Unity through fear? How hegemonic perceptions influence public support for the European Union.
Euroscepticism: Causes, Consequences, and Responses Author: Axel Noorman
Leiden University
Political Science: International Relations and Organizations, Year 3, Block 4 Thesis Supervisor: Dr. Nikoleta Yordanova
Second Reader: Dr. Paul Meerts
Table of Contents
Introduction ... 1
Background and Concepts ... 2
Review of the Literature ... 4
Research Design ... 10
Analysis... 13
Conclusion ... 19
References ... 21
1
Introduction
The study of Euroscepticism and its relation to foreign influences has been a notable topic for years, with scholars each focusing on different angles and ideas. Literature has predominantly focused on the subject of migration and threat perception from big states like Russia. Interestingly enough, the latter only looks at an individual country or regional analysis, whereas the entire European Union as a case is often overlooked. Moreover, the influences of other hegemonic states like the United States on individual support of the European Union are very scarcely researched. These ideas would be considered of high importance in light of events of the past years, most notably the Russian annexation of the Crimea and the surprise election of Donald Trump in the United States. Following these issues, this research shall thus focus on the question: ‘Under what conditions do individual perceptions of hegemonic states in the world impact public support for the European Union within its member states?’.
Individual perceptions can be considered as a broad concept. In order to provide an answer to the research question, this paper shall focus on one of the most extreme forms of perception: the experience of threat. Concerning perceptions of hegemonies, perceived threat is considered to be the subject brought forward most extensively in the literature in the connection to support for the European Union. Drawing on this previous research and international relations theory, it is to be expected that the findings would demonstrate two important conditions connected to the relationship between individual perceptions of hegemonic states and public support for the European Union. First, it is argued that identity plays a role in this relationship, where it is argued that a perceived different identity from another country would lead to more threat perception. Practical examples then have shown that this idea in practice has led to more support for the European Union. Second, it is argued that the strength of the relationship between increased perceived hegemonic threat and increased public support for the European Union varies per country, if not region. This idea has its origins in events such as the Crimea annexation, and some literary examples.
To investigate these theories, three hypotheses shall be tested. The first two
hypotheses focus on the first theoretical idea mentioned in the previous paragraph, taking the examples of Russia and the United States to test these theories. The last hypothesis will focus on the second theory, investigating the differences in strength of the relationship between increased perceived hegemonic threat and increased public support for the European Union in both Western Europe and Eastern Europe. This paper shall make use of the Eurobarometer
2
data and employ logistic regression analyses to investigate these hypotheses. The findings demonstrate that both the identity-based theory and region-based theory are supported, and thus it is therefore confirmed that these are important conditions for the establishment of more support for the European Union as a consequence of perceived hegemonic threat. Ultimately, this paper will provide a good contribution to the field of security in relation to European integration.
This paper is divided into several sections. First, I shall discuss the background of this project, including underlying ideas of recent events, and concepts. This is followed by a discussion of the pre-existing literature on the subject. Next, I will discuss the underlying theories with connected hypotheses. After this section, the discussion of the theoretical narrative will be completed, meaning that the analysis will follow. I shall first discuss the design of the study, followed by the results of the analysis. Once this is completed, I shall conclude with a discussion of these results and a brief overview of the relevance of the findings of the paper, as well as its limits.
Background and Concepts
In the spring of 2014, after Euromaidan, Russian troops were suddenly spotted in Crimea. It was clear that this was not routine (Russia has a naval base in Sevastopol), but rather an actual military operation. Not soon after it was confirmed that Russia had de facto annexed the Crimea peninsula. This action was met with worldwide responses of disapproval and in Europe, Ukraine was supported by both NATO and to an extent from the European Union (Akinyemi, 2014). It must be remarked that Russia is not the only big state in the world whose actions have received negative responses in recent years. In 2016, the surprise election of Donald Trump caused quite a shock across the world. In Europe, Donald Trump’s victory and policies have led to criticism. Especially his stances on NATO, where the
president wishes that European states pay more for the military alliance, are considered to be most straining (Dombrowski & Reich, 2017, p. 1026). Besides criticism from European states, there was also support from them. Eurosceptic parties and politicians reacted very pleased to the news (Weaver, 2016), and so this victory in the hegemony across the Atlantic Ocean presented a shivering moment for supporters of the European Union.
The previous paragraph hinted towards the subject of support for the European Union, but how is support for the European Union actually defined? The opposite of support was hinted at in the previous paragraph: Euroscepticism. What actually conforms as support or skepticism towards the European Union has been hotly debated amongst scholars. The most
3
prevalent issue focusses on the question of whether support for the European Union is as simple as a dual classification, where support is either defined as: ‘you either support the union, or you do not’ (see for instance: Treib, 2014). Other authors have defined support for the European Union as something being a more complex issue. An often used categorization does take into account full support and full skepticism, but also includes categories that are in between. De Vries (2018) follows such a categorization. Beyond the categories of full
support and no support at all, De Vries (2018) also uses the categories of policy skepticism and regime skepticism (pp. 79-80). Policy skepticism refers to an individual’s skepticism towards the policy conducted by the European Union, and regime skepticism refers to an individual’s skepticism towards the institutions of the European Union (De Vries, 2018, pp. 79-80). It is also important to remark that in both cases mentioned, individuals adhering to these categories do not advocate for their respective countries to leave the European Union (De Vries, 2018, p. 80). This definition is the one that would be most preferred, as it is a more narrow and realistic than a simple dual classification.
