External actors may shape conflict resolution. Recently, they have helped implement power-sharing in Bosnia-Herzegovina, Iraq, Lebanon, Northern Ireland and Macedonia, and have encouraged or facilitated power-sharing in Burundi, Cyprus, the Democratic Republic of Congo, Liberia and Nepal. Compared with the classic cases of consociational power-sharing (Lijphart Reference Lijphart1977), the engagement of external actors in recent cases of post-conflict power-sharing has brought about a ‘new wave’ of power-sharing democracy (Taylor Reference Taylor2009: 7). Existing research suggests that external agents can be positive – that is, they can motivate domestic elites to share power, help implement or run new institutions, or engage in coercive state-building measures (Kerr Reference Kerr2005: 7). When international organizations and states intervene they usually promote integration or accommodation, including consociational power-sharing (O'Leary Reference O'Leary2005: 33). Although integration is usually the preferred international option, experience shows that the international community ‘has been prepared to back accommodation where that has been demanded, if only, and unfortunately, after rebellion has threatened order’ (McGarry et al. Reference McGarry, O'Leary and Simeon2008: 87). Yet a key question is: how effective are external actors in inducing political parties to share power after violent conflict? Comparing Bosnia-Herzegovina and Northern Ireland allows us to investigate how external actors help form and maintain power-sharing through their socialization efforts. The argument here relies on but adds to work on rationalist mechanisms of socialization; that is, the promotion of power-sharing by external actors and the adoption of power-sharing by the internal contenders can be explained by strategic behaviour. The latter's strategic behaviour is based on positive and negative incentives. These incentives will likely probably be more effective when they uphold a peace agreement that satisfies groups’ preferences on constitutional issues. External incentives can, under certain conditions, lead to the internalization and the potential ‘habitualization’ of power-sharing as norm-conforming behaviour (see Risse et al. Reference Risse, Ropp and Sikkink1999). Yet the strategy of external actors will be less effective when their socialization efforts are inconsistent and coercive, viewed as threatening to one or more of the contenders.
Socialization and power-sharing
The international relations and European Union literature on compliance by states helps explain the interaction between external and internal actors over policy questions. The rationalist approach focuses on cost-–benefit calculations and material incentives, while constructivists highlight persuasion and social learning. Efforts have also been made to adopt a synthetic approach ‘encompassing both instrumental choice and social learning’ (Checkel Reference Checkel2001: 581). Jeffrey Checkel (Reference Checkel2005: 804) defines socialization as ‘the process of inducting actors into the norms and rules of a given community’, leading to ‘sustained compliance based on the internalization of these norms’. Socialization occurs when actors switch from ‘a logic of consequences’ to ‘a logic of appropriateness’, which operates without direct underpinning by material incentives or sanctions. He suggests two types of internalization of norms: when conscious instrumental calculations are replaced by role playing; and when conscious instrumental calculations are replaced by ‘taken-for-grantedness’ (Checkel Reference Checkel2005: 804). Others, by contrast, stress rationalist mechanisms of socialization, including bargaining and reinforcement (Schimmelfennig Reference Schimmelfennig2000, Reference Schimmelfennig2005; for an earlier discussion, see Ikenberry et al. Reference Ikenberry and Kupchan1990). Ultimately, the outcome of socialization is norm adoption and rule-guided conduct.
This compliance literature can inform the field of conflict resolution. We can investigate whether political parties in peace processes have been subject to socialization efforts by external actors and whether they have internalized power-sharing as an appropriate post-conflict institutional arrangement. Frequently suggested mechanisms of socialization include role playing, normative suasion and strategic calculation (see Checkel Reference Checkel2005). In the promotion of power-sharing there may be second-track efforts to encourage workshops in which role playing occurs, but the focus should be on rationalist mechanisms of strategic calculation, whereby elites ‘carefully calculate and seek to maximize given interests, adapting their behaviour to the norms and rules’ favoured by the external agent (Checkel Reference Checkel2005: 809). I will draw on the work of Frank Schimmelfennig (Reference Schimmelfennig2005: 830) on ‘reinforcement’, who suggests that external actors may ‘reward norm-conforming behavior and punish norm-violating behavior’ and the subjects of such socialization efforts ‘conform with the norms and rules in order to avoid punishment and gain rewards’. The targeted endgame is ‘the habitualization of norm-conforming behaviour’, which can be explained as ‘the outcome of a stable, structurally determined equilibrium’ in which actors simply have no incentive to deviate from the institutionalized norms’ (Schimmelfennig Reference Schimmelfennig2000: 119).
