Electoral Tourism, Normative Power or Instrumentalised Practice? European Union Election Observation Missions in Africa

Since 1994, the European Union (EU) has participated in 142 Election Observation Missions (EOMs) and Election Expert Missions (EEMs) (EEAS, 2016). Yet, even though they are significant, there has been limited scholarly work on EU-EOMs. This article seeks to close this gap. By offering a comprehensive analysis, this article argues that EU-EMOs exemplify how the EU and African governments see election monitoring as an instrumentalised tool. The discussion reveals that the practice of intrumentalisation significantly hinders norm diffusion through EU-EOMs. Introducing various subject areas, the first section illustrates how the EU’s role in international relations and with it EU-EOMs are marked by the EU’s dichotomous character and inherent tensions between its normative and strategic commitments. Building on these insights, the second section challenges constructivist concepts of norm diffusion. The third section projects the instrumental nature of election monitoring through the quantitative and case study analysis. This mixed methods approach aims to integrate factors that explain why actors engage with each other through EU-EOMs, notably trade, aid and legitimacy. It further scrutinises what occurs during EU-EOMs and how these interactions can lead to the internalisation of norms. The conclusion draws on the theoretical and empirical findings to discuss why the EU is unable to effectively diffuse norms through EU-EOMs and how these findings alter our understanding of election monitoring.


Introduction
Since 1994, the European Union (EU) has participated in 142 Election Observation Missions (EOMs) and Election Expert Missions (EEMs) (EEAS, 2016). 1 Early EU observers were stamped as electoral tourists (Carothers, 1997). However, over the years the EU has increasingly enhanced rules and procedures guiding its missions to promote its image as an international actor. Embodied within the EU's normative agenda and a 'naïve faith in the power of elections' (Cheeseman, 2015, p. 117), today EU-EOMs are a fundamental aspect of the EU's efforts to promote democratic norms. Moreover, media outlets, government agencies distributing development assistance and academics regularly reference EU-EOMs reports. Yet, even though they are significant, there has been limited scholarly work on EU-EOMs. This article seeks to close this gap. Simultaneously, the geographic focus on Africa interrogates the role African governments play in inviting EU delegations. The quotes above highlight the different perceptions of election monitoring held by African leaders and the EU. Chris Patten's quote reflects the inherent normative agenda present in EU democracy promotion and David Martin emphasises the supposed neutrality of EU observers. Conversely some African leaders, including Robert Mugabe, have come to oppose the idea of foreign observers openly judging their country's elections (Dorman, 2005). This raises many questions. Why do the EU and African governments participate in election monitoring? How do the interactions between the EU and host governments shape this mutual process? Does the EU actually diffuse democratic norms through EU-EOMs? And what is the impact of election monitoring on the overall conception of the EU in the international arena?
By offering a comprehensive analysis, this article argues that EU-EMOs exemplify how the EU and African governments see election monitoring as an instrumentalised tool. The discussion reveals that the practice of intrumentalisation significantly hinders norm diffusion through EU-EOMs. In African Studies instrumentalisation is a commonly applied conceptual tool in explaining political behaviour. In what they call the 'political instrumentalisation of disorder', Chabal and Daloz (1999, p. 1) refer to 'the process by which political actors in Africa seek to maximise their returns on the state of confusion, uncertainty, and sometimes, even chaos, which characterises most African polities'. Similarly, Whitfield and Fraser demonstrate how negotiating aid is 'strategic game in which the two parties take note of each other's behaviour ' (2009, p. 32). Subsequently, the notion of instrumentalisation weakens the assumptions of Normative Power Europe (NPE) -a popular concept describing the EU's role as a 'norm changer' in world politics (Manners, 2002). NPE's conceptualisations of norm diffusion, based on constructivism, are deeply Western-centric, verging on neo-colonial practices. In problematising NPE, the theoretical framework applies an adaptation of Sikkink and Risse's (1999) model of norm socialisation to allow more space for agency and actors' instrumental behaviour. Whilst the proposed model is also derived on constructivist scholars, it engages with its weaknesses by presenting norm diffusion in EU-EOMs as a mutual engagement of election observation shaped by the rational endeavours of both actors over aid, trade and legitimacy. Both, the EU, by sending observer teams and African governments, by inviting international observers, determine the conditions under which they engage through election monitoring and aim to benefit from the interaction.
Introducing various subject areas, the first section illustrates how the EU's role in international relations and with it EU-EOMs are marked by the EU's dichotomous character and inherent tensions between its normative and strategic commitments. Building on these insights, the second section challenges constructivist concepts of norm diffusion. Inspired by Acharya's emphasis on agency, this article proposes an adaptation of Risse and Sikkink's (1999) norm socialisation model. The third section projects the instrumental nature of election monitoring through the quantitative and case study analysis. This mixed methods approach aims to integrate factors that explain why actors engage with each other through EU-EOMs, notably trade, aid and legitimacy. It further scrutinises what occurs during EU-EOMs and how these interactions can lead to the internalisation of norms. The conclusion draws on the theoretical and empirical findings to discuss why the EU is unable to effectively diffuse norms through EU-EOMs and how these findings alter our understanding of election monitoring.
A majority of EU-EOMS take place in Africa, 62 out of 120 elections the EU observed between 2003 and 2016 were in African countries (EEAS, 2016). Between 2003 and 2011, the EU sent 42 EOMs to African countries. The time frame has been chosen because it is only since 2003, that delegations have published final reports and because the analysis investigates whether there is a link between the final report and re-observation at the following election. EU-EOMs final reports, which include the delegations' overall assessment of the electoral process are available on the EEAS website and are published in English or French and the host countries' official languages (EC, 2008, p. 110). However, such reports are worded carefully and it is an impossible task to objectively evaluate if an election observation report is positive or negative. I follow Hyde in classifying the official final reports as negative only if they 'seriously question the winner of the election or the legitimacy of the process' (2011). Reports are negative if they outright propose that the elections ' cannot be considered credible' (EU-EOM, Nigeria 2007a) or EU observers believe that it 'fell short of key international and regional standards' (EU-EOM, Kenya 2007b).
Vigorously analysing the norm socialisation model, the second section further incorporates the qualitative accounts of two case studies: Côte d'Ivoire's 2010 presidential election and Ethiopia's 2005 parliamentary election. Case studies 'permit a more in-depth understanding of how international norms interact with very different domestic structures' (Risse and Ropp, 2013, p. 4), allow more creative thinking and are thus better suited to studying norm diffusion (Gilardi, 2012, p. 471). To explore the dynamics between countries and EU delegations, the discussion also includes the pre-and post-election context. Importantly, the cases share similarities, as polling days were followed by post-election violence, and differ, as the EU re-observed Ethiopia's 2010 election despite a negative report, whereas following a positive report, it did not send a delegation to Côte d'Ivoire in 2015. The cases deliberately highlight distinct factors to emphasise the ambiguous nature of elections and varying domestic contexts. In addition to EU documents, the analysis relies on academic and newspaper articles, government statements and reports from human rights organisations. Moving beyond quantitative data, the aim is to explore the dynamics between the EU and African governments, which is consequently linked to scrutinise NPE.
One has to note that as elections are inherently ambiguous and are marked by particular domestic complexities, it is problematic to generalise the effect of EU-EOMs on democratising countries. Therefore, this analysis largely relies on qualitative data, which allows an examination of the interactions and engagements between host governments and observer teams. This includes assessing their motivations, strategic agendas and particular domestic circumstances. Election observation is often presented as an instrument governments employ as part of a neutral 'governance' agenda, and therefore the 'relationships between donors and African states are not presented in terms of realpolitik' (Dorman, 2005, p. 172). Studying EU-EOMs through the instrumentalisation lens allows confronting such presumptions. Not only is it close to impossible for observers to be neutral, but their good intentions can also have devastating effects. In fact, 'the more donors struggled to fix African democracies, the more they become involved in controversial projects of political and social engineering that risked undermining their own legitimacy' (Cheeseman, 2015, p. 116).
Overall, this article has two purposes. The theoretical purpose is to develop a model of norm diffusion, which emphasises the importance of agency and accounts for the rational tendencies of actors. The empirical part contextualises this model applying it on EU-EOMs to gain a greater understanding of why the EU and African government participate in election monitoring, which has become a core tool for promoting democracy. Accounting for the instrumental nature of EU-EOMs, the analysis relies on literature that has emerged from election monitoring, norm diffusion, the EU's role as an international actor and African Studies. Drawing ideas from these different areas enriches the analysis. Based on an integrated approach involving a quantitative study of EU-EOMs final reports and detailed accounts of the 2010 election in Côte d'Ivoire and the 2005 election in Ethiopia, rigorous empirical evidence further guides this study of EU-EOMs. The cases both illustrate how the interactions between EU observers and respective governments influenced the monitoring process and determined a possible re-invitation to the following elections. Equally they highlight the impact international actors can have on domestic politics and the role of elections in consolidating democracies.

