 275–94.
Borrás, S. and Radaelli, C.M. (2010) ‘Recalibrating the Open Method of Coordination: Towards
Diverse and More Effective Usages’. Swedish Institute for European Policy
Studies, Report No. 7.
Europeanization mechanisms and process tracing 349
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd
Büchs, M. (2008) ‘The Open Method of Coordination as a “Two-level Game”’. Policy & Politics,
Vol. 36, No. 1, pp. 21–37.
Checkel, J.T. (2001) ‘Why Comply? Social Learning and European Identity Change’. International
Organization, Vol. 55, No. 3, pp. 553–88.
Checkel, J.T. (2005) ‘International Institutions and Socialization in Europe: Introduction and
Framework’. International Organization, Vol. 59, No. 4, pp. 801–26.
Checkel, J.T. (2006) ‘Tracing Causal Mechanisms’. International Studies Review, Vol. 8, No. 2,
pp. 362–70.
Copeland, P. and Ter Haar, B. (2013) ‘A Toothless Bite? The Effectiveness of the European Employment
Strategy as a Governance Tool’. Journal of European Social Policy, Vol. 23, No. 1,
pp. 21–36.
Council of the European Union (2010) Council Decision of 21 October 2010 on guidelines for the
employment policies of the Member States, 2010/707/EU.
De La Porte, C. and Pochet, P. (2012) ‘Why and How (Still) Study the Open Method of Coordination
(OMC)’. Journal of European Social Policy, Vol. 22, No. 3, pp. 336–49.
De Ruiter, R. (2013) ‘Full Disclosure? The Open Method of Coordination, Parliamentary Debates
and Media Coverage’. European Union Politics, Vol. 14, No. 1, pp. 95–114.
Dobbin, F., Simmons, B. and Garrett, J. (2007) ‘The Global Diffusion of Public Policies’. Annual
Review of Sociology, Vol. 33, pp. 449–72.
European Commission (2010) An agenda for new skills and jobs: a European contribution towards
full employment, COM(2010) 682 final.
European Commission (2013) Annual growth survey 2014, COM(2013) 800 final.
Falkner, F., Hartlapp, M. and Treib, O. (2007) ‘Worlds of Compliance: Why Leading Approaches
to European Union Implementation Are Only “Sometimes-True Theories”’. European Journal
of Political Research, Vol. 46, No. 3, pp. 395–416.
Featherstone, K. (2003) ‘Introduction: In the Name of Europe’. In Featherstone, K. and Radaelli,
C.M. (eds) The Politics of Europeanization (Oxford: Oxford University Press).
George, A.L. and Bennett, A. (2005) Case Studies and Theory Development in the Social Sciences
(Cambridge, MA: MIT Press).
Graziano, P., Jacquot, S. and Palier, B. (eds) (2011) The EU and the Domestic Politics of Welfare
State Reforms (Basingstoke: Palgrave Macmillan).
Hall, P.A. (1993) ‘Policy Paradigms, Social Learning and the State: The Case of Economic
Policymaking in Britain’. Comparative Politics, Vol. 25, No. 3, pp. 275–96.
Hall, P.A. (2008) ‘Systematic Process Analysis: When and How to Use It’. European Management
Review, Vol. 7, No. 3, pp. 24–31.
Haverland, M. (2006) ‘Does the EU Cause Domestic Developments? Improving Case Selection in
Europeanization Research’. West European Politics, Vol. 29, No. 1, pp. 134–46.
Heclo, H. (1974) Modern Social Politics in Britain and Sweden: From Relief to Income Maintenance
(New Haven/London: Yale University Press).
Hedström, P. (2005) Dissecting the Social: On the Principles of Analytical Sociology (Cambridge:
Cambridge University Press).
Hedström, P. and Ylikoski, P. (2010) ‘Causal Mechanisms in the Social Sciences’. Annual Review
of Sociology, Vol. 36, pp. 49–67.
Heidenreich, M. and Bischoff, G. (2008) ‘The Open Method of Co-ordination: A Way to
the Europeanization of Social and Employment Policies?’ JCMS, Vol. 46, No. 3,
pp. 497–532.
Hodson, D. (2004) ‘Macroeconomic Co-ordination in the Euro Area: The Scope and Limits of the
Open Method’. Journal of European Public Policy, Vol. 11, No. 2, pp. 231–48.
James, O. and Lodge, M. (2003) ‘The Limitations of “Policy Transfer” and “Lesson Drawing” for
Public Policy Research’. Political Studies Review, Vol. 1, pp. 179–93.
Kyriakos Moumoutzis and Sotirios Zartaloudis350
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd
Kröger, S. (2009) ‘The Open Method of Coordination: Underconceptualisation, Overdetermination,
Depoliticisation and Beyond’, European Integration Online Papers, Special
Issue 1, Vol. 13, Art. 5.
López-Santana, M. (2006) ‘The Domestic Implications of European Soft Law: Framing and Transmitting
Change in Employment Policy’. Journal of European Public Policy, Vol. 13, No. 4,
pp. 481–99.
Machamer, P., Darden, L. and Craver, C.F. (2000) ‘Thinking about Mechanisms’. Philosophy of
Science, Vol. 67, No. 1, pp. 1–25.
Mahoney, J. (2007) ‘Qualitative Methodology and Comparative Politics’. Comparative Political
Studies, Vol. 40, No. 2, pp. 122–44.
Mahoney, J. and Goertz, G. (2006) ‘A Tale of Two Cultures: Contrasting Quantitative and Qualitative
Research’. Political Analysis, Vol. 14, No. 3, pp. 227–49.
