INTRODUCTION
With his successful coup d’état on 25 July 1996, the former Burundian president Pierre Buyoya brought the ongoing regional mediation efforts to a standstill. After the military takeover, violence increased in the second half of 1996: the number of battle-related casualties and refugees rose, and the fighting spread geographically (Hoskins et al. Reference Hoskins, Nutt and Watson1997; Ngaruko & Nkurunziza Reference Ngaruko, Nkurunziza, Collier and Sambanis2005: 47–61; UCDP 2014). The neighbouring countries reacted swiftly and agreed to ‘exert maximum pressure on the regime in Bujumbura including the imposition of economic sanctions’ (Second Regional Summit on the Burundi Conflict 1996). They demanded the immediate restoration of Burundi's National Assembly and the reinstatement of political parties as well as peace negotiations. Simultaneously, the international donors froze all assistance other than emergency aid.
The subsequent unfolding of the sanctions episode against Burundi was characterised by two seemingly contradictory developments: On the one hand, the regime resisted the external coercion, to the surprise of the sanction senders, who, given Burundi's economic and geopolitical vulnerability, had expected to quickly force Buyoya into negotiations. The embargo's initial effect on Burundi was indeed harsh; it further paralysed the already war-torn economy (Khadiagala Reference Khadiagala and Boulden2003; Kamungi et al. Reference Kamungi, Oketch, Huggins, Huggins and Clover2005). This raises the question of how the Buyoya government managed to withstand the pressure of one of the most comprehensive sanction regimes since the end of the Cold War. In addition to sanction-busting activities, the regime launched a vocal campaign against the sanctions that (1) helped it to lobby for exemptions, (2) contributed to it regaining some international legitimacy so that the international donor community renewed its engagement, and (3) shifted the blame for economic problems to the embargo.
On the other hand, Buyoya eventually engaged in the kind of regionally mediated unconditional negotiations with all parties to the conflict that he had previously rejected. According to the assessment of several scholars, the sanctions contributed significantly to reviving the peace process because of their severe economic consequences (Ndikumana Reference Ndikumana2000; Lemarchand Reference Lemarchand, Khadiagala and Lyons2001: 91; Maundi Reference Maundi, Sriram and Wermester2003: 341). However, the fact that the regime agreed to the Arusha negotiations when the sanctions' economic impact had withered casts doubt on this interpretation. This paper argues that the controversies surrounding the sanctions shaped political contestation between the Buyoya government and the political oppositionFootnote 1 beyond the embargo's mere financial effect. The Buyoya government became argumentatively ‘self-entrapped’ in its own diplomatic campaign against the sanctions, which stressed the government's willingness to engage in peace talks. This paradox underscores the need to go beyond the assessment of how costly sanctions coerce regimes into compliance and to seriously examine their signalling dimension. Against the backdrop of the embargo's severe economic consequences, the sanctions against Burundi serve as a ‘critical case’ (Flyvbjerg Reference Flyvbjerg2006: 230) with which to explore the importance of the signals sent by sanctions.
The paper's findings are based on 34 semi-structured interviews with diplomats, policy-makers, military personnel and journalists conducted in Burundi in August and September 2013. These interviews were crucial to understanding the respective actors' perceptions of sanctions as guiding their responses to the regional and international pressure. Press releases from political parties and sympathiser groups as well as official documents collected in Burundi were used to supplement the analysis whenever they were available.Footnote 2 Although the sanctions were imposed on Burundi during the late 1990s, they still constitute an integral part of the country's (diverse) contemporary historical narrative(s) and were thus remembered by all interviewees in great detail. Moreover, former political elites who are no longer in office were able to speak more openly about past events than they could have done at the time the sanctions were in place (see also Vorrath Reference Vorrath, Thomson, Ansoms and Murison2012).
The reminder of the paper is organised as follows. The first section sketches limitations of the conventional sanctions literature with its emphasis on economic coercion and proposes an alternative framework that focuses on signalling, symbolic and political aspects of sanctions. The next section explores how the Buyoya regime initially withstood the economic pressure of the embargo with an aggressive and fairly successful anti-sanctions campaign but then became increasingly self-entrapped in the campaign's rhetoric. Finally, the conclusion discusses the findings in light of past studies on the embargo against Burundi and the current research on sanctions more generally.
INTERNATIONAL SANCTIONS: ECONOMIC COERCION OR POWERFUL SIGNALS?