EU-EOMs Between Normative and Strategic Commitments
Acknowledging the rapid spread of election monitoring and the EU's growing role in international relations, a discussion of EU-EOMs is highly relevant. EU-EOMs exemplify how the EU's foreign policy is characterised by a split between normative and strategic commitments. Moreover, it is often difficult to distinguish between norms and interests (Diez, 2005), a consideration, which is a crucial factor in deconstructing the EU's normative agenda.

The Rise of Election Monitoring and the EU's Engagement
Since election monitoring has become a widespread phenomenon, the role of EU-EOMs requires further scrutiny. Exhibiting normative and strategic interests, election observation is a compelling case through which to scrutinise the EU's nature as a supra-national actor. International election observers are members of ' official delegations of foreigners invited to observe and report on the credibility of the electoral process' (Hyde, 2011, p. 5). Among the many international and non-governmental organisations that participate are the United Nations (UN), the Organisation for Economic Co-operation and Development, the EU, the Carter Centre and the National Democratic Institute (Kelley, 2009). Today, international observers monitor 80 percent of unconsolidated democracies (Hyde, 2011;Kelley, 2012). Historically, election monitoring emerged in the 1970s (Merloe, 2015) and experienced most growth around the 1990s with the end of the Cold War (Kelley, 2009, p. 61;2013, p. 16). However, little is known about the effects of election monitoring (Kelley, 2012, p. 9). Early studies on the practice were largely critical. Criticism addressed poor standards, procedures and allegiances (Geisler, 1993) and the many amateurs among observers fostering a sense of ' electoral tourism' (Carothers, 1997, p. 22). Recent literature has relied on larger data sets and observing trends over time. There seems to be some consensus that under certain conditions the presence of international observers can reduce fraud (Hyde, 2007) and that 'it is associated with improved election quality and increased turnover' (Kelley, 2013, p. 130). Studies have also shown links between the presence of monitoring groups and opposition boycotts (Kelley, 2011), post-electoral violence (Daxecker, 2012) and electoral fiscal manipulation (Hyde and O'Mahony, 2010). This debate conclusively highlights that EU-EOMs deserve greater attention.
Triggered by the post-Cold War era and deepening European integration, the EU first sent a delegation to observe Russia's 1993 election. Since then, the EU has become a prominent advocate of election monitoring (Kelley, 2013, p. 35). Through its Handbook for EU Election Observation and by establishing the European Instrument for Democracy and Human Rights (EIDHR), which funds EU-EOMs, the EU has established a coherent monitoring framework, co-managed by the European Commission, the European Parliament (EP) and the European External Action Service (EEAS). The European Commission is also signed the UN Declaration of Principles for International Election Observation (2005). Today, around 8 percent of the EU's budget supports democracy promotion (Cheeseman, 2015, p. 115) and the current 2014-2020 EIDHR covers €1,332,752,000 (European Commission, 2016).

The EU's Normative Agenda: Policy Narratives and Contradictions
Moreover, the EU has taken an increasingly normative turn. In 2012 José Manuel Barroso postulated that 'together we have the power, and the scale to shape the world into a fairer, rules based and human rights abiding place' (2012). In this context EU-EOMs are presented as a 'vital component of EU activities to promote democracy worldwide' (European Commission, 2008) and mirrors the EU's perception that observing elections 'strengthens democratic institutions, builds public confidence in electoral processes and helps deter fraud, intimidation and violence' (Kelley, 2012, p. 9). Yet, due to restricted financial means, the EU accepts invitations to observe a country's election according to their political usefulness (European Commission, 2008). This selection bias implies that the EU only 'visits countries, in which it deems it can have an influence on the process, but also in which it has political interest' (Kelley, 2013, p. 161). Interpreting this normative rhetoric against rational international endeavours, scholars have long struggled over how to position the EU's role in international relations.
Due to limited space but acknowledging the rich literature on what kind of 'power' the EU is, this analysis focuses on Manners' NPE. Contributing to this vibrant debate this study takes a critical stance towards NPE. Election monitoring allows challenging Manners' account of norm diffusion and thereby displaying the EU's instrumental use of EOMs. Normative theorisations 'have become a mainstream approach to the study of the EU' (Whitman, 2013) -manifest in the fact that Manners' article has been cited nearly 600 times (Web of Science, 2016). NPE relies on the idea that, given its normative basis, more attention should be given to the EU's ' ability to shape conceptions of normal' (Manners, 2002, p. 329). Manners (2002, p. 252) claims 'that the most important factor shaping the international role of the EU is not what it does or what it says, but what it is'. However, NPE requires further qualifications, as the notion takes a provocatively moral stance in talking ' about practices of legitimation, normalisation and, in a sense, 'making right' certain values or policies' (Keene, 2013, p. 277). Furthermore, whilst Manners postulates how NPE can potentially act as a ' civilising force', this is deeply problematic considering Europe's colonial past. Moving away from normative theorising requires looking at what the EU actually does and 'understanding the external relations of the EU as a merger of the conceptual and the empirical' (Haarstrup, 2013, p. 32). One ought to ask 'how can we know, as it is implied, that acting in a normative way is a good thing?' (Sjursen, 2006, p. 236). Empowering others is not an altruistic practice and the EU often merely mimes normative power, which has been 'in line with the consolidation of its own power and influence' (Holden and Warren, 2015, p. 58). This demands a constant engagement with the EU's Eurocentric assumptions, while deconstructing both EU rhetoric and actions (Onar and Nicolaïdis, 2013).
The contradictions between the EU's normative agenda and its broader strategic interests are particularly prominent on the African continent (Del-Biondo, 2014, p. 2). Its normative endeavours are most prominent within the good governance agenda. Since the 1990s, Brussels has aimed to promote democratic reforms beyond its borders (Smith, 2014) and provide development assistance as a single actor (Youngs, 2009, p. 895;Carbone, 2010, p. 26). However, questioning the reality of the donor-recipient relationship, one has to recognise Africa's growing strategic importance (Mangala, 2013, p. 2), manifested in the 2007 Joint Africa-EU Strategy. The European Commission mostly distributes aid through the European Development Fund (EDF), which is negotiated on a 5 years basis. Yet, the EU continues the be the dominant partner, as the European Commission and member states effectively manage its relationship with African countries to work in its favour (Wade and Meunier 2010), notably through 'the extent of its resources, the length of its integration experience and position in prior European-African relations' (Haarstrup, 2013, p. 43). Overall, whilst the EU has set itself up as a norm diffuser by observing elections in democratising countries, when studying EU-EOMs, a constant deconstruction of EU actions is required to balance Brussels Eurocentric rhetoric.

