by Markus Kornprobst
INTRODUCTION
Most of what has been written on war to peace transitions focuses on conflict resolution techniques. The outcome of a peace process is considered a function of the conflict resolution techniques employed. The significance of the conflict situation to which the techniques are applied either are not systematically taken into account, downplayed, or even completely neglected. There are two other strands of the literature that question this virtually exclusive focus on conflict resolution techniques: one set focuses on conflict situations; the other explores the fit between conflict resolution techniques and conflict situations.
This study scrutinizes the explanatory power of these three strands of literature by testing the most influential hypotheses each has generated against the Angolan and the Mozambican experiences. These cases are chosen partly because they pose an interesting empirical puzzle. Although the two cases are similar, the outcomes were radically different. In Angola, two peace agreements - the Bicesse Agreement in 1991 and the Lusaka Protocol in 1994 - were concluded in order to terminate the war between the government, the Movimento Popular de Libertaçao de Angola (MPLA), and the insurgents, the Uniao Nacional para Independência Total de Angola (UNITA). The results of these two efforts were the same. The country relapsed into full-scale war after a brief period of negative peace. In the Mozambican case, by contrast, a peace accord - the Rome Agreement - was concluded between the government, the Frente de Libertaçao de Moçambique (Frelimo), and the insurgency movement, the Resisténcia Nacional Mocambicana (Renamo) in 1992. Despite initial difficulties, the agreement was successfully implemented after two years, and the war came to an end. More importantly, Angola and Mozambique are chosen as cases because many of the arguments that the literature on war to peace transitions has developed rely on evidence from these two cases. Revisiting the Angolan and Mozambican experiences in sufficient depth offers the opportunity to scrutinize the inferences drawn from these cases.
Contrary to most of what has been written on these two transitions, and contrary to the conflict resolution technique and the conflict situation strands of the literature, I argue that the outcomes of the Angolan and Mozambican peace processes stemmed from the incompatibility and compatibility, respectively, of the conflict situations with the conflict resolution techniques that were employed. The cases at hand suggest that conflict resolution techniques are not suitable or unsuitable as such, but adequate or inadequate for a peculiar conflict situation.
This argument is developed in three steps: first, I identify three sets of literature on war to peace transitions and the main hypotheses that they generated. Second, I give an overview of the Angolan and the Mozambican transition periods, and outline the research design of this study. Third, I test the applicability of the hypotheses that I identified in the literature review.
LITERATURE REVIEW
This section identifies three strands of literature on war to peace transitions and the main arguments that each set has generated. The three sets are as follows: literature on conflict resolution techniques, on conflict situations, and on the relationship between conflict resolution techniques and conflict situations.
Conflict Resolution Techniques
Most of the literature emphasizes the causal relationship between conflict resolution techniques (i.e. strategies that are designed to end internal war through a compromise between the conflict parties and not through total victory on the battlefield) and the success or failure of war to peace transitions. Different cases of war to peace transitions are compared and lessons are drawn for other cases without, however, systematically taking into account different conflict situations. The irrelevance of the peculiarities of a certain conflict situation is usually assumed, but in some cases explicitly argued for.1 This set of literature has generated two main arguments to explain the success and failure of war to peace transitions in general, and the Angolan and Mozambican peace processes in particular; these are power-sharing agreements versus winner-takes-all elections, and the extent of third-party intervention.
Marina Ottaway argues that successful war to peace transitions require power-sharing agreements. Democratization, she suggests, is a disruptive process, because the rules of the political game that are established in a consolidated democracy are absent during and immediately after a war to peace transition. In the absence of these rules, there are no checks and balances for the competition between the conflict parties. Ottaway therefore suggests that long-term national conferences are needed to prevent the disruptive effects of democratization.2
Caroline Hartzell argues in a similar vein. Her contention is that the most extensively institutionalized peace settlements - i.e. agreements that provide political and economic security for each conflict party after elections have been held - are most likely to be implemented successfully. Hartzell distinguishes between three degrees of institutionalization: proportional representation after the elections is the lowest degree; administrative proportional representation on the basis of a formula representative of the size of groups is the medium degree; and the division of political power in a federal system or the granting of regional autonomy constitutes the highest degree of institutionalization. In short, the more the former war parties disperse power among them, the more secure they are, and the more likely the war to peace transition is to succeed.3
Many analysts propose that large-scale third-party intervention makes a war to peace transition more likely to be successful. Fen Osler Hampson contends that a "successful peace process depends upon a lot of outside help and assistance, not only with the negotiation of a peace agreement, but also with its implementation."4 The importance of large-scale third-party intervention also features prominently in the studies of other authors, such as Chester Crocker, Charles William Maynes, Barbara Walter, Rafael López-Pintor, and Nicole Ball. Success and failure of war to peace transitions are regarded as functions of the nature of external assistance.5
Most of the literature that focuses on Angola and Mozambique echoes the more general arguments outlined above, including the neglect of the conflict situations. Several scholars attribute the failed transition in Angola to the lack of power-sharing agreements and the winner-takes-all elections. J. Lewis Rasmussen, for example, argues that the "lack of a powersharing option was a major determinant in Jonas Savimbi's decision ... to breach the 1991 settlement that temporarily ended Angola's civil war ...."6 Norrie MacQueen and Fen Osler Hampson make the same argument.7
Another typical way of viewing the failure in Angola and the success in Mozambique is to point to differences in the nature of third-party intervention. Hampson, explaining the failure in Angola, maintains that "the Achilles' heel of the agreement [Bicesse] was the lack of an adequate and effective third-party presence and monitoring mechanism to supervise and assist with implementation."8 The United Nations in particular is blamed for an ineffective intervention in Angola. It is argued that they committed too few resources to be able to facilitate the war to peace transition. The success in Mozambique, by contrast, is explained by the extent of resources and the comprehensiveness of the mandate of the United Nations field mission.9 Some studies on Angola also deal with the roles of the United States and the Soviet Union. Victoria Brittain and Fernando Gonvalces contend that insufficient superpower pressure was, in addition to the insufficient role played by the United Nations, a primary cause for the collapse of the Angolan peace process.10
Conflict Situation
Compared to the impact of conflict resolution techniques on war to peace transitions, there is surprisingly little written on the impact of conflict situations. Three main arguments that focus - albeit not exclusively - on conflict situations as explanations for success and failure of war to peace transitions can be identified: ripe for resolution, external conflict dimensions, and economic dynamics.
I. William Zartman developed the concept of "ripe for resolution" in order to identify the particular moment in time in which a conflict situation is configured in a way that negotiation and mediation can succeed. Negotiations and mediation, according to Zartman, are likely to be successful if there is a hurting stalemate. Stalemate means that it is clear to all conflict parties that they cannot unilaterally win the conflict. Hurting means that the conflict parties "feel uncomfortable in the costly dead end."11 Although this concept is designed to explain success and failure of negotiation and mediation (i.e. it was originally designed to explain why a war to peace transition begins, not why its implementation succeeds or fails), it has become an important analytical tool for the study of war to peace transitions.
Several scholars contend that the external dimension of internal wars affects the outcome of war to peace transitions. If the antagonism between external conflict parties declines or even disappears, a settlement becomes more likely. If, by contrast, an external conflict in which an internal war is embedded is exacerbated, a settlement becomes less likely. This applies to global conflicts, such as the Cold War, as well as to regional conflicts.12
Finally, Mats Berdal and David Keen focus on the economic aspect of internal war. They argue that the continuation of internal war is often a function of the economic interest of one or more conflict parties. According to the authors, the current literature focuses too much on political violence and neglects economic causes of war and continuation of war. Analyzing economic origins of violence helps us to understand internal wars that seem senseless when they are understood as political conflicts only.13
Thus, with regard to the conflict situation, three causes for the outcome of peace processes recur in the literature on war to peace transitions: hurting stalemate, changes in the external dimension of internal war, and economic dynamics. The hurting stalemate, operationalized as a peculiar military situation, and external conflict dynamics feature prominently in studies that aim to explain why the peace agreements in Angola and Mozambique were concluded, but not why their implementation failed or succeeded.14 The economic component of the Angolan and Mozambican conflict situations, by contrast, has been addressed in several studies. The contention of these studies is that the termination of war was an economic necessity in Mozambique, whereas the Angolan conflict parties had the resources to continue the war.15
Compatibility of Conflict Situation and Conflict Resolution Technique
The compatibility of conflict situation and conflict resolution technique is an under-researched phenomenon. This strand of the literature on war to peace transitions is clearly the least developed. Among the few scholars who address the relationship between the two is Stephen John Stedman.16 Analyzing spoiler problems in war to peace transitions, he develops a typology of spoilers that consists of three types: limited, greedy, and total. A limited spoiler is a conflict party that pursues limited goals during the peace process, such as the achievement of a power-sharing solution, but does not oppose the overall peace process in principle. A greedy tries to profit from a peace process as much as possible, but is also not principally opposed to peace. A total spoiler, by contrast, is opposed to the transition from war to peace, because it pursues the goal of attaining absolute power.