Traditional approaches suggest that the purpose of sanctions is ‘to bring about policy change in the targeted nation through imposing the severest possible economic harm’ (Kaempfer & Lowenberg Reference Kaempfer and Lowenberg1988: 786; see also Barber Reference Barber1979; Doxey Reference Doxey1980). The economic pain inflicted upon the target has indeed been identified as one of the most reliable predictors of sanctions success. Hufbauer et al. (1990: 47–8) found that high costs to the target – operationalised as a percentage of the GNP – are positively correlated with sanctions effectiveness, which was repeatedly confirmed in subsequent large-N research (Drury Reference Drury1998; Morgan & Schwebach Reference Morgan and Schwebach1997; Nooruddin Reference Nooruddin2002; for a diverging finding, see Tsebelis Reference Tsebelis1990). In an attempt to quantify the impact of the financial consequences on sanctions success, Morgan et al. (Reference Morgan, Bapat and Krustev2009) estimate the success rate of sanctions, which stands at 38 per cent when costs are minor, to increase to 94 per cent when the costs to the target are severe. Likewise, Allen (Reference Allen2005) shows that comprehensive sanctions reduce the time period after which the target concedes compared with less severe measures. Further specifying these insights, scholars argue that the economic leverage of the sender rather than the cost to the target as such increases the effectiveness of sanctions (Dizaji & van Bergeijk Reference Dizaji and van Bergeijk2013). In that context, the degree to which a target previously relied on the sanctions sender for imports appears to affect its vulnerability (Dashti-Gibson et al. Reference Dashti-Gibson, Davis and Radcliff1997; Hart Reference Hart2000). In contrast, third-party assistance from so-called black knights reduces the exposure to external pressure (Hufbauer et al. Reference Hufbauer, Schott and Elliott1990: 45–6; see also Drury Reference Drury1998; Early Reference Early2009; McLean & Whang Reference McLean and Whang2010). Finally, economically health targets are less susceptible to international sanctions (Drury Reference Drury1998; Jing et al. Reference Jing, Kaempfer and Lowenberg2003).
Two mechanisms of how costly sanctions induce a target to change its policies can be discerned. On the one hand, higher costs place stronger pressure on the target's government (inter alia Dashti-Gibson et al. Reference Dashti-Gibson, Davis and Radcliff1997; Hart Reference Hart2000; Dizaji & van Bergeijk Reference Dizaji and van Bergeijk2013). If leaders of the targeted regime act on the basis of rational cost-benefit analyses, they should ‘find it harder to justify resisting the sender's pressure as the costs of the sanction go up’ (Nooruddin Reference Nooruddin2002: 69; see also McGillivray & Stam Reference McGillivray and Stam2004). One the other hand, the economic harm caused by severe sanctions is thought to bring about political disintegration because the targeted society is only willing to bear the financial damage up to a certain level (Kerr & Gaisford Reference Kerr and Gaisford1994). Recent approaches argue that the costs imposed on a country's ruling coalition rather than the overall economic consequences are crucial, particularly in non-democratic regimes (Lektzian & Souva Reference Lektzian and Souva2007; Escribà-Folch & Wright Reference Escribà-Folch and Wright2010).
In a nutshell, this research can be summarised as follows: When sanctions work, this should be attributable to the economic pain inflicted upon the target. Furthermore, sanctions should be most effective when they hurt key support groups. In other words, if targeted regimes, especially authoritarian ones, manage to protect their selectorate from financial damage (Bolks & Al-Sowayel Reference Bolks and Al-Sowayel2000; Escribà-Folch & Wright Reference Escribà-Folch and Wright2010), they should be able to resist the pressure from sanctions.
When the sanctions against Burundi were implemented, the Regional Initiative's expectation was indeed that, given Burundi's economic vulnerability, the weight of the embargo would force the government to resume the type of unconditional negotiations they demanded (Khadiagala Reference Khadiagala and Boulden2003: 227; Khadiagala Reference Khadiagala2007: 126–127). Most scholars agree that sanctions ultimately played a major role in pressuring the Buyoya government into the Arusha negotiations due to their severe costs. Ndikumana (Reference Ndikumana2000: 435) exemplarily attributes the sanctions' success to ‘mounting domestic frustration with deteriorating economic conditions’ (see also Lemarchand Reference Lemarchand, Khadiagala and Lyons2001: 92; Maundi Reference Maundi, Sriram and Wermester2003: 341).Footnote 3 However, the regime initially managed to withstand the embargo's economic pressure with the help of an anti-sanctions campaign that has been characterised as highly dynamic, aggressive and fairly effective (Bunting et al. Reference Bunting, Mwansasu and Bugoya1999). The regime's crucial supporters, particularly urban businessmen and (former) high-ranking civil servants, were shielded from the costs of sanctions or even profited from the opaqueness surrounding the trade of now scarce resources (Ngaruko & Nkurunziza Reference Ngaruko and Nkurunziza2000).
Despite its ability to adapt to the economic constraints set by the embargo, the government eventually responded to the sanction senders' key demand to engage in unconditional, inclusive and regionally mediated peace talks. Hoskins et al. (1997: 81) hypothesise that the ‘progress eventually made may be connected less to economic sanctions than to Burundi's isolation from the international community’. In a similar manner, Khadiagala (Reference Khadiagala2007: 124) stresses that sanctions not only affected the regime economically but also signalled disapproval of the coup and thereby denied the new government legitimacy without, however, spelling out how this development contributed to bringing about the regime's concessions.