The strategy of ‘reinforcement by reward’ was used in attaching conditions to membership of the European Union (Schimmelfennig et al. Reference Schimmelfennig, Engert and Knobel2003). Its success has been attributed to material bargaining mechanisms (membership in return for compliance) and to low domestic political costs to the politicians in accession countries (Schimmelfennig et al. Reference Schimmelfennig, Engert and Knobel2003). Judith Kelley (Reference Kelley2004: 40) argues that ‘the bottom line is that conditionality will only work if the cost–benefit calculations come out in favor of a policy change.’ In this sense, the external actor may act and be seen as the ‘gatekeeper for resources in the social environment which the [domestic] actor needs or desires to have’ (Schimmelfennig Reference Schimmelfennig2000: 117).
Extrapolating this thinking to a peace process, prospective resources and benefits for current antagonists in a conflict may include access to power and patronage, funding for post-war reconstruction, and the reputational – that is, rewards which peacemakers may win. External agents engaged in promoting and implementing power-sharing structures may be considered as a ‘socialization agency’. On the basis of their own cost-–benefit calculations, these external actors (international organizations or states) may decide that power-sharing offers the best option (as part of a wider settlement) for securing sustainable peace. Conditionality or ‘reinforcement by reward’ occurs when the external actor links a political reward to the adoption, maintenance or reform of power-sharing. Importantly, external actors may have different mandates, especially when there are several such agents, and may work towards different goals. What matters, however, is that they seek to induce cross-community power-sharing. Comparing their efforts may let us explain better the effectiveness of incentives. In the analysis which follows, it is assumed that internal antagonists will act instrumentally, calculating whether the benefits of power-sharing outweigh the disadvantages of conformity. They are likely to calculate how rewards can be won at minimum cost. Drawing therefore on rationalist discussions of socialization and compliance, as well as the literature on post-conflict institution-building, I hypothesize that power-sharing executives are formed and maintained when external actors’ rewards for compliance meet domestic actors’ institutional preferences.Footnote 1 In these circumstances, power-sharing may become internalized, leading to the ‘locking-in’ of power-sharing practice over time and the acceptance of the system as ‘taken for granted’.
To evaluate this understanding, the roles of external actors in two case studies are investigated here, namely, Northern Ireland and Bosnia-Herzegovina (BiH). Both have been considered instances of ‘complex consociation’ involving external actors in power-sharing promotion and peace implementation (O'Leary Reference O'Leary2005). Though the external actors in the two cases have different mandates, they have arguably operated similar strategies of socialization through ‘reinforcement’ in inducing internal elites to share power. Yet, while the British and Irish governments have celebrated the potential for sustainable peace arising from the Good Friday Agreement of 1998 and the St Andrews Agreement of 2006, external actors in Bosnia-Herzegovina continue to be embroiled in the pursuit of cross-community cooperation since the signing of the Dayton Peace Agreement (DPA) of 1995. It is surely worth seeking to explain why external agents appear to have had more success in one case than the other.
Northern Ireland
Following 30 years of significant armed conflict, Northern Ireland has had radically reduced levels of violence since 1999. A power-sharing settlement was negotiated and eventually implemented as the result of strategic interaction among the internal parties (unionists and nationalists) and two external actors – the British and Irish governments.Footnote 2 Though successive British government pursued inconsistent policies between 1968 and 1984, London and Dublin progressively reconverged in their policies towards the region after 1985, and eventually facilitated the power-sharing Good Friday Agreement. The intergovernmental role in bringing about a settlement has been widely recognized (Clancy Reference Clancy2010; Kerr Reference Kerr2005; O'Duffy Reference O'Duffy2007). They had attempted to do so before, in the brief and eventually failed power-sharing Sunningdale communiqué of 1973. But after the 1998 agreement they persisted; despite the British government's unilateral suspension of the new institutions the two governments endured and, with mild modifications to the 1998 agreement, appear to have generated a stable power-sharing executive.
In 1972, following the transfer to direct rule from Westminster, the British government published a discussion paper, The Future of Northern Ireland (Northern Ireland Office 1972). It foreshadowed power-sharing and an institutionalized role for cooperation between the two jurisdictions in Ireland. Following the June 1973 Northern Ireland assembly elections, Secretary of State for Northern Ireland William Whitelaw called the parties to Belfast for negotiations. For the Irish government and the nationalist Social Democratic and Labour Party (Reference KerrSDLP), power-sharing and the Council of Ireland were key features of a political settlement.Footnote 3 Unionists were offered a return to power within Northern Ireland, albeit shared with nationalists. Much debate took place over the distribution of seats in the executive, with unionists ultimately securing a majority and allowing party leader Brian Faulkner to win (slim) approval from the Ulster Unionist Council to participate in power-sharing (Kerr Reference Kerr2011: 104). In December the parties and the two governments convened in Sunningdale, England, to reach agreement on the details of the Council of Ireland. Faulkner found himself under pressure from Prime Minister Edward Heath, Irish Minister for Foreign Affairs Garret FitzGerald and the Social Democratic and Labour Party's John Hume. Though Faulkner (Reference Faulkner1978: 237) insisted that ‘nothing agreed at Sunningdale infringed on the powers of the Northern Ireland Assembly’ and that the deal did nothing to threaten Northern Ireland's constitutional position within the UK, the final agreement proved difficult to sell to the unionist population. Pressure mounted from the wider unionist community opposed to power-sharing and the Council of Ireland, climaxing with the Ulster Workers’ Council strike of May 1974. The British government's failure to address the strike, support the executive and defend the Sunningdale communiqué irrevocably damaged the power-sharing system (see Kerr Reference Kerr2011: ch. 11; O'Leary Reference O'Leary2004).