The Importance of Agency: Re-thinking Norm Diffusion Theory
Norm diffusion offers a suitable approach to engage with EU rhetoric and actions. Norms are inherently linked to the practice of election monitoring, as it is ' a fascinating example of how global norms change and how world order evolves despite recalcitrant opposition' (Kelley, 2012, p. 18). The constructivist school, defining norms as the 'standard of appropriate behaviour for actors with a given identity' (Finnemore and Sikkink, 1998, p. 891) has largely shaped the debate on various mechanisms of norm diffusion. In proposing a model, which diverts from constructivism, this article asserts the importance of agency, conceptualised as norm entrepreneurs and norm takers. Following Acharya (2004;, emphasising agency allows adequately analysing the mutual engagement between observers and observed. The following part develops the theoretical argument by highlighting the weaknesses of constructivist models and scrutinising Manners' norm diffusion mechanisms. Re-thinking this process, a new model based on Risse and Sikkink's (1999) process of norms socialisation is presented. Through this model one can determine whether norm diffusion has been successful and assess NPE during EU-EOMs. Ultimately, the model suggests that instrumentalisaiton is present during EU-EOMs.

The Conventional Model of Norm Diffusion
Whilst constructivism has been influential in accounting for the increasing acceptance of human rights related norms, it fails to incorporate those who receive norms or norm takers (Acharya, 2013). Norm diffusion is not a topdown process and norm takers tend to adopt norms in a way that fits their local settings. According to constructivists, norm diffusion only takes places in the international arena, when a country complies with a new norm and ultimately internalises it. This occurs because of changing circumstances, which then triggers a three-stage process, namely norm emergence, cascade, and internalisation (Finnemore and Sikkink, 1998). Facilitating the emergence of norms, norm entrepreneurs form the basis of Finnemore and Sikkink's argument (1998, p. 897), as 'norms are actively built by agents having strong notions about appropriate or desirable behaviour in their community ' (1998, p. 896). During the 'norm life cycle', socialisation leads to norms cascades in which a critical mass of norm conforming states are accumulated (Finnemore and Sikkink, 1998, p. 901). Risse and Sikkink (1999, p. 5) propose a similar process of norm socialisation. Figure 1 illustrates how strategic bargaining and moral consciousness-raising interact and ultimately instigate the institutionalisation and internalisation of norms. Whilst instrumental adoption of pressure remains crucial at the beginning of the socialisation process, the final stage, internalisation, is characterised by norms 'implemented independently from the moral consciousness of actors' and simply being 'taken for granted' (Risse and Sikkink, 1999, p. 17). However, with its Western centric focus, constructivism tends to downplay local contexts and underestimates the often problematic and ambiguous nature of norms (Engelkamp and Glaab, 2015, p. 203). Most importantly, it does no account for subsidiarity: 'local actors creating rules with a view to preserve their autonomy from dominance, neglect, violation, or abuse by more powerful central actors' (Acharya, 2011, p. 97). Overall, both the Finnemore/Sikkink and the Risse/Sikkink models lack a sufficient emphasis on agency and therefore fail to Magnes: Electoral Tourism, Normative Power or Instrumentalised Practice? European Union Election Observation Missions in Africa Art. 2, page 5 of 17 account for the strategic bargaining between observers and governments.

Norm Diffusion through Election Monitoring
During election monitoring, the interactions between observers and observed are paramount, as by influencing whether a country re-invites observers, interactions notably determine successful norm diffusion. In the election monitoring literature, Kelley (2008) and Hyde (2011) sought explanations for when norm diffusion occurs, offering notable contributions to this debate. But they equally fail to sufficiently engage with agency by neglecting the continued engagement between observers and observed and cannot accurately explain the lack of internalisation. Adopting the Finnemore/Sikkink's 'norm life cycle', Kelley argues that the norm of inviting observers emerged out of an ' evolving set of norms related to democracy, elections, and human rights' and reached a 'tipping point around the end of the Cold War' (Kelley, 2008, p. 225). This further led to a norm cascade as refusing monitors became increasingly costly for incumbents (Kelley, 2008, p. 226). In avoiding the emphasis on norm entrepreneurs, Hyde (2011) offers a compelling alternative of norm emergence, focusing on the role of incumbent governments. Accordingly, Hyde posits that political elites initiated the norm of election monitoring by inviting foreign observers in the first place. Not all leaders were committed to democracy, but 'gaining the ' democracy' label matters for a country's status in the world (Merloe, 2015, p. 83). Indeed, allowing observers into a country can be seen as an effort by incumbents to gain increased international support (Hyde, 2011, p. 11). Derived from the idea of a signalling game -once some countries invite and receive endorsement from monitoring delegations, the practice is recognised as a signal by governments of their commitment to holding democratic elections (Hyde, 2011, p. 40). Over time, inviting observers became common practice, as pro-democracy actors are more inclined to support countries on the road towards democratisation. Conversely, non-compliance with the norm can be costly, as incumbent governments believe they 'would be better off by signalling their commitment to democratisation' (Hyde, 2011, p. 73). Hyde acknowledges the rationalist side involved in decisionmaking (2011, p. 23) and highlights the deep imbalances between observers and observed, with the latter depending on the former in terms of financial assistance and international recognition. Hyde's signalling game, however, cannot explain the interactions between observers and incumbents after the latter have issued an invitation. Furthermore, ' as it is questionable whether governmental elites have truly internalised the fundamental democratic norms that underlie the practice of election monitoring' (Kelley, 2008, p. 232) both models cannot explain why this internalisation has not occurred.