The United States and Russia, both mentioned in the background narrative, share a certain characteristic. Both states have historically been hegemonic powers. Whilst the United States still is a hegemony, Russia is not. Hegemonies can be defined as dominant states who exert (partial, if not full) authority over one or multiple states (Lake, 2006, p. 25). It has been often debated whether the United States is being surpassed as the world’s
hegemony. The most notable case of a surpassing hegemony would be linked to China. Christensen (2015) argues that China would not pass the United States any time soon, that the threat of China is more regional rather than global, and that it is better if the United States and China both contribute to global order. What is to be taken from this is that United States hegemony can be questioned, and so rising powers can also be taken into account for the conceptualization of hegemony in this paper. Russia is another example. After the fall of the Soviet Union, Russia lost influence in world affairs, and much of the late 1990s were spent focusing on internal affairs and essentially bringing the country back on its feet (MacFarlane, 2006). However, nowadays Russia has been notable in challenging United States’ hegemonic character (David, 2017), thus making it clear that it is a rising power and would fit the
character of this research. Hence, both powers in place and rising powers are to be considered hegemonic.
As mentioned, hegemons can exercise authority, and authority can sometimes be exercised through threats. As threats are used to measure individual perception, it is another central piece of this research. But how is it defined? Definitions often vary, but there are
4
some characteristics of the concept that many definitions share. Scholtz (2000) considers that such overarching characteristics of a threat are its future-orientation, uncertainty, and the generation of negative emotions due to the perception of the concept (p. 24). In political science, the concept can take a slightly different turn. Threats can be considered as a means of exercising influence, more specifically, using coercive measures or statements as a means to influence how a specific actor behaves (Davis, 2000, p. 10). This conceptualization clearly takes a more institutional approach. Similarly, whilst there is a messenger, there is equally a receiving side of the threats. How one reacts towards these threats, or whether one feels threatened can be defined as the perceived level of threat. The recipients of the threat can often be widespread, but this research puts a specific focus on the individual level. Hence, both concepts mentioned in this brief section are key for understanding the perceived level of threat that an individual can experience. For this paper, perceived threat is the level of threat that an individual experiences based on the actor, the hegemonies, enacting the perceived threat.
Review of the Literature
The literature on the subject of support for the European Union and perceived level of external threat is quite widespread. Some articles mentioned in this section are connected to each other, whilst others take completely different angles. One of those major articles that takes a different angle is written by Russo (2016). In this article, the author finds that among European officials, there can be support for further European integration in terms of a common European foreign policy when faced with an existing threat from a foreign country, in this case, Russia (Russo, 2016). The findings of this article thus suggest that the
relationship between the variables established in the research question would indeed exist. The research in this paper however focusses on investigating the relationship between the variables from a top-down perspective, rather than taking a bottom-up perspective (Russo, 2016, p. 197). Whilst the findings do suggest an existing link between the variables
mentioned, this article would still leave open an important gap that requires further investigation.
Taken the previous point into account, we must then turn to the literature that focusses more on an individual citizen analysis. Before we turn to external threat and support for the European Union in more detail, it is important to equally establish the basics of this potential relationship. An article written by Rousseau and Garcia-Retamero (2007) brings forward an interesting point. The authors of this article found that a difference in identity can be related
5
to higher levels of threat perception (Rousseau & Garcia-Retamero, 2007, p. 766). This article is focused on foreign affairs rather than internal affairs discussed earlier. Regarding threat perception, Rousseau and Garcia-Retamero (2007) conduct a survey experiment where respondents from Spain have to give answers to questions related to Russia, which lead to the aforementioned results (p. 761). Whilst the findings are interesting however, they are only limited to Spain. An analysis of more European states would be more preferable.
Differences in identity have indeed also been brought forward in other articles. Two articles, in this case, stand out, which are those written by Carey (2002) and McLaren (2002). Both articles include research that demonstrates that national identity can lead to more
skeptical views of the European Union (Carey, 2002, p. 388; McLaren, 2002, p. 564). Whilst this might seem like similar conclusions, it must be remarked that Carey (2002) puts a focus on national identity in the sense of patriotism, whilst McLaren (2002) focusses more on the perceived threat from other cultures, notably migration (McLaren, 2002). Whilst the findings are interesting, they are mostly related to cultural factors. Especially in the McLaren (2002) article, the perceived threat is cultural, rather than institutional. The perceived level of threat that is experienced is related to migrant groups rather than states. This is thus something that requires further analysis. The subject of national and EU support as mentioned by Carey (2002) takes a different angle in an article written by Armingeon and Ceka (2014). The authors of this article establish that support for the European Union is often related to support for the national government, meaning that citizens take support for the national government as a proxy for support of the European Union (Armingeon & Ceka, 2014, p. 103). Whilst these findings are interesting, they do hold issues. The authors argue that the existence of this positive relation is valid in the case of growing support. In other words, as support for the national government increases, support for the European Union increases (Armingeon & Ceka, 2014, p. 94). The opposite side of this relationship, however, where support decreases rather than increases, is not touched upon by the authors. This is then left for future research. Other flaws include the fact that the article is mostly internally focused, and not necessarily on foreign factors of potential European Union support.
