Negotiated Settlements of Civil Wars vs. Victories

Some civil war scholars counter-argue the practice of halting wars with peace agreements, given that, in many cases, either these agreements failed to last or were unable to produce a quality peace. Monica Duffy Toft strongly argues in this line and shows in her article “Ending Civil Wars: A Case for Rebel Victory?” in International Security that victories pave the way for a lasting peace and democracy, along with economic prosperity, as opposed to any agreed settlements (Toft, 2010a). She shows that rebel victory in particular, as opposed to government victory, results in a quality peace and democracy.

Robert Harrison Wagner has also argued that, in a negotiated settlement, rebels tactically utilize peace agreements to re-gather their strength so that they could go back to a full-fledged war; whereas a victory weakens fighters up to that extent that they could never come back to a full-fledged war situation (taken from Licklider, 1995). Alexander B. Downes also argues that negotiated settlements are not correct solutions to all kinds of civil wars. He shows that settlements that ended ethnic civil wars or identity wars failed two-thirds of the times, whereas the ideological wars also ending in negotiated settlements were hardly repeated (Downes, 2004). William Zartman has also asserted that some fights are worth fighting for (Zartman, 2007). Nonetheless, it should be noted that Roy Licklider has found that military victory is likely to be followed by acts of genocide (in the identity wars cases) (Licklider, 1995).

Toft, and also Caroline Hartzell (Hartzell, 1999), claims that the practice of signing agreements significantly increased after the end of the Cold War (Toft, 2010a). However, Toft has come to her conclusion using the dataset of both Cold War and post-Cold War – data from 1940s to 2000s (see Toft, 2010a; Toft, 2010b; Toft, 2010c). Therefore, taking into account the shift in the world order after the two power blocks lost their ideological interests in civil wars (Stedman et al., 2002), the question arises whether those claims can be empirically verified if the dataset is to be disaggregated into post-Cold War data only. Against this backdrop, this study tests the above-mentioned hypotheses or claims on the evidence.  Table 1 below contains the disaggregated data of Toft (Toft, 2010b) which shows the figure of post-Cold War cases. As the population of the data is rather small to be able to apply any statistical tool, the simple tabulation of data is analyzed – based on percentage calculation.

Table 1

1990 Onwards Civil War Terminations Data:Sigdel-table-final

Source: Data Appendix for Monica Duffy Toft (Toft, 2010b)

According to Toft’s criteria, when violence re-erupts within the period of 5 years after the signing of peace agreements, a war is taken as recurred. She only includes those wars in her data set that have 1,000 battle deaths per year in average (for details, see Toft, 2010a). Incidentally, UCDP (Uppsala Conflict Data Program) includes wars which have 25 battle-related deaths per year and, as a result, has many more war terminations included, compared to Toft (see Kreutz, 2010).

Here, the author disaggregates victory cases in rebel victory and government victory, which strikingly shows that not a single case of civil war that ended in a rebel victory has seen war recurrence. Toft’s hypothesis holds. However, with the available data, no relationship between victory and recurrence can be established. Besides, one major anomaly of this hypothesis is that, among the 8 rebel victory cases, one is the case of Afghanistan. Afghanistan took a path of becoming a failed state when Mujahedeen first, and then Taliban, held sway over almost all parts