In the years following the collapse of the 1974 power-sharing executive, there was some policy variation, even incoherence, on the part of successive British governments towards power-sharing in Northern Ireland. Secretary of State Merlyn Rees adopted a ‘distancing’ policy towards debates within the elected Constitutional Convention of 1975 and refused to reject the convention's report that recommended a return to majoritarianism (see Kerr Reference Kerr2011: ch. 10). Subsequent Secretary of State Roy Mason made no effort to find a political solution ‘lest they heighten expectations and create ambiguity’ (O'Leary and McGarry Reference O'Leary and McGarry1993: 202). The Conservative Party manifesto for the 1979 Westminster election proposed Northern Ireland's greater integration in the UK rather than power-sharing with an Irish dimension. Once in power, the Conservatives eventually pursued policies that considered power-sharing, as in Secretary of State James Prior's proposals of 1982 for ‘rolling devolution’ – that is, incremental devolution. O'Leary (Reference O'Leary1997) suggests that, despite the policy contradictions under Margaret Thatcher's premiership, there was some policy learning; after the emergence of a significant Sinn Féin electoral presence in Northern Ireland, a role for the Republic of Ireland and power-sharing were reassessed. The Anglo-Irish Agreement of 1985 provided a formal legal role for the Republic of Ireland and encouraged a power-sharing devolved government. By the 1990s the two governments had arrived at a ‘bi-national, intergovernmental conflict regulation strategy’ (O'Duffy Reference O'Duffy2007). In the 1998 agreement, following the text of the Downing Street Declaration of 1993, they entrenched a compromise on self-determination, incentivizing both communities:
[I]t is for the people of the island of Ireland alone, by agreement between the two parts respectively and without external impediment, to exercise their right of self-determination on the basis of consent, freely and concurrently given, North and South, to bring about a united Ireland, if that is their wish, accepting that this right must be achieved and exercised with and subject to the agreement and consent of a majority of the people of Northern Ireland. (British–Irish Agreement 1998)
Among the novelties in the Good Friday Agreement is its historic compromise on the contending groups’ constitutional preferences. It is arguable that the negotiations, the drafting of the agreement and its implementation displayed ‘reinforcement by reward’ because the two governments offered the parties assurances, group protection and side-payments. The Ulster Unionist Party supported the agreement to cement Northern Ireland's constitutional status and end Irish Republican Army (IRA) violence. The agreement cemented the principle that Northern Ireland's status as part of the UK would not change without there being majority support among the people of Northern Ireland, a principle agreed by the Irish government and the two major Northern Irish nationalist parties, the Social Democratic and Labour Party and Sinn Féin. The Irish constitution was amended, so that the unification of Ireland was to be sought through the mutual consent of both parts of Ireland rather than claimed as an existing legal right (Coakley Reference Coakley2001: 236). When Ulster Unionist Party leader David Trimble looked ready to stall on permitting power-sharing to start before the Irish Republican Army decommissioned its weapons, Tony Blair drafted a side-letter promising to allow Sinn Féin exclusion from the institutions should Irish Republican Army decommissioning not take place within the first six months of the assembly (Blair Reference Blair2010: 175). For the Social Democratic and Labour Party, the Good Friday Agreement contained crucial features: internal power-sharing and an ‘Irish dimension’, the North–South Ministerial Council (NSMC). Republicans, too, were incentivized to support power-sharing. Police reform was promised. The two governments gave assurances regarding a wider equality agenda and the release of prisoners. Blair even made a ‘private oral assurance’ that the British government would bring forward prisoner release from two years to one (Blair Reference Blair2010: 175). Sinn Féin secured ‘side-payments’, including police reform and demilitarization (the withdrawal of the British army to normal garrison levels). They also could not be blocked from participation in government in the North while campaigning to strengthen their party in the South. Republicans felt able to make the argument that the power-sharing institutions offered a potential transition to a united Ireland. As Sinn Féin president Gerry Adams wrote (Reference Adams2003: 367–8), a united Ireland remained the future prize, ‘We knew from the parameters of the talks laid down by the two governments that Irish unity would not come out of this phase of the negotiations, but we set ourselves the task of weakening the British link while defending Irish national rights.’ Beyond internal power-sharing, the Good Friday Agreement created the North–-South Ministerial Council, designed to incentivize nationalists, and the British-–Irish Council to incentivize unionists.Footnote 4
