The roots of the Angolan civil war can be traced back to the country’s struggle for independence (which forms a separate case study), during which multiple guerrilla groups struggled against each other as well as against the Portuguese forces. After the collapse of the Portuguese colonial empire in January 1975, the three revolutionary movements—the Marxist Popular Movement for the Liberation of Angola (MPLA), the Ugandan Peoples Army (UPA) which merged with the National Front for the Liberation of Angola (FMLA) in 1962, and the National Union for the Total Independence of Angola (UNITA)—signed the Alvor Accords with the post-Salazar Portuguese government. The agreement granted Angola independence and established a three-way power sharing government. However, collaboration and trust between the three groups was never established, and the country descended into a civil war that was to last almost three decades.[i]
The Angolan civil war initially degenerated into a proxy war between Cold War opponents that intervened to support the different warring sides. With the assistance of Cuban troops and Soviet support, the MPLA managed to win the initial phase of conventional fighting in 1975-1976, oust the FNLA from Luanda and become the de facto Angolan government. The FNLA disintegrated, but the US and South Africa-backed UNITA continued its guerilla warfare against the MPLA-government from its base in the east and south of the country.[ii] Despite the waning of international interest and involvement in the conflict towards the end of the Cold War, UNITA continued its insurgency against the MPLA government until its leader Jonas Savimbi’s death in 2002. The MPLA was able to sustain its rule due to the country’s growing oil revenues, while UNITA forces controlled the diamond trade in the country’s hinterland.[iii]
The Angolan government faced a second insurgency in the oil-rich Angolan exclave Cabinda, where the pro-independence movement FLEC (Front of Liberation of the State of Cabinda) fought for secession. Initially suppressed by the Portuguese and the MPLA, the FLEC engaged in armed struggle against the Angolan state throughout the 1990s and 2000s. However, the number of civilian deaths in this conflict was comparatively small.[iv]
Political histories of the conflict often divide it into three phases: 1975 – 1991, 1992 to 1994, and 1998 – 2002. The end date of each of these phases is marked by a period of peace, however short. However, delineating the history by data on civilian fatalities as described by this project suggests a slightly different periodization of the conflict. We conclude the period of the independence conflict with the consolidation of MPLA in the capitol, Luanda following a massacre in 1977 (separate case study).
This case study begins with the escalation of fighting in the 1980s. Throughout the country’s history from the struggle for independence through 2002, it experienced large-scale violence that clustered in phases with intensive civilian carnage followed by less violent periods.[v] Until the later 1990s, there is no well-disaggregated data on civilian fatalities, so it is very difficult to parse the conflict by year. There is an apparent lull in violence 1994 – 1998, followed by a significant spike in violence against civilians 1998 – 2002.
We also note that the data confirm that considerably more civilians died than were directly killed because of the devastation of the country’s medical, food, livelihood, infrastructure and housing over more than three decades of conflict.
What follows below is a rough outline of how the conflict was fought over its duration in an attempt to describe patterns of violence which might help shed light on the risks facing civilians.
There is very little information about the precise patterns of violence against civilians during the 1980s. By all accounts, the scale of the conflict escalated during this decade, with funding, weapons, and forces pouring into the country from neighboring countries and the US, USSR and Cuba. While accounts of military actions provide a glimpse into the nature of organized violence during this period, documentation of its civilian toll is sparse. Many accounts suggest that the conflict was largely fought with conventional forces, and with efforts to move civilians out of the main battlefields; hence, civilians who were killed are treated as “collateral damage.”[vi] This study includes civilians killed as “collateral damage,” and we note that large-scale fighting can be extremely costly in terms of civilian lives, but in this case, it is simply not clear.
The Angolan War emerged as a key site of proxy war during the 1980s, intensifying after 1986, with the involvement of many countries, neighboring and international. The Angolan President Agostinho Neto died in 1979 and was followed by Jose Eduardo dos Santos. UNITA having re-grouped with support from South Africa, began to overrun government strongholds, striking military and economic targets. On December 24, 1982, their assault on a Huambo suburb marks the beginning of a low-level guerilla insurgency.[vii] The continued success of combined UNITA/South African offensives through 1983 provoked a strong response by the government/Cubans, including air raids and use of heavy weaponry. The conflict took on a semi-conventional nature. UNITA launched two more offensives in 1984.