Another strand of research, which has developed mostly in isolation of the quantitative studies' emphasis on the financial pain inflicted upon the targets, highlights that sanctions also serve the function of expressing the senders' (moral) disapproval of the targeted regime (Galtung Reference Galtung1967; Lindsay Reference Lindsay1986; Nossal Reference Nossal and Väyrynen1999). This signalling impact of sanctions has recently received revived attention (Crawford & Klotz Reference Crawford and Klotz1999; Grebe Reference Grebe2010; Grauvogel & von Soest Reference Grauvogel and von Soest2014). In addition to coercing a target into compliance or constraining it, sanctions send a message to targets and audiences (Giumelli Reference Giumelli2010, Reference Giumelli2011). However, both past and current accounts of sanctions symbolism are based on a problematic distinction between instrumental and signalling purposes of sanctions (inter alia Lindsay Reference Lindsay1986; Morgan & Schwebach Reference Morgan and Schwebach1997; for an early critique, see Baldwin Reference Baldwin1999: 102). Going beyond this dichotomy, Lopez & Cortright (Reference Lopez and Cortright2002: 16, emphasis added) argue in pointed terms that ‘the symbolic or signaling purposes of sanctions may be less measurable (…), but they can be important to achieving the sender's goals’. In other words, the signals conveyed by sanctions may also contribute to the target's behavioural change. This potential mechanism of how sanctions work has not yet been sufficiently explored. In the following, I propose a framework for exploring how messages sent by sanctions may result in concessions by the targeted regime that draws on the notion of argumentative self-entrapment.
Such an ‘argumentative self-entrapment’ (Risse Reference Risse1999: 531) occurs when states that are facing international criticism and demands to act in line with certain norms begin ‘talking the talk’, meaning that they discursively comply with an internationally backed and promoted discourse of democratic governance. Governments confronted with international pressure as a result of norm-violating behaviour initially tend to increase domestic repression and deny the validity of international claims while stressing national sovereignty to discourage international interference. When the pressure mounts, national governments adjust strategically to (re-)gain access to foreign aid and secure their rule vis-à-vis domestic opposition movements – for example, by making limited concessions to those advocating better protection of human rights and democratic governance (Risse & Sikkink Reference Risse, Sikkink, Risse, Ropp and Sikkink1999: 10). Tactical commitments to human rights and democracy initially appear to be a low-cost strategy for responding to criticism, especially ‘for repressive states that come under close scrutiny’ (Hafner-Burton et al. Reference Hafner-Burton, Tsutsui and Meyer2008: 117). Accordingly, many states exhibit a certain degree of norm-conforming rhetoric or behaviour in order to silence demands for real change (Hathaway Reference Hathaway2002).
However, limited concessions have been found to trigger developments that ultimately induce the governments to offer changes in behaviour from which they then find it hard to extricate themselves (Risse & Sikkink Reference Risse, Sikkink, Risse, Ropp and Sikkink1999). A state that ‘faces (actual or potential) sanctions and suffers a process of ‘shaming’ in which norm-violating states are portrayed as “pariah states”’ is increasingly forced to justify its behaviour in front of international and domestic audiences (Muñoz Reference Muñoz2009: 43). When norm-violating governments accept the norms rhetorically in order to decrease the international and domestic pressure on them, they embark upon a process in which they are measured against their verbal concessions at later stages. Instrumental commitments then backfire because domestic or international actors are able to hold them accountable, especially when the governments no longer deny the validity of the international demands per se, but rather claim that international actors misinterpret the domestic situation and hence engage in a public debate regarding the latters' critiques. Superficial commitments may translate into profound concessions over time when initial reforms spark further demands for change (Goodman & Jinks Reference Goodman and Jinks2008), especially if such reforms are interpreted as a partial success by opposition groups, which strengthens them and motivates additional demands (McAdam Reference McAdam2010).
THE REGIONAL EMBARGO AGAINST BURUNDI
The struggle surrounding the sanctions against Burundi was shaped by long-standing internal and regional dynamics. From 1966 to 1993 Burundi was governed by three military regimes under the unitary party Union pour le Progrès national (UPRONA, Union for National Progress). The ruling elite, composed exclusively of the Tutsi ethnic minority, repressed its citizens to sustain its grip on power (Uvin Reference Uvin1999). After another violent outbreak against the Hutus in 1988, the international community increased pressure on the then-president Buyoya, who eventually paved the way for the introduction of a multiparty system in the early 1990s (Laely Reference Laely1997; Ngaruko & Nkurunziza Reference Ngaruko and Nkurunziza2000).
After losing the 1993 elections, which were widely praised as free and fair, Buyoya accepted defeat and ceded power to Melchior Ndadaye, the candidate of the major opposition party, Front pour la Démocratie au Burundi (FRODEBU, Front for Democracy in Burundi) (Young Reference Young1996). The new government made attempts at coalition and confidence building with the former ruling elite (Reyntjens Reference Reyntjens1993; Lemarchand Reference Lemarchand1994; Sullivan Reference Sullivan2005), but also initiated the ‘Frodebisation’ of the civil service (Reyntjens Reference Reyntjens2000: 13) and promoted the redistribution of national resources (Reyntjens Reference Reyntjens1993: 579) as well as the return of Hutu refugees (Prunier Reference Prunier1994), thereby creating anxiety among the Tutsi population.
Moreover, the transition process was undermined by the former elite's continued control over the army, which became an important veto player in domestic politics and hampered democratic gains (Reyntjens Reference Reyntjens1993; Lemarchand Reference Lemarchand1994; Bratton & van de Walle Reference Bratton and van de Walle1994). After the elections, elements of the armed forces assassinated President Ndadaye on 21 October 1993. These events resulted in a ‘creeping coup’ (Reyntjens Reference Reyntjens2006: 117), which de facto shifted power back into the hands of the military. Subsequently, the new Hutu-dominated rebel movements Conseil National pour la Défense de la Démocratie (CNDD, National Council for the Defence of Democracy) and its armed wing Front pour la Défense de la Démocratie (FDD, Front for the Defence of Democracy) as well as the Union pour la liberation nationale (ULINA) were created in 1994 and 1996, respectively, and existing groups renewed their activity (Ndikumana Reference Ndikumana2000).