After the referendums of May 1998 and the first assembly election in June 1998 the two governments continued their strategy of inducing power-sharing between unionists and nationalists when implementation difficulties were encountered. This strategy was hampered, however, by unilateral and coercive measures carried out by the British government, which at their best might have been construed as ‘reinforcement by punishment’. The key example was Secretary of State for Northern Ireland Peter Mandelson's suspension order of 2000. Recalling, in his memoirs, the decision to suspend the institutions, Mandelson (Reference Mandelson2010: 291) refers to his ‘nuclear option’ threat to Sinn Féin that in the absence of Irish Republican Army decommissioning, he would ‘reinstate direct rule from London, and would lay the responsibility for it on the IRA and Sinn Féin’. Mandelson had in fact opted for suspension to ‘save Trimble’ – that is, to prevent the need for fresh elections if the unionist leader made good his threat to resign as First Minister unless the Irish Republican Army decommissioned its weapons. Although Mandelson (Reference Mandelson2010: 297) concluded that there was ‘no alternative to suspension if I was to retain the trust of unionists’, the decision provoked protest from nationalists as well as discomfort for the Irish government. Irish Taoiseach Bertie Ahern (Reference Ahern2000) commented that suspension damaged the institutions, and stressed that the British–Irish Agreement did not provide for suspension: ‘In that context, suspension raises issues of concern for the Government and any significant extension of it could make the situation more difficult.’ For Sinn Féin, suspension precipitated a crisis ‘because a government has unilaterally and illegally broken a commitment’.Footnote 5 The Social Democratic and Labour Party warned that Mandelson had prioritized the Ulster Unionist Party's concerns ahead of other parties and that suspension ‘fundamentally weakened’ the likelihood of Irish Republican Army decommissioning.Footnote 6 The British government was severely criticized by nationalists for its handling of police reform, particularly over the May 2000 Police Bill designed to implement the recommendations made by the Patten Commission.Footnote 7 For Sinn Féin (2000), the Police Bill was ‘a major departure’ from the recommendations of the Patten Report, and the Social Democratic and Labour Party (2000) reiterated its ‘disappointment at the failure to implement unambiguously’ the Patten recommendations. Though the Social Democratic and Labour Party was ready to support the police service by August 2001, Sinn Féin continued to criticize the British government for reneging on its commitments. This controversy led to a stand-off over policing, demilitarization and decommissioning until the UK completed fully the agenda of the independent police commission, the Irish Republican Army executed its ‘acts of completion’ in 2005 and Sinn Féin officially supported the reformed police in 2007. Postponing Northern Ireland assembly elections was another instance of the British government's unilateral coercive measures (again out of step with Dublin) and was viewed by Ahern as creating ‘more problems than it solves’.Footnote 8 These examples demonstrate how inconsistencies and contradictions on the part of external actors can impair relations with the domestic parties and also aggravate inter-group relations over sensitive policy issues.
Following further suspension of the power-sharing institutions in 2002 and the Democratic Unionist Party (DUP) victory as the largest unionist party in 2003, the two governments’ strategy focused on incentivizing the latter to accept the terms of the 1998 agreement and share power with Sinn Féin, now the largest party on the nationalist side.Footnote 9 In turn, Sinn Féin was required to support the new (and improved) policing institutions. Blair's then Chief of Staff Jonathan Powell (Reference Powell2008: 276) recalls the decision to put pressure on the parties: without progress by 24 November 2006, ‘we would suspend the Assembly, cancel elections and move to a Plan B of greater Anglo-Irish cooperation.’ In the absence of agreement, Secretary of State Peter Hain threatened, ‘It is either a functioning executive with powers devolved on 26 March or it's dissolution and the politicians stop getting paid, Stormont closes down and we pack up Northern Ireland politics, maybe for years to come.’Footnote 10 At around the same time, Hain announced unpopular policies on education, local government and water charges. Though it is difficult to assess the impact of these threats, Democratic Unionist Party leader Ian Paisley suggested that this pressure did have an important bearing on his party's decision to share power with Sinn Féin, ‘We were told that if we didn't do this then it was going to be curtains for our country . . . How would I have faced my people if I had allowed this country to have the union destroyed and the setting up of a joint government by the south of Ireland?’Footnote 11 For the Democratic Unionist Party, a new agreement was enabled by Irish Republican Army decommissioning in 2005. It was this changed political and security context, bolstered by incentives, that encouraged the Democratic Unionist Party to share power with republicans (Wilford Reference Wilford2010). For Sinn Féin, proper power-sharing would have to include the devolution of policing and justice powers from Westminster. The party eventually convened a conference on policing in January 2007 with a motion to support ‘a police service which is representative of the community it serves, free from partisan political control and democratically accountable’.Footnote 12 With around 90 per cent of delegates in support, Adams declared the policy change as part of the republican strategy to achieve Irish unity.