The previous literature focused little on the relationship between the variables, but more on topics that seem rather similar. Which literature then is more relevant? An article written by Kostadinova (2000) gives more insight into the question of foreign threat, as this specific article places its focus on how a threat from Russia and how it contributes to the desire for Eastern European states to integrate more in western institutions (pp. 246-247). The findings presented actually demonstrated that the threat of Russia did account for
pro-6
integrationists stands for Eastern European member states to join NATO (Kostadinova, 2000, p. 245). Unfortunately, the question of integration regarding the European Union and
integration due to a foreign threat was little touched upon in this article, but the findings of this article do lead to that specific question. Fortunately, there is literature that focusses more on this specific question. Caplanova, Orviska, and Hudson (2004) conducted a study where respondents were asked questions regarding the reasoning for them to join both the European Union and NATO. Similar to the previous article, this research was limited to Eastern
Europe. The findings concluded that support for integration into the European Union was slightly higher, but that reasons for joining the European Union were focused mostly on economic reasons and, to an extent, security reasons, whereas NATO was mostly serving a security reason (Caplanova et al., 2004, p. 284). The survey design was not optimal, however, as survey questions regarding the European Union and its security benefits were not taken into account. Other articles that focus on the subject present a more detailed analysis of a specific country. Ehin (2001) conducts research in their article on the subject of European integration in the context of the Baltic states, and whilst the results are worthy of
consideration, the topic of the European Union and its security benefits from other states that tend to act in an intimidating manner (considering Russia, see for instance: Cichock, 1999) are little touched upon. The specific question has been researched in another article on the Baltic states, where in the case of Estonia it has been shown that there indeed is a motivation to integrate into the European Union to avoid a Russian threat (Kuus, 2002). Nevertheless, the research conducted is only limited to one country. A very recent article comes from Favero (2020). This article focusses on the topic of gender and support for the European Union, and the findings indicate that support for the European Union among Polish female students is higher than male students (p. 29). Moreover, the author highlighted that one of the reasons why participants are satisfied with European Union membership is that they see the European Union as a community that can protect the country from potential threats from Russia (Favero, 2020, p. 29). Indeed, it is clear that these findings are highly relevant, but the author also admits that the findings are very limited in applicability across the European Union (Favero, 2020, p. 32). Considering that most articles mentioned in this paragraph are limited to Eastern Europe, the question regarding the nature of European Union public support in relation to foreign threats across all European Union member states still stands.
Articles mentioned before mostly focus on the case of the foreign threat being Russia, but what about other cases? More specifically, has there been any literature on the United States? The existence of articles that consider this specific case are very limited, and those
7
that do exist might require more context for the argument to apply. One of these articles that refers to such context is written by Eichenberg and Dalton (1993). In their article, they illustrate an example of how under the threat of the Soviet Union, European countries would indeed prefer to integrate into NATO, rather than focusing on a European community (Eichenberg & Dalton, 1993, p. 515). Whilst this is an example of the past, considering the stances of Donald Trump towards NATO and his more isolationist ideals (see for instance: Schreer, 2019), it can be questioned whether in Europe they still wish to be reliant on US defense rather than doing it themselves. Further research on Donald Trump does exist. Minkus, Deutschmann, and Delhey (2019) investigate how the election of Donald Trump affected support for the European Union. The finding presented that the election of Donald Trump led to an increase of support for the European Union (Minkus et al., 2019, p. 409). These findings are highly relevant for this paper, but the scope of the research is small. The research is only conducted over a short time-span, which focusses around the election time in 2016 (Minkus et al., 2019, p. 404). How the policy conducted by the Trump administration has influenced European Union support is not clear. With the gaps identified, we turn to the potential theories that have arisen from these works.
Theory and Hypotheses
Research prior to this paper discussed in the previous section has pointed out that identity can play a role in threat perception. Most prevalent in social constructivist theory, it is argued that the way states view each other does not necessarily have to depend on just raw strength and anarchy, but also on the identity perception that they have for one another (Wendt, 1992). As mentioned, in previous research it has been shown that a difference in identity contributes to a higher level of threat perception amongst individuals (Rousseau & Garcia-Retamero, 2007). This article demonstrates this by focusing on the case of Russia (Rousseau & Garcia-Retamero, 2007, 762). But how would a higher level of threat perception play out in terms of support for European integration? It can be argued that a higher level of threat perception would in fact lead to an increase in support for the European Union. Past cases have shown this to be especially true in individual European countries, where the threat and intimidation of another state presumed to be more hostile and potentially to hold more military power would lead to a desire for further integration in the European Union to safeguard themselves from said intimidation (Kuus, 2000; Kostadinova, 2000; Caplanova et al., 2004). One could thus expect a negative relationship between the perceived threat from a
8
hegemonic state and public support for the European Union. Therefore, the first hypothesis is formulated as followed:
H1: An increase in the perceived level of threat from Russia will lead to an increase in public support for the European Union.
To comply with social constructivist theory mentioned above, this hypothesis is taking the case of Russia, as one could argue that Russia holds a different identity than member states of the European Union (Rousseau & Garcia-Retamero, 2007).
Considering that a different identity would lead to higher threat perception, one could also argue that this relationship can be reversed. A similar identity would lead to less threat perception (Rousseau & Garcia-Retamero, 2007). This is where the United States of America comes into the picture. One could argue that the United States has a similar identity to
Europe, namely a western identity (Hill, 1992, p. 85) Would this then mean that less threat perception would lead to less support for the European Union? One would first have to question whether European citizens feel at all threatened by the United States. Considering the lack of theoretical backing, this is unlikely. In the literature mentioned earlier, a rally effect in the past (namely, the election of Donald Trump) has led to a surge in European Union support (Minkus et al., 2019). Interestingly enough, this specific rally effect stands out. A rally effect is often defined as a moment in time where national or international crisis leads to a sudden surge in support for the incumbent leadership of a given state, where citizens are able to set aside their differences and act more in unison (Hetherington & Nelson, 2003, p. 37; Baker & Oneal, 2001, p. 664). The rally effect of Minkus et al. (2019) does stand out, however, considering that it does not fit the definition above. Clearly, the election of Donald Trump was unexpected and sudden, but it does not fit the criteria of a crisis, either
domestically or internationally. Rather, it is more likely that the sudden surge could be connected to a more negative view of the United States as a consequence of the election of Donald Trump, rather than an actual feeling of threat. If this is taken into account, then one could argue that it is rather negative views of the United States that lead to more support for the European Union. This is evident in policy leadership, as actions by Donald Trump have often been the target of international criticism. Indeed, often cited events include the stance of Trump towards NATO (Dombrowski & Reich, 2017, p. 1026), and the recent withdrawal of contribution payments to the World Health Organization by the United States government has been notable for its international criticism that it has received (Smith, 2020). If these points are taken into account, one could question the public responses to such events and how this
9
would impact European Union support. Therefore, hypothesis two has been defined as the following:
H2: An increase in negative views towards the United States of America will lead to an increase in public support for the European Union.