In reaction to this conflict and the accompanying threat to the region's stability, the presidents of Rwanda, Uganda, Tanzania and Zaire launched the ‘Great Lakes Regional Peace Initiative for a negotiated peace in Burundi’ (henceforth, Regional Initiative) in November 1995 and appointed the former Tanzanian president Julius Nyerere as mediator. At the same time, UN Secretary-General Boutros-Ghali proposed the deployment of a standby force of 25,000 soldiers in Burundi (Dupont Reference Dupont1997). This proposal was turned down on 5 March 1996 by the UN Security Council in Resolution 1049, which expressed preference for regionally led peace efforts (Neethling Reference Neethling2005). In June and July 1996, the regional mediation culminated in a series of summits, which advocated a negotiated settlement and national reconciliation but also considered implementing an arms embargo and visa bans for those undermining the peace process (Daley Reference Daley2007). The situation in Burundi escalated when President Ntibantunganya, fearing a military takeover of power, fled to the US embassy (New York Times 27.7.1996). On 25 July 1996, the army installed Buyoya in a bloodless coup d’état (Nsanze Reference Nsanze2003: 423).
Only six days after the coup, the heads of state of Ethiopia, Kenya, Rwanda, Tanzania, Uganda, Zaire and Zambia as well as Secretary-General Salim Ahmed Salim of the Organization of African Unity (OAU) met in Arusha. Using the framework of the Regional Initiative, they strongly condemned the coup and imposed sanctions to achieve a return to constitutional order and ‘immediate and unconditional negotiations with all parties to the conflict’ (Second Regional Summit on the Burundi Conflict 1996). Nyerere, who had already suggested sanctions against Burundi before the coup and felt that the military takeover had undermined his mediation efforts, forcefully articulated the pro-sanctions case (Wolpe Reference Wolpe2011). The regional leaders affirmed his approach as they feared a further spill over of the conflict, which had already led to an influx of Burundian refugees to the neighbouring countries (Nimubona Reference Nimubona, Deslaurier and Juhé-Beaulaton2007: 505). In addition, most heads of state had grown impatient with the former (Tutsi) elite's reliance on unconstitutional means to hold on to power (Wolpe Reference Wolpe2011). A regional sanctions coordinating committee (RSCC) was established to monitor the sanctions' effects and humanitarian impact (ReliefWeb 1996).
The international community at first supported the Regional Initiative. In line with preceding international endorsements of the regional attempts to address the crisis in Burundi, the UN Security Council adopted a resolution on 30 August 1996 that welcomed the regional peace endeavours (UN Security Council 1996). Western governments also condemned the coup d’état and initially supported the sanctions (Prunier Reference Prunier1997: 16). US Department of State spokesman Nicholas Burns stressed that his government would ‘work to isolate any government that emerges in Bujumbura by extra-constitutional means’ (Inter Press Service 31.7.1996). Accordingly, the United States refused to recognise the new government and suspended bilateral assistance (Lomasney & Halperin Reference Lomasney and Halperin1998). Likewise, the EU expressed ‘support for the regional leaders’ (European Union 1996) and cancelled its financial assistance to the post-coup government (Ngaruko & Nkurunziza Reference Ngaruko, Nkurunziza, Collier and Sambanis2005: 53–54). This stance was still maintained at the beginning of 1997, when the EU envoy to the Great Lakes region, Aldo Ajello, told reporters after a meeting with Buyoya that ‘the behaviour of the authorities in Burundi had not encouraged the international community to remove the sanctions’ (ReliefWeb 1997).
Burundi's geographic location, its close economic ties with those who implemented the embargo and its dependency on bi- and multilateral aid magnified the sanctions’ economic impact. Being landlocked, Burundi exported and imported most goods through the port of the Tanzanian capital Dar es Salaam and, to a lesser extent, through the Kenyan port of Mombasa. Furthermore, Burundi depended heavily on official development assistance, which accounted for almost 30% of its gross national income in the three years preceding the sanctions (World Bank 1996). Dwindling revenues from the blocked export of Burundi's major cash crops, coffee and tea, additionally limited the government's financial room for manoeuvre (Banque de la République du Burundi 1996). According to a survey carried out in the capital Bujumbura, the cost of living rose ‘alarmingly […] due to the economic blockade’Footnote 4 (Banque de la République du Burundi 1996: 31). The sanctions and the ongoing civil war led to a 36% increase in the general price index during the first year after their imposition (Kamungi et al. Reference Kamungi, Oketch, Huggins, Huggins and Clover2005). However, the sanctions’ economic impact did not trigger large-scale protest against the Burundian government as anticipated by the sanction senders and the domestic opposition (Ndayisaba 2013 int.)Footnote 5 nor did the financial pain cause the government to yield to the senders’ demands at first.