Stedman argues that the conflict resolution techniques have to be appropriate to the type of spoiler. The strategy of inducement is apt to make a limited spoiler stay in and conclude a war to peace transition. Greedy spoilers need to be socialized into a new set of norms that defines acceptable behavior. In order to make the spoiler abide by the norms, custodians regularly persuade the spoiler, and provide rewards and punishment. Coercion, finally, is the strategy suitable to overcome the problems posed by a total spoiler. In sum, Stedman contends that conflict resolution techniques have to be employed according to the demands of a certain conflict situation. These demands are defined by the type of spoiler, or types of spoilers, that a war to peace transition is troubled with.
Very few studies on Angola and Mozambique deal with the fit or non-fit of conflict resolution techniques and conflict situation. Those that address this issue tend to emphasize the interplay between the economic conflict situation component and conflict resolution techniques. According to this research, Angola's war could not be terminated because the techniques employed were not apt to make UNITA give up its rich resource base. Mozambique's war ended because the conflict resolution techniques used by the international community were the conflict parties' means to overcome economic exhaustion.17
RESEARCH DESIGN
For the purposes of this study, I define a war to peace transition as a process that is aimed at transforming a violent mode of intra-state conflict between two or more conflict parties, through the implementation of a negotiated peace agreement, into a non-violent mode that is carried out within the systems of conflict management of a newly created polity. A war to peace transition is concluded successfully if this transformation occurs, i.e. if the conflict between the former war parties is carried out within the systems of conflict management, containment, and de-escalation of a new polity. A war to peace transition fails if the conflict continues to be carried out in a violent mode. I use the terms peace process and war to peace transition interchangeably.
Explaining the outcomes of the Angolan and Mozambican transitions involves solving a three-fold puzzle: first, despite many similarities between the Angolan and the Mozambican cases, the outcomes of the peace processes were radically different. Both countries had been ravaged by civil war for decades. In Angola, the MPLA government and the UNITA rebels had fought each other since the early 1970s. In Mozambique, the Frelimo government and the Renamo insurgents had battled since the late 1970s. Both countries embarked on a peace process in the early 1990s. MPLA and UNITA concluded the Bicesse Accord, Frelimo and Renamo the Rome Agreement. Both agreements sought to achieve peace through democratization, most importantly through free and fair elections. Both agreements were concluded in a changed international environment where the Cold War had come to an end, and where South Africa was preparing its transition from an Apartheid state to a democracy. Finally, the United Nations deployed multifunctional peace missions in both Angola and Mozambique, in order to facilitate the implementation of the peace agreements. These field missions included military, civilian, and police functions, as opposed to the limitation of traditional peacekeeping and observer missions to a military component. In spite of all these similarities, the Angolan transition failed, whereas the Rome Agreement was successfully implemented in Mozambique. What explains these different outcomes despite the similarities between the two cases?
Second, the Angolan transition consisted of two distinct phases, both of which were unsuccessful. The first phase involved the implementation of the Bicesse Agreement. After this process had failed, MPLA and UNITA signed a new peace accord, the Lusaka Protocol. There are significant differences between these two accords. The Lusaka Protocol, for example, contains a power-sharing agreement between MPLA and UNITA, whereas the Bicesse Accord does not. Moreover, the United Nations considerably stepped up its presence: the United Nations Angola Verification Mission (UNAVEM) II was replaced by UNAVEM III. Nevertheless, the outcome of the Bicesse and the Lusaka implementation process was the same: return to war. What explains this persistent failure of the war to peace transition in Angola?
Third, there were also two distinct phases in Mozambique, both overseen by the United Nations Operation in Mozambique (ONUMOZ): The first phase lasted from the beginning of the implementation process in 1992 to mid-1993. The entire implementation process of the Rome Agreement was behind schedule from the very beginning, because Renamo obstructed the work in the political commissions. Initially, progress in Mozambique was even slower than in Angola. The second phase started in mid-1993 and lasted until the end of the transition period. In mid-1993, Frelimo and Renamo successfully resolved the outstanding issues. The political commissions became the motor of the war to peace transition, as envisaged in the Rome Agreement. Why were these two phases so different? What happened in mid-1993 in Mozambique?
Hence, in attempting to explain the outcomes of the Angolan and Mozambican peace processes, this study asks three questions: first, what explains the variance of the war to peace transitions across the two cases, i.e. Angola and Mozambique?; second, what accounts for the lack of variance within the Angolan case?; and third, why was there a significant variance within the Mozambican case?
Each of the hypotheses outlined in the literature review are tested according to these three questions. Which of those contentions can give an answer to all three questions? The hypotheses are clearly specified and do not pose any difficulties for testing. There is one exception: Stedman's argument on the compatibility of conflict situation and conflict resolution techniques does not specify why a party is an absolute or a limited spoiler. This study uses the three components of the conflict situation as outlined in the literature to determine what kind of spoiler a party is: external conflict dimension, military stalemate, and economic situation.
COMPARING CONFLICT RESOLUTION TECHNIQUES IN ANGOLA AND MOZAMBIQUE
Are the outcomes in Angola and Mozambique due to conflict resolution techniques, irrespective of the conflict situation? This section tests the two main arguments that the literature on conflict resolution techniques has generated: power-sharing agreement and large-scale external intervention.
Power-sharing Agreements Versus Winner-takes-it-all Elections
Is a war to peace transition, independent of the conflict situation, more likely to be successful if power-sharing agreements between the rival parties are included? The Angolan and the Mozambican cases suggest that the answer is no. It fails all three tests. First, it cannot explain the lack of variance during the Angolan war to peace transition. Angola's UNITA rejected the electoral results, independent of whether the elections were winner-takes-all elections or whether there was a power-sharing agreement. The Bicesse Agreement of 1991 formulated the principles for a winner-takes-all election.18 After UNITA went back to war, a second agreement was concluded, the Lusaka Protocol, that stipulated the rules for a Government of National Unity and Reconciliation (GURN), and provincial and local responsibilities for UNITA officials.19 Yet, again UNITA returned to war.20 The presence or absence of power-sharing agreements, therefore, fails to explain the persistent failure to bring the Angolan war to an end.
Second, the power-sharing argument cannot explain why the Mozambican peace process consisted of two distinct phases: failure before June 1993, and success thereafter. Power-sharing agreements were never agreed upon in the Mozambican peace process. The elections were winner-takes-all elections - an arrangement that causes the failure of war to peace transitions according to the power-sharing argument.
Third, the power-sharing hypothesis cannot explain the difference between the Angolan and the Mozambican experiences. Nowhere in the Rome Agreement, concluded between Frelimo and Renamo, are power-sharing agreements mentioned for the time after the elections. Indeed, when Frelimo won the elections, it was not even prepared to appoint Renamo governors for the provinces in which Renamo had won the elections. In contrast to Angola's power-sharing provisions in the Lusaka Protocol, the elections in Mozambique were winner-takes-all elections. Nevertheless, Mozambique's war to peace transition was successful, whereas Angola's was not.
Nature of Third-Party Intervention I: External Pressure
Does "extensive third-party involvement,"21 independent from the conflict situation, cause successful war to peace transitions? This section on third-party involvement focuses on external pressure by state and non-state actors. The next section deals specifically with United Nations peace missions.
For the first time in its history, the United Nations Security Council, acting explicitly under Chapter VII of the Charter, imposed an arms embargo against a party to an internal war as a reaction to UNITA's breaches of the Bicesse Agreement.22 Yet this pressure had no effect. The Lusaka Protocol could not be implemented despite this pressure. No such pressure occurred against any of the Mozambican parties. The arms embargo, a highly significant type of third-party pressure, therefore, cannot explain the lack of variance during the Angolan peace process and the different outcomes of the Angolan and Mozambican transitions. On the contrary: it makes them even more puzzling.
It has been correctly emphasized in the literature that superpower pressure on the Angolan parties declined significantly after the Bicesse Agreement was concluded.23 Pressure by the Soviet Union and the United States to end the war was crucial for the successful negotiations that led to the agreement. Yet, after the agreement was signed, pressure declined. The Soviet Union collapsed in December 1991, and Moscow was compelled to concentrate on its domestic situation rather than to continue its superpower role in world politics. United States involvement was ideologically motivated to such a degree that Washington initially continued to assist UNITA, even after it had begun to wage a new war. It was only eight months after Angola's elections that the United States recognized the MPLA government.