At the same time (1983 – 4), insurgents in Cabinda also launched an armed uprising.
The war escalated considerably in 1985 – 1987, as increased international military support to both sides changed the capacities of the armed groups. By 1986, the USSR and Cuba were heavily arming and supporting the Angolan government and the US arming UNITA with stinger surface-to-air missiles, among other equipment. Another large battle, noted as the second largest in Africa since World War II, occurred in January – March 1988. With both sides claiming to have pushed the other backed, negotiations resulted in a ceasefire on August 8, 1988. As part of that agreement, the United Nations Angola Verification Mission (UNAVEM) was created. Nonetheless, fighting continued.
As the Cold War wound down, a series of international talks paved the way for the 1991 Bicesse Accords. Included in this agreement was a cease-fire, conditions for the departure of both South African and Cuban forces from Angola, and a political transition to multi-party elections under UN supervision.
However, Savimbi’s UNITA refused to acknowledge its 1992 electoral defeat, choosing instead to proceed with guerilla warfare. As Linda Heywood argues, this marked the spread of violence to areas previously untouched, and she notes that 1992 – 2002 represented “the most sustained period of military and state violence in Angola.”[viii] The MPLA retaliated with indiscriminate killing of UNITA supporters and civilians from the Ovimbundu ethnic group. In Luanda, people “armed with kitchen knives, pistols, clubs, machetes, went from door to door massacring anyone they suspected of belonging to UNITA.”[ix] MPLA supporters living in the UNITA-controlled regions in the central highlands fled to the coast, fearing retribution. UNITA and MPLA forces began planting mines in contested rural regions, causing additional civilian casualties (the Angolan war created 77,000 amputees).[x] Between 1993 and 1994, UNITA bombed strategic cities such as Kuito, Malange and Huambo. Both sides engaged in tactical violence against civilians. The war thus re-escalated between 1992 and 1994, leading to greater devastation than had occurred throughout the previous two decades.[xi]
Fighting in this period was quieted with the signing of the Lusaka Protocol in 1994, which included a cease-fire and program for demobilization and disarmament and the deployment of a bigger UN peacekeeping mission (UNAVEM II). By some reports, however, fighting between the two parties continued.[xii] Schneider et al. argue that in mid-September 1995 there was a weekly death toll of more than 100 victims.[xiii] Overall, however, reports indicate that the years 1994 – 1998 brought an uneasy peace between the two sides.
The 1997 Government of National Reconciliation failed when Savimbi rejected the position of Vice-President. In 1998, President Dos Santos thus initiated another armed offensive against UNITA strongholds. Both the Angolan government and UNITA rebels engaged in scorched-earth offensives, siege warfare and other tactics that primarily targeted civilians. UNITA in particular aimed to push civilian populations into government-held cities in order to stress the government’s capacity to protect its citizens.[xiv] It also engaged in particularly gruesome tactics against civilians, including torture, mutilation, death by crucifixion and abductions.[xv] By 1999, there were 10 million land mines in the country, and Angola topped UNICEF’s child risk measure, which examines risk of death, malnutrition, abuse and development failure for children around the world. A report by the UN in January 1999 warned that the country was facing a potentially catastrophic breakdown.[xvi] By 1999, the number of people displaced by the fighting exceeded 600,000.[xvii]
Ziemke describes the final period of the war as the most lethal for civilians. She writes:
Many quantitative assessments, scholarly articles, United Nations reports, humanitarian accounts, newspaper reports, and firsthand witnesses to the war explicitly mention that something about the conflict changed from 1998 to 2002 (Sites and Learning 2002; Brinkman 1999; see raw dataset). For example, Sites and Learning suggest that as the fighting grew in intensity during 1998, so too did the deliberate, brutal, and direct targeting of civilians (2002). Even in light of the massive humanitarian tragedies that unfolded during 1993-1995, nothing prepared these observers for the seemingly deliberate change in strategy from conventional warfare to terrorizing civilians.[xviii]
She further notes an increase in the visibility and brutality of violence, with widespread reports of torture and mutilation and increases in reporting of rape.[xix]
Best estimate of civilians killed by violence: at least 50,000. The below chart indicates the range of numbers put forward by researchers.