Withstanding Costly Sanctions
The regime launched a campaign against the embargo that targeted both domestic constituencies and the international community. The Committee for the Defence of Economic and Social Interests (CODIES – Comité pour la Défense des Intérêts Économiques et Sociaux), which comprised crucial economic actors including the Chamber of Commerce, the Belgian-Burundian and Franco-Burundian chambers of trade, various unions, and the Association of Burundian Banks, initiated and coordinated a number of activities. A press agency was created to inform the national and international public ‘about […] the new government's objectives’ as well as the disastrous effects of the embargo (Comité pour la Défense des Intérêts Économiques et Sociaux 1996). The CODIES also discussed filing a case before the International Court of Justice and presenting official complaints at the UN Security Council and the World Trade Organization. To support these efforts, the staff at the permanent missions in Geneva, New York and Brussels should be expanded and provided with argumentative guidelines (Chambre de Commerce d'Industrie, d'Agriculture et d'Artisanat du Burundi 1996). Domestically, the new government denounced the sanctions as an unjust punishment and appealed to the Burundians’ national solidarity and pride. For example, it defiantly nicknamed a new boulevard constructed in the capital during this period ‘Avenue de l'embargo’ (Buyoya 2013 int.).
The campaign against the sanctions put forth two major arguments: First, the process of deciding and implementing the embargo was condemned as a breach of international law because ‘only the United Nations can impose sanctions’ (Mbonimpa 2013 int.). According to this reading, the sanctions constituted an exclusively Tanzanian initiative, which allegedly pursued a secret agenda of weakening ‘the proud and ancient nation of Burundi’ (Nyamoya Reference Nyamoya1997). Nyerere's disproportionate engagement for the FRODEBU and the perception that he was using the sanctions to bring Buyoya back not to the peace process in general but specifically to his regional negotiation table in Arusha attracted major criticism (Peterson Reference Peterson2006; Daley Reference Daley2007). Consequently, the Buyoya government refused to accept Nyerere as the mediator (Mthembu-Salter Reference Mthembu-Salter2002).Footnote 6 By shifting the blame for the imposition of sanctions to Tanzania and to Nyerere's desire to punish the post-coup government for interrupting his mediation efforts, Buyoya made sure that his core constituencies – namely, the army and the urban Tutsi elite – blamed them rather than his coup for the sanctions (Economist Intelligence Unit 1997).
In response, the Regional Initiative sought the Western governments’ support to demonstrate the broad international approval of the embargo (ReliefWeb 1996; Second Regional Summit on the Burundi Conflict 1996). Likewise, FRODEBU activists and other proponents of the embargo stressed its genuinely regional nature (Bamvuginyumvira 2013 int.; Ndorimana 2013 int.; Ngendahayo 2013 int.) and the fact that ‘the region could not have taken the decision [to impose sanctions] without the support of the United Nations’ (Mbonerane 2013 int.).
Second, the regime stressed the humanitarian impact of the sanctions. According to the government, the sanctions disproportionately hit the poor, leading to a ‘shortage of essential elements for the[ir] survival’ (Banyiyezako 2013 int.). The campaign also emphasised the embargo's impact on the broader population, stressing that ‘FRODEBU claimed to represent eighty per cent, but these eighty per cent were suffering from sanctions’ (Ntahuga 2013 int.). In line with this strategy, a state-controlled NGO, the Association for the Preservation of Peace in Burundi, filed a case before the African Commission on Human and Peoples’ Rights denouncing the sanctions’ lack of proportionality (Olinga Reference Olinga2005). In addition, sanctions were made responsible for all economic and social difficulties in an attempt to divert attention away from the regime's own mismanagement (Hoskins et al. Reference Hoskins, Nutt and Watson1997), a military budget that absorbed about half of the government's revenues and the disastrous effects of the civil war (Economist Intelligence Unit 1998).
The government campaign regarding the humanitarian impact of sanctions was increasingly echoed by NGOs and UN agencies. Despite exemptions for the import of humanitarian goods, the embargo complicated their work and turned them into key players in the anti-sanctions discourse (Bruderlein Reference Bruderlein1998; Khadiagala Reference Khadiagala and Boulden2003: 230). In response to the growing concern about the humanitarian impact of sanctions, the RSCC created a list of humanitarian exemptions on 24 October 1996 (Regional Sanctions Coordination Committee 1996). On 16 April 1997, further items were added to the list (Hoskins et al. Reference Hoskins, Nutt and Watson1997). Moreover, multilateral assistance continued to a certain extent. The World Bank channelled its funds through domestic NGOs such as Twitezimbere (Nyandwi 2013 int.), and many programmes officially shifted their focus towards emergency assistance in order to continue their work (Nduwimana 2013 int.). Furthermore, Western governments began to voice doubts about the embargo's ability to positively affect the situation in Burundi (Daley Reference Daley2007; Wolpe Reference Wolpe2011). As one of the first European countries, France supported Buyoya in demanding that regional actors should lift the embargo (Cornwell & de Beer Reference Cornwell and de Beer1999; Manirakiza Reference Manirakiza2007: 61–76) and by mid-1998, the UN, the EU and the US also suggested that the sanctions should be reconsidered (Mthembu-Salter Reference Mthembu-Salter1999).