Bosnia-Herzegovina
Strategic interaction between external and internal actors has also been a key feature of power-sharing in Bosnia-Herzegovina. Heavy external engagement by the US, under Richard Holbrooke, led to the Dayton Peace Agreement. Subsequently, external actors assumed considerable responsibility for peace implementation, including the operation of power-sharing among the three groups or ‘constituent peoples’. A major problem, however, is that the underlying dispute about self-determination is now said to be ‘frozen’ (Bieber Reference Bieber2008). Bosniaks see the entity structure, the Federation of Bosnia and Herzegovina and the Serb Republic (Republika Srpska – RS) as a reward for ethnic cleansing. They want to create a new constitution based on a unitary state. Bosnian Serbs, by contrast, resist any attempt to erode the authority of the Serb Republic and have, on occasions, called for a referendum on independence. Bosnian Croats make occasional calls for the creation of their own entity but emphatically want to retain equal power to protect their group rights. Despite the engagement of a host of external actors and a longer period of engagement, the incentives they have used to induce behavioural shifts among the protagonists in Bosnia have been largely ineffective in securing stable power-sharing, compared with Northern Ireland.Footnote 13 Indeed, the strategy of external actors has progressively shifted from implementation of the Dayton structures to seeking to reshape those structures, and that has hampered relations between the external actors and the domestic elites (particularly Bosnian Serb representatives) and among the contending groups.
The Dayton Peace Agreement is widely recognized as the outcome of external coercion. Equally, it is known that some of its domestic signatories had little influence over its content (Chandler Reference Chandler2006: 31). But there were socialization efforts on the part of the external actors and instrumental calculations by the parties representing the groups.Footnote 14 US negotiator Richard Holbrooke (Reference Holbrooke1998: 232) recalled his objectives, ‘to turn the sixty-day cease-fire into a permanent peace’ and to ‘gain agreement for a multiethnic state’. The talks were essentially about ending the war. Any settlement would have to be settled on the basis of ‘wartime structures’ with ‘two functioning levels of government’ between Sarajevo and two entities, a Bosniak-Croat Federation and a ‘Bosnian Serb entity with no claims to sovereignty’ (Holbrooke Reference Holbrooke1998: 241). External incentives to make a settlement were based on rewards for an agreement and threats of negative consequences in its absence. The Bosniak delegation sought a single state, recognizing the Serb Republic was a difficult compromise for them. Having held out on the position of the contested city of Brčko, they finally accepted a solution following pressure from the US. While Holbrooke spelled out Bosniak achievements, including a World Bank package (Holbrooke Reference Holbrooke1998: 274–5), US Secretary of State Warren Christopher warned that in the absence of a settlement, President Clinton would withdraw his support (Holbrooke Reference Holbrooke1998: 275). For the Croats, the carrots included regaining eastern Slavonia from the Serbs, cementing the federation and securing equality at the state level in the presidency and the senate. When Croat President Franjo Tudjman had stalled over agreement on the federation, in which the Croats did not enjoy full parity with the Bosniaks, the US team had reassured him they would insist upon power-sharing at the state level (Holbrooke Reference Holbrooke1998: 262–3). The Bosnian Serbs, represented by Serbian President Milosevic, sought recognition, the end of sanctions and retaining a 51/49 per cent territorial split of Bosnia between the federation and the republic, respectively. Incentives included not just an end to war but also the establishment of institutions with reward of resources or patronage for their respective supporters. Pressure was put on Milosevic. With no deal, the US would ‘shut down without an agreement’, sanctions would continue and conflict could resume (Holbrooke Reference Holbrooke1998: 280).