|Authors/Source||Dates considered||Total number of civilian deaths||Notes|
|Ziemke (2008)[xx]||1961 – 2002||56,359||Civilian deaths (based on newspaper reports)|
|Lacina and Gleditsch (2005) [xxi]||1975 – 2002||160,476||“Battle-deaths” (civilians and soldiers)|
|Lacina and Gleditsch (2005)||1975 – 2002||1.5 million||Total number of deaths, with 1,339,525 indirect deaths.|
|Querido (2009) [xxii]||1975 – 2002||1,100,000||Total number of deaths; likely to include military/battle-deaths|
|Brogan (1998)[xxiii]||1976 – 1998||500,000||“These are all very approximate totals.”|
|Sivard (1996)[xxiv]||1975 – 1996||500,000||At least 300,000 since 1992.|
|Heldt, Wallensteen and Nordquist (1992)[xxv]||1975 – 1991||66,666e||Total of 100,000 with 1/3 being military deaths. Estimate total nr of war-related deaths: 300,000 – 500,000.|
|Laffin (1994)[xxvi]||1975(?) – May 1991||350,000||Total number of deaths?|
|Leitenberg (2003)[xxvii]||1980 – 1995||1,000,000||“War-related deaths.” Estimate based on private archives.|
|Wilson and Wallensteen (1988)||1975 – 1985||>11,000||“Casualties” – unclear if civilian & soldiers?|
|Lindgren at al. (1990)[xxviii]||1975 – 1989||25,600||“Casualties” – unclear if civilian & soldiers?|
|Leitenberg (2003)||1995 – 2000||100,000||Total war-related deaths|
It is difficult to match the very high totals with specific time periods (especially of 5 years or less). The only dataset that goes year-by-year and only counts civilian deaths, the Uppsala One Sided Violence database, arrives at significantly lower numbers, though it only starts counting in 1989 and is likely to be an under-estimate.[xxix]
Relying primarily on press reports, Ziemke (2008) searched and coded all of the key battle events, territorial transfers and one-sided massacres that took place in Angola over 41 years of war. She coded a total of 9,216 violent events between 1961 and 2002. For every event she included, when possible, a fatality estimate and the identity of those killed.[xxx] According to her research, the Angolan civil war shows a general trend toward the escalation of violence against civilians over time (see graph), with massacres occurring as combatants lose territory and try to prevent civilian defection to the other side. [xxxi]
Schneider et al. (2011) examine patterns of one-sided violence (directed against civilians and unarmed combatants) in the course of the Angolan civil war. Their data, which is drawn from the Konstanz One-Sided Violence Event Data Set (KOSVED), indicates an increase in violence against civilians towards the end of the 1990s.[xxxii]
Schneider et al. 2011, p. 64.