In addition to officially granted exemptions, the increase in sanction-busting activity played a major role in reducing the economic impact of the sanctions. Air corridors were established between Bujumbura and other African countries, as well as Belgium,Footnote 7 with the help of the government, which endowed Air Burundi with the exclusive right to explore the possibilities for the regular transport of commodities and passengers (Bulletin Officiel du Burundi 1997a). These air corridors were supplemented by a significant amount of cross-border smuggling with neighbouring countries (Hoskins et al. Reference Hoskins, Nutt and Watson1997; Mthembu-Salter Reference Mthembu-Salter1999). Additionally, Burundi continued its trade with non-COMESA (Common Market for Eastern and Southern Africa) states, so that virtually all goods were available again by 1997, albeit at higher prices (IRIN 1997a). Due to insufficient capacity, as well as an increasing lack of political will of some countries, the RSCC did not efficiently monitor let alone enforce the sanctions regime (International Crisis Group 1998). Commercial activity between Nairobi and Bujumbura was resumed in July 1997 following the granting of permission by the Kenyan government (ReliefWeb 1997), and Eritrea restarted trade with Burundi in the same year (Bentley & Southall Reference Bentley and Southall2005; Khadiagala Reference Khadiagala2007). The readiness to enforce the remaining sanctions further decreased when a decision by the regional foreign ministers to end the sanctions was overruled at a meeting of the regional heads of state in Kampala, Uganda, on 21 February 1998 in an opaque process. According to the rumours, the decision was only taken when those heads of state critical of the embargo had already left (Wohlgemuth Reference Wohlgemuth, Wallensteen and Staibano2005: 135).
The Burundian government also used the scarcity of goods to its own advantage. The disruption of the economy and the widespread smuggling resulting from sanctions facilitated the restoration of clientelist networks among ‘urban businessmen [who] were more affected than the general population’ (Bihute 2013 int.). Like the leaders of the two preceding military regimes, Buyoya had made use of extensive patronage networks during his first rule from 1987 to 1993 (Uvin Reference Uvin1999). After 1996, many of the private firms that belonged to (former) high-ranking civil servants (Ngaruko & Nkurunziza Reference Ngaruko and Nkurunziza2000) benefited from sanction-busting activities. Speculative importers included senior members of the military and the civil service, and some evidence also points to the involvement of cabinet ministers (Economist Intelligence Unit 1997). Local businessmen, who were generally supportive of the Buyoya government, also profited from the clandestine trade (Calmeyn et al. Reference Calmeyn, Musabyimana and Reychler2000: 89; Daley Reference Daley2007).
Lastly, the regime managed to adapt to the economic constraints set by the embargo. The ‘café-carburant’ (coffee for fuel) initiative was launched to secure the supply of petrol and the export of Burundi's most important source of foreign revenues. Moreover, a ‘solidarity fund’ was set up to generate alternative revenues for the regime's war efforts in the face of foreign exchange shortages (Buyoya Reference Buyoya2011: 147). The Central Bank created a list of prioritised items to be pursued with the limited amount of foreign exchange available (Banque de la République du Burundi 1996: 60). This list was later amended to include selected enterprises of national interest such as the fuel-trading Société Interpetrol (Bulletin Officiel du Burundi 1997b). Private banks were obligated to transfer all remaining foreign exchange to the Central Bank (Banque de la République du Burundi 1996: 60). Finally, the serving of external debt was suspended in response to the growing budget deficit (Girukwigomba 2013 int.).
The Buyoya regime managed to reduce the economic pain caused by the sanctions, so that the economy slowly recovered. After the nation's GDP decreased sharply in 1996 due to the sanctions and the civil war, it remained stagnant in 1997 and grew by 7·4% in the following year (Hoskins et al. Reference Hoskins, Nutt and Watson1997; Kamungi et al. Reference Kamungi, Oketch, Huggins, Huggins and Clover2005). In a nutshell, economic development ‘improved notably as regards internal production and public finances and […] inflation decreased remarkably thanks to the regular provision of goods of primary necessity’ (Banque de la République du Burundi 1998: X).
Argumentative Self-Entrapment
Despite this economic recovery, the regime ultimately responded to the Regional Initiative's most important requirement for the lifting of sanctions: unconditional negotiations with all parties to the conflict. At first,Footnote 8 the regime denied that its seizure of power could be characterised as a coup and argued that the sanctions thus lacked any well-founded reason. The toughness of the sanctions took Buyoya and his supporters, who had relied on Buyoya's international reputation for having paved the way to multiparty democracy, by surprise (Uvin Reference Uvin1999). This standing was undermined by the signals of regime disapproval conveyed by the sanctions: The Regional Initiative referred to the new government as the ‘Buyoya regime’ rather than talking about the Burundian government (Second Regional Summit on the Burundi Conflict 1996; Sixth Regional Summit on the Burundi Conflict 1998). Moreover, Nyerere characterised the sanctions as a signal of ‘support [for] a democratically elected government’ (Le Soir 29.8.1996). Responding to this challenge, the Buyoya government initiated its diplomatic campaign, during which it became increasingly entrapped in its own strategy of denouncing the sanctions as undermining the regime's efforts to negotiate.