Again, however, it is problematic to conclude from this that the lack of external pressure, in this case superpower pressure, caused the failure of conflict resolution in Angola. The United Nations sanctions have been mentioned already. In addition to this, the United States and Russia resumed their strong pressure shortly after the failure of the implementation of Bicesse. At the beginning of 1998, this led analysts such as Norrie MacQueen to conclude that the "qualified success,"24 as the author put it, of UNAVEM III and MONUA was caused by strong pressure from the United States and Russia. History proved this statement wrong. The latter half of 1998 saw Angola relapse into full-scale war - despite strong pressure from the United States and Russia to settle the conflict. Hence, regardless of the extent of superpower pressure, the Angolan peace process failed.
The Mozambican conflict was less "nurtured" through external pressure than the Angolan one. This had already begun with the peace negotiations in which the Roman Catholic Church and Zimbabwe featured as the most important mediators. The involvement of the United States and the Soviet Union, and later Russia, was confined to a non-formal role. Zimbabwe's mediation efforts proved important during the peace process, as did the efforts of South Africa, Malawi, Portugal, and the United States. These efforts, however, are by no means comparable to the pressure exerted by the United States and the Soviet Union/Russia in Angola that culminated in the first-ever imposition of sanctions against an insurgency movement by the Security Council, acting under Chapter VII of the Charter. Diplomatic pressure, therefore, cannot explain the different outcomes in Angola and Mozambique. Contrary to what the hypothesis on third-party intervention predicts, the transition with more diplomatic pressure was unsuccessful. It also cannot explain the two different phases of the Mozambican transition. Diplomatic pressure in the unsuccessful first phase was not significantly different from the successful second phase.
Nature of Third-Party Intervention II: United Nations Peace Missions
Much of the literature on war to peace transitions in general and on the Angolan and Mozambican cases in particular focuses on the United Nations field missions. It is proposed that the more resources a field mission has, the more likely it is to be successful. Here, the three tests yield mixed results. While the hypothesis gives a plausible explanation for the two different phases of the Mozambican transition, it cannot explain the lack of variance in Angola and the differences across the Angolan and the Mozambican cases.
The hypothesis cannot explain the Angolan case. The Angolan war to peace transition failed independently of the extent of third-party intervention. The United Nations have successively deployed four peace-support operations in Angola. UNAVEM I was a small observer mission designed to monitor and verify the withdrawal of Cuban troops. After this mission had been completed successfully, UNAVEM II was deployed in order to oversee the implementation of the Bicesse Agreement. UNAVEM II was again only a small observer mission. Its maximum strength consisted of slightly more than 1,000 personnel: 350 military observers, 126 United Nations Civilian Police (UNCIVPOL), and the rest civilian personnel, most of whom were electoral observers. Not only the human resources were very limited, but also the monetary resources. Less than US$200 million is indeed a very small amount of money for a peace-support operation that is designed to oversee a war to peace transition. The United Nations spent double the amount for the war to peace transition in Namibia, three times the amount in Mozambique, and ten times in Cambodia.25
The limited mandate of UNAVEM II corresponded to its scarce resources. According to the Bicesse Agreement, the United Nations had the mandate to monitor and verify the ceasefire, the performance of the government's and UNITA's police, as well as the cantonment and demobilization process.26 This monitoring function, however, was confined to monitoring the joint government-UNITA monitoring teams. The United Nations did not have a mandate for direct monitoring and verification. Only with regard to the elections did the United Nations have the mandate for direct monitoring.27 The mandate of the United Nations was limited in two further aspects. First, its role in the political commissions was tenuous because the United Nations representative was only an "invited guest."28 Moreover, the United Nations did not have a mandate with regard to the creation of a new defense force. Given the scarce resources and its limited mandate, it seems logically sound to conclude that the failed implementation of the Bicesse Agreement was due to the insufficient third-party intervention by the United Nations as so often stated in the literature.
The failed implementation of the Lusaka Protocol, however, contradicts this conclusion. A new peace mission was deployed: UNAVEM III. The mandate of UNAVEM III was much more comprehensive than the mandate of UNAVEM II. Only the electoral mandate remained the same: monitor and verify. All other functions went well beyond those stipulated in the Bicesse Agreement. According to the Lusaka Protocol, the United Nations mandate with regard to the ceasefire implementation surpassed mere monitoring and verification, in particular beyond the indirect monitoring mandate for UNAVEM II. The Protocol states that the "overall supervision, control and verification of the re-established cease-fire will be the responsibility of the United Nations . . . with the participation of the Government and UNITA." The United Nations, therefore, assumed primary responsibility for overseeing the ceasefire. Cantonment and demobilization were to be monitored directly as was the performance of the local police, both in contrast to the indirect mandate for UNAVEM II. UNAVEM III had the mandate to monitor and verify the integration of government and UNITA forces into one new national defense force. UNAVEM II did not have a mandate in this subject area at all with the exception of participation in the work of political commissions as "invited guest." UNAVEM III also had a much broader mandate with regard to these commissions. The United Nations assumed chairpersonship in the only commission provided for by the Lusaka Protocol, the Joint Commission, and even had a mandate for good offices and mediation.29
Corresponding to the enlarged mandate were the resources allocated to UNAVEM III. The overall expenses of the peace-support operation amounted to almost US$900 million, more than four times the expenses of UNAVEM II. The maximum strength of UNAVEM III was more than 7,000 persons, which is seven times the number of UNAVEM II's personnel. Perhaps most significantly, UNAVEM III was a peacekeeping operation including more than 6,000 armed troops. UNAVEM II, by contrast, was an observer mission without blue helmets.30
Thus, the hypothesis of third-party intervention fails to explain the dynamics of the Angolan peace process. UNAVEM III had a much more comprehensive mandate and much more human and monetary resources than UNAVEM II. Nevertheless, and in contradiction to the hypothesis of third-party intervention, the result was the same: neither UNAVEM II, nor UNAVEM III were able to fulfil their objectives.
The argument of the significance of external intervention, irrespective of the conflict situation, becomes even more problematic when UNAVEM III and ONUMOZ are compared. The two missions had the same mandate as far as the work in political commissions, cantonment and demobilization, elections, and the performance of the local police are concerned.31 Yet the mandate of UNAVEM III went beyond ONUMOZ's mandate in the subject areas of the performance of local police, the creation of the new armed forces and, in particular, the ceasefire. The Rome Agreement did not provide for the monitoring of the local police. Only one year after the peace agreement was concluded, the conflict parties requested the United Nations to directly monitor and verify the performance of the local police. One month later, in November 1993, the UNCIVPOL component of ONUMOZ arrived in Mozambique. The monitoring mandate of UNAVEM III, by contrast, was included in the Lusaka Protocol. As a consequence, police observers were deployed more promptly than in Mozambique.32 Moreover, the Rome Agreement did not provide for a United Nations mandate for the creation of a new defense force, except for the political work in the commissions. The Lusaka Protocol, by contrast, gave the United Nations the mandate to monitor and verify the integration process of the former rival military organizations of MPLA and UNITA. Finally, the function of ONUMOZ was confined to monitoring and verifying the ceasefire, whereas the mandate of UNAVEM III included the supervision and control of the ceasefire with the participation of the conflict parties; that is it included the legal grounds to request changes of the parties' behavior.33 In short, the mandates of UNAVEM III and ONUMOZ were similar with the former's being slightly more comprehensive than the latter's.
As far as resources are concerned, UNAVEM III and ONUMOZ are again similar, but again slightly more resources were allocated to UNAVEM III. The human resources were alike. The maximum strength of UNAVEM III was well over 7,000, more than 6,500 of which were troops. The maximum strength of ONUMOZ was more than 9,000, of which more than 6,000 were peacekeepers.34 The slight difference in personnel was due to the fact that ONUMOZ oversaw an electoral process, something that was envisaged in the Lusaka Protocol, but no elections were held. The monetary resources allocated amounted to almost US$900 million in the case of UNAVEM III, but only to slightly more than US$500 million in the case of ONUMOZ.
Thus, UNAVEM III had a more comprehensive mandate and allocated more resources to the peace process than ONUMOZ. In Mozambique, there was only one peace mission, the costs of which amounted to slightly more than US$500 million. In Angola, four peace missions, UNAVEM I, UNAVEM II, UNAVEM III, and MONUA cost US$1.3 billion.35 Nevertheless, and contrary to what much of the literature on conflict resolution techniques would predict, UNAVEM III failed to facilitate Angola's war to peace transition, whereas ONUMOZ succeeded.