The EU as a Norm Entrepreneur
As NPE is rooted in constructivism, its shortcomings are of a similar nature to the discussed approaches. NPE is overly focused on the EU and there is no engagement with how norms diffused by the EU are perceived and accepted by those who receive them. Thus, problematising NPE is a crucial step to understanding and assessing norm diffusion through EU-EOMs. Implying that it acts as a norm entrepreneur, Manners' (2002) notion of NPE argues the EU is able to define 'what passes for normal in world politics' (Manners, 2002, p. 253). The ontological quality, most importantly its nature as a hybrid polity, its historical context and legal foundations, allow it to be a 'norm changer' through six mechanisms: contagion, informational diffusion, procedural diffusion, transference, overt diffusion and the cultural filter. Most notably overt diffusion or 'the physical presence of the EU in third states' can be related to EU-EOMs (Manners, 2002, p. 245). By visiting polling stations, holding press conferences and through social Art. 2, page 6 of 17 media channels, EU monitoring delegations visibly assert their presence in host countries. Transference, or the exchange of goods and the cultural filter, referring to 'the interplay between the construction of knowledge and the creation of social and political identity' (Manners, 2002, p. 245) can also be traced to EU-EOMs because sending delegations is often related to increased development assistance and trade. Since Manners' publication, numerous studies have come to diverging conclusion of whether the EU asserts normative power. This requires a closer evaluation of EU actions and engagements with cooperating countries. Normative rhetoric has become an appealing strategy for Brussels, but in reality, trade-offs between the EU's will and its actual record as a progressive actor are common (Carbone, 2010, p. 26). Scheipers and Sicurelli (2008) argue that NPE played a role in African countries signing the Kyoto Protocol and the statue on the International Criminal Court (ICC). But conversely, NPE had limited impact on its Mediterranean partners (Pace, 2007, p. 665) as 'the EU has failed to bring about democratic transitions in its Southern neighbours' (van Hüllen, 2012, p. 117). This disjuncture between the EU's normative endeavours and normative impact suggests that the mechanisms of norm diffusion are more complex than laid out by Manners. Notably, there is an explanatory gap in looking at domestic politics such as the degree of political liberalisation, as it shapes the engagement with the EU, which consequently determines the effectiveness of EU democracy promotion (van Hüllen, 2012, p. 126). For instance, different domestic situations in Tunisia and Morocco seem paramount in explaining divergent 'survival strategies' and therefore different levels of willingness to cooperate with the EU (van Hüllen, 2012, p. 128). Equally, Manners' notion of norm diffusion lacks a reflexive angle, as 'EU actions are either Eurocentric, unrelfexively inclusive or both' (Bicchi, 2006, p. 298). Accordingly, one has to determine a model that fosters a comprehensive discussion of when the EU's efforts to act as a norm changer are successful and contribute to internalisation.

An Adapted Model of Norm Diffusion: Norm Takers and Internalisation
To determine NPE in EU-EOMs this article reconceptualises norm diffusion based on an understanding of constructivism's weaknesses, especially given the role of 'non-Western actors has been side-lined in the literature on norm diffusion' (Acharya, 2004, p. 270). Visualising the importance of agency, Acharya's framework of norm diffusion focuses on norm takers and their engagements during norm adaption, congruence-building and localication (2004, p. 241). In contextualising this process on election monitoring and to identify internalisation, one has to scrutinise the dynamics and contestations between norm entrepreneurs and norm takers. Accordingly, Figure 2 introduces an adaptation of Risse and Sikkink's (1999) process of norm socialisation. The original model 'under-theorised the process leading from commitment to compliance (Risse and Ropp, 2013, p. 11). Therefore, the below analysis includes norm takers and norm entrepreneurs, and emphasises their continued interactions through strategic bargaining and moral consciousness-raising with double arrows. Depending on the outcome of these engagements, norms can become institutionalised and internalised. Impor- tantly, the process is not top-down, as because of the omnipresent bargaining process, institutionalisation does not automatically lead towards internalisation.
Following Williams, the two-way arrow between institutionalisation and internalisation accounts for the 'intersubjective nature of this process ' (2007, p. 260). This further connects norm diffusion with rational behaviour, which has been accepted yet not sufficiently elaborated by constructivism. Norms relate to ' a strategic social construction' in which 'rationality cannot be separated from any politically significant episode of normative change (Finnemore and Sikkink, 1989, p. 888). Norms always lead to some criticism, as a norms' strength is measured by 'the level of opprobrium community members attract from their peers for engaging in behaviour that violates the norm' (Williams, 2007, p. 258). In fact, resistance to a norm triggers 'redefinition, contextualisation and localisation' illustrating the agency-aspect of norm diffusion (Acharya, 2013, p. 470). Aware of this ingrained contestation, the proposed model diverts from an overly normative and paternalistic approach and integrates considerations of efficiency and utility with moral principles present among all stakeholders (Acharya, 2004, p. 242). Figure 3 applies the modified process of norm socialisation to EU-EOMs. Strategic bargaining denotes the politics of inviting observers: the EU cannot observe all elections it is invited to, thus it decides which countries are more important to launch a mission in and governments can chose to invite among multiple observer teams. Similarly, both actors have moral reasons to participate. The EU wants to be perceived as democracy-promoting institution and incumbents have much to gain from showing commitment to conducting transparent elections. As election monitoring does not feature institutionalisation, this step is simply seen as a mission being conducted. During this time, different developments can notably impact interactions between incumbents and EU-EOMs. Was the EU's assessment positive or negative? Did EU-EOMs detect fraud or were the elections followed by post-election violence? These factors are suggested to influence internalisation, which indicates if a government re-invites the EU to observe its election and the EU decides to accept the invitation and therefore marks successful norm diffusion.
The research indicates a significant degree of instrumentalization present in election monitoring. Outweighing the costs and benefits of engaging with the monitoring regime, both actors seek to minimise challenges and gain from election monitoring, whilst showing commitment to democratic principles. This suggests that EU-EOMs are instrumentalised by both observers and observed. A such account diverts from traditional constructivist scholars, who believe that 'instrumentally or strategically motivated adaption' (Risse and Sikkink, 1999, p. 10) only plays a minor role in the latter stages of norm diffusion. Yet, it does not inherently reject constructivism, but adjusts its premises to a more representative model.

The Practice of Instrumentalising EU-EOMs
Empirically applying the adapted model of norm diffusion, this article suggests that instrumental behaviour is an omnipresent factor of EU-EOMs. In following the adapted model laid out above, this following part, offers a brief account of the 2010 Ivorian and the 2005 Ethiopian elections and engages with the moral consciousnessraising versus strategic bargaining process, analyses what happens during a mission and illustrates instrumentalisation by investigating internalisation.

