Version classiqueVersion mobile

Poverty and Peace in the Portuguese Speaking African Countries

Cristina Rodrigues Udelsmann
Ana Bénard da Costa

The Nature of the Parties on the Prospects of Power-Sharing in the Angola Peace Processes

Ricardo de Sousa


This paper identifies the use of power-sharing dimensions in peace negotiations for the Angola civil war and their success or otherwise. It concludes that a series of dynamics were at play preventing successful power-sharing solutions and adds one factor to the current literature. In particular, the monolithic nature of each contending party made it practically impossible for the Mpla and Unita to cooperate under a shared structure. This is illustrated through a historical revision of the conflict's power-sharing provisions and characteristics of each party.

Texte intégral


  • 1 The paper was prepared as part of the “Poverty and Peace in Portuguese Speaking African Countries” (...)

1Power-sharing agreements have become a relevant topic in today’s international affairs and a common component for negotiated solutions to conflicts. Of the 38 civil wars resolved via a process of negotiations between 1945 and 1998 (a sub-set of the overall universe, which also includes settlements imposed by conflict victors), only one agreement failed to have any form of power-sharing, the short lived 1989 Gbadolite Accord for Angola (Hartzell & Hoddie, 2003).

2The civil war in Angola started in 1975 and ended in 2002. Throughout, a series of significant changes occurred to the conflict and several peace processes were executed. This paper both analyses whether power-sharing provisions were used and also identifies key constraints in the process leading to peace and democracy.

  • 2 Monolithic, something which is single, massive and unchangeable (Oxford Dictionary, 1989).

3We start with a review of the power-sharing concepts before then presenting a historical description of the conflict in Angola and of the agreements reached. There follows a description of the nature of each party (Mpla and Unita) throughout the conflict before proceeding with an identification of the main factors limiting power-sharing present in the conflict as identified in the literature. In the next section, the analysis proposes that an additional factor, the monolithic2 nature of each party, is necessary to fully explain the dynamics described. The paper finishes with a conclusion.

Power-sharing, conflict resolution and democracy

4Generally speaking power-sharing agreements are political architectures aimed at guaranteeing potential warring factions a role in the country's government, through either securing one's inclusion or the competitor’s exclusion from specific areas, and hence decreasing the stakes of political contestation.

5Gates and Strom (2007) identify how the capacity of power-sharing agreements to promote peace depends on the relative military capacity of each side to the conflict as well as on the potential role of spoilers. The ideal environment for peace to be successfully achieved is when the sides are evenly balanced and the costs of war are relatively high. In their analysis, one of the greatest threats to peace are “spoilers” - leaders and parties that have the capacity and will to resort to violence and to subvert peace processes through the use of force.

6Hartzell and Hoddie (2003) categorize power-sharing into four types depending if it intends to divide power along political, territorial, military or economic dimensions. They conclude that when resolving civil wars through negotiated processes the greater the number of power-sharing dimensions included, the greater the chances of an enduring peace.

7Binningsbo (2005) and Reynal-Querol (2002) have identified that the Lijphart model is, in general, suitable for post-conflict societies. Lijphart’s (1977) model on consociational democracy (in Gates & Strom (2007)) has four main components: a grand coalition, autonomy for each ethnic segment in all matters not of common concern, mutual veto rights and proportionality in political representation, civil service appointments and the allocation of public funds.

  • 3 Additionally, Jarstad (2006) identifies that in war torn societies there can be long term negative (...)

8But several shortcomings have been identified to power-sharing agreements. Besides the classical transaction costs, adverse selection and moral hazard (Gates & Strom, 2007), Spears (2000) lists a series of challenges to power-sharing3, specifically that: it interferes with the option of total power offered by competitive elections; it is normally integrated by parties into a strategy to augment military and political power; it requires otherwise incompatible individuals and groups to co-operate; one of the groups is required to relinquish some power, either the stronger to level the power field (conceding more power than would be gained through electoral competition or military victory) or the weaker to become integrated in the political game (for instances releasing claims for regional autonomy through integration into the government); groups fear their power can be jeopardised in the future; and there are varying degrees of commitment to a strategy.

9As is presented in the next section almost all of these challenges were present in the Angola conflict at one stage or another of trying to reach a power-sharing solution able to bring about peace and democracy.

The Angola conflict and power-sharing solutions

10The case of the Angola conflict is paramount to this field given it spans a long period of time where power-sharing was either not attempted or failed to be implemented both with limited and extended formulations. Only a military victory in 2002, together with negotiations and power-sharing provisions, would establish lasting peace in the country.

  • 4 As a reference to the relative dimensions of each movement’s ethnic constituents, one can highlight (...)

11The independence process was established by the Alvor accords signed in 15 January 1975. The accords agreement committed the three liberation movements: the National Front for the Liberation of Angola (Fnla); Popular Movement for the Liberation of Angola (Mpla); and National Union for the Total Independence of Angola (Unita)4 to form a transitional coalition government, a National Defence Commission which would integrate troops from each side and the development of a constitution to which elections were to follow in October 1975, with 11 November 1975 as the date for independence (George, 2005).