Yet the hypothesis of third-party intervention sheds light on the difference between the two Mozambican phases. When Renamo obstructed the political commissions in mid-1993, the United Nations established a Trust Fund that was designed, inter alia, to provide the necessary resources for Renamo's transformation into a political party. Renamo resumed its work in the political commissions. After the United Nations had begun to pay Renamo's leader, Alfonso Dhlakama, US$300,000 in order to overcome Renamo's logistical difficulties, the work in the political commissions speeded up, and the successful phase of the Mozambican transition began.
In sum, we are left with a new puzzle: comparing the Angolan and the Mozambican cases, hypotheses on conflict resolution techniques predict that failure in Mozambique is more likely than in Angola. Comparing the two phases of the Angolan peace process, they predict that the implementation of the Lusaka Agreement is more likely to succeed than the implementation of the Bicesse Agreement. Both predictions are clearly wrong. Yet conflict techniques can explain the difference between the two distinct phases of the Mozambican transition. This finding points to the need for a more refined analysis that is able to explain when a certain conflict resolution technique is likely to be successful and when not.36
COMPARING THE ANGOLAN AND MOZAMBICAN
CONFLICT SITUATIONS
Do different conflict situations explain the difference between the Angolan and the Mozambican outcomes? Does a change of the Mozambican conflict situation help us to understand why a phase of success followed a phase of failure? Does an unchanging, intractable conflict situation explain the failure in Angola? This section tests the three key variables found in the conflict situation literature: external conflict dimension, hurting military stalemate, and economic dynamics.
External Conflict Dimension
Until the 1990s, the Angolan and the Mozambican wars were embedded in the regional conflict between Apartheid-South Africa and the Front Line States north of South Africa who fought Apartheid and Pretoria's occupation of South-West Africa (now Namibia). In addition to this, the Angolan war was intertwined with the superpower antagonism of the Cold War. Both aspects of the external conflict situation were resolved by the time the peace processes in Angola and Mozambique began.
The Angolan war was highly internationalized from the very beginning. Foreign aid was the principal tool through which MPLA and UNITA sought to increase their capabilities to fight each other, and, before the Portuguese withdrawal, also to fight the colonial power. Foreign aid was forthcoming and the Angolan war became firmly embedded in the Cold War and in the regional conflict between South Africa and the Front Line States. The Soviet Union, trying to expand its sphere of influence into the Third World, supported the MPLA. Cuba, in accordance with its interventionist communist ideology, provided combat troops. The United States, following the Cold War logic of countering the Soviet threat everywhere in the world, extended its support to UNITA.37 South Africa assisted UNITA with material and human resources as soon as Portugal began to withdraw from Angola in 1974. Pretoria's original objective was to prevent the coming to power of the MPLA, a party that was allied to the South West African People's Organization (SWAPO) and the African National Congress (ANC). When this objective failed, Pretoria used UNITA as a tool to weaken the antagonistic government in Luanda, and to prevent the infiltration of SWAPO and ANC guerrillas into South Africa's de facto province of South West Africa. UNITA established its provisional capital, Jamba, in south-east Angola, and prevented the infiltration of SWAPO and ANC guerrillas into the north-east of South West Africa. South African cross-border raids prevented an infiltration in the west.
The Cold War dimension of Angola's war was the first external dimension to be resolved. With the coming to power of Mikhail Gorbachev, a dètente between the superpowers began that ultimately culminated in the end of the Cold War. In 1986, spokesmen reported from the Twenty-seventh Party Congress of the Communist Party of the Soviet Union that Moscow favored the political settlement of regional conflicts. Cuba's withdrawal was due to Soviet pressure and the resolution of the regional conflict dimension. In 1988, South Africa signed the Brazzaville Protocol according to which it would withdraw from Namibia within one year. The Namibian-Angola border, therefore, lost its strategic significance for Pretoria in the late 1980s. More importantly, of course, Nelson Mandela's release from prison in 1990 was a strong indication that Apartheid, along with foreign policies aimed at protecting it, was about to come to an end.
In contrast to the Angolan war, the Mozambican war was not part of the Cold War, but was caught up in the regional conflict between Apartheid-South Africa and the Front Line States.38 In 1977, Rhodesian intelligence formed, out of a nucleus of exiled Frelimo opponents, a terror movement soon to be named Renamo.39 When the Lancaster House Agreement put an end to white minority rule in what was soon renamed Zimbabwe in 1979, South Africa took over as Renamo's principal patron. In 1978, P.W. Botha had come to power and from then on his notion of a "Total Onslaught" of communism against South Africa shaped Pretoria's foreign policies.40 South Africa began to destabilize Mozambique by supplying Renamo with arms and other equipment from mid- 1979 onward.
In 1984, Mozambique, as a last resort after almost a decade of devastating war, concluded the Nkomati Accord with South Africa. In Maputo's perception, the war was caused by a foreign power and the same foreign power had the key to its resolution. South Africa and Mozambique agreed to cut off any assistance for Renamo and the ANC, respectively. Initially, South Africa did not abide by the agreement. In 1988, however, Mozambique and South Africa decided to revive Nkomati. From then on, Pretoria abided by the agreement. Two years later, as outlined above, South Africa had finally taken the road to freedom for all its people, and Pretoria was about to change fundamentally not only its domestic but also its foreign policies.
Hence, the resolution of the external conflict dimension seems to having been a necessary condition for the peace processes in Angola and Mozambique to be initiated. Yet it cannot account for their different outcomes. Although the overarching external conflicts were resolved in both cases, the outcomes of the peace processes were radically different. Additionally, the first phase of the Mozambican transition was a failure although the external conflict dimension was resolved.
Hurting Military Stalemate
Hurting military stalemates facilitated the conclusion of the Bicesse Agreement and the Lusaka Protocol for Angola, and the Rome Agreement for Mozambique.41 During the Lusaka implementation, the hurting stalemate continued. During the Bicesse and Rome implementations, by contrast, the military situation changed in favor of the insurgents.
Before the Bicesse implementation started, there was a hurting military stalemate between the Angolan conflict parties. MPLA and UNITA had fought each other for over 25 years, first as rival insurgencies and then as government and insurgents, but had been unable to defeat each other. From 1985 onward, the military situation changed critically in favor of the MPLA. By July 1985, government forces had encircled the region around Jamba, UNITA's provisional capital in south-eastern Angola, and threatened UNITA's main logistic base at Mavinga from nearby Cuito Cuanavale. Yet all the government's attempts to take the base failed, although UNITA remained under siege in 1986 and 1987. At the end of 1989, the MPLA attacked Mavinga again. UNITA, however, successfully reverted to guerrilla tactics and defended this critical base. Mavinga showed a military stalemate at a particular moment of time. The MPLA was clearly stronger than UNITA in military terms - in the very contrast to UNITA it could even threaten UNITA's stronghold in the south-east - but the government could not eliminate the insurgency movement on the battlefield.
This military situation was not only a military stalemate, but also a hurting military stalemate. UNITA's very existence had been threatened by the government offensives. The insurgents could only narrowly escape military defeat. Moreover, they had lost much of their territory. In order to remove the strong military pressure, UNITA had to fundamentally change its interaction with the MPLA.42 The MPLA, for its part, was not in a winning situation either. The momentary military situation favored the government, but the future looked bleak. The MPLA faced dramatically dwindling Soviet support. In addition to this, Cuban troops, much more crucial for the MPLA than South African troops for UNITA, had stopped fighting on the side of the government. United States support for UNITA, on the contrary, was increasing. The longer the war lasted, the weaker the MPLA and the stronger UNITA would become.43 This hurting military stalemate was one of the main reasons the two rivals began negotiations and concluded the Bicesse Agreement.
The implementation of the peace treaty, however, critically changed the military situation in favor of UNITA. The ceasefire as well as the cantonment and demobilization process took away the military pressure from UNITA. The government troops had to give up strategically important positions and withdraw to cantonment areas. UNITA, therefore, ceased to be the beleaguered insurgency movement on the brink of military defeat. After UNITA lost the elections, it took advantage of the new favorable military situation. The insurgents went back to war and made rapid gains throughout the country. It took almost two years before the government could stop UNITA's advance on all fronts.