Norm Entrepreneurs Meet Norm Takers: EU-EOMs in Côte d'Ivoire and Ethiopia
Upon the invitation of the Ivorian government and the electoral commission the EU observed Côte d'Ivoire's 2010 presidential election. After 11 years of political instability and turmoil, EU officials saw the 2010 elections as a turning point for Côte d'Ivoire. However, after a relatively calm first round, in which 85.11 percent of the electorate cast their vote (EU-EOM, 2010b), this positive attitude drastically changed. Instead of a peaceful democratic process, the 2010 elections resulted in a fierce struggle of power between the incumbent Laurent Gbagbo and the leading opposition candidate Alassane Ouattara. EU observers described the second round as 'tense' and noted several violent incidents around the city of Abidjan (EU-EOM, 2010c). The statement specifically condemned Gbagbo's aggressive language, describing his opponent as a 'maker of war'. When the final results were published, Ouattara had won a majority. But clinging on to his presidency, Gbagbo initiated a constitutional coup, which ensured his victory. In the following months, at least 3,000 civilians were killed, 250,000 citizens were internally displaced or fled to neighbouring countries and 150 women were raped during violent clashes between Gbagbo's and Ouattar's forces (Wormington, 2015;Strauss, 2011). By April 2011 Ouattara, with the support of the UN and the French army defeated Gbagbo and assumed the presidency. In 2015, Ouattara was re-elected for a second term, without the presence of EU observers. This case is a compelling example of the domestic complexities and international entanglement involved in election monitoring. But, why did Gbagbo expose himself to the attention of the international community? And why did the EU accept the invitation, considering that Gbagbo was unlikely to accept an electoral defeat? Similar questions arise when analysing the 2005 elections in Ethiopia. Observers described the campaign period as the most open and transparent in Ethiopia's history (EU-EOM, 2005), as in the run-up to the elections, the governing Ethiopian People's Revolutionary Democratic Front (EPRDF) marginally liberalised political space (Abbink, 2006). The EPRDF and opposition parties also sponsored peaceful rallies (Harbeson, 2005, p. 150) and on polling day itself, turnout was high, as 'people felt that change was possible' (Abbink, 2006, p. 183). As a result, the major opposition parties, notably the Coalition for Unity and Democracy (CUD) and the Union of Ethiopian Democratic Forces (UEDF) significantly increased their vote share. Yet, before the counting process officially terminated, EPRDF claimed victory, arrested opposition candidates and journalists of private media outlets (EU-EOM, 2015) and through a state media campaign accused the opposition of disloyalty (Abbink, 2006, p. 185). During post-election protests, which occurred despite a ban on public demonstrations, more than 200 demonstrators were killed and 20,000-30,000 opposition members arrested (Tronvoll, 2009, p. 455). Because of the deteriorating human rights situation, the final EU-EOM report declared that ' overall, the elections fell short of international principles for genuine democratic elections' (2005, p. 1). Whilst trying to act as a mediator between government and opposition in the post-election period (Hackenesch, 2015, p. 92), the EU delegation became vocally involved in the post-election struggle, which caused significant tensions in Ethiopia-EU relations. But, why did the government under Prime Minister Meles Zenawi invite the EU, if his party was determined to use all available means to stay in power? And if 'losing power in 2005, even in one or two parts of the federation, does not seem to have been an option that the ruling party contemplated' (Vaughan, 2011, p. 632), why did the EU decide to observe Ethiopia's election?
Legitimacy is a significant factor when scrutinising these questions. Legitimacy applies to the EU and incumbents, who seek international recognition and want to appear as credible leaders within their respective countries. In Côte d'Ivoire incumbent president Gbagbo sought legitimacy after a prolonged period of political instability, in which its relations to the international community had suffered. Originally scheduled for 2005, elections were postponed six times by Gbagbo. Throughout this period, which was marked by continued conflict between northern regions, controlled by rebels and the South, held by the Government, 'politics revolved around the terms under which the country could be reunified and elections could be held' (Boone and Kriger, 2010, p. 192). Thus, by 2010 international stakeholders saw the election as a rescue from crisis for Côte d'Ivoire (Banegas, 2011, p. 457) and with the presence of UN peacekeeping troops, Gbagbo must have had little choice other than inviting international observers. After commissioning several opinion polls, he also felt convinced that he was able to win the election (Piccolino, 2011, p. 21).
Similarly, in contesting the first open polls in Ethiopia's history, the EPRDF pursued a strategy of consolidating its domestic and international legitimacy (Gagliardone, 2014, p. 292). Despite constant pressure from the EU, the Ethiopian government had previously seen elections as a purely domestic matter (Kelley, 2012, p. 23). Reflected in a statement from 2000, Prime Minister Meles Zenawi declared that exercising self-determination, Ethiopia has to conduct its own elections and that 'if there is the assumption that election is not democratic unless foreign observers monitored the process, this is a distorted outlook' (Hyde, 2011, p. 85). Only in the run-up to the 2005 election the EPRDF agreed to strengthen cooperation with the EU and invited EU observers (Hackenesch, 2015, p. 92). Domestically, a notable factor that had let to this decision was a split within in the Central Committee of the ruling party in 2001, which 'significantly pressured the leadership and arguably opened political space to competitors in the 2005 election' (Hackenesch, 2015, p. 91;Vaughan, 2011, p. 620). The EPRDF also could not 'translate its significant developmental achievements into true legitimacy' (Lefort, 2013, p. 463) and general instability following the Ethiopian-Eritrean war (1998)(1999)(2000) further

Magnes: Electoral Tourism, Normative Power or Instrumentalised Practice? European Union Election Observation Missions in Africa
Art. 2, page 9 of 17 encouraged and gave a sense of acknowledgment to opposition parties (Vaughan, 2011, p. 631).
Likewise, for the EU election monitoring is more than a mere altruistic endeavour. Brussels has an interest in observing elections it believes are politically decisive and that will boost its image as an international actor. The EU seeks to construct a positive image of its external actions, not least as 'what the world thinks about the EU, is an important factor in facilitating or opposing EU achievements' (Elgström and Chaban, 2015, p. 22). By conducting EU-EOMs, the EU gains insights into the domestic politics of a particular country and can assert visible presence. The latter is not only important for Africans, who see EU observers at polling stations, but also Europeans, as EU-EOMs are frequently quoted by media outlets in the West. Both Côte d'Ivoire's 2010 and Ethiopia's 2005 elections were opportunities for the EU to display its capacity in monitoring elections. As most international actors, the EU saw the Ivorian elections as a crucial step towards democratisation as well as a sign for the country's normalising relations with the international community, which was marked by stark diplomatic tensions at the time (EU-EOM, 2011, pp. 4,9). The 2010 elections were also widely followed, not least due to the presence of 9,150 peacekeeping personnel during the election period, contributing to security arrangements (UN, 2010). Overall, being part of Côte d'Ivoire's expected transition towards democratisation was an appealing factor to the EU. As for Ethiopia, the EU saw the country as a crucial strategic and development partner. In a 2008 report, the European Commission even asserted that 'reflecting its population size, economic strength, resource endowment and its role as a regional service centre, Ethiopia commands a significant economic and geopolitical position in the Horn of Africa' (Ethiopia/European Commission, 2008a, p. 12). In aiming to legitimise aid flows to Ethiopia pre-2005, the EU had reluctantly been trying to engage with Ethiopia on good governance issues. When invited to observe the 2005 election, one can easily understand why the EU agreed to send a mission.
Aspects of aid are particularly gripping when scrutinising Ethiopia-EU relations. As more than 14 percent of its gross national income is foreign aid (Hackenesch, 2015, p. 89), the impact of donor-countries in stabilising the Ethiopian government has been significant (Abbink, 2006, p. 196). Accordingly, the Ethiopian elites also 'wants to have good relations with the donor community' (Abbink, 2006, p. 187). Since the EU was invited to observe Ethiopia's elections, EDF disbursements have increased steadily: from €122 million in 2005 to €215 million in 2007 (Ethiopia/European Commission, 2008b). Even in the latest 2014-2020 round of the EDF, Ethiopia was allocated €745.2 million, an increase of €100 million compared to the 2008-2013 financial cycle (Ethiopia/European Commission, 2014). Immediately after 2005, donor countries were disillusioned (Abbink, 2006, p. 174) and overall the EU and other donors halted more than $375 million in budgetary support (Hyde, 2011, p. 90). Meles Zenawi remained firm, however, and ' after a few weeks the international donors caved in and resumed full development assistance to Ethiopia' (Tronvoll and Aalen, 2009, p. 204). The simple fact that the Ethiopian government had invited EU observers seemed sufficient enough to continue development cooperation. As for trade, Ethiopia is one of the fastest growing nonoil producing economies in Sub-Saharan Africa. Despite increasing investments from China, the EU continues to be Ethiopia's largest trading partner (Hackenesch, 2015, p. 89), which helps explaining why both parties engaged through election monitoring.
Côte d'Ivoire illustrates a slightly different case. The country had been isolated from the international community and thus did not receive significant development assistance until the 2007 Ouagadougou Peace Agreement. However, inviting EU observers seems to have had some impact, as since the regime change, Côte d'Ivoire has received €119 million in humanitarian aid and development assistance from the EU (ECHO, 2016). Furthermore, whereas under the 10 th round of the EDF the EU only supported Côte d'Ivoire on designated projects; in the latest 2014-2020 financial period, the EU designated Côte d'Ivoire €273 million of EDF funding (Côte d'Ivoire/EU, 2014, p. 2). More importantly, economic factors considerably shape Ivorian-EU relations. With trade worth €5,564 million in 2014, the EU is the country's largest trading partner absorbing 36 percent of Ivorian exports and providing 26 percent of imports (EEAS 2016;EC 2015). These numbers are largely influenced by Ivorian cocoa production, as Côte d'Ivoire is the worlds' largest cocoa exporter, 'producing around 40 percent of global cocoa harvests' (Bax and Wearden, 2011). Overall, Ethiopia and Côte d'Ivoire exemplify how trade, aid and legitimacy affect the bargaining process that leads to a country's election being observed by an EU delegation.