Lacina and Gleditsch (2005) emphasize the unusually large gap between violent and non-violent deaths in the Angolan civil war.[xxxiii] Many civilians died due to famine, landmines, disease and malnutrition as a direct consequence of war. This aspect of the conflict becomes clearer in the 1990s when reporting on the fate of civilians increased—whether this is a result of increased international presence during this same time period or if the plight of civilians actually worsened precipitously is not entirely clear. Regardless, the scale of suffering is enormous. For instance, UN estimates of civilian deaths between October 1992 and December 1993 at 450,000-500,000. In mid-1993, the UN counted about 1,000 war-related deaths per day.[xxxiv] Wallensteen and Axell (1994) note that there were 16,000 civilian deaths in 1993 alone.[xxxv] The World Bank Violent Conflicts dataset lists 49,383 deaths between 1991-1995, but it is not clear whether this number refers to civilians or combatant deaths or both.[xxxvi]
UNITA’s leader Jonas Savimbi was killed by government forces in February 2002. UNITA and the Government of Angola subsequently signed a ceasefire agreement on April 4, 2002, whereby UNITA conceded military defeat and signed the Luena Memorandum, which re-enacted the Lusaka accords.[xxxvii] The UN Security Council in December 2002 lifted the international sanctions applied against UNITA in 1997, and UNITA began its transformation into a political party.[xxxviii] Savimbi’s death thus brought military and state violence to an end, although inter-social violence against women, the poor and marginalized continued.[xxxix]
However, UNITA had already experienced increasing military difficulties starting in the mid-1990s. After having disregarded the 1992 election results Savimbi lost most of his international legitimacy and support. The end of the Cold War era and the demise of apartheid in South Africa had already weakened UNITA’s external financial and military support networks. New UN sanctions on UNITA’s diamond trade further depleted Savimbi’s primary source of revenue.[xl] By August 1999, Angolan President Dos Santos announced that he would no longer negotiate with Savimbi and instead aim for a clear military victory. The Angolan army then relatively regained several of UNITA’s core territories.[xli] Koloma Beck therefore argues that UNITA was already “in a state of decomposition” when Savimbi died in early 2002.[xlii] As Ziemke notes, however, even as their capacity weakened, UNITA increasingly resorted to targeting civilians.
We code the primary ending in this case as a defeat of the primary perpetrators (a non-state actor, UNITA) by a domestic force, the government. Nonetheless, we also note that the decline of international support for UNITA precipitated and contributed to its decline in capacity. Strategic shifts among international actors were important, although not causal, both in terms of the international withdrawal as well as the UN’s role in helping to oversee mediation efforts. We code this as having multiple victim groups, as civilians associated with both the insurgent and government side were targeted.
Beck, Teresa Koloma. 2012. The Normality of Civil War: Armed Groups and Everyday Life in Angola. Frankfurt am Main: Campus Verlag.
Brogan, Patrick. 1998. World Conflicts: Where and Why They are Happening. Bloomsbury Publishing.
Eck, Kristine and Lisa Hultman. 2007. “Violence Against Civilians in War.” Journal of Peace Research vol. 44 no. 2.
Schneider, Gerald, Lilli Banholzer, and Roos Haer. 2001. “Cain’s Choice: Causes of One-Sided Violence Against Civilians,” in War: An Introduction to Theories and Research on Collective Violence,” edited by Thor G.Jakobsen. Hauppauge, NY: Nova Science Publishers.
Heldt, Biger, Peter Wallensteen and Kjell-Åke Nordquist. 1992. “Major Armed Conflicts in 1991,” in SIPRI Yearbook 1992: World Armaments and Disarmament. Oxford: Oxford University Press, 417–456.
Heywood, Linda. 2011. “Angola and the Violent Years 1975 – 2008: Civilian Casualties” Portuguese Studies Review, 19:1 – 2, 311 – 332.
Human Rights Watch. 1994. “Angola: Arms Trade and Violations of the Laws of War,” Human Rights Watch.
Human Rights Watch. 1999. “Angola Unravels: The Rise and Fall of the Lusaka Peace Process,” Human Rights Watch.
Lacina, Bethany and Nils Peter Gleditsch. 2005. “Monitoring Trends in Global Combat: A New Dataset of Battle Deaths,” European Journal of Population 21: 145-166.
Laffin, John. 1994. The World in Conflict: War Annual 6. London: Brassey’s.
Leitenberg, Milton . 2003. “Death in Wars and Conflicts Between 1945 and 2000,” Peace Studies Program Occasional Paper #29. Ithaca, NY: Cornell University.
Lindgren, Karin, Kenneth Wilson, Peter Wallensteen and Kjell-Åke Nordquist. 1990. “Major Armed Conflicts in 1989,” in SIPRI Yearbook 1990: World Armaments and Disarmament. Oxford: Oxford University Press.
Matloff, Judith. 1997. Fragments of a Forgotten War. Johannesburg: Penguin, 1997.