Eventually, another stalemate brought the MPLA and in particular UNITA back to the negotiating table. UNITA had made enormous gains throughout the country in the aftermath of the collapse of the Bicesse implementation process. The rebels, however, were far from militarily defeating the government. UNITA's control had been extended to new areas in the countryside, in particular to eastern Angola. Yet the urban centers, with very few exceptions, remained under government control. Moreover, there were indications that UNITA's military success was about to be reversed. First, UNITA was territorially overstretched and was not able to meet the logistical demands of defending its newly occupied territories. Second, the government was about to launch a powerful counteroffensive that had been prepared through the massive purchase of arms and the hiring of mercenaries. Third, international pressure on UNITA peaked. UNITA had isolated itself internationally by breaking the Bicesse Agreement.44 This had, as mentioned above, even led to the first-ever Chapter VII measure against an insurgency movement by the United Nations. The MPLA, on its part, was further away from military victory against UNITA than ever before. And more importantly, in contrast to the costs of continued military confrontation, the costs for a settlement for the MPLA were low. The main objective of the Lusaka Protocol was to implement the electoral result of Angola's first free and fair elections that had been won by the MPLA.
In contrast to the Bicesse implementation, the Lusaka implementation did not fundamentally change the military situation, at least not in favor of the insurgents. This time it was not the government, but UNITA that had to withdraw its troops from positions they had just gained. In addition to this, UNITA lost its most important regional ally during the Lusaka implementation. In what was soon to be renamed the Democratic Republic of the Congo, UNITA's ally, Mobutu Sese Seko was ousted, and replaced by the MPLA's ally, Laurent Kabila. This threatened UNITA's supply routes through the Congo that had become even more important since Namibia had gained its independence from South Africa. Despite this unfavorable military situation, however, UNITA obstructed the peace process and, finally, resumed the war as it had done at the end of the Bicesse implementation process. UNITA, therefore, returned to war, whether the military situation was favourable or not.
In Mozambique, Frelimo and Renamo fought each other for 15 years. From 1979 to 1980, the Frelimo government seemed capable of militarily eliminating the insurgency group. From 1981 to 1985, however, Renamo expanded from its strongholds in Zambézia, Manica, and Sofala provinces in central Mozambique, and extended its military operations into every province except Cabo Delgado in the north. Attempts by Frelimo - from 1982 onward also by Zimbabwe - to effectively counter Renamo, failed. Due to Zimbabwean support, the military balance shifted again in favor of the government in 1985. By July, Zimbabwean troops had started their military campaign in Renamo's strongholds in the Manica and Sofala provinces. In August 1985, Zimbabwean troops overran casa banana, Renamo's headquarters in the Gorongosa mountains.45 In 1986, a highly successful offensive against Renamo in Zambézia began. By the end of 1987, Renamo had managed to maintain a strong presence in the province only in the areas of Upper Zambézia which were difficult to access. A similarly effective offensive was carried out in Manica and Sofala provinces in 1989. A year later, Renamo no longer held a single town and its control of "Free Mozambique" was confined to a few small and isolated difficult access areas in central Mozambique.
At the end of the 1980s, a hurting military stalemate developed. Not only was Renamo incapable of defeating the government, but the rebels had also lost most of their territories, and even their headquarters. Even worse, Renamo had lost its only patron, South Africa. The government's advance against Renamo-controlled areas, however, did not defeat Renamo. The insurgents increased their acts of large-scale terror and sabotage, targeting the agricultural and industrial centers in central and southern Mozambique. Neither was the Mozambican government capable of military victory, nor was it able to remove Renamo's pressure on its rule, caused by acts of sabotage and large-scale terror. In addition to this, the international community pressed for a peaceful settlement of the dispute.
The military situation in Mozambique changed in much the same way as the Angolan one during the Bicesse implementation. Renamo's military situation improved during the peace process. Most importantly, it was better at the beginning of the successful second than at the beginning of the failed first phase of the implementation process. The ceasefire as well as cantonment and demobilization not only put an end to a highly threatening military situation, but also enabled Renamo to return to its strongholds in central Mozambique. Nevertheless, Renamo did not resume the war. Nevertheless, the second phase of the Mozambican peace process was successful, whereas the first phase was not.
Hence, the research puzzle has again become more puzzling. The presence or absence of a hurting stalemate helps to explain why peace agreements were concluded in Angola and Mozambique, but not why the two processes collapsed in Angola, and why the Mozambican transition was successful. In Mozambique, the implementation process of the Rome Agreement changed from failure to success, although the military situation changed in favor of Renamo during the implementation of the peace agreement. In Angola, UNITA went back to fighting whether or not the military situation improved during the peace process.
Economic Situation
There is a significant difference between the Angolan and the Mozambican conflict situation as far as the economic dynamics of the conflict are concerned. Complete economic exhaustion was a major cause for peace negotiations and conflict resolution in Mozambique. The Angolan parties, by contrast, not only had plentiful resources to continue the war, but also much to lose in case of an electoral defeat.
Angola, as often emphasized in the literature, is potentially one of the richest countries in Africa due to its abundant natural resources, in particular oil and diamonds. From the late 1970s onward, the oil revenues for the government grew enormously. Between 1979 and 1983, for example, annual revenues of crude oil export increased from US$1 billion to US $1.5 billion.46 This trend continued. At the time of the peace process, Angola's oil fields in the north-west of the country and in the Cabinda enclave pumped 500,000 barrels annually. UNITA financed its war with the diamonds in Angola's east. Immediately after the Bicesse implementation collapsed, for instance, UNITA shipped diamonds worth more than US$100 million to Europe and Israel within only two months.47
Throughout the war, neither the MPLA nor UNITA had been able to threaten the resources of the other. Roughly speaking, Angola was divided into two parts: west and east. UNITA dominated the east, the government the west. The insurgents, although determined to put pressure on the government through terror and sabotage, never succeeded in threatening the government's industrial assets in western Angola. The MPLA failed to rid UNITA of its resource base in the east, in particular the central-east and south-east.
Angola's mineral riches were a twofold problem for the peace process: first, the conflict parties had the resources to fight a war, even without or with less external aid than during the Cold War; and second, both parties had much to lose in case the other won the elections. The extension of state administration, laid down in the Bicesse Agreement and the Lusaka Protocol, would encompass the turn over of economically extremely profitable territories to the rival. It would mean losing plentiful resources, and making the rival stronger.
The economic situation in Mozambique was very different. Beginning in the mid-1980s, Frelimo and Renamo were running out of resources needed to wage war for two reasons. First, Mozambique is poorly endowed with natural resources and is one of the least industrially developed countries in the world. Second, Frelimo and Renamo, in contrast to MPLA and UNITA, successfully targeted each other's already scarce resource bases. Renamo succeeded in establishing itself in a strategically important location, the Gorongosa mountains, close to the Beira corridor in central Mozambique, from where it could put enormous pressure on the economy. The insurgents systematically targeted the corridor's agricultural and industrial centers as well as the road and railway traffic to and from the port of Beira. Moreover, the insurgency movement was also very effective in disrupting the economy in the south. Even industrial production on the outskirts of Maputo was hit by Renamo. The insurgents, however, were under enormous economic pressure as well. The joint Zimbabwean-Frelimo offensives were successful in retaking Renamo-controlled areas in central Mozambique, including the rebel's headquarters, thereby further limiting Renamo's resource base.
While the economic situation had been precarious since the outbreak of war, Mozambique's economic situation grew desperate at the beginning of the 1990s. During a severe drought from 1990 to 1992, mass starvation could no longer be averted. The provinces particularly affected were Nampula, Zambèzia, Manica, Sofala, Gaza, and Inhambane. Thus, both government and Renamo-held areas were affected. The Gross National Product (GNP) per capita helps to illustrate the desperate situation. The GNP per capita had dropped dramatically from US $270 in 1985 to US $170 in 1986. Another landslide drop occurred between 1986 and 1987, when it further decreased to US $80. After it had remained at approximately the same level from 1987 to 1991, it fell to a dramatic low of US $65 in 1992.48 Since the country was in a state of complete exhaustion the fighting between the two parties became a low-intensity affair at the beginning of the 1990s. The foes even reached agreement about food supplies to Renamo areas. A fundamental change of interaction between Frelimo and Renamo was an economic necessity.
The economic dynamics offer a plausible explanation why the Angolan transition was a failure. The resource endowment of Angola is a structural factor that did not change during the transition period. UNITA, the loser of the elections, would have lost at least a considerable part of its access to resources, had it accepted the conclusion of the peace process. The economic argument also offers a plausible explanation as to why the outcome of the peace process in Angola was so different from Mozambique. In the latter, peace was an economic imperative, whereas in the former peace was an economic disadvantage for one party. Yet the hypothesis on economic dynamics does not explain why there was a phase of failure and a phase of success in Mozambique. There was economic exhaustion in both phases.