Mailand, M. (2008) ‘The Uneven Impact of the European Employment Strategy on Member
States’ Employment Policies: A Comparative Analysis’. Journal of European Social Policy,
Vol. 18, No. 4, pp. 353–65.
March, J.G. and Olsen, J.P. (2004) ‘The Logic of Appropriateness’, ARENA Working Paper 04/09.
May, P.J. (1992) ‘Policy Learning and Failure’. Journal of Public Policy, Vol. 12, No. 4, pp.
331–54.
Moumoutzis, K. (2011) ‘Still Fashionable Yet Useless? Addressing Problems with Research on
the Europeanization of Foreign Policy’. JCMS, Vol. 49, No. 3, pp. 607–29.
Olson, M. (1971) The Logic of Collective Action: Public Goods and the Theory of Groups
(Cambridge, MA: Harvard University Press).
Radaelli, C.M. (2000) ‘Policy Transfer in the European Union: Institutional Isomorphism as a
Source of Legitimacy’. Governance, Vol. 13, No. 1, pp. 25–43.
Radaelli, C.M. (2003a) ‘The Europeanization of Public Policy’. In Featherstone K. and Radaelli C.
M. (eds) The Politics of Europeanization (Oxford: Oxford University Press).
Radaelli, C.M. (2003b) ‘The Open Method of Coordination: A New Governance Architecture for
the European Union?’. Swedish Institute for European Policy Studies, Report No. 1.
Radaelli, C. M. (2006) ‘Europeanization: Solution or Problem?’ In Cini, M. and Bourne, A.K.
(eds) Palgrave Advances in European Union Studies (Basingstoke: Palgrave Macmillan).
Radaelli, C.M. (2008) ‘Europeanisation, Policy Learning and New Modes of Governance’. Journal
of Comparative Policy Analysis: Research and Practice, Vol. 10, No. 3, pp. 239–54.
Radaelli, C.M. (2012) ‘Europeanization: The Challenge of Establishing Causality’. In Radaelli, C.
M. and Exadaktylos, T. (eds) Research Design in European Studies: Establishing Causality in
Europeanization (Basingstoke: Palgrave Macmillan).
Risse, T., Cowles, M.G. and Caporaso, J. (2001) ‘Europeanization and Domestic Change: Introduction’.
In Cowles, M.G., Caporaso, J. and Risse, T. (eds) Transforming Europe:
Europeanization and Domestic Change (Ithaca, NY: Cornell University Press).
Rohlfing, I. (2012) Case Studies and Causal Inference: An Integrative Framework (Basingstoke:
Palgrave Macmillan).
Rohlfing, I. (2014) ‘Comparative Hypothesis Testing Via Process Tracing’. Sociological Methods
and Research, Vol. 43, No. 4, pp. 606–42.
Sabatier, P.A. (1987) ‘Knowledge, Policy-Oriented Learning and Policy Change: An Advocacy
Coalition Framework’. Science Communication, Vol. 8, No. 4, pp. 649–92.
Schmidt, V.A. (2002) ‘Europeanization and the Mechanics of Economic Policy Adjustment’.
Journal of European Public Policy, Vol. 9, No. 6, pp. 894–912.
Schmidt, V.A. and Radaelli, C.M. (2004) ‘Policy Change and Discourse in Europe: Conceptual
and Methodological Issues’. West European Politics, Vol. 27, No. 2, pp. 183–210.
Thomson, R. (2009) ‘Same Effects in Different Worlds: The Transposition of EU Directives’.
Journal of European Public Policy, Vol. 16, No. 1, pp. 1–18.
Europeanization mechanisms and process tracing 351
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd
Trubek, D.M. and Trubek, L.G. (2005) ‘Hard and Soft Law in the Construction of Social
Europe: the Role of the Open Method of Co-ordination’. European Law Journal, Vol. 11,
No. 3, pp. 343–64.
Van Evera, S. (1997) Guide to Methods for Students of Political Science (Ithaca, NY: Cornell
University Press).
Van Rie, T. and Marx, I. (2012) ‘The European Union at Work? The European Employment Strategy
from Crisis to Crisis’. Journal of Common Market Studies, Vol. 50, No. 2, pp. 335–56.
Visser, J. (2005) ‘The OMC as Selective Amplifier for National Strategies of Reform’. In Zeitlin,
J., Pochet, P. and Magnusson, L. (eds) The Open Method of Coordination in Action: The
European Employment and Social Inclusion Strategies (Brussels: Peter Lang).
Zartaloudis, S. (2013) ‘Wielding Soft Power in a World of Neglect: The Europeanization of Greek
and Portuguese Public Employment Services’. JCMS, Vol. 51, No. 6, pp. 1178–95.
Zeitlin, J. (2005) ‘Conclusion’. In Zeitlin, J., Pochet, P. and Magnusson, L. (eds) The Open
Method of Coordination in Action: The European Employment and Social Inclusion Strategies
(Brussels: Peter Lang).
Zeitlin, J., Barcevičius, E. and Weishaupt, J.T. (2014) ‘Institutional Design and National Influence
of EU Social Policy Coordination: Advancing a Contradictory Debate’. In Barcevičius, E.,
Weishaupt, J.T. and Zeitlin, J. (eds) Assessing the Open Method of Coordination: Institutional
Design and National Influence of EU Social Policy Coordination (Basingstoke: Palgrave
Macmillan).
Kyriakos Moumoutzis and Sotirios Zartaloudis352
© 2015 The Author(s) JCMS: Journal of Common Market Studies © 2015 John Wiley & Sons Ltd