12Nevertheless during that year the transition process collapsed and on the day of independence a long civil war began opposing the three movements (while the Fnla would soon cease to represent a significant party to the conflict). On the one side, there was the Mpla supported mainly by Cuba and the Soviet Union and on the other side there was Unita supported mainly by the United States of America (USA) and South Africa.

13The period that was dawning, and particularly from the 1980s onwards, would see the territory divided into an area of Mpla control, where a one-party-state system was developed, and an area of Unita control also characterised by an autocratic, militarist structure of power relying considerably on traditional power structures for the management of populations within their territories. According to Rothchild and Hartzell (1995), the end to direct external involvement would only occur, among other changes, after a military stalemate around 1987 and 1988 made it clear for South Africa that the balance of forces had changed and the cost of war now exceeded its anticipated benefits.

  • 5 Additionally, it would implicitly mean the end to South Africa's incursions into Angolan territory (...)

14The subsequent New York Accords, signed on December 22 1988, marked the end of the internationalized Cold-War status of the conflict with the withdrawal of Cuban troops from Angola and the independence of Namibia5. Nevertheless, the New York Accords contained no provisions for addressing the internal conflict in Angola or additional commitments by external actors to end their assistance to the parties (Hartzell & Hoddie, 2007).

15The accords were the culmination of the lengthy Chester Crocker (United States Assistant Secretary of State for African Affairs) negotiations process initiated seven years earlier and which, with the initial policy paper approved by Ronald Reagan in March 1981, also included political reconciliation between Unita and the Mpla as an objective. The fact that such an objective was dropped from the negotiations is supposedly due to Mpla fears over Unita, reported by the Soviets when questioned by the US (Crocker, 1993). At this stage, the parties to the conflict were simply too far apart for any solution, or in other words, the parties lacked the ripeness for resolution (Hampson, 1996). Furthermore, the internal Angolan conflict was subordinate to the overall regional solution and did not jeopardise achieving the latter.

16The New York Accords closed the first Cold-War sub-period of the Angola conflict, which started in 1975 and inaugurated the second sub-period characterised by a non-internationalized status of the Angola conflict, which would terminate with the end of the Cold War and the Bicesse peace process in 1991.

17Therefore, although the conflict in Angola continued, negotiations between the Mpla and Unita beginning in early 1989 would lead to the Gbadolite Accord of June. The accords did not achieve peace, nevertheless the process surrounding it was important because it identified national reconciliation in Angola as an objective, recognized Zairian president Mobutu Sese Seko as mediator and created some regional peer pressure to reach an agreement. Moreover, it gave Unita and its leader increased respectability and legitimacy (Rothchild and Hartzell, 1995).

  • 6 Bicesse is a small village in the council of Cascais in Portugal, where the accords were signed.
  • 7 Since the 1960s, FLEC has been fighting a low intensity guerrilla campaign for the independence of (...)

18Only after the war had reached a military stalemate did both parties come to a peace agreement. In May 1991, the Bicesse Accords were signed in the context of the Cold- War ending6. The Bicesse Accords stipulated an immediate ceasefire, the creation of a national army and elections (James, 2004) in accordance with a semi-presidential democratic model. It is reported that the rapport between the parties at this stage went as far as being involved in joint operations against the Front for the Liberation of the Enclave for Cabinda (FLEC)7 in the oil rich enclave of Cabinda (MRP, 2005).

19The accords had significant military power-sharing provisions including the establishment of a national military force in parity and some political power-sharing provisions through verification commissions. Its significant shortcoming was that executive power was modelled on a presidential “loser-takes-nothing” (Gates & Storm, 2007) structure. This was most evident in the absence of an agreement on a framework for decentralization, covering the structure for regional and local government, which was only to be decided after the elections (Rothchild & Hartzell, 1995).

20Implementation of the agreement faltered as although the ceasefire held, troop assembly and demobilization lagged behind, especially for Unita troops, amongst warning signs that Jonas Savimbi might be contemplating the scenario of returning to conflict (Hartzell & Hoddie, 2007). Its implementation was supported by UNAVEM II (United Nations Angola Verification Mission), a mission considered to be insufficient in size and capacity for the challenge (Hodges, 2001).

21Despite these signs, national elections were held on September 29 and 30, 1992. Unita lost the legislative (34 percent against 54 percent for the Mpla) and presidential elections, with José Eduardo dos Santos achieving 49.7 percent of the votes and Jonas Savimbi 40 percent even if a second round of balloting, never carried out, for the presidential election would be required as neither of the two main presidential candidates had achieved a clear majority. Despite being considered “generally free and fair” by the international community, Unita claimed electoral fraud and resumed conflict.