During a Mission: EU Assessments and Electoral Fraud
Once the mechanisms are set in place for the EU to observe an election, the institutionalisation stage will codetermine whether norm diffusion is successful. Therefore one has to thoroughly analyse what happens during a mission. Was EU assessment positive or negative? Did incumbents rig the election? Answering these questions further reveals the instrumental nature of election monitoring and reflects the inherent political gains that are at stake. As discussed, the impact of international observers on elections remains a controversial issue. It is clear that despite election monitoring, governments employ electoral fraud, as autocrats are conflicted between the pressure to hold internationally certified elections and a desire to guarantee their hold on power' (Hyde, 2007, p. 41). Yet, Kelley highlights that even if detected cheating 'international criticism might not be much worse than if the incumbent had outright refused monitors' (Kelley, 2012, p. 31). This also applies to the EU. Whilst the EU's Handbook for Election Monitoring states in detail how EU-EOMs should be conducted, in practice the EU is very careful about how it frames its assessments. Looking at EU final reports in African countries between 2003-2011, one can observe that only in a minority of cases, EU assessment is negative. Out of 42 observed elections, only 8 final reports clearly state that an election did not meet international standards, whereas 25 reports remain fairly positive (the remaining elections were EEM or no report was available). Of course this does not mean that EU observers did not note any irregularities, but overall the genuine will of voters is reflected in the outcome of the election. Assessments are also adjusted to the particular domestic situation of a country. Kelley notes for instance that monitoring organisations are more likely to assess an election positively when there is a turnover of power (Kelley, 2012, p. 159). This is confirmed through EU-EOMs, as all 8 negative reports coincided with a victory of the ruling party. Côte d'Ivoire is particularly interesting in this regard, as the EU's final report noted irregularities during both rounds but re-affirmed the election outcome, which saw opposition candidate Ouattara as a victor. Thus, the report was overall positive because the electoral process adhered to international standards. As discussed, Ethiopia is one of the few times an EU observer team issued a negative report.
Despite divergent assessments, Côte d'Ivoire and Ethiopia are both examples of when incumbents severely interfered with the electoral process. The simple presence of EU observers clearly did not deter fraud. In the Ivorian case Gbagbo engineered his victory following polling day, which triggered a civil war and Ethiopia's ruling EPRDF refused to publish official results after declaring its electoral success. In Côte d'Ivoire Ouattara had initially received 54.1 percent of the vote. But unwilling to accept a defeat, Gbagbo filed a complaint to the Constitutional Council, 'which alleged that violence, intimidation, and other irregularities in the north had prevented his supporters from voting' (Bassett, 2011, p. 477). Under significant pressure, the Council gave in and annulled 664,405 votes, the equivalent of 16% of the electorate, which ensured Gbagbo won the election with 51.45 percent compared to Ouattara's 48.55 percent (Human Rights Watch, 2015;Bassett, 2011, p. 478). Following a constitutional strategy by relying on the Council, Gbagbo defended his actions by arguing that he was defending Côte d'Ivoire's sovereignty (Banegas, 2011, p. 459;Bassett, 2011, p. 479). In Ethiopia, the final count gave the EPRDF 371 out of 547 seats. This result was announced despite opposition claims that they had won the elections but were denied victory (Abbink, 2006, pp. 176,183). The EU-EOM also expressed its concerns, notably in an internal report, that was leaked to the Ethiopian press. It noted that 'the exit polls did not correspond with earlier election results' released by the electoral commissions (Daxecker, 2012, p. 508). In addition to carefully orchestrated manipulation and despite EU longterm observers still being in the country, the Ethiopian elites used coercive means to shut down any public protest against the government. This included intimidation of opposition supporters, tighter control of the press and mass arrests of journalist and non-EPRDF politicians. Overall, in both examples the presence of EU-EOMs did not contribute to free and fair elections.
More importantly, Côte d'Ivoire and Ethiopia also reflect how influential international presence can be on domestic politics. Even though the EU argues that 'its observers will not interfere in the election process' (European Commission, 2008, p. 25) the actions of EU observers, including their official assessments, played a notable role in the cases of Côte d'Ivoire and Ethiopia. From the moment Gbagbo's announced his victory, UN and EU election observers asserted that his allegations were distorted. The EU's final report, released in January 2011 clearly put the blame for the escalation on the presidential camp and affirmed that despite several irregularities, Ouattara had won the second round of the elections (EU-EOM, 2011, pp. 5,6). Additionally, the UN, EU, AU and ECOWAS ' quickly deployed, albeit ineffectively, a wide range of measures designed to make him back down' including diplomatic isolation and economic, trade and financial embargos (Banegas, 2011, p. 458). From the outset, Gbagbo perceived this international coalition as a threat and did not back down (Piccolino, 2011, p. 5). The international community, notably France with a mandate from the UN Security Council, reacted by employing a military intervention and supporting Ouattara's forces. The EU-EOM team probably had little influence on these developments. But by openly denouncing the validity of the Ivoirian Constitutional Council, the EU-EOM legitimised UN actions, as the Security Council agreed with EU observers that Ouattarra was the rightful winner (Bellamy and Williams, 2011, p. 832). Furthermore, the leading role played by one EU member state, France, cannot be neglected. French troops decisively backed Ouattara's forces in arresting Gbagbo, who by April 2011 had retreated into a bunker within his compound (Banegas, 2011, p. 466). Following the regime change, the intervention and the impact of France's economic interests, which were at stake have been sharply criticised for substantially determining Ivorian politics (Helly, 2013, p. 150;Mbeki, 2011).
Whereas in Côte d'Ivoire, its involvement was more indirect, the EU delegation publicly responded to the post-election crisis in Ethiopia. Undoubtedly, the EU-EOM 'gave a well-founded critical evaluation of the elections' (Abbink, 2006, p. 188). However, the delegation went beyond its task of merely observing the electoral process. An EU-EOM statement printed in an Ethiopian newspaper ' emphasised serious concern regarding threats and intimidation against supporters of opposition parties including isolated instances of murder' (Daxecker, 2012, p. 508). EU Chief Observer Ana Gomez also wrote to the Ministry of Information and demanded the release of several CUD politicians (EU-EOM, 2005, p. 41). In response, Prime Minister Zenawi published a letter openly accusing the EU delegation. Zenawi asserted that 'many Ethiopians feel that the EU-EOM made a significant contribution to the June events by leaking a wildly speculative report which suggested that the opposition had won the elections' (Zenawi, 2005). He also attacked Ana Gomez arguing that it was not her business making recommendations and calling her ' an election observer turned self-appointed colonial viceroy' who 'twists the arms of the government to force it to accept her dictates' (Zenawi, 2005). Even though the Prime Minister's reactions were seen as disproportionate, the developments illustrate 'international organisations contributed to the contestation of electoral manipulation by non-state actor' (Daxecker, 2012, p. 508). Since 2005, Ethiopia has seen a step-by-step closing of political space, severely threatening liberal values and politics of plurality (Tronvoll and Aalen, 2009, p. 194). In an attempt to consolidate its power, EPRDF initiated a party re-energising process, which boosted party membership from 760,000 in 2005 to 4.5 million in 2008 (Tronvoll, 2009, p. 469;di Nunzio, 2014). Furthermore, the government trialled politicians from opposition parties -former CDU leader Birtukan Medeksa received a life sentence but was released in 2010 -introduced 'repressive measures against independent journalists and began implementing one of the most pervasive systems for filtering online political content in sub-Saharan Africa' (Gagliardone, 2014, p. 280).