One could also argue that views towards other countries can differ more from country to country. Despite the similarities that for instance can present themselves in an overarching western identity, it must be remarked that identity can also be shaped at a different level, for instance, the national level. Different countries can have different views of a certain
hegemon. These identities can be shaped by past events, for instance. More present-day issues that caused distrust were mentioned in earlier sections, namely the downing of Malaysian Airlines flight 17 and the annexation of Crimea. With the former, one could definitely argue that this particular case sparked an international outcry towards Russia, especially after an international investigation deemed the latter responsible (Walker, 2018). However, considering the majority of the victims held Dutch nationality, this disaster could be particularly felt in the Netherlands (Walker, 2018). The annexation of Crimea also led to an international outcry, but one could argue that in Eastern Europe this event could have set a reminder to the Soviet occupation of the past. Even after the collapse of the Soviet Union, former states in Eastern Europe (either Soviet-occupied or satellite states) still continued to argue that Russia continued to pose a threat, whilst in Western Europe, there was more desire for cooperation with Russia (Daalder, 2017, p. 36). Hence, it could be argued that the level of perceived threat can differ from country to country within the European Union. This in turn would then lead to different levels of support for the European Union in relation to the former. Unfortunately, investigating this relationship per European Union member state would exceed far beyond the length constraints of this paper, so the investigation of this relationship requires to be investigated on a different level. It can be argued that there still would be noticeable differences between Eastern and Western Europe in terms of this relationship. This is something that Daalder (2017) already hinted towards. Furthermore, the literature discussed earlier already hinted towards the existence of the relationship in Eastern Europe (Kuus, 2000; Kostadinova, 2000; Caplanova et al., 2004). Theories for Western Europe however, beyond the ones mentioned in this section, are largely non-existent.
Therefore, this relationship is still interesting to investigate in terms of the regional difference between Eastern and Western Europe. Therefore, the third and last hypothesis has been formulated as followed:
10
H3: An increase in public support for the European Union in relation to an increase in the perceived level of threat from Russia will be higher in Eastern Europe than in Western Europe.
Taking the literature aforementioned into account, and so keeping in mind the idea of the past threat of the Soviet Union, the boundary between East and West in this research shall follow the line of the past Iron Curtain.
Research Design
To investigate the hypotheses given in the previous section, this paper shall make use of the Eurobarometer data, specifically Eurobarometer Dataset 88.1. The data from this dataset was collected during fieldwork in late September 2017 and early October 2017 and contains 27881 respondents (European Commission and European Parliament, 2018). This dataset comprises not only of usual Eurobarometer questionnaires but also includes a special questionnaire, focusing on the future of the European Union where several questions focus on the views of countries outside of the European Union (European Commission and European Parliament, 2018), which makes it appropriate for this research. It must be remarked that there has been research that has shown that the reliability of the Eurobarometer has been put into question (see for instance: Höpner & Jurczyk, 2015), but due to constraints on existing data and time, this paper shall continue with this specific dataset.
The previous literature on the subject mentioned earlier already hinted towards the case selection. This research shall primarily focus on the United States and Russia. The reasoning behind picking these two states for the research is that they both fit the concept of hegemony mentioned earlier. Not only hegemonies in place shall be researched, but also rising powers. Of course, if not only hegemonic states but also rising powers fit the criteria for the research, then other rising powers like China or India could also be used for the research. This paper however shall refrain from using other cases, due to time and research constraints. Russia has been chosen as the rising power for this research due to theoretical reasons brought forward in the literature. The variables used in this research are
operationalized using survey questions from the dataset. The dependent variable, support for the European Union, shall be measured using the following question in the dataset (European Commission and European Parliament, 2018):
11
‘Generally speaking, do you think that (OUR COUNTRY)'s membership of the EU is...?’
Respondents were able to answer this question using the options ‘good’ (recoded as 1), ‘bad’ (recoded as 0), ‘neither good nor bad’, or ‘do not know’(European Commission and
European Parliament, 2018). To allow for better interpretation of the results, this variable shall be recoded to dismiss the categories of ‘neither good nor bad’ and ‘do not know’. The categories of good and bad show clearly what views the respondent holds towards the European Union. On the contrary, the option of ‘neither good nor bad’ makes it less clear to what extend a respondent supports/is skeptical towards the European Union. To provide clearer results, this option has therefore been dismissed. Following this process of recoding, the variable itself becomes binary, which means the analysis shall consist of a logistic regression analysis. Unfortunately, this does not exactly follow the conceptualization provided earlier, considering a binary classification of European Union support is not
desirable. However, due to data constraints, this variable must be accepted. The independent variable, perceived threat, is measured using a proxy variable which follows this question in the dataset (European Commission and European Parliament, 2018):
‘As regards each of the following countries or group of countries, do you have a positive or a negative view about it?’