22Unita is normally considered an inside spoiler (because it was a willing participant in the process) of the peace process of 1991/1992 in Angola. Stedman (1997) argues that an important factor in the decision to reengage in conflict was a conviction by Jonas Savimbi and his generals that a military victory was possible at that time. And, in fact, in just a short period, Unita was able to take about 70% of the territory. Nonetheless when a year later, in November 1993, Jonas Savimbi returned to negotiations it was “after the rearmed Angolan army had rolled back Unita's gains, the United States had granted diplomatic recognition to the Angolan government, the United Nations had imposed sanctions, and 300,000 Angolans had been killed” (Stedman, 1997, 39).

23A negotiated process aiming at an expanded power-sharing agreement began in 1993 leading to the relatively successful Lusaka Protocol of 1994. This protocol augmented the power-sharing provisions of the Bicesse Accord to include not only executive power sharing (allegedly including a non-written agreement on a vice-presidential position for Jonas Savimbi) but also extended to the police force and the territorial level with the provision of governorships and municipalities.

24In the following years of 1996 and 1997, peace was almost reached with the support of the international community through UNAVEM III but Unita’s implementation of the protocol faltered. For instance, the formation of the Government of Unity and National Reconciliation (GURN) only occurred in April 1997, when some Unita disenfranchised deputies took up their seats in parliament (Hodges, 2001). In September 1998, the government suspended the coalition accusing Jonas Savimbi of continually reneging on his commitments and specifically of holding onto his strongholds (especially Andulo and Bailundo) while secretly rearming his army (Hartzell & Hoddie, 2007; Vidal, 2006) and shortly after declared the peace process annulled and that the only path to peace was war.

25The conflict would only finish with the victory of the Mpla over Unita in 2002, after Jonas Savimbi’s death and the signing of the April 2002 Luena Memorandum of Understanding, which represented an addendum to the Lusaka Protocol. In particular, it includes provisions for the integration of Unita officers and soldiers into the national army and the rest to be demobilized. The success of this agreement confirms the study of Licklider (1995), which identifies how among civil wars that have ended, the settlements imposed by a conflict’s victor prove more durable than negotiated agreements (in Hatzell & Hoddie, 2003). Nevertheless the negotiation process at this stage together with the power-sharing provisions of the Lusaka protocol were also important in guaranteeing peace within Unita.

The nature of the parties

26Key actors in the time-line process described above are the Mpla and Unita parties, wherein their characteristics definitely shaped the events and choices made. For this reason, it is important to look at the history of both the Mpla and Unita (the Fnla is not analysed here as it did not constitute a significant military force as from the early 1980s) and identify their main characteristics.


27The Mpla was formed in the 1950s in the coastal and urban areas of the centre north, gaining support mainly from the Mbundus of Luanda and Malanje and from mixed-race intellectuals. The role of each ethnic group and of socialism in the party has been contested ever since, but the driving force of this group at the time was opposition to the Portuguese colonial government (Spikes, 1993).

28Around independence, the party fragmented into three groups, one led by President Agostinho Neto and two contenders, who would lose the contest: Daniel Chipenda and the Andrade brothers. Soon after independence a new challenge to Agostinho Neto would come from within his party with an attempted coup by Nito Alves in 1977. Again, Agostinho Neto emerged victorious but this time executed a process of rectification, which extended well beyond the party, Luanda and that year (Hodges, 2001).

29In the aftermath of the attempted coup d’état, the Mpla would increase presidential powers, create a frightening state security system, purge the party from 110,000 members to 31,000 with the establishment of rigorous selection processes and military political control of the judiciary system (Vidal, 2006). Integrated in this process was the transformation of the Mpla into a Workers Party in December 1977, establishing the vision of a one-party-state inspired on the Marxist-Leninist model.

30According to Hodges (2001), these initiatives created conditions for an uncontested succession of Agostinho Neto by José Eduardo dos Santos as president in 1979 (when the former died of illness during treatment in Moscow), but essentially the rectification process gave birth to a generalized culture of fear, conformism, lack of initiative and submission in society.

31The characteristics of authoritarianism, rectification, inter-penetration of the state and party structures and the political control of the judicial system continued after 1979. Most of all, a process was initiated concentrating powers away from the party and into the presidency, which would be justified by the war besetting the country. In the early 1980s, the Office of the President was created to deal with foreign business, most importantly to control oil revenues (Vidal, 2006). In December 1982, the Central Committee afforded special powers to the president empowering him to reshuffle both the Politburo and the Mpla Armed Forces for the Liberation of Angola (FAPLA) (George, 2005) and in 1983 and 1984 creating a kind of parallel martial government responding directly to the Commander in Chief of the Armed Forces, which was the President in office (Vidal, 2006).

32Although progressively concentrating power in himself, the president needed to carefully negotiate policy changes within his group. In 1984, Crocker (1993) could still grasp some vulnerability in the position of dos Santos in the stances taken by the Mpla in negotiations. Additionally, in the early 1990s, during the Bicesse accord negotiations, the Mpla had to carry out a military and cabinet reshuffle, most likely to remove potential critics of conciliation efforts with Unita (Hatzell & Hoddie, 2007) from positions of influence. Or later, in early September 1991, on the eve of elections when “Santos privately expressed interest [in a power-sharing agreement with Unita] but felt he could not publicly commit to such a deal” (Stedman, 1997, 38).