In the post-agreement stage, the strategy of external actors has been based on a ‘reinforcement’ mechanism of socialization. Norm-violating behaviour, as viewed by the Office of the High Representative (OHR), is met with policy imposition. It is perhaps unsurprising that external actors have relied on robust incentives, particularly given the acknowledgement that the Dayton Peace Accord was made by ‘international decree’ and that ‘[n]o-one thought it wise to submit the constitution to any sort of parliamentary or other similar proceeding’ (Bildt Reference Bildt1998: 139). In particular, the Office of the High Representative has enjoyed considerable leverage through subsequent acquisition of ‘the Bonn powers’, awarded at the 1997 meeting of the Peace Implementation Council (PIC). Under these powers the high representative is entitled to take interim measures when parties are unable to agree, as well as to take actions against politicians or officials ‘in violation’ of the Dayton Peace Agreement (Peace Implementation Council 1997). Initially, the Bonn powers were used extensively to drive ‘reform’. High Representative Wolfgang Petritsch commented that he was under pressure to use the powers, and he ‘had to act as the most interventionist High Representative ever’ (quoted in Bose Reference Bose2006: 16). High Representative Paddy Ashdown continued with significant use of the Bonn powers; indeed, it has been suggested that Ashdown ‘wielded the Bonn powers, boldly, leaving behind perhaps the strongest legacy of any HR’ (ICG 2007: 6).Footnote 15 Christian Schwarz-Schilling, Ashdown's successor, announced a shift in policy. He pledged to use the Bonn powers to ‘maintain peace and security or to further BiH's [Bosnia-Herzegovina's] cooperation with the ICTY [International Criminal Tribunal for the former Yugoslavia]. I have made it equally clear that I will not use the Bonn powers for anything else’ (quoted in ICG 2007: 6). The use of the Bonn powers has fluctuated over time and across different categories of decisions, with the peak during Ashdown's tenure and a reduction since the Peace Implementation Council's announcement in 2006 of its intention to close the Office of the High Representative (Szewczyk Reference Szewczyk2010: 35–9). The reliance on the Bonn powers has led to the suggestion that the high representative has acted like a ‘European Raj’, and to sustained opposition from domestic politicians, especially Bosnian Serbs (Knaus and Martin Reference Knaus and Martin2003). As in Northern Ireland, more coercive measures by external actors have had an adverse effect on relations between the external and internal actors. As discussed below, the strategy employed by the Office of the High Representative, including the imposition of policy reform, has been strongly criticized by domestic elites, particularly Bosnian Serb parties.
An important feature of external actors’ socialization strategy in Bosnia-Herzegovina has been the effort to incentivize elites to ‘reform’ political structures, expressed as the need to ‘improve functionality’ of the power-sharing system.Footnote 16 Many of these reforms, however, are not neutral because they propose the removal or the amendment of protections sought and won by one of the parties at Dayton. Such pressure has been bound up with the promise of accession to the EU (Noutcheva Reference Noutcheva2009). For Ashdown, progress was defined as ‘moving out of the era of Dayton and into the era of Brussels’ (OHR 2003, emphasis added). Stability and Association Agreement (SAA) negotiations with the EU were made conditional on meeting 16 policy areas, a key condition being police reform.Footnote 17 Since 2005, external efforts to reform the constitution have gathered pace. In 2006 most parties agreed on a package of amendments, dubbed the ‘April package’.Footnote 18 But the legislation failed to secure the required two-thirds majority in parliament by two votes. High Representative Schwarz-Schilling stated that agreement was necessary to ‘better equip Bosnia-Herzegovina for EU integration, and better equip government to serve its citizens’ (OHR 2006). Importantly, messages from the EU, cloaked in the language of ‘functionality’, about the transfer of competences from the entities to the state-level, favour Bosniak preferences for a stronger, centralized state. Bosniak politicians (including former member of the presidency Haris Silajdzic) have regularly suggested abolishing the entities in favour of a unitary state based on individual citizenship, their original negotiating goal at Dayton. Indeed, he rejected the ‘April package’ of 2006 on the basis that the proposals did not go far enough to reduce the powers of the entities. In addition to the accession carrot, external actors have threatened ‘reinforcement by punishment’ should agreement on reform fail to materialize. In September 2007, when a deal collapsed, High Representative Miroslav Lajčák declared that reform ‘will determine whether integration or isolation are in store’ (OHR 2007a). Lajčák imposed a decision in October 2007 to revise decision making, making it more difficult for a party or entity to block the Council of Ministers and parliament (OHR 2007b). Following an agreement among the parties in the Mostar Declaration of October 2007, the Stability and Association Agreement with the EU was signed in June 2008. Yet, despite this pressure, policing structures have not changed much, arguably demonstrating the limitations of EU conditionality in Bosnia-Herzegovina.Footnote 19