To justify the takeover of power, Buyoya's supporters argued that the previous president's flight to the US embassy had created a power vacuum, which left the stumbling nation without political leadership in a deepening security crisis (Mworoha 2013 int.). In such a situation, Buyoya assumed his responsibility as an experienced statesman in a ‘political change dictated by the events’ (Buyoya 2013 int.) that was guided by ‘humanitarian motives’ (Nijimbere 2013 int.) and prevented a takeover by Bagaza and other more extremist elements in the Tutsi community. In short, the regime disputed that the takeover of power could be characterised as a coup d’état, which would have potentially justified a strong regional and international response, during this ‘denial phase’ (Risse Reference Risse1999: 545). In contrast, FRODEBU stressed that the coup had overthrown a democratically elected government (Misago 2013 int.; Nduwimana 2013 int.). According to the party's leading politicians and diplomats, the region's heads of state, who had applauded the democratic 1993 elections, could not accept military interference in Burundi once again (Mbonerane 2013 int.; Ntibantunganya 2013 int.).
In a second phase, the Buyoya government then turned to criticising the sanctions for undermining the true intentions behind the coup – namely, to re-establish democracy and revive the peace process (Longman Reference Longman1998: 20). The regime responded to some of the senders’ demands in order to be able to argue that the sanctions were no longer necessary. Conscious of the senders’ pressure, it reopened the National Assembly and lifted the ban on political parties before the first review of economic sanctions in Arusha in mid-October 1996 (Mthembu-Salter Reference Mthembu-Salter1999; Khadiagala Reference Khadiagala2007). Nonetheless, the political significance of these reinstated institutions remained extremely limited. Political parties were only allowed under the vague requirement that they ‘positively contribute to Burundi's development’ (Economist Intelligence Unit 1997), which gave Buyoya the freedom to outlaw them as he pleased. Moreover, the Assemblée Nationale lacked clearly defined legislative powers in the absence of the constitution, which had been suspended after the coup (Hoskins et al. Reference Hoskins, Nutt and Watson1997).
In addition to responding to the senders’ demands to restore constitutional rule, the regime stressed its willingness to engage in peace talks (Buyoya Reference Buyoya2011: 142–3). According to this discourse, the prolonged sanctions undermined the government's attempts to find an internal solution to the crisis and ‘handicapped the negotiations’ (Girukwigomba 2013 int.). To prove the regime's preparedness to advance a negotiated settlement of the conflict, the government leaked information about secret talks with the rebel movement CNDD under the mediation of the Italian Catholic organisation Community of Sant'Egidio in Rome at the end of 1996 (Dupont Reference Dupont1997; Wolpe Reference Wolpe2011) and circulated the agenda, which included the restoration of the constitutional order, a ceasefire and the involvement of other political parties in the process (Nijimbere Reference Nijimbere2001; Buyoya 2013 int.). However, the talks subsequently collapsed due to pressure from the Tutsi elites on the regime after the information had been leaked.
Meanwhile, the former ruling party and the major opposition party after the coup in Burundi, FRODEBU, repeatedly made reference to the sanctions. This not only served as a constant reminder of the regime's illegal assumption of power (Bamvuginyumvira 2013 int.) but also legitimised and empowered FRODEBU's claims vis-à-vis the government (for a similar theoretical argument, see Risse & Sikkink Reference Risse, Sikkink, Risse, Ropp and Sikkink1999). Put differently, FRODEBU wanted to keep the sanctions in place because they added authority to its demands (Ntibantunganya 2013 int.). The Hutu-dominated diaspora in Eastern Africa and Europe enhanced these lobbying efforts by paying official delegation visits to the European Commission, the European Parliament, the United Nations and Western governments and by lobbying the public (Ngendahayo 2013 int.). In Germany, for instance, the so-called Burundi Büro, financed by the German non-profit organisation Aktion Courage,Footnote 9 coordinated this work. Its Burundi news bulletin, issued weekly, criticised the irregular nature of Buyoya's rule and emphasised the need for sustained international pressure, including sanctions (Burundi Büro Reference Burundi1997). Major rebel groups such as the CNDD-FDD also supported the sanctions as a means to constrain the regime economically and militarily (Ndiho 2013 int.).
In a third phase, initial concessions spurred further demands for politically more meaningful steps. Following a meeting of the Regional Initiative in Kampala on 18 August 1997, Tanzanian foreign affairs minister Jakaya Kikwete announced that the sanctions would remain in place until they achieved their aim of fully restoring democratic rule in Burundi (Inter Press Service 19.8.1997). The government initiated a National Debate, reflecting Buyoya's preference for the domestic peace processes (Brachet & Wolpe Reference Brachet and Wolpe2005). Such an internal political solution was meant to ease the regional pressure (Maundi Reference Maundi, Sriram and Wermester2003: 341). The National Debate consisted of 15 presidential meetings and round tables for representatives of the political parties, members of parliament, public officials, the army, the police and civilian groups. At the beginning of 1998, the internal debate led to the conclusion of a partnership agreement that stipulated the enlargement of the National Assembly from 81 to 121 members under a new constitution and the reshuffling of the government, with FRODEBU's Frédéric Bamvuginyumvira to become prime minister (République du Burundi 1998). The regime depicted this as further evidence that the Buyoya government was really seeking to resolve the crisis through negotiations (Girukwigomba 2013 int.). In exchange, the partnership agreement stated that the sanctions constituted an undeniable obstacle for the route towards peace and that the transitional institutions should launch initiatives for the resumption of bilateral and multilateral cooperation (République du Burundi 1998). While FRODEBU's internal wing thenceforth demanded the lifting of sanctions, the external wing accused it of having become ‘Buyoya's marginalised lapdog’ (Misago 2013 int.). The 1998 partnership hence manifested the de facto split between the exiled faction of FRODEBU and those who had remained in Burundi after the coup (International Crisis Group 1999).