  • 8 There are several definitions of neo-patrimonialism. In this case, it intends to refer to a situati (...)

33Throughout the 1980s, a new, select group of young politicians and technicians were promoted through the ranks of the system and owing their ascension to power significantly to the president (as occurred also in the FAPLA and the party) (Vidal, 2006). The decade and a half of inconsistent economic plans, where reform would be followed by counter reform (Hodges, 2001) and also the incapacity in the 1990s to reach agreement with the World Bank and International Monetary Fund (IMF) can well be explained by the resistance that the established system made to significant changes to neo-patrimonialism8 and the president's needs to accommodate the demands of his power base.

34Correspondingly, the second half of the 1980s and early 1990s saw significant changes with the formal elimination of the one-party-state system, introduction of a multiparty system, some basic democracy type laws and economic reforms to partially open up the system. The party itself was also revitalized in the run up to the 1992 elections with a reorganization of its structure, expansion in the distribution of material rewards, political recovery of traditional authorities and an increase in membership from 65,362 in 1990 to 544,639 at the end of 1992 (Vidal, 2006).

35But the regime did essentially keep its power concentrated in the president (for instance, there was no decentralization) and supported by two core pillars: the Mpla party, an extraordinary instrument in the service of the president, and the army, militarily efficient and politically loyal (Chabal, 2006).

36The results of the elections afforded José Eduardo dos Santos and Angola international prestige, which gained a significant landmark in 1993 with US recognition of the Mpla government. Simultaneously, this lowered the risk of a presidential challenge from within the Mpla (Hatzell & Hoddie, 2007) thereby certainly allowing more flexibility in the negotiation period leading up to the Lusaka protocol and the envisioned GURN coalition government.

37In the second half of the 1990s, there is an intensification of the co-optation strategy of the Mpla. It not only spread extensively to any form of civil society initiative but also to the opposition parties, a phenomenon categorized as “Renewed”. The most symbolic is that of Unita where there was a split in the leadership and the “Renewed Unita” emerged and assumed seats in parliament reserved for Savimbi’s party, although without credibility.

38In summary, ever since its inception, the Mpla is a party with significant incidence of eliminating contestants rather negotiation and dilution. A system of neo-patrimonialism began to be implemented as early as the creation of a one-party-state-system under the Cold War umbrella. But the gravitational power quickly shifted from the party to the president. The extent this system of clientelism would reach into society depended on the overall economic situation provided by the oil revenues. Although forced to formally change due to the war and economic collapse and to go through elections, the main historic operating structures of the regime (for instance, the party-state relationship, army or executive workings) were never dismantled (Messiant, 2006), because the elections were won and then change was either not implemented or limited in scope. This organism, with deep roots going back before independence, carefully selected the new generation of leaders, co-opting and buying in opponents, has been led by the same president for over two decades. There is little predisposition to seriously release power to the other party in the conflict or open up the oil revenues to new potential rivals. This is particularly, so after 1992-1993 when the regime was granted international recognition and Unita started to be seen unanimously by the international community as the spoiler of peace in Angola.


39Unita was founded in the 1960s by Jonas Savimbi in the south of Angola mainly supported by the Ovimbundu people of the central plateau. During the struggle for independence, Unita allegedly collaborated with the Portuguese through the provision of information. Its organization was initially based on the structure of the protestant church and traditional popular leaders.

40In the aftermath of the Mpla independence military victory, Unita was weakened, particularly so because it had not achieved its strategic objective of securing the Benguela railway, which might have gained it recognition from Zambian President, Kenneth Kaunda, and had temporarily lost significant support from South Africa. In early 1976, Unita resorted to guerrilla warfare, regrouping in the jungle before later in the decade establishing its operational headquarters in Jamba, in Moxico region.

41During the 1980s, Jonas Savimbi managed to develop Unita into a “quasi-state” dimension able to exercise the monopoly of force in its areas. A diamond economy developed and established centralized territorial control with a governance structure including a president, government and service institutions. The authority system was institutionalised according to a patrimonial functional logic, according to which economic resources and power positions were distributed along patron-client relations. The system was stabilised through the brutal use of force, applied especially for the purpose of eliminating challengers to the authority of Jonas Savimbi (Bakonyi & Stuvoy, 2005).

42In the late 1980s, Unita is believed to rule a region in Jamba with between 8,000 to 10,000 people, 80,000 to 100,000 in the surrounding regions and with about 30,000 troops in 1984 (George, 2005).

43The 1990s are shaped by an attempt by Unita to transform itself from a guerrilla force into a political party with parliamentary and governmental responsibilities. These changes lead to significant instability inside the party. The first sign of it was a high- profile defection in the run-up to the elections with accusations over an assassination plot against Jonas Savimbi (Stuvoy, 2002).