To induce reform, a number of conditions were put in place to end the mission of the Office of the High Representative. In addition to five policy objectives, two conditions had to be met: the signing of the Stability and Association Agreement and a positive assessment of the political situation ‘based on full compliance with the DPA [Dayton Peace Agreement]’ (Peace Implementation Council 2008). In November 2008 the leaders of the three main parties made progress on state property, the census, constitutional reform and the status of Brčko. Again, the Office of the High Representative (2007a) hailed the potential to deliver ‘the only important strategic issue, one that everyone in Bosnia-Herzegovina agrees on, and it is the road to Euro-Atlantic integration’. In October 2009 dialogue led by the EU presidency and the US, the ‘Butmir process’, intended to make progress on constitutional reform and meeting conditions for closing the Office of the High Representative; however, it did not materialize. It has been suggested that the EU and US teams were ill-prepared, that the timing was inadvisable given the run-up to elections in 2010, and that there was too much belief in the carrot of enlargement to overcome domestic elites’ political preferences.Footnote 20 Though the European Commission (2009) has stated that reform ‘to permit [state-wide] institutions to function effectively is necessary before the Commission can recommend the granting of candidate status’, the EU delegation also professes that no part of the acquis communautaire mandates a particular institutional configuration in member states, and that centralized and decentralized structures can complement one another provided the system works effectively.Footnote 21 Nevertheless, divergent preferences among the groups make efforts to ‘improve functionality’, never mind wholesale constitutional change, a serious challenge. Even in the aftermath of the European Court of Human Rights ruling in December 2009, which found the constitutional restrictions upon access of ‘others’ to the collective presidency negotiated at Dayton to be in violation of the European Convention of Human Rights, progress has remained elusive. An EU official suggests that the parties have dragged their heels on complying with the ruling as it would open a ‘Pandora's box’, necessitating major revision of the Dayton structures.Footnote 22
The efforts by the EU and the international community to ‘move beyond Dayton’ and realize the ‘dream of Europe’ have, of course, been resisted by Bosnian Serb parties, but also by Bosnian Croats. Bosnian Serbs view calls for institutional reform as overt attempts to strengthen state institutions and to weaken their group rights and the autonomy of the Serb Republic secured under Dayton. They resisted police reform that would abolish the republic's force and create police regions across entity boundaries. Interpreting the proposals as a threat to their sovereignty, Serb Republic President Dragan Cavic declared: ‘the RS is a result of a four-year fight of the Serb people . . . not a gift of the international community’ (quoted in ICG 2005: 8). The republic's Prime Minister Milorad Dodik announced that his position on police reform would remain unchanged ‘even if that means giving up the SAA [Stability and Association Agreement] negotiations’ (quoted in ICG 2007: 19). Bosnian Serb reaction to external incentives has been increasingly characterized by confrontation between Dodik's party, the Alliance of Independent Social Democrats (Savez Nezavisnih Socijaldemokrata – SNSD) and the EU's office. Bosnian Serb politicians have become increasingly robust, shown in Dodik's threat to bring 50,000 Serbs to Sarajevo to demonstrate against the high representative's threat to remove him from power (ICG 2007: 7). In October 2007 the imposition of reform by High Representative Lajčák led to the resignation of the Bosnian Serb chair of the Council of Ministers, Nikola Spiric. In June 2009, following pressure by the Office of the High Representative on the Serb Republic to retract declarations opposing the transfer of competences, Dodik threatened a walk-out from institutions and a referendum on the Bonn powers (ICG 2009a: 3). In September 2009 Dodik again threatened to remove Serb politicians from the state institutions should the Office of the High Representative seek to impose further measures (ICG 2009a: 2). In the context of the Butmir process, he made progress conditional ‘on agreement the RS has the right to secede after a three-year waiting period’ (ICG 2009b: 5). Adding to the stand-off, the Serb Republic National Assembly passed a resolution approving a referendum on the state court and the Office of the High Representative's authority. It was only with a visit to Banja Luka in May 2011 by the EU's High Representative for Foreign Affairs Baroness Ashton that Dodik appeared to back down and the situation was (at least temporarily) diffused.
Conclusions
Overall, these case studies corroborate the hypothesis that power-sharing is formed and maintained when external actors’ rewards for compliance meet internal actors’ minimum constitutional preferences in the post-conflict environment. In both cases the use of ‘reinforcement by reward’ by external actors to induce power-sharing is apparent. In Northern Ireland, despite earlier variation (the British government's lack of commitment to the Sunningdale communiqué and contradictions in policy during the mid- to late 1970s and early 1980s), the two sovereign governments have been crucial in arriving at a power-sharing settlement, and in maintaining it. For unionists, the 1998 Agreement secured Northern Ireland's position in the UK on the basis of majority consent. For nationalists and republicans, the agreement offered inclusion in political institutions and the possibility of securing Irish unity in the future. Following Irish Republican Army decommissioning in 2005 and the St Andrews Agreement of 2006, the Democratic Unionist Party was incentivized to enter a power-sharing executive including Sinn Féin (see Mitchell et al. Reference Mitchell, Evans and O'Leary2009). In turn, Sinn Féin secured the transfer of policing and justice powers to Northern Ireland, portrayed by the party as another step in Irish self-determination, the removal of British rule in Ireland and ultimately to a united Ireland. The Northern Ireland case suggests that effective external incentives can lead to norm internalization and potential habitualization over time as power-sharing appears largely ‘taken for granted’. As Rick Wilford notes, even though the Democratic Unionist Party might favour a different form of power-sharing to that under the 1998 agreement, party leader Peter Robinson ‘is pragmatic enough to realise that it is not on the horizon and that his colours and those of his party are therefore nailed indefinitely to the fully inclusive power-sharing mast’ (Wilford Reference Wilford2010: 153).