This operationalization follows a certain logic, as one can argue that one views an object negatively if one feels threatened by it (Scholtz, 2000, p. 14). This idea has been further explored by Rousseau and Garcia-Retamero (2007), who argued that a similar identity would lead to more positive views of a country (p. 750). This relationship would then be equally true in a reversed nature for a different identity. Considering that the same authors have established that threat perception and identity are indeed correlated (Rousseau & Garcia-Retamero, 2007, p. 766), this proxy would thus suffice. The respondents were given a certain country and then asked to rate their views of this specific country (European Commission and European Parliament, 2018). In the dataset, this question is divided into several variables where the division is based on the country specified. The variables are categorical, and thus to provide better interpretation of the results, the variables shall be recoded from a categorical to a scale variable, where all respondents who did not give a satisfying answer (the category ‘do not know’) will be coded as missing. The scale that has been applied is positive, meaning that if the number on the scale increases, the answer of the respondent is more positive.
12
For both theoretical reasons and to provide more accurate results, the research shall also make use of control variables. It has been argued in the past that economic factors can play a role in how one views the European Union. A study conducted by Gabel (1998) clearly shows that higher income groups tend to show more support for the European Union than lower-income groups (p. 52). Income groups shall thus be taken into account for the research, which in the dataset is proxied as social class. Education is also often considered for support of European integration. Similar to the previous control variable, Gabel (1998) also found that higher levels of education tend to be more supportive of European integration than lower levels of education (p. 52). The level of education will thus also be added to the equation. This variable is operationalized using a proxy variable, namely the years of education. This variable has been recoded to dismiss all answers that range from refusals to the category of ‘’do not know’’, but it also excludes those who are still studying. These respondents are excluded as they do not provide an exact answer to the years that they already have been studying, whilst this is desired. The results of the research conducted by Rohrschneider (2002) also demonstrated that age can be a significant factor in support of the European Union (p. 471), and so this control variable will also be included. Knowledge of the European Union is another particularly interesting control variable. It has been argued that an individual who knows more about the European Union is more likely to form a more informed opinion about the European Union based on the institution itself rather than taking the national
government as a proxy (Hobolt, 2012, p. 99; Armingeon & Ceka, 2014, p. 95). Unfortunately, the data does not allow for direct measurement of this variable, so the research shall use a proxy variable that focusses on interest in European affairs. One can assume that someone who does not demonstrate an interest in European affairs is also likely to have less
knowledge of the European Union, as their lack of interest would prevent them from
becoming more informed. Next to interest in European Union affairs, this research shall also control for interest in political affairs in general. Lastly, gender is also included as a control variable, as past research has indicated that differences in gender could account for support for the European Union (see for instance Favero, 2020; Liebert, 1999). Ideology is another control variable that has often been used in similar research (see for instance: Rohrschneider, 2002), but the Eurobarometer data on ideology placement along the left-right scale does not provide satisfactory details of an individual’s placement on this scale when one takes into account the subject of Euroscepticism. Treib (2012) stated that Eurosceptic views can be prevalent in parties that identify themselves as either right- or left-wing (p. 1543). The
13
variables in the dataset that measure ideology however do not specifically take this fact into account. For this reason, this specific control variable has been left out.
Analysis
Preliminary results of the analysis have demonstrated that the model violates assumptions such as linearity of the logit1. Hence, all the analyses conducted have been
applied with bootstrapping to provide accurate results. Table 1 shows the frequencies of the variables. The discussion of the results also takes into account certain statistics that are not displayed in the table, which are displayed more detailed in the datafiles.
Table 2 demonstrates the results of the conducted analysis2. Model 1 indicates the
relationship between the variables without controlling for other variables, and model 2 does include control variables. Whilst model 1 can be used for a comparison of the overall model fit with model 2, attention will focus on model 2. The model itself is presented as reasonable in relation to the data, as it can account for a small portion of the variance, Nagelkerke’s R2 =
0.144. Whilst it does not explain a large portion of the variance, the model does seem to predict the cases correctly quite well, 83.7%. Whilst this percentage is indeed good, prediction of the correct cases of each group individually does hold issues. The modal category appears to predict a large portion correctly, 99.1%, whereas the other category does not, 8.6%3. Nevertheless, the model does fit the data, χ2(8) = 1360.820, p < 0.001.
The coefficients in table 2 are all odds ratios, indicating the increase or decrease of odds that a certain outcome will occur as one increases or decreases the predictor. The results from the variable that focusses on Russia4 demonstrates an odds ratio of 0.758. This indicates
a negative relationship, as an odds ratio below 1 tells us that the odds of support of the European Union decreases with 0.242 as views of Russia increases with one unit (ceteris paribus). This is also true in the reversed nature, where one can argue based on these results that (ceteris paribus) as one’s views of Russia decrease with one unit, the odds of approval for the European Union would increase by 24.2%.
1 See the data files for more details.
2 In the analyses conducted in this section, all levels of significance and standard errors are based on the coefficients with robust bootstrapping applied.
3 See data files
4 This variable and the variable measuring views towards the USA both have a positive scale, meaning that as it increases with one unit, respondents are more positive towards the given country.
14 Table 1.
N Minimum Maximum Mean Standard
Deviation Support of the EU 18871 0 1 0.8306 0.37513 Views towards the USA 25761 1 4 2.4802 0.86634 Views towards Russia 25257 1 4 2.2423 0.85377 Eastern European countries 27881 0 1 0.3858 0.48680 Views towards Russia × Eastern European countries 26914 0 4 0.8818 1.27984 Social class 26666 1 5 2.3608 0.99565 Education 25559 2 71 19.5679 5.27474 Age 27881 15 99 51.59 18.314 Gender 27881 0 1 0.4482 0.49731 Interest in European Union affairs 27730 1 4 2.5642 0.87455 Interest in politics 27881 1 4 2.5921 0.97140
15
As this odds ratio in the model is highly significant, p < 0.01, these findings thus support the idea that an increase in the level of the perceived threat from Russia would lead to an increase of public support for the European Union (H1).