44The incapacity to transform into a political party and the electoral defeat saw a series of factions challenging the Jonas Savimbi leadership. The first significant break away faction came from the Renewed-Unita which took up seats in both the parliament and GURN in 1997. In the following year, 1998, a new Unita Chivukuyuku faction would emerge, but failed to recruit influential supporters. By 2001, five factions could be identified although the militaristic Jonas Savimbi wing continued to be the dominant (Stuvoy, 2002).

45The monolithic structure could not adapt and the disintegration of authority from the mid-1990s led to a situation in 1998 where Unita administrative structures had nearly dissolved (Bakonyi & Stuvoy, 2005) and with it also the patrimonial system. Furthermore, the aggressive military stance of the Mpla after 1998 led to recurrent military losses for Unita.

  • 9 Jonas Savimbi first travels to the US in 1981 and in that year returns to be met by Reagan administ (...)

46Only with the death in combat of Unita leader, Jonas Savimbi, could peace be reached. His leadership was able to create the rebel group and party to the extent of long being an effective threat to the Mpla government and having also acquired international status in the 1980s9.

47Once again, the Cold War setting bred a rebel group, without concern as to whether its organizational structures would be conducive to a form of shared-power, but rather concerned about its capacities to progressively challenge the Soviet like state. The autocratic monolithic system of the organization was maintained throughout and incapable of adapting to a post-election culture and began to break up under military, political and economic pressure.

48Additionally, the Unita party showed no signs of being able to incorporate internal dissident voices in 1992 and therefore it does not seem likely that there would be many chances to integrate into a structure with opposing voices. For instance, from the late 1990s onwards, it was probably more profitable for the Unita leadership to gain entrance into the Mpla neo-patrimonial system and negotiate access to its own revenue sources than to continue fighting. Some of the high-ranking Unita leaders did this by establishing some break away factions. Nevertheless, the militaristic faction of Jonas Savimbi did not. Eventually, this Unita faction, which continued to defy the elections, may be considered a case of illegitimate representation of their constituent interests - by this time the value of peace had increased significantly.

Limitations of the power-sharing model in Angola

49Several factors can be found in the literature for the failure to reach or implement the power-sharing agreements identified above in the Angolan conflict’s history. They can be grouped into: inappropriate external pressure, a lack of a structural power-sharing solution, leadership characteristics and ambitions, mistrust after a prolonged war and the role of resources in shaping incentives for the parties.

  • 10 One cautionary note is in place here. Just as Chester Crocker (1993) highlights that regional power (...)

50External pressure10 influenced the conflict in several ways. Indirectly but fundamentally, the political and economic collapse of the Soviet Union system required the Mpla regime to readjust. Superficially when regional actors (Mobutu Sese Seko) pressured the Mpla and Unita into signing the Gbadolite accords in 1989, which did not reflect the real intentions of the parties (Hartzell & Hoddie, 2007) but did engage them in a negotiated process. By the incapacity to halt the conflict given the international community: lacked consensus after Unita’s 1992 return to war and therefore did not send clear signals to local parties as to how unacceptable some solutions to a country's challenges are (Spears, 2000); and lacked effective implementation capacity to perform its assigned role, where due to the “inability of third parties, notably the United Nations, to provide resources needed to implement the 1991 peace accords doomed them to failure in the face of widespread cheating and non-compliance” (Hampson, 1996, 88).

51In addition, a broader and swifter array of power-sharing and power-dividing institutions have been identified as important in order to increase the likelihood for a negotiated peace (Hartzell & Hoddie, 2007). For instance, Stedman (1997) identifies the lack of power-sharing provisions in the 1991 Bicesse accords as a reason for the resumption of the conflict, although this author extensively identifies moments before and after the elections where power-sharing proposals were unsuccessfully presented both to the Mpla and Unita by a range of external actors, in particular the United States and South Africa.

52However this factor links to the next, regarding leadership, in the argument of Jarstad (2006). It is claimed that the 1994 Lusaka power-sharing provisions were good because they were able to attract some Unita factions, isolating the militarist Jonas Savimbi group, eventually leading to this group's defeat and its leader’s death in combat. When that occurred, the implementation of the agreement could resume without additional militaristic factions emerging rendering Lusaka, and its provisions, a positive contribution to peace. Nevertheless, this argument is based on the assumption that Savimbi's group would not settle for peace in any case. Even if the inducement only opened the Unita appetite for power in 1992, as Stedman (2007) argues, Jonas Savimbi’s behaviour can be considered that of a “greedy spoiler” where a heavy dose of coercion combined with extremely high costs for non-compliance, might have been a better option for achieving peace. According to Stedman (1997) Ambassador-designate Edmund De Jarnette identified Savimbi’s personality and his hegemonic ambitions for Angola as the problem. Furthermore, other leadership incompatibilities included the personalities of the leaders being a source of limitations at the time of elections (Anstee, 1996, 147) and their track record of leading their parties against each other in the midst of fervent denunciations (Spears, 2000).