COMPATIBILITY OF CONFLICT SITUATION AND
CONFLICT RESOLUTION TECHNIQUE
The compatibility of the conflict situation and the conflict resolution techniques explains the research puzzle of this study. Due to a lack of compatibility, Angola's Bicesse and Lusaka implementation were unsuccessful. Mozambique's transition, by contrast, succeeded, because the conflict resolution techniques used in Mozambique's second implementation phase matched the needs of the conflict situation.
During and immediately after the Bicesse implementation, the custodians of the peace process, in particular the United Nations and the United States, tried to use a strategy of inducement to make UNITA abide by the peace agreement. Even after UNITA had returned to war, the United States interpreted its actions as understandable. The result of the failure of these inducements was the most destructive phase of the Angolan war that ended only with the conclusion of the Lusaka Protocol.
UNITA was a total spoiler for whom strategies of inducement were inadequate. UNITA had its own state-within-a-state and abundant resources. To give up the occupation of its territories and to transform into a political party would have meant surrendering political and enormous economic power to the arch-enemy. To be sure, some mistakes made during the Bicesse implementation were not made during the Lusaka implementation. Most notably, the United Nations tried to use a strategy of coercion in a number of instances. The arms embargo by the Security Council is the best example for this. However, coercion was limited to a few actions by the United Nations, and even these actions were incomplete. Again, the arms embargo, which was largely ineffective, is the best example for this.
Renamo, by contrast, was not a total, but a limited spoiler. Given the enormously precarious economic situation that the Mozambican rebels faced, there was hardly an alternative to peace, but this peace had to pay off. Renamo did not obstruct the peace as such, but tried to profit from it as much as possible. The inducements by the United Nations were apt to make Renamo abide by the Rome Agreement. Most of these inducements started only in mid-1993, the beginning of the successful phase of the peace process. In mid-May, a Trust Fund for the implementation of the Rome Agreement was created that was designed, inter alia, to provide resources for Renamo's transition into a political party. Italy alone contributed nearly US$6 million to this fund. From September onward, Renamo received payments of US$300,000 a month, finally growing to US$3.9 million by October 1994, when the war to peace transition was successfully concluded. Finally, during his visit to Mozambique, the United Nations Secretary-General, Boutros Boutros-Ghali, used US$1.2 billion in promised grants and loans as inducement for Renamo. He announced that these funds would be held until Renamo and the government had achieved substantial progress in the peace process. With the beginning of these inducements, the peace process was put back on track. By November 1993, the political commissions, previously obstructed by Renamo, had become, as Richard Snyge put it, "the cockpit of the peace process."49
CONCLUSION
The dominant strand of literature on war to peace transitions and most studies on the Angolan and Mozambican peace processes argue that the outcome of war to peace transitions is a function of the conflict resolution techniques employed. By comparing across and within the Angolan and the Mozambican peace processes, this study develops the argument that not only do conflict resolution techniques alone fail to explain the different outcomes in Angola and Mozambique, but they even predict the opposite of what actually happened. According to hypotheses on conflict resolution techniques, the Mozambican transition was more likely to fail than the Angolan peace process. Likewise, the two cases cannot be explained by focussing on the conflict situations only. Instead, the outcomes in Angola and Mozambique are best explained by a non-fit and fit between conflict resolution techniques and conflict situation, respectively. In other words, evidence from the Angolan and the Mozambican experiences strongly suggests that conflict resolution techniques have to match a given conflict situation in order to be successful.
Yet this issue of compatibility has been neglected in the literature. Three points are essential to enhance our understanding of the relationship between conflict resolution techniques and conflict situations. First, studies of war to peace transitions require an in-depth analysis of the case or the cases under scrutiny. It is only through detailed analysis that the peculiar configuration of a certain conflict situation can be identified, and repercussions of conflict resolution techniques can be meaningfully assessed. Second, research on war to peace transitions should draw pragmatically from existing theoretical perspectives to understand the interplay of conflict situation and conflict resolution techniques. Rationalist theories, for example, study the logic of choice under the constraints of a given situation without inquiring into issues of identity and interest formation. This perspective may be useful for gaining insights into conflicts that are predominantly about resources. Yet it may be too limited for the study of ethnic conflicts in which the actors' identities play an important role. Constructivist approaches, inquiring into actors' attachment of meaning to a material world, may be more apt to help explain the dynamics of such conflicts. Third, more research is needed about what conflict resolution techniques, whether successful or not, actually do. This requires researchers to go beyond plausible correlations and inquire into generative mechanisms: how do conflict resolution techniques influence the actors during a peace process? What techniques persuade them and how? What conflict resolution norms are internalized, under what conditions, by what actors? Is mere coercion sufficient to make a party abide by an agreement? If so, under what conditions?
To be sure, these three suggestions do not make the task of the researcher easier. On the contrary, they point to a more complex research agenda. Yet it may be an agenda that helps us to considerably improve our understanding of the dynamics of peaceful conflict resolution - an understanding that is so highly important in today's world.
Endnotes
I would like to thank Sarah Hipworth, Sheila Rusike, and two anonymous referees for valuable suggestions to earlier versions of this paper. In particular, I am very grateful to Annette Seegers, without whose support and in-depth knowledge about Angola and Mozambique this paper would never have been written. The shortcomings of this paper, needless to say, are mine.
1. See, for example, Patrick M. Regan, "Conditions of Successful Third-Party Intervention in Intrastate Conflicts," Journal of Conflict
Resolution 40, no. 2 (June 1996), pp. 5-53.
Return to Article
2. Marina Ottaway, "Democratization in Collapsed States," in I. William Zartman, ed., Collapsed States: The Disintegration and
Restoration of Legitimate Authority (Boulder, CO: Lynne Rienner, 1995), pp. 235-50.
Return to Article
3. Caroline Hartzell, "Explaining the Stability of Negotiated Settlements to Intrastate Wars," Journal of Conflict Resolution 43, no. 1
(February 1999), pp. 3-22.
Return to Article
4. Fen Osler Hampson, "Why Orphaned Peace Settlements Are More Prone to Failure," in Chester A. Crocker, Fen Osler Hampson with
Pamela Aall, eds., Managing Global Chaos: Sources and Responses to International Conflict (Washington, DC: United States Institute of
Peace Press, 1996), p. 535. See also Fen Osler Hampson, Nurturing Peace: Why Peace Settlements Succeed or Fail (Washington, DC: United
States Institute of Peace Press, 1996).
Return to Article
5. Chester Crocker and Fen Osler Hampson, "Making Peace Settlements Work," Foreign Policy, no. 104 (Fall 1996), pp. 54-71; Charles
William Maynes, "Containing Ethnic Conflict," Foreign Policy no. 90 (Spring 1993), pp. 3-21; Barbara F. Walter, "The Critical Barrier to
Civil War Settlement," International Organization 51, no. 2 (Summer 1997), pp. 335-64; Rafael López-Pintor, "Reconciliation Elections:
A Post-Cold War Experience," in Krishna Kumar, ed., Rebuilding Societies After Civil War: Critical Roles for International Assistance
(Boulder, CO: Lynne Rienner, 1997), pp. 43-62; Nicole Ball with Tammy Halevy, Making Peace Work: The Role of the International
Development Community (Washington, DC: Johns Hopkins University Press, 1996).
Return to Article
6. J. Lewis Rasmussen, "Peacemaking in the Twenty-First Century: New Rules, New Roles, New Actors," in I. William Zartman and J.
Lewis Rasmussen, eds., Peacemaking in International Conflict: Methods and Techniques (Washington, DC: United States Institute of Peace
Press: 1997), p. 34. Jonas Savimbi was the founder and leader of UNITA.
Return to Article
7. Norrie MacQueen, "Peacekeeping by Attrition: The United Nations in Angola," Journal of Modern African Studies 36, no. 3 (September
1998), pp. 399-422; Hampson, "Why Orphaned Peace Settlements Are More Prone to Failure," pp. 537-41.