The results from the United States tell us a different story. The odds ratio for the variable that focusses on this country is demonstrated to be 1.451. As this number is above 1, this would indicate a positive relationship between the variables. This means that (ceteris paribus as favorable views towards the United States increase with one unit, the odds of public support for the European Union would increase with 45.1%. Similarly, in a reversed fashion (ceteris paribus) a decrease in views towards the United States would mean a decrease in support for the European Union. Considering that the odds ratio is highly significant, p < 0,01, it must be concluded that an increase in negative views towards the United States will not lead to an increase in public support for the European Union. Hypothesis two is thus not supported.
What would then explain the idea that the relationship between views towards the United States and public support for the European Union? As mentioned earlier, it was
theorized that a similar identity would lead to a decrease in threat perception from one state to the other (Rousseau & Garcia-Retamero, 2007). Considering that it can be argued that the United States and member states of the European Union do share similarities in their identity (for instance, the fact that both share Western views), it could account for the positive
relationship that was brought forward in the results. The disapproval of European leaders regarding policy measures that Donald Trump has taken in his presidency was used as a theoretical justification for the initial hypothesis. Nevertheless, this theoretical backing does not have to be incorrect. If the theory of a similar identity is correct, then one could also discuss identity-formation. The process of identity formation is one that is in a constant process, and whether identities change rapidly or not could depend on its stability (Wendt, 1994, p. 386). Therefore, it can be argued that potential changes in identity have not yet emerged, considering that Donald Trump has not been president for a very long time (at the time of publication, he will not have finished his first term yet).
Next, the research turns to the analysis that examines whether support for the European Union in relation to an external threat will be higher in Eastern Europe than in Western Europe (H3). to analyze whether the reaction will be higher in Eastern Europe, a dummy variable which focuses on Eastern European countries will be added, followed by an interaction variable measuring views towards Russia pinned together with the Eastern European countries. See table 3 for the results.
16 Table 2.
Model 1 Model 2
(Constant) 4.045** 0.276**
Respondent’s views towards Russia
0.717** [0,680; 0,757]
0.758** [0.716; 0.802] Respondent’s views towards
the USA 1.473** [1.396; 1.553] 1.451** [1.372; 1.535] Social class 1.124** [1.071; 1.043] Education 1.032** [1.021; 1.043] Age 0.999 [0.996; 1.001] Gender (ref, 1 = male) 0.754** [0.689; 0.825] Interest in European Union
affairs 2.118** [2.000; 2.244] Interest in politics 0.996 [0.946; 1.048] -2LL 13504.324 12402.833
Cox and Snell’s R2 0.017 0.086
Nagelkerke’s R2 0.028 0.144
Overall % correct 83.00 83.70
Modal % correct 100.00 99.10
N 15106 15106
Note: odds ratios with 95% confidence intervals in brackets *** p < 0.001, ** p < 0.01, * p < 0.05
Similarly to the previous analysis, only model 2 shall be taken into account. The model itself again demonstrates the ability of explaining a small amount of the variance, Nagelkerke’s R2 = 0.147. Next to the variance, The model demonstrates that the percentage of
correctly predicted cases equals 83.7%. Similar comments to the correctly predicted cases in the previous analysis apply in this model as well. Furthermore, the model also appears to
17
make a good fit with the data, χ2(10) = 1389.075, p < 0.001.
Table 3 reports the B coefficients rather than the odds ratios as done in the previous analysis. This is because reporting the B coefficients is more relevant for calculating the effect of Eastern European respondents on the relationship5. The B coefficient of the
interaction effect (-0.162) indicates a negative relationship similarly shown in the results of the previous analysis. The coefficient for the variable measuring respondents views towards Russia equals -0.233. A negative number would again imply an inverse relationship. The coefficient for the interaction variable equals -0.162. If we are to determine the actual effect of the interaction, the two coefficients need to be summed together. This result would be -0.395, which would indicate a stronger effect opposed to measuring the relationship between the variables in an EU-wide context. Furthermore, this number can be converted back to an odds ratio:
𝑒!".$%& = 0.674
The odds ratio would indicate an inverse relationship between the variable. As (ceteris paribus) positive views towards Russia decreases with one unit, the odds for support of the European Union would increase with 32.6%. This effect appears to be stronger than when the relationship is measured in an EU-wide context (0.793, equaling a percentage of 20,75).
Considering that the coefficients are statistically significant, these findings thus support the idea that the relationship between perceived threat from Russia and support for the European Union would be stronger in Eastern Europe than in Western Europe (H3).
In summary, the findings demonstrate that in the European Union a negative
relationship between perceived threat from Russia and public support for the European Union can be proven (H1), whilst a negative relationship between public perception of the United States and public support for the European Union is not supported (H2). The evidence however does support the theory that a negative relationship between perceived threat from Russia and public support for the European Union would be stronger in Eastern Europe than in Western Europe (H3).