53Mistrust and the results of decades of war led to a relationship of in-depth antagonism between the groups. Messiant (2006) identifies the issue of mistrust between the parties as limiting the implementation of the 1994 Lusaka accord. In this case, Unita would not demilitarize until it gained power and the Mpla was determined to limit the effectiveness of the GURN. One always needs to factor in that after almost two decades of conflict supporters from both Unita and the Mpla, as well as the general population, had known no other modus operandis than conflict (with brief periods of relative peace and differing intensities depending on location) with each other. What is more, concerns over the fragile peace process, leading to six successive amnesties between 1981 up to the late 1990s, which allowed the development of an environment of impunity in society conducive to human rights violations, further exacerbated the lack of trust between the groups (Hodges, 2001).

  • 11 See for instance the accounts of Billon (2001a and 2001b) on the role of oil and diamonds in the co (...)

54Finally, the key role of resources in shaping the incentives of the parties is almost unanimously identified in the literature as a factor contributing to the prolonged continuation of the conflict11. Although this factor cannot totally explain the conflict it had a singular role in funding it, while during the Cold War, support was also derived from the respective external supporters. The existence of sources of funding generated the lack of pressure to concede more and reach shared solutions.

An additional cause for failure – the monolithic nature of the parties

55Nevertheless, these factors do not completely address the underlying question of the lack of pre-disposition to enter into power-sharing solutions, conducive both to peace and eventually to democracy. One additional factor is required to better explain the shortcomings to the overall peace process. This underlying factor is the nature of the two parties in the conflict. Their similarities in terms of their monolithic structures and cultures allowed little organic flexibility to merge or share structures with another similar party.

56This is present throughout the conflict. In the 1989 negotiations, the achievement was the recognition by the Mpla of Unita as a party with which to negotiate, even though still pushing to maintain the one-party-state system. In the 1992 elections, there were two mono-party systems facing each other (Mabeko-Tali’s, 2006): one the party-state of the Mpla and the other a rebel movement dominated by its military structure. The 1994 extended power-sharing solutions proved insufficient to convince the Unita leadership to integrate into the state system.

57Looking at one of the most determinant events of the conflict's history in more detail, to a certain extent the “winner-takes-all” solution of the 1992 elections was in synchrony with the internal logics of both the Mpla and Unita monoliths. There was little prospect of success, even if properly enforced, due to their own philosophies – each movement’s expected non acceptance of integration into the other. Even later when the GURN was established and Unita parliamentary seats occupied, on the one hand Unita members of the government would not be allowed to make independent decisions and on the other hand the Mpla majority in parliament would block any possibility of alternative action. In this short lived period of power-sharing, naturally, the culture was not about reaching consensus but most importantly the vision of what the future should be was also not considered.

58As suggested by Messiant (2006), the nature of the constitutional changes of 1991 and 1992, which seem significant, were more of form than of substance as the neo- patrimonialist patron-client type system was kept intact while functioning within a slightly different architecture. Following on from this perspective, the fact that the solution envisioned at Bicesse in May 1991 did not foresee enough provisions for power-sharing is then a result of the incapacity to move the Mpla into a more flexible and accommodating position, which would in fact change their power structure. As would later be confirmed, the opening up of the political system was only to the extent that it did not jeopardise control by the incumbent and allowing for the restraining of the system should the situation so require and as indeed would happen when conflict intensified in 1992 and 1998. It is the sheer incapacity of the Mpla to accommodate such demands that is at stake in the context of a belligerent challenger, which was, according to most analysts, determined to achieve full power or make its price for peace very high.

  • 12 The economic reforms attempted could also have progressively allowed for the emergence of new group (...)

59One conclusion is that the change process that was initiated in the middle of the 1980s should have included provisions not only for constitutional or economic reforms but also reform envisioning democratic practices within both parties12. It is considered that the monolithic nature of the Mpla produced only slight, limited and, sometimes, only superficial changes to the party-state apparatus and that Unita disintegrated when faced with the need to integrate into this system as the defeated party in the elections.


60In Angola, 27 years on from independence, 14 years after the New York Accords, more than 500,000 deaths, tens of thousands of persons mutilated by anti-personnel mines and the displacement of approximately 4.1 million people took place before peace was secured.

61From a historical perspective, one can identify that in the case of Angola reality linked both parties to power. Unita was linked to the central plateau and the Ovimbundu possibly as much as the Mpla was linked to Luanda. The Mpla’s openness seems directly linked to feelings of insecurity towards Unita and Jonas Savimbi in particular. The link of one neo-patrimonialist system and its leadership to the other was eventually the “blessing” for peace and the “curse” for democracy, associated with each party’s link to oil and diamonds.

62In searching for a solution to the conflict, power-sharing provisions were increasingly considered. Starting from the 1989 absence of any provisions to the 1991 significant military and lighter political power-sharing provisions, in 1994 power-sharing provisions were extended to the executive level. Several factors referenced in the literature and identified in this paper either prevent a power-sharing deal being reached sooner or inhibit its success.

63This paper adds an additional significant factor limiting the success of power-sharing provisions: the monolithic nature of the parties. This is demonstrated by both analysis of the conflict’s history and the nature of both organizations – Mpla and Unita.