From Institutionalisation to Internalisation
The norm of election monitoring is institutionalised if a government decides to re-invite international observers. However, the interdependence between institutionalisation and internalisation is more complex than issuing a simple invitation. In the case of the EU, African governments may decide to invite international observers, but exempt the EU and similarly the EU can reject invitations. The EU clearly specifies that ' an observation may not be sent to a country where from the outset an election cannot be expected to fall substantially short if international standards, or where a country's democratic practices are considered to be generally sound' (European Commission, 2008, p. 7). Plus, the EU's decision whether or not it accepts an invitation is also informed by domestic factors and geo-strategic, economic and security concerns (Hackenesch, 2015, p. 87). Internalisation of election monitoring is therefore an inter-subjective process informed by stakeholders' instrumental tendencies.
A brief overview of all EU-EOMs missions in Africa suggests that the EU's final reports have little influence on whether the EU re-observed an election. In 16 cases (see Figure 4), a positive report resulted in another mission, but equally, in 11 countries, a positive report did not result in the EU observing the following election. Yet, it is striking that the EU re-observed elections in all 8 countries that previously had a negative assessment. This suggest a that a closer analysis is necessary to reveal the involved interdependencies.
Côte d'Ivoire and Ethiopia are both fascinating examples of the instrumental tendencies involved in re-observing elections. In 2015, Côte d'Ivoire held its second consecutive election. The EU only sent a delegation of independent electoral experts to monitor and analyse the electoral process. With over 80 percent of the vote share, Ouattarra was re-elected for another 5-year term (Abidjan, 2015). In the absence of an EU delegation, observers from the AU and ECOWAS declared the election as free and fair (Human Rights Watch, 2015, p. 3). But why did the EU not send a full EU-EOM? One can only speculate. However, after 2010 much criticism emerged, which accused international stakeholders of their over-enthusiasm leading up to the elections. Former South African President Mbeki argued that 'the 2010 Ivorian presidential elections should not have been held when they were held, as it was perfectly foreseeable that they would further entrench the very conflict it was suggested they would end' (Mbeki, 2011). Indeed it must have been clear that neither Gbagbo nor Ouattara, who have been core political figures since the 1980s, were likely to accept a defeat. Gbagbo comes from the wealthier Southwest of the country whereas Ouattara, a Muslim from the North-East of Côte d'Ivoire served as Prime Minister under Houphöuet-Boigny. But, due to tightened citizenship laws that excluded people whose parents were born outside Côte d'Ivoire, Ouattara -his parents are from Burkina Faso -was barred from standing in the 2000 elections. He continued to receive support from Northern parts of the country, and 'had become a symbol for the grievances of the disenfranchised northern political, economic, ethnic and regional grouping' (Collett, 2006, p. 627). Moreover, despite Gbagbo currently facing trial at the ICC, he still enjoys support in Southern parts of the country and Human Right Watch have noted that the atrocities committed in early 2011 have not yet sufficiently been dealt with (2011). Plus, there remains resentment among the Ivorian population, as pro-Ouattara forces were also involved in committing violent crimes (Strauss, 2011). Thus, its' previous endeavours, which were received critically and the continued fragility in Côte d'Ivoire could have informed the EU's decision not to reobserve the elections.
Ethiopia has had two parliamentary elections since 2005, both of which were won by the EPDRF. The EU observed the 2010 election, but in 2015, a misunderstanding between Ethiopian officials and Brussels meant that no EU observers assessed the electoral process. 2 Similar to Côte d'Ivoire, it seems as though the step from institutionalisation to internalisation is not straightforward. As discussed, the EU delegation's reaction to the 2005 crisis seriously distorted Ethiopian-EU relations. Thus, the Ethiopian government was hesitant to re-invite an EU-EOM in 2010. When it did, the Ethiopian government ' decided not to allow the observer mission to present its final results in Addis Ababa, which has never happened

Assessment
Re-observed Countries Positive Re-observed 16 Positive Not re-observed 11 Negative Re-observed 8 Negative Not re-observed 0 EEMs or no report available 7 before in the history of EU election observer missions' (Hackenesch, 2015, p. 93). The government's main focus in 2010 was to sustain full control and prevent a repeat of the 2005 election (Di Nunzio, 2014, p. 412), which resulted in increasing repression and authoritarian tendencies, clearly present in the April 2008 local elections (Tronvoll and Aalen, 2009, p. 194). Post 2005, the EU also became more cautious and a 2008 report even claims that 'respect for human rights in today's Ethiopia has improved' (Ethiopia/ European Commission, 2008, p. 14).
After the 2010 elections, the EU-EOM's preliminary report notes that 'while several positive improvements have been introduced, the electoral process fell short of certain international commitments, notably regarding the transparency of the process and the lack of a level playing field' (2010d, p. 1). In 2015, it was unclear whether Ethiopia had not invited the EU, or whether the EU had declined an invitation. In an interview before the 2015 elections Ethiopia's new Prime Minister Hailemariam Desalegn, following the foot steps of his predecessor, emphasised his country's sovereignty during elections: 'Let our people judge us, not someone from outside -after all, a democratic, credible, fair and free elections should be credible in the eyes of our people.' These inherent tensions accentuate the dynamic process of bargaining and contestation within EU-EOMs.