18 Table 3. Model 1 Model 2 (Constant) 1.294** (0.084) -1.423** (0.184) Respondent’s views towards
the USA
0.373** (0.030)
0.346** (0.032) Respondent’s views towards
Russia
-0.278** (0.038)
-0.233** (0.040) Eastern European countries
(ref, 1 = East)
0.419** (0.149)
0.614** (0.155) Views towards Russia ×
Eastern European countries
-0.151* (0.057) -0.162* (0.060) Social class 0.126** (0.026) Education 0.031** (0.007) Age -0.001 (0.001) Gender (ref, 1 = male) -0.269** (0.046) Interest in European Union
affairs 0.763** (0.033) Interest in politics -0.010 (0.025) -2LL 13495.457 12374.578
Cox and Snell’s R2 0.018 0.088
Nagelkerke’s R2 0.029 0.147
Overall % correct 83.00 83.70
Modal % correct 100.00 98.90
N 15106 15106
Note: binary logistic regression coefficients with standard errors in brackets. *** p < 0.001, ** p < 0.01, * p < 0.05
19
Conclusion
This paper focused on determining the existence of a relationship between individual hegemonic perception and support for the European Union. Individual hegemonic threat perception served as a good measurement for general perception, and thus the findings can be connected to the main question of the research. Whilst the hypotheses all took different angles, it can be concluded that as individual perception from a hegemonic state becomes more negative, support for the European Union would indeed increase. Based on the findings, there are important condition that applies to this relationship. First, as demonstrated in the findings that support H1, a different identity is key. The results demonstrated that as threat perception from Russia increased, support for the European Union increased. Equally, the results of H2 regarding the United States demonstrated that countries with a similar identity would not lead to more support for the European Union if the views of that specific country were to decrease. Second, regional background is key. The findings of H3 demonstrated that in Eastern Europe, the negative relationship demonstrated in H1 is stronger in this region than in Western Europe. This thus shows that regional location is equally an important factor in the strength of the relationship investigated.
This paper was set out to contribute to the field of security and the European Union integration. Suggestions for further research on this topic would include a more narrow analysis of individual countries. Indeed, previous sections of this paper already hinted towards this. More specifically, H3 already provided a leading direction towards this
question. Considering both the literary and real world examples provided in this paper, this is something that I would recommend future research to focus on. Another angle focusses on the European Union Foreign Action Service. Would citizens support further integration of European foreign policy? Furthermore, because of this paper it is interesting to ask how European politicians would respond to the phenomenon investigated in this paper. How would Eurosceptic politicians get the voters on their side in light of a hegemonic threat? Would different political parties employ a similar strategy, or is it rather dependent on the party itself? This question can also be asked for pro-integrationist parties.
The research conducted in this paper however is not without its flaws, and are important to highlight. An important issue focusses on the fact that threat perception was used extensively to measure for individual perception of states. Of course, there are equally other factors that could contribute to negative views which could account for the established relationship between the variables. One could, for instance, think of policies conducted by a certain state that cause a negative response from another state, similarly to which the second
20
hypothesis focused on. Considering that the findings of this hypothesis also differed from the findings of the hypotheses that focused on threat perception, it is therefore highly
recommended that future research puts more attention on other factors that could contribute to negative hegemonic perceptions and support for the European Union. Other flaws have been hinted towards in previous sections. The biggest challenge this research faced were the data and time constrains, which led to some notable issues. As mentioned, the reliability of the Eurobarometer has been put into question. This research continued using this data however due to constrains on existing data. The dataset provided a good combination between the variables, and for this reason it seemed appropriate to go ahead the specific dataset used. However, should more data on the subject become available in the future, I would highly encourage replication of this research. Next, the cases studied were also limited. As mentioned, the conceptualization of hegemony used in this research allows for the
investigation of other rising states, such as China and India. Unfortunately, length constrains did not allow for further investigation of other (potential) hegemonic states. Of course, this does not mean that similar results as demonstrated in this paper would occur in other cases. Nevertheless, this is something that shall be left for future research to touch upon.
21
References
Akinyemi, A. (2014, March 2). Ukraine Crisis: World Leaders React to Unfolding Disaster in Crimea. International Business Times UK. https://www.ibtimes.co.uk/ukraine-crisis-world-leaders-react-unfolding-crisis-crimea-1438553
Armingeon, K., & Ceka, B. (2014). The loss of trust in the European Union during the great recession since 2007: The role of heuristics from the national political system. European
Union Politics, 15(1), 82–107.
Baker, W. D., & Oneal, J. R. (2001). Patriotism or opinion leadership? The nature and origins of the ’ ‘rally“round the flag”’ effect. Journal of Conflict Resolution, 45(5), 661–687.
Caplanova, A., Orviska, M., & Hudson, J. (2004). Eastern European Attitudes to Integration with Western Europe. JCMS: Journal of Common Market Studies, 42(2), 271–288.
Carey, S. (2002). Undivided loyalties: Is national identity an obstacle to European integration?
European Union Politics, 3(4), 387–413.
Christensen, T. J. (2015). The China challenge: Shaping the choices of a rising power. W. W. Norton & Co.
Cichock, M. A. (1999). Interdependence and manipulation in the Russian-Baltic relationship: 1993–97. Journal of Baltic Studies, 30(2), 89–116.
Daalder, I. H. (2017). Responding to Russia’s Resurgence: Not Quiet on the Eastern Front America’s Forgotten Wars. Foreign Affairs, 96(6), [ii]-38.
David, M. (2017). Russia’s challenge to US hegemony and the implications for Europe. In
American hegemony and the rise of emerging powers (1st ed., pp. 198–215). Routledge.
Davis, J. W. (2000). Threats and Promises: The Pursuit of International Influence. John Hopkins University Press.
22
De Vries, C. E. (2018). Euroscepticism and the Future of European Integration. Oxford University Press.
Dombrowski, P., & Reich, S. (2017). Does Donald Trump have a grand strategy? International
Affairs, 93(5), 1013–1037.