Having started negotiations and reinstated the political parties as well as the National Assembly, the government claimed in 1998 that it had fulfilled the region's conditions for lifting the sanctions (Mbonimpa 2013 int.). After the summit in Kampala on 21 February 1998, however, the Regional Initiative stated that no significant progress had been achieved towards a peaceful settlement of the conflict, requested unconditional negotiations including all rebel movements under the mediation of Nyerere, and threatened to uphold the embargo otherwise (IRIN 1998). After arguing more and more vocally that peace talks with both Hutus and Tutsis were its key intention (New York Times 10.8.1997; Dagne Reference Dagne1998), the Buyoya administration had increasing difficulty justifying its reluctance to engage in the revitalisation of the Arusha negotiations. By mid-1998, most parties to the conflict – namely, 17 political parties and rebel movementsFootnote 10 – eventually embarked on the Arusha process, with the first round of negotiations taking place from 15 to 21 June 1998. The process strove to find a solution to the civil war and to lay the foundation for a transitional government that would incorporate representatives of all the principal factions (Lemarchand Reference Lemarchand2007). In January 1999, the remaining trade restrictions were finally lifted because the Burundi government had proved its irrevocable commitment to negotiations (Bentley & Southall Reference Bentley and Southall2005).
CONCLUSION
The regional embargo in combination with the Western countries’ decision to freeze development assistance initially hit the small, land-locked Burundi harshly. However, this financial pain only led to limited cooperation on the part of the sanctioned regime. It responded to the demands to reopen the National Assembly and re-legalise political parties, but it only endowed these bodies with a very restricted mandate. In the meantime, the Buyoya government consolidated its grip on power. Economically, earnings from black market activities and the potential benefits of sanction-busting activities were used to cater to patronage networks. Politically, a successful campaign against the sanctions helped the regime to regain some recognition among the international community. However, this diplomatic campaign eventually backfired as Buyoya became entrapped within this strategy of emphasising the regime's willingness to engage in peace talks. While the regime managed to address the economic constraints posed by the sanctions to a considerable extent, its campaign to fight off challenges to its legitimacy – amplified by the sanctions – in order to reengage important donors ultimately failed.
This should not suggest that the regime's argumentative self-entrapment in its campaign against sanctions is sufficient to explain the embargo's effectiveness. Clearly, the Arusha process also built upon preceding negotiations (Maundi Reference Maundi, Sriram and Wermester2003). However, Buyoya's willingness to participate in unconditional, inclusive negotiations was initially limited. He stressed both in public announcements and during private talks with Nyerere that no negotiations would take place until the embargo was lifted (Boyer Reference Boyer1996; Hoskins et al. Reference Hoskins, Nutt and Watson1997: 77–85). In addition to conditioning peace talks on the lifting of sanctions, Buyoya at first ruled out negotiations with the rebel movements and also sought to exclude them from the national debate (IRIN 1997b). Lastly and perhaps most importantly, ‘his actual political decisions seem […] to be war-oriented’, as the Special Rapporteur for the UN Commission on Human Rights put it in his second report (UN Commission on Human Rights 1997).
Second, the conflict parties’ war weariness and a mutually painful stalemate (Cunningham Reference Cunningham2011: 178) contributed to the major political initiative that occurred in mid-1998, even though the fact that neither side was near victory had already been established in 1996 (Economist Intelligence Unit 1996) without prompting successful negotiations. In addition, internal divisions within the Hutu and Tutsi camps had complicated the peace process. As Buyoya faced intraparty resistance to negotiating with FRODEBU from the military wing of UPRONA, he was only willing to further engage in the peace process once he had established himself domestically against this militant wing (Wolpe Reference Wolpe2011). While negotiations might have occurred one way or the other, the sanctions forced the government to negotiate with more openness and according to the conditions set by the Regional Initiative (Bihute 2013 int.).
Going beyond the case of Burundi, this paper shows how the signals sent by sanctions that delegitimise the targeted regime may become so deeply ingrained in domestic controversies that they profoundly affect political struggles, particularly how the incumbent regime and its opponents position themselves vis-à-vis each other and the external coercion. The predominant focus on the sanction's financial impact of past research has tended to obscure the possibility that a state ‘may initiate sanctions not simply to compel action on the part of the target, but to communicate its preference, support allies, deter others from engaging in similar activity, and dissuade the target from expanding its objectionable activity’ (Kirshner Reference Kirshner1997: 34). The embargo against Burundi hence sheds light not only on how targeted regimes manage to withstand economic pressure from sanctions in the short run, but also on how campaigns against sanctions – while initially helping the regime to sustain its power – can force targeted regimes to grant major political concessions if they start ‘talking the talk’ of sanction senders. In Burundi, the former ruling party and main opposition party after the coup, FRODEBU, continuously referred to the sanctions, which served as a reminder of the Buyoya government's irregular assumption of power, in order to uphold pressure on the regime. The embargo played such a key aspect in this strategy that FRODEBU launched a pro-sanctions campaign (Ngendahayo 2013 int.). How and under which conditions sanctions effectively enable domestic opposition to the sanctioned regime hence emerges as a crucial question that merits further research.