Petrini, Benjamin. 2010. “Violent Conflict Dataset, 1991-2008.” Washington, D.C.: World Bank, Social Development Department. Available at: http://siteresources.worldbank.org/EXTCPR/Resources/407739-1267651559887/Violent_Conflict_Dataset_combined.pdf.
Prendergast, John. 1999. “Angola’s Deadly War: Dealing with Savimbi’s Hell on Earth.” Special Report, October 12. United States Institute of Peace. Available at: http://www.usip.org/publications/angolas-deadly-war-dealing-savimbis-hell-earth
Querido, Chyanda M. 2009. “State-Sponsored Mass Killing in African Wars – Greed or Grievance?” Institute for Advanced Economic Research 15.
Schneider, Gerald, Lilli Banhozer, and Roos Haer. 2011. “Cain’s Choice: Causes of One-Sided Violence Against Civilians” in War: an introduction to theories and research on collective violence, ed. Tor G. Jakobsen. HAuppage, NY: Nova Science Publi, 57 – 82.
Sivard, Ruth. 1987. World Military and Social Expenditures 1987-88. World Priorities.
Toft, Monica Duffy. 2010. Securing the Peace: The Durable Settlement of Civil Wars – Data Appendix. Princeton University Press.
Wallensteen, Peter. 1994. and Karin Axell, “Major Armed Conflicts,” in SIPRI Yearbook 1994: World Armaments and Disarmament. Oxford: Oxford University Press.
Weigert, Stephen L. 2011. Angola: A Modern Military History, 1961 – 2002. New York: Palgrave Macmilan.
Ziemke, Jennifer. 2008a. “From Battles to Massacres,” Paper prepared for the 3rd annual Harvard-Yale-MIT Graduate Student Conference on Order, Conflict and Violence, Yale University, April 18-19.
[i] Toft 2010, 40.
[ii] Toft 2010, Data Appendix, 40.
[iii] Ziemke 2008a, 26.
[iv] Uppsala University, “UCPD Conflict Encyclopedia,” http://www.ucdp.uu.se/gpdatabase/gpcountry.php?id=4#.
[v] Schneider, Banholzer, Haer 2001, 64.
[vi] Ziemke 2008b, 82.
[vii] Weigert 2011, 74.
[viii] Heywood 2011, 313.
[ix] Heywood 2011, 322.
[x] Heywood 2011, 323.
[xi] James Ciment, Angola and Mozambique: Postcolonial Wars in Southern Africa (New York: Facts on File, 1997), cited in Ziemke 2008b, 26.
[xii] Matloff 1997, 113, 126, 142, 201-3.
[xiii] Schneider, Banholzer, Haer 2001, 64.
[xiv] Prendergast 1999.
[xv] Human Rights Watch 1994, pp. 4, 108; and Human Rights Watch 1999, 3.
[xvi] Ziemke 2008, 29.
[xvii] Heywood, 325.
[xviii] Zimeke 2008, 101.
[xix] Zimeke 2008, 102.
[xx] Ziemke 2008b, 51.
[xxi] Lacina and Gleditsch 2005.
[xxii] Querido 2009, 352.
[xxiii] Brogan 1998.
[xxiv] Sivard 1987.
[xxv] Heldt, Wallensteen and Nordquist 1992.
[xxvi] Laffin 1994, 14.
[xxvii] Leitenberg 2003.
[xxviii] Lindgren, Wilson, Wallensteen and Nordquist 1990.
[xxix] Eck and Hultman 2007.
[xxx] Ziemke 2008b, 5.
[xxxi] Ziemke 2008b, 12.
[xxxii] Schneider, Banholzer, Haer 2001, 64.
[xxxiii] Lacina and Gleditsch 2005, 14.
[xxxiv] Human Rights Watch 1994, 1.
[xxxv] Wallensteen and Axell 1994.
[xxxvi] Petrini 2010.
[xxxvii] Beck 2012, 22.
[xxxviii] Beck 2012, 22.
[xxxix] Heywood, 315.
[xl] Ziemke 2008b, 27.
[xli] Ziemke 2008b, 27.
[xlii] Beck 2012, 108.