Ehin, P. (2001). Determinants of public support for EU membership: Data from the Baltic countries. European Journal of Political Research, 40(1), 31–56.
Eichenberg, R. C., & Dalton, R. J. (1993). Europeans and the European Community: The
dynamics of public support for European integration. International Organisation, 47(4), 507– 534.
European Commission and European Parliament. (2018). Eurobarometer 88.1 (2017) (ZA6925) [Dataset]. GESIS Data Archive. https://doi.org/10.4232/1.12959
Favero, A. (2020). The influence of gender on attitudes towards the EU among the Polish ‘winners of European integration’. European Political Science Review, 12(1), 19–34.
Gabel, M. J. (1998). Interests and Integration: Market Liberalization, Public Opinion, and
European Union. University of Michigan Press.
Hetherington, M. J., & Nelson, M. (2003). Anatomy of a Rally Effect: George W. Bush and the War on Terrorism. Political Science and Politics, 36(01), 37–42.
Hill, S. (1992). The West and the rest: Discourse and power. In Race and Racialization, 2E:
Essential Readings (pp. 85–95).
Hobolt, S. B. (2012). Citizen Satisfaction with Democracy in the European Union. JCMS: Journal
of Common Market Studies, 50(s1), 88–105.
Höpner, M., & Jurczyk, B. (2015). How the Eurobarometer Blurs the Line between Research and Propaganda. PIfG Discussion Paper, 15(6), 27.
23
Kostadinova, T. (2000). East European public support for NATO membership: Fears and aspirations. Journal of Peace Research, 37(2), 235–249.
Kuus, M. (2002). European integration in identity narratives in Estonia: A quest for security.
Journal of Peace Research, 39(1), 91–108.
Lake, D. A. (2006). American Hegemony and the Future of East-West Relations. International
Studies Perspectives, 7(1), 23–30.
Liebert, U. (1999). Gender politics in the European Union. European Societies, 1(2), 197–239. MacFarlane, S. N. (2006). The ‘R’ in BRICs: Is Russia an emerging power? International Affairs,
82(1), 41–57.
McLaren, L. M. (2002). Public support for the European Union: Cost/benefit analysis or perceived cultural threat? The Journal of Politics, 64(2), 551–566.
Minkus, L., Deutschmann, E., & Delhey, J. (2019). A Trump Effect on the EU’s Popularity? The U.S. Presidential Election as a Natural Experiment. Perspectives on Politics, 17(02), 399– 416.
Rohrschneider, R. (2002). The Democracy Deficit and Mass Support for an EU-Wide Government. American Journal of Political Science, 46(2), 463.
Rousseau, D. L., & Garcia-Retamero, R. (2007). Identity, power,and threat perception: A cross-national experimental study. Journal of Conflict Resolution, 51(5), 744–771.
Russo, F. (2016). The ‘Russia Threat’ in the Eyes of National Parliamentarians: An Opportunity for Foreign Policy Integration? Journal of European Integration, 38(2), 195–209.
24
Schreer, B. (2019). Trump, NATO and the Future of Europe’s Defence. The RUSI Journal,
164(1), 10–17.
Smith, A. (2020, April 15). ‘Not the time’: Trump faces global criticism over move to end WHO funding. NBC News. https://www.nbcnews.com/news/world/global-criticism-grows-trump-move-end-who-funding-amid-pandemic-n1184146
Treib, O. (2014). The voter says no, but nobody listens: Causes and consequences of the
Eurosceptic vote in the 2014 European elections. Journal of European Public Policy, 21(10), 1541–1554.
Walker, S. (2018, May 25). MH17: Australia and Netherlands accuse Russia of complicity. The
Guardian.
http://www.theguardian.com/world/2018/may/25/mh17-australia-and-netherlands-accuse-russia-of-complicity
Weaver, M. (2016, November 9). Rightwing populists first to congratulate Trump on historic upset. The Guardian. http://www.theguardian.com/us-news/2016/nov/09/far-right-first-to-congratulate-donald-trump-on-historic-upset
Wendt, A. (1992). Anarchy is what states make of it: The social construction of power politics.
International Organization, 46(2), 391–425.
Wendt, A. (1994). Collective identity formation and the international state. The American Political
25
Appendix: Bonus OLS Regression Analysis
Model 1 Model 2 Model 3
(Constant) 1.422** (0.017) 0.622** (0.032) 0.602** (0.034) Respondent’s views
towards the USA
0.097** (0.006) 0.087** (0.006) 0.084** (0.006) Respondent’s views towards Russia -0.089** (0.006) -0.070** (0.006) -0.066** (0.034) Eastern European countries (ref, 1 = East) 0.061* (0.028) Views towards Russia × Eastern European countries -0.015 (0.012) Social class 0.040** (0.005) 0.041** (0.005) Education 0.006** (0.001) 0.007** (0.001) Age 3.241E-5 (0.000) 5.254E-5 (0.000) Gender (ref, 1 = male) -0.031** (0.009) -0.029** (0.009) Interest in European Union affairs 0.206** (0.006) 0.207** (0.006) Interest in politics 0.013* (0.009) 0.012* (0.006) R2 0.017 0.104 0.105 Adj. R2 0.017 0.104 0.104 N 21557 21557 21557
Note: OLS regression coefficients with standard errors in brackets. *** p < 0.001, ** p < 0.01, * p < 0.05
26
This brief section contains a bonus OLS regression analysis which includes the
middle category ‘’neither good nor bad’’ of the dependent variable used in this research6. The
findings appear to match the results of this paper, with the exception for H3. However, similar implications for the paper apply to these findings as well.