Conclusion: Assessing the Significance of EU-EOMs
Neither ' electoral tourism' nor norm power mechanism; the EU and African governments have increasingly used EU-EOMs as an instrumentalised tool. This instrumentalisation offers an explanation of why the EU and African governments cooperate through election monitoring.
Remarkably little is still known about the effects of international election observers. In contributing to this emerging literature, this article seeks to deliver a nuanced but comprehensive discussion of EU-EOMs in African countries and reveals their instrumental nature. Employing a model of norm socialisation, which focuses on the interactions of all involved stakeholders, it challenges constructivist concepts of norm diffusion in terms of questioning its deterministic focus on norm entrepreneurs. Focusing on agency possessed by all involved actors, this article assumes that both African governments and the EU have the potential to engage in processes of norm diffusion. Yet, in acknowledging limited statehood (Risse and Ropp, 2013) and the fact that the EU is not a single actor, as its' actions emerge out of a dynamic process between member states and the European Commission, restrictions and constraints on agency should be considered in future research. Projecting the proposed model on an empirical analysis, quantitative and qualitative analyses demonstrate the practise of instrumentalising EU-EOMs through. Norm diffusion is demonstrably not a straightforward process determined by norm entrepreneurs, pushing for new rules and procedures. Instead, all involved actors including their benefit-maximising strategies determine the outcome of norm diffusion.
The findings also question the success of NPE in EU-EOMs. In the presented cases of Ethiopia and Côte d'Ivoire and Ethiopia, NPE is not the primary power mechanism through which the EU operates. Whilst Manners (2002) mechanisms of norm diffusion do not specify when this process is successful, the proposed model of norm socialisation determines this to be the case, when norms are internalised. During EU-EOMs this occurs when the EU is re-invited and accepts this invitation to re-observe a country's election. But, because of the instrumental behaviour of both actors, the step from institutionalisation to internalisation is not straightforward. The fact that the EU did not observe the most recent elections in Ethiopia and Côte d'Ivoire demonstrates these involved complexities.
Overall, three reasons integral to its instrumental behaviour challenge the EU's ability to diffuse norms when observing elections abroad. First, NPE would require the EU to accept all invitations, regardless of their 'political usefulness'. Yet, despite observing more elections over time, the EU is more likely to accept an invitation if that country receives foreign aid or is a politically important partner (Kelley, 2010, p. 169) and simply does not have the capacity to observe all elections it is invited to. To avoid such biases, the EU needs to strengthen its financial means. Thus, EU-EOMs are yet another example of Hill's (1993) capabilities-expectations gap, haunting EU foreign policy and future research would do well to examine whether the EU can effectively assert ' actorness' through EU-EOMs.
Second, questions around the EU's impartiality interfere with NPE. When examining EU-EOMs, one has to understand monitors predisposition, as 'the notion of neutral election observers is a myth' (Kelley, 2010, p. 168). The most prominent examples, confronting the 'supposed neutrality' of election observation were the 2000 and 2002 elections in Zimbabwe (Dorman, 2005, p. 156). As observers were drawn into the political developments, they 'had become part of the story and therefore could not be neutral' (Dorman, 2005, p. 168). Similarly, in 1999 the EU was widely criticised for assessing Nigeria's election too positively in an attempt to restore normal relations with the country (Kelley, 2010, p. 166). EU chief observer Ana Gomez also arguably stretched her mandate as a neutral observer following polling day in Ethiopia. Writing to government ministers to ask for the release of opposition candidates was not listed under the tasks of EU-EOMs. More recently, EU observers have become more careful. For example, during Ghana's 2008 election, the EU delegation did not critically engage with fraudulent voter registers (Jockers, Kohnert and Nugent, 2010, p. 103). This accentuates that there is a clear demand for greater consistency, as only if EU-EOMs evaluations are well founded, norm takers respond to the EU as a norm changer.
Third, more engagement by the EU nor African governments with the EU-EOMs recommendations are required. Just observing elections arguably has little impact on democratization processes. Even Brussels is increasingly aware of the fact that elections are set within electoral cycles, which greatly determine the conditions on election-day. Hence in each report, the EU lists a set of recommendations, which the European Commission 'will seek to follow-up on the implementation of the re-commendations ' (2008, p. 179). But commitment to follow-up on these recommendations is emerging slowly, with the EU sending follow-up missions only to selected countries (European Parliament, 2012). For instance, whilst Nigeria has invited the EU to observe every national election since 1999, electoral processes have been described as lacking international standards in 2003 and 2007 due to mismanagement by the electoral commission and statesponsored manipulation. Equally, despite regular election support, Malawi's elections have been described as 'short of fully meeting commitments to international and regional standards for elections' (EU-EOM, 2009, p. 3). Not least, Ethiopia exemplifies how recommendations can be a sensitive issues. Referring to EU-EOM recommendations, Meles Zenawi outright declared that 'merely discussing the ideas, let alone accepting them is unthinkable' (Meles Zenawi, 2005). This also proposes that further research could look at popular and elite opinions (Keuleers, 2015;Bachmann and Müller, 2015) on EU-EOMs to investigate how the EU's normative narratives are perceived in African countries.
Furthermore, one has to ask how instrumentalisation influences election monitoring. Revealing further research avenues, an approach informed by instrumentalisation indicates how host countries can choose to invite different organisations and in return these delegations can accept or decline an invitation. This reveals the instrumental behaviour involved in inviting observers and demonstrates that African elites have taken a sense of ownership of international election monitoring. Whilst the EU did not observe the most recent elections in Ethiopia and Côte d'Ivoire, other organisations were present. Different delegations also produce puzzling contradictory statements. Kelley, for example notes that 'in roughly a third of cases, monitoring missions disagreed with one another about their overall assessments ' (2010, p. 158). Especially the AU and ECOWAS 'reach conclusions about election quality that often clash with the verdicts of Western observers, not to mention the beliefs of the losing political parties' (Bratton, 2013, p. 24). For instance, in 2015 the AU was the only international observer group present in Ethiopia and therefore extremely careful about its judgement. Likewise, at the 2015 Ivorian elections ECOWAS observers, led by chief observer and former Nigerian president Olusegun Obasanjo, described the ' elections as extremely visible' (Obasanjo, 2015). Whilst AU and ECOWAS observers might have a better local understanding of elections one cannot neglect that both organisations have not developed concrete assessment standards, do not publish full reports and its chief observers are often former politicians, who have been accused of manipulating polls. Obasanjo was a military ruler in Nigeria and won the widely criticised 1999 and 2003 elections. However, the EU also openly supports AU and ECOWAS. In 2015 the EU assigned €6 million in 'support of AU election observation missions to meet international standards, and thus contribute to inclusive, transparent and credible elections in Africa' (AU/EU, 2015).
Once more, this shows the significance of election monitoring in the context of Africa-EU relations. Overall, analysing EU-EOMs though an instrumentalisation lens has deconstructed Europe's normative agenda, revealed the 'neutrality' disguise of election monitoring and emphasised the crucial role African governments play during this process. Like Desalegn, many African leaders remain critical towards international involvement in their elections but depend on the EU development assistance. Thus, the future of EU-EOMs is likely to be as contested as its' past.

Notes
1 Because EEMs are made up of small teams of election experts, the focus of this article lies on EU-EOMs, which are larger delegations that comprehensively assess elections in countries, which the EU has been invited to. 2 Interview 17.02.2016. (Interviewee wishes to remain anonymous).

Additional File
The additional file for this article can be found as follows: • Appendix. Assessments of EU-EOMs Final Reports. DOI: https://doi.org/10.22599/ujpir.13.s1