Return to Article
8. Hampson, Nurturing Peace, p. 220.
Return to Article
9. For comparisons between Angola and Mozambique see, Assis V. Malaquias, "UN Peace Operations in Lusophone Africa: Contrasting
Strategies and Outcomes," Journal of Conflict Studies 18, no. 2 (Fall 1998), pp. 66-88; Thania Pfaffenholz, "Angola und Mosambik -
Kriegsbeendigung und Friedenskonsolidierung zwischen Fehlschlag und Erfolg," in Volker Matthies, ed., Vom Krieg zum Frieden:
Kriegsbeendigung und Friedenskonsolidierung (Bremen: Edition Temmen, 1995), pp. 139-64; Joseph Hanlon, "The Ghost of Angola," Work
in Progress, no. 87 (March 1993), p. 29; for Angola, see Margaret Joan Anstee, Orphan of the Cold War: The Inside Story of the Collapse
of the Angolan Peace Process, 1992-93 (London: Macmillan, 1996); Margaret Joan Anstee, "Angola: A Tragedy Not To Be Forgotten," The
World Today 52, no. 7 (July 1996), pp. 190, 191; Hampson, Nurturing Peace, p. 125; Yvonne C. Lodico, "A Peace That Fell Apart: The
United Nations and the War in Angola," in William J. Durch, ed., UN Peacekeeping, American Politics, and the Uncivil Wars of the 1990s
(New York: St. Martin's, 1996), pp. 103-34; Horace Campbell, "The United Nations and Angola," Southern Africa Political & Economic
Monthly 7, no. 2 (November 1993), pp. 46-49; Phyllis Johnson, "Lessons Unlearnt," in Work in Progress 89 (June 1993), p. 7, Christopher
Pycroft, "Angola: The Forgotten Tragedy," Journal of Southern African Studies 20, no. 2 (June 1994), pp. 241-62.
Return to Article
10. Victoria Brittain, "When Democracy is not Enough: Denying Angola's Electoral Result," Southern Africa Report 8, no. 3-4
(January/February 1993), pp. 41-46; Fernando Goncalves, "Angola's Senseless War," Southern Africa Political & Economic Monthly 7, no.
2 (November 1993), pp. 11-13.
Return to Article
11. The quote is taken from I. William Zartman and Saadia Touval, "International Mediation in the Post-Cold War Era," in Crocker and
Hampson, eds., Managing Global Chaos, p. 453. The concept was first developed in I. William Zartman, Ripe for Resolution: Conflict and
Intervention in Africa (Oxford: Oxford University Press, 1985). Since to win an internal war unilaterally means to win it militarily, the focus
of most studies that use the concept of ripe for resolution is the military conflict situation. In this study, therefore, the term "hurting stalemate"
is used synonymously with the term "hurting military stalemate." For a similar operationalization of the concept, see Roy Licklider, Stopping
the Killing: How Civil Wars End (New York: New York University Press, 1993), p. 310.
Return to Article
12. For the significance of great power relations for the ending of internal wars, see, for example, Roger E. Kanet and Edward A. Kolodziej,
eds., The Cold War as Cooperation: Superpower Cooperation in Regional Conflict Management (Baltimore, MD: Johns Hopkins University
Press, 1991); for the significance of regional politics for the ending of internal wars, see, for instance, Carrie Manning, "The Collapse of Peace
in Angola," Current History 98, no. 628 (May 1999), pp. 208-12; and also Victoria Brittain, Death of Dignity: Angola's Civil War (London:
Pluto Press, 1998), pp. 96-100.
Return to Article
13. Mats Berdal and David Keen, "Violence and Economic Agendas in Civil Wars: Some Policy Implications," Millennium 26, no. 3 (1997),
pp. 795-818; for other authors dealing with the economic dimension of conflict, see Stephan Haggard and Robert R. Kaufman, The Political
Economy of Democratic Transitions (Princeton, NJ: Princeton University Press, 1995); Paul Collier, Civil War and the Economics of the Peace
Dividend (Oxford: Centre for the Study of African Economies, 1995).
Return to Article
14. Notable exceptions are the studies by Manning and Brittain. Both argue that regional conflict dynamics are primary causes of the collapse
of the Angolan peace process. MPLA and UNITA clashed as allies of Laurent Kabila and Mobutu Seso Seko in Zaire during the peace process.
This, according to the authors, made the Angolan war to peace transition collapse. See Manning, "The Collapse of Peace in Angola," pp.
208-12; and Brittain, Death of Dignity, pp. 96-100. This argument cannot explain, however, why not only the Lusaka but also the Bicesse
implementation process failed. No external disputes interfered with the latter.
Return to Article
15. Several largely descriptive accounts place much emphasis on the economic component of the conflict situations. See, for example,
Brittain, Death of Dignity; Alex Vines, Peace Postponed: Angola Since the Lusaka Protocol (London: Catholic Institute for International
Relations, 1998). So do some of the more analytical studies, for example, José Gonçalves, Economics and Politics of the Angolan Conflict:
The Transition Re-negotiated (Bellville: Centre for Southern African Studies, 1995); Barry Munslow, "Angola: The Search for Peace and
Reconstruction," in Oliver Furley and Roy May, eds., Peacekeeping in Africa (Aldershot: Ashgate, 1998), pp. 179-205.
Return to Article
16. Stephen John Stedman, "Spoiler Problems in Peace Processes," International Security 22, no. 2 (Fall 1997), pp. 5-53. Some authors, such
as T. David Mason, Joseph P. Weingarten, and Patrick J. Fett, allow for an inclusion of both conflict resolution techniques and conflict
situations, without, however, explicitly dealing with the fit between the two: T. David Mason and Patrick J. Fett, "How Civil Wars End: A
Rational Choice Approach," Journal of Conflict Resolution 40, no. 4 (December 1996), pp. 239-68; T. David Mason, Joseph P. Weingarten,
and Patrick J. Fett, "Predicting the Outcome of Civil Wars," Political Research Quarterly 52, no. 2 (June 1999), pp. 239-68.
Return to Article
17. Alex Vines, "The Business of Peace: 'Tiny' Rowland, Financial Incentives and the Mozambican Settlement," in Jeremy Armon, Dylan
Hendrickson, and Alex Vines, The Mozambican Peace Process in Perspective (London: Conciliation Resources, 1998), pp. 66- 74; Alex
Vines, No Democracy Without Money: The Road to Peace in Mozambique (London: Catholic Institute for International Relations, 1994).
See also Martinho Chachiua and Mark Malan, "Anomalies and Acquiescence: The Mozambican Peace Process Revisited," African Security
Review 7, no. 4 (1998), pp. 18-28; and Stedman, "Spoiler Problems in Peace Processes," p. 5-53.
Return to Article
18. Attachment IV/I of the Bicesse Agreement. The Bicesse Agreement can be found at:
http://www.incore.ulst.ac.uk/cds/agreements/pdf/ang1.pdf.
Return to Article
19. Annex 6/II of the Lusaka Protocol. For the Lusaka Protocol, see http://www.incore.ulst.ac.uk/cds/agreements/lusaka.html.
Return to Article
20. According to Hartzell's coding scheme, the institutionalization of security guarantees in Mozambique was the lowest possible. The
Angolan one - as far as the Lusaka implementation is concerned - was arguably the highest one, but at least medium. See Hartzell, "Explaining
the Stability of Negotiated Settlements to Intrastate Wars," pp. 3-22. Despite the higher institutionalization in Angola, however, the Angolan
war did not come to an end, and the Mozambican one did.
Return to Article
21. Hampson, Nurturing Peace, p. 208. Note that this argument is made without taking conflict situations into account at all, although these
situations differed markedly. The conflict in Cyprus has many characteristics of an inter-state conflict; Namibia's war to peace transition
transformed a de facto colony into an independent state; finally, the peace processes in Cambodia, El Salvador, and Angola aimed at ending
internal wars between governments and insurgents. Also note that the author does not include South Africa in his list. The South African
transition was one of the most remarkable ones of our times - and successfully concluded without "extensive third-party involvement."
Return to Article
22. S/RES/834, 1 June 1993; sanctions were used as a means to compel UNITA more than once: in 1997, the Security Council threatened
sanctions against UNITA again; S/PRST/1997/3, 30 January 1997.
Return to Article
23. Anstee, Orphan of the Cold War, p. 533; Brittain, "When Democracy is Not Enough," pp. 41-46; Fernando Goncalves, "Angola's
Senseless War," pp. 11-13; Inge Tvedten, "The Angolan Debacle," Journal of Democracy 4, no. 2 (April 1993), pp. 108-18.
Return to Article
24. MacQueen, "Peacekeeping by Attrition," pp. 399-422.
Return to Article
25. These and the following figures are taken from United Nations, The Blue Helmets: A Review of United Nations Peace-keeping (New
York: United Nations Publication, 1996), pp. 689-776. The UN spent almost US$400 million for the United Nations Transitional Assistance
Group (UNTAG) in Namibia, more than US$500 million for the United Nations Mission in Mozambique (ONUMOZ), and more than US$1.6
billion for the United Nations Advance Mission in Cambodia (UNAMIC), and the United Nations Transitional Authority in Cambodia
(UNTAC) combined.
Return to Article
26. Cantonment means the assembly of soldiers at designated assembly points. The cantonment stage is followed by their demobilization.
Return to Article
27. For the mandate in the subject areas of cantonment and demobilization, see: Attachment I/I (4) of the Bicesse Agreement; ceasefire:
Attachment III/II (2) and Attachment III/III (3, 4) of the Bicesse Agreement; police performance: Attachment I/I (B) of the Bicesse Agreement;
elections: Attachment III/6 of the Bicesse Agreement.