64In the end, one of the parties would need to be integrated into the other in order to survive. In effect, it was the neo-patrimonialist nature of the parties which allowed the Mpla to appropriate the Unita factions into their client networks and, finally reach peace through victory in 2002 over the sole remaining and already isolated Jonas Savimbi led militaristic faction. Nevertheless, this also required negotiation and the implementation of the 1994 power-sharing solutions for peace actually to be reached.

65It is hereby proposed that among the several post-conflict social engineering initiatives, one that could have positively contributed towards results would have been reform of the party structures to move them away from autocratic-monolithic characteristics.

66The analysis presents the parties in an almost unidimensional perspective, without exploring the several axes connecting and separating them. Future research could incorporate the dimensions of ethnicity, occupation, class, religion, culture, language, region, urban/rural and others into the analysis and see to what extent they contributed to process outcomes.


Anstee, Margaret J. (1996) Orphan of the cold war: the inside story of the collapse of the Angolan peace process, 1992-93. Basingstoke, New York, Palgrave Macmillan

Bakonyi, Jutta; Stuvoy, Kirsti (2005) “Violence & social order beyond the state: Somalia & Angola”, Review of African Political Economy, 32:104, p.359 – 382

Billon, Philippe Le (2001a) “The political ecology of war: natural resources and armed conflicts”, Political Geography 20, p.561-584. DOI: 10.1016/S0962-6298(01)00015-4

Billon, Philippe Le (2001b) “Angola’s Political Economy of War: the Role of Oil and Diamonds 1975-2000”, African Affairs, 100, p.55-80. DOI: 10.1093/afraf/100.398.55

Binningsbo, Helga M. (2005) “Consociational Democracy and Post-conflict Peace. Will Power- Sharing Institutions Increase the Probability of Lasting Peace after Civil War?”, Paper prepared for presentation at the 13th Annual National Political Science Conference, Hurdalsjøen, Norway, 5–7 January, 2005

Chabal, Patrick (2006) “Introdução - Transições políticas em Angola: et pluribus unum”, in Vidal, N. and Andrade, Justino P. (eds.) O processo de transição para o multipartidarismo em Angola. Lisboa: Ed. Firmamento

Crocker, Chester (1993) High Noon in Southern Africa: Making Peace in a Rough Neighbourhood, W.W. Norton

Ferreira, Manuel E. (2006) “Angola: Conflict and development, 1961-2002”, The Economics of Peace and Security Journal, [online] Vol. 1, No.1, pp.25-29

Gates, Scott and Strom, Kaare, (2007) “Power-sharing, Agency and Civil Conflict - Power- sharing Agreements, Negotiations and Peace Processes”, Policy brief of a project entitled ‘Power-sharing Agreements, Negotiations and Peace Processes’ of the Center for the Study of Civil War, Oslo (CSCW)

George, Edward (2005). The Cuban Intervention in Angola, 1965-1991 - From Che Guevara to Cuito Cuanavale, London, Frank Cass

Hampson, F. Osler (1996) Nurturing peace: why peace settlements succeed or fail, US Institute of Peace Press

Hartzell, Caroline A. and Hoddie, Matthew (2007) Crafting Peace: Power-sharing Institutions and the Negotiated Settlement of Civil Wars, Penn State Press

Hartzell, Caroline A. and Hoddie, Matthew (2003) “Institutionalizing Peace: Power-Sharing and Post-Civil War Conflict Management”, American Journal of Political Science, Vol. 47, No. 2, April 2003, p. 318–332. DOI: 10.1111/1540-5907.00022

Hodges, Tony (2001). Angola from Afro-Stalinism to Petro-Diamond Capitalism, Oxford, James Currey

James, William M. (2004) Historical dictionary of Angola, New Ed. Oxford: The Scarecrow Press, Inc

Jarstad, Anna K. (2006) “Power-sharing for Peace and Democracy”, Paper prepared for presentation at the 47th annual meeting of the International Studies Association, San Diego, USA, 22-25 March 2006, and at the Vail Symposium 27-29 March 2006

Lijphart, Arend (1977) Democracy in Plural Societies: A Comparative Exploration. New Haven: Yale University Press.

Licklider, Roy (1995) “The Consequences of Negotiated Settlements in Civil Wars, 1945–1993”, American Political Science Review, 89(3), 681–90. DOI: 10.2307/2082982

Mabeko-Tali, Jean-Michel (2006) “Capítulo XX– Exclusão e estratégias de sobrevivência no Estado-nação”, in Vidal, N. and Andrade, Justino P. (eds.) O processo de transição para o multipartidarismo em Angola, Lisboa: Ed. Firmamento

Messiant, Christine (2006) “Capítulo XX – Transição para o multipartidarismo sem transição para a democracia”, in Vidal, N. and Andrade, Justino P. (eds.) O processo de transição para o multipartidarismo em Angola, Lisboa: Ed. Firmamento