The evidence from the cases also demonstrates that external actors’ efforts in inducing power-sharing will be hampered if the groups’ institutional and constitutional preferences are denied, and if the external agents are seen systematically to favour one of the domestic parties’ preferred model of the state. Any incentives they propose will be less effective when they are perceived to be pushing an institutional reshaping against the preferences of one of the protagonists. In Bosnia-Herzegovina, the key external actors, the EU and the US, have shifted in their socialization efforts from upholding the Dayton structures to promoting (and sometimes imposing) institutional transformations that significantly modify the core compromises in the Dayton Agreement. While the external actors initially sought to incentivize the effective workings of the Dayton Peace Agreement (often using the Bonn powers to do so), they have increasingly sought to move beyond Dayton, transferring powers from the entities to the state and seeking to modify the rules governing power-sharing in the state presidency and the senate. Thus, their efforts appear to be weakening the federal and consociational structures under Dayton (which secures group rights) to a more integrationist agenda, in which majority-rule democracy is seen as the European norm. Power-sharing, far from becoming ‘habitualized’, has been de-stabilized because Bosniaks have felt incentivized to challenge the power-sharing arrangements, and Serbs have gone back to proposing secession, partially in response. Symptomatic of these difficulties have been the Serb Republic National Assembly's resolution to hold a referendum on the Office of the High Representative's authority, the delay in executive formation following the 2010 elections and the decision of two Croat parties to create a Croat national assembly within the federation (see ICG 2011). The carrot of accession to the EU has proven ‘insufficiently tasty to break the mould in Bosnia’. All significant ‘reforms’ have come about as the result of imposition (ICG 2007: 19). Strained inter-group relations are brought into stark relief by the international community's difficulty in achieving reform, with one community opposed to any tinkering of their rights under Dayton and another hoping for greater intervention to do just that. External incentives will continue to be stretched as Bosnian Serbs question the legitimacy of the international community and demand the closure of the Office of the High Representative while Bosniaks favour the maintenance of international engagement and the creation of a centralized state.
Turning to broader policy lessons, the evidence suggests that the use of negative incentives, even coercive measures that may be termed ‘reinforcement by punishment’, have significant limitations unless they are perceived as balanced. Though it is difficult to assess the impact of negative threats, it is clear that more coercive measures are a risky strategy for an external actor engaged in promoting power-sharing and peace implementation. Serious political problems may arise when external forces are seen to break their commitments and take illegitimate decisions, unauthorized by the agreement, including policy imposition. Such coercion is likely to lead to deadlock and stand-offs between the external and internal actors and the potential destabilization of what might already be a fragile peace. In Northern Ireland, the British government's coercive decisions added to political tensions, exemplified by the debacle surrounding Mandelson's suspension legislation in 2000. Interestingly, Blair (Reference Blair2010: 189) acknowledges the problem of government incoherence: ‘there were fudges, things said and done that had little intellectual or political consistency except that of seeing us through each set of obstacles.’ In Bosnia-Herzegovina, the use of the Bonn powers has progressively worsened relations between the international community and Bosnian Serb elites.
Consistency obviously matters too. In Northern Ireland, the British government's decision to suspend the power-sharing institutions (Mandelson's ‘nuclear option’) and its initial decision not to implement the full set of proposals on police reform aggravated an already difficult political situation and hampered its relations with nationalists and republicans. In Bosnia-Herzegovina, external actors have also been inconsistent. The adoption and usage of the Bonn powers modified the initial agreement; the shift from the Office of the High Representative to EU-led engagement has transferred the locus of the lead external actor; and the internationals have shifted from implementing the Dayton agreement to modifying it unilaterally. External engagement has not been fully coherent, influenced by fluctuating political demands, the preferred approach of successive high representatives and the fluid arrangements in moving towards the closure of the Office of the High Representative.Footnote 23 The dynamics of this shift have led some domestic actors, particularly politicians from the Serb Republic, to question the legitimacy of the international community. These cases also enable us to appreciate the socialization mechanism of strategic calculation or ‘reinforcement’. External actors’ engagement may be inconsistent, among themselves, and over time, and such inconsistency (however unintentional) is likely to lead to political crisis and a slowing down of progress on peace implementation. While external actors may employ ‘socialization through reinforcement’, this depiction of agents in the literature may be overly rationalist. The politics of ‘muddling through’ (Lindblom Reference Lindblom1959) is evident in both Northern Ireland and Bosnia, and though muddling through may have its political successes, it also has its failures.
Acknowledgements
Part of this research was carried out while I was a Sawyer Mellon Postdoctoral Fellow at the Penn Program in Ethnic Conflict at the University of Pennsylvania. I wish to thank the Andrew W. Mellon Foundation for funding this fellowship, research and travel.