Return to Article
28. This formulation is used to describe the role of the United Nations in the Joint Political- Military Commission (CCPM) which had the
primary political responsibility during the implementation process of the Bicesse Agreement: Attachment II/2 of the Bicesse Agreement.
Similar formulations are used to describe the United Nations mandate in subsidiary commissions.
Return to Article
29. For the mandate in the subject area of elections, see: Annex 7/II (6) and 8/A (II) of the Lusaka Protocol; ceasefire: Annex 2/I (4) of the
Lusaka Protocol; cantonment and demobilization: Annex 2/II (4), II (5,6) and 8/A (2) of the Lusaka Protocol; local police: Annex 5/II (1) and
8/A (II) of the Lusaka Protocol; integration of forces: Annex 8/A (II) of the Lusaka Protocol; political commissions: Annex 8/A (II) and C
(1) of the Lusaka Protocol. The Joint Commission had the competence to create subsidiary commissions.
Return to Article
30. UNAVEM III was followed by the fourth United Nations peace-support mission. When UNAVEM III was withdrawn after the seemingly
successful inauguration of the GURN in 1998, the United Nations Mission of Observers in Angola (MONUA) was deployed in 1998. All
figures are taken from United Nations sources; expenses for UNAVEM II and III: http://www.un.org/Depts/DPKO/Missions/unavem_f.htm;
troops strength: United Nations, Blue Helmets, pp. 263-65.
Return to Article
31. According to Protocol V/II (1) and Protocol IV/VI (I) of the Rome Agreement, the United Nations were to assume chairpersonship in
two crucial commissions: the Supervision and Control Commission (CSC), and the Cease-fire Commission (CCF). Later in the peace process,
the government and Renamo requested the United Nations also to assume chairpersonship in the Reintegration Commission (CORE), and in
the Commission for the Joint Armed Forces for the Defense of Mozambique (CCFADM). The direct monitoring role of the United Nations
with regard to cantonment and demobilization is laid down in Protocol VI/I of the Rome Agreement and with regard to the elections in Protocol
III of the Rome Agreement. See http://www.incore.ulst.ac.uk/cds/agreements/pdf/moz4.pdf.
Return to Article
32. The deployment began in March 1995, four months after the signing of the Lusaka Protocol. For the mandate of UNAVEM III in the
subject area of police, see Annex 5/II (1) and 8/A of the Lusaka Protocol.
Return to Article
33. UNAVEM III's mandate with regard to the integration of armed forces is laid down in Annex 8/A (II) of the Lusaka Protocol; its mandate
with regard to the ceasefire in Annex 2/I (4) and Annex 8/A (II) of the Lusaka Protocol. For ONUMOZ's mandate in the subject area of
ceasefire, see Protocol VI/I (2) of the Rome Agreement.
Return to Article
34. The status of the peacekeepers in Angola was clear from the very start, but not in Mozambique. According to the Lusaka Protocol, the
means to facilitate the war to peace transition included armed peacekeepers: Annex 8/A (II) of the Lusaka Protocol. The mission requested
in the Rome Agreement, by contrast, was an observer mission, consisting of unarmed military observers and civilian personnel. The United
Nations and the international community put pressure on the Mozambican government to request a peacekeeping contingent, but the latter's
reluctance allowed the deployment to begin not before March 1993, five months after the peace agreement was concluded. The status-of-forces
agreement, allowing the peacekeepers free movement all over the country without previous approval by the government, was not concluded
before May 1993, seven months after the beginning of the transition process.
Return to Article
35. All figures are taken from United Nations sources; for the United Nations expenses in Angola, see
http://www.un.org/Depts/DPKO/Missions/unavem_f.htm; and also http://www.un.org/Dept/DPKO/Missions/monua_f.htm; not included in
the calculation of the overall costs of the United Nations involvement in Angola are the costs of a small United Nations presence that followed
MONUA: United Nations Office in Angola (UNOA); for the expenses of UNOMOZ, see United Nations, Blue Helmets, p. 725.
Return to Article
36. In addition to these two main arguments, some authors have contended after the collapse of the Bicesse implementation that one of the
reasons for failure in Angola and for success in Mozambique were flexible timetables. See, for example, Marina Ottaway, "Angola's Failed
Elections," in Krishna Kumar, ed., Postconflict Elections, Democratization, and International Assistance (Boulder, CO: Lynne Rienner, 1998),
pp. 133-52. In the light of the experience of the Lusaka implementation, however, this argument does not hold either. It is correct that the
Bicesse implementation was characterized by inflexible timetables. The transition proceeded more or less as scheduled whether crucial steps
were implemented or not. It is also true that this was different in Mozambique. The implementation deadline in critical areas was re-scheduled
several times, in particular with regard to cantonment and demobilization. The same, however, and even to a higher degree, applies to the
Lusaka implementation in Angola. The cantonment and demobilization process was declared as concluded one year later than scheduled, the
Government of National Unity and Reconciliation was inaugurated months later than scheduled, and there was also great flexibility as far as
the extension of state administration is concerned.
Return to Article
37. The United States temporarily ceased to play an influential role due to the Clark Amendment in 1975 that put an end to covert assistance
to UNITA.
Return to Article
38. Given its anti-imperialistic ideology and its attempt to expand its sphere of influence, the Soviet Union supported Frelimo during its
struggle against Portuguese colonial rule. This support continued to be significant until the late 1980s, although its scale was far less than in
Angola. In 1986, a series of cuts began. The Soviet Union cut its aid by two-thirds in this year alone. The United States never countered the
Soviet presence in Mozambique. This is partly due to Renamo's conduct of war: under pressure from right-wing groups, the United States
government requested Robert Gershony to report on the Mozambican war. The report describes in detail Renamo's flagrant violations of ius
in bello principles. See Robert Gershony, Summary of Mozambican Refugee Accounts of Principally Conflict-Related Experience in
Mozambique, report submitted to Ambassador Jonathan Moore, Director of the Bureau for Refugee Programs, and Dr. Chester A. Crocker,
Assistant Secretary of African Affairs, Washington, 1988.
Return to Article
39. This was Salisbury's response to Frelimo's policies. Before and after Mozambique's independence, Frelimo supported the Zimbabwe
African National Union (ZANU) and the Zimbabwe African People's Union (ZAPU) in their struggle against white-minority rule in Rhodesia.
When Frelimo came to power, Mozambique became a Front Line State and supported the United Nations boycott against Rhodesia.
Land-locked Rhodesia was cut off from its outlet to the sea, the Mozambican port of Beira.
Return to Article
40. Total Onslaught was Botha's explanation of the replacement of white-minority rule by Soviet sponsored liberation movements in Angola,
Mozambique, and Rhodesia. Pretoria perceived itself as the potential next victim of this alleged onslaught.
Return to Article
41. Concerning the methodology of this section, Zartman, who invented the concept of a hurting stalemate, ties the stalemate to the
perception of the parties concerned. Perceptions, however, are often difficult to test empirically, and this applies in particular to situations of
internal war where very little information - as opposed to propaganda - about the decision-making of the conflict parties is known. Due to
this lack of data, I use two objective criteria instead of perceptions: the military situation within the country and external support for the
conflict parties.
Return to Article
42. In addition to this, UNITA's most important patron, the United States, pressured for a new, peaceful interaction within a democratic
polity.
Return to Article
43. The MPLA tried to re-define its position in the international system, in order to compensate for the loss of Soviet financial assistance.
Yet apart from gaining membership in IMF and World Bank in September 1989 - against initial resistance by the United States - these efforts
did not yield the success that the MPLA had hoped for.
Return to Article
44. For a good overview of this situation, see Paul Hare, Angola's Last Best Chance for Peace: An Insider's Account of the Peace Process
(Washington, DC: United States Institute of Peace Press, 1998), p. 18.
Return to Article
45. Half a year later, Renamo retook its former headquarters, then ineffectively guarded by government soldiers. Zimbabwean troops,
however, seized control again in April 1986.
Return to Article
46. These figures are taken from African Contemporary Record XVII, 1984/85, B 612-B 625.
Return to Article
47. James Ciment, Angola and Mozambique: Postcolonial Wars in Southern Africa (New York: Facts on File, 1997) pp. 211-12.
Return to Article
48. This data are taken from United Nations Statistical Yearbook 50 (1993), p. 163.
Return to Article
49. Richard Snyge, Mozambique: UN Peacekeeping in Action (Washington, DC: United States Institute of Peace, 1997), p. 92.
Return to Article