MRP - Minorities at Risk Project (2005) Minorities at Risk Dataset College Park, MD: Center for International Development and Conflict Management. [online] Available at: [Accessed 15 July 2008]

Oxford Advanced Learners Dictionary (1989), Chief Editor A. P. Cowie, Oxford University Press, Fourth Edition

Wright, George (2001) A destruição de um País – A política dos Estados Unidos para Angola desde 1945, Lisbon: Caminho

Reynal-Querol, Marta (2002) “Ethnicity, Political Systems, and Civil Wars”, Journal of Conflict Resolution, 46(1), 29–54. DOI: 10.1177/0022002702046001003

Rothchild, Donald and Hartzell, Caroline (1995) “Chapter 8 - Interstate and intrastate negotiations in Angola”, in I. William Zartman (ed.) Elusive Peace: negotiating an end to civil wars, pp.175- 203, Brookings Institution Press

Spears, Ian S. (2000) “Understanding inclusive peace agreements in Africa: the problems of sharing power”, Third World Quarterly, 21(1), 105-118

Spikes, Daniel (1993) Angola and the politics of intervention: from local bush war to chronic crisis in Southern Africa, Jefferson, North Carolina: McFarland and Company

Stedman, Stephen J. (1997) “Spoiler Problems in Peace Processes”, International Security 22(2), 5-53

Stuvoy, Kirsti (2002) “War Economy and the Social Order of Insurgencies - An Analysis of the Internal Structure of Unita’s War Economy”. Arbeitspapier no. 3, 2002, Hamburg University, IPW

Vidal, Nuno (2006) “Capítulo 1 – processo de transição para o multipartidarismo em Angola - Multipartidarismo em Angola”, in Vidal, N. and Andrade, Justino P. (eds.) O processo de transição para o multipartidarismo em Angola, Lisbon: Ed. Firmamento


1 The paper was prepared as part of the “Poverty and Peace in Portuguese Speaking African Countries” project funded by the Foundation for Science and Technology (FCT) of the Portuguese Ministry of Science, Technology and Higher Education - research grant: PTDC/AFR/64207/2006. I would like to thank participants at the “Southern Africa in the Cold War era Working Seminar” organised by the Instituto Português de Relações Internacionais (IPRI) and the London School of Economics (LSE) Ideas at Fundação Luso Americana (FLAD) in Lisbon on 8 and 9 of May 2009 for their feedback and additionally to Inge Ruigrok and Gerhard Seibert for their valuable comments on an earlier draft of this paper. Any remaining errors are my own.

2 Monolithic, something which is single, massive and unchangeable (Oxford Dictionary, 1989).

3 Additionally, Jarstad (2006) identifies that in war torn societies there can be long term negative implications of power-sharing deals for both peace and democracy.

4 As a reference to the relative dimensions of each movement’s ethnic constituents, one can highlight the Bakongo with 12 per cent, Mbundu with 24 percent and Ovimbundu with 32 per cent, more or less associated with the Fnla, Mpla and Unita respectively. But these should not be considered as directly reflecting the movements or the parties or their respective organizational capacities at the time or in future periods.

5 Additionally, it would implicitly mean the end to South Africa's incursions into Angolan territory and an informal agreement is believed to have included the closing of African National Congress (ANC) training camps in Angola (Rothchild & Hartzell, 1995).

6 Bicesse is a small village in the council of Cascais in Portugal, where the accords were signed.

7 Since the 1960s, FLEC has been fighting a low intensity guerrilla campaign for the independence of the oil rich Angolan enclave and would not be involved in any of the agreements. In August 2006, a ceasefire was signed between FLEC-Renewed and the Angolan Government.

8 There are several definitions of neo-patrimonialism. In this case, it intends to refer to a situation of patrons using state resources in order to secure the loyalty of clients in the general population, and is indicative of informal patron-client relationships that can reach from very high up in state structures, down to individuals in say, small villages.

9 Jonas Savimbi first travels to the US in 1981 and in that year returns to be met by Reagan administration officials. In 1986, Jonas Savimbi is officially received in Washington by President Ronald Reagan and meets with the Secretaries of Defence and State. In 1988, Jonas Savimbi returns to the US and meets the presidential candidate (and vice-president) George Bush. In 1991, he would return to the US and meet President George Bush.

10 One cautionary note is in place here. Just as Chester Crocker (1993) highlights that regional powers (like the parties in control in South Africa or Cuba) do not hang around awaiting superpower instructions to execute their policies, in the same manner non regional powers like the parties in Angola have a will of their own and act as much out of the external constraints and pressures presented to them as from inner needs, organic requirements and local constrains and pressures.

11 See for instance the accounts of Billon (2001a and 2001b) on the role of oil and diamonds in the conflict and in Angola and Ferreira (2006) for an overview of the economic conditions from 1961 to 2002. For an account of external interventions see Wright (2001) for the United States and George (2005) for Cuban involvement.

12 The economic reforms attempted could also have progressively allowed for the emergence of new groups in the political spectrum but such is not directly analysed here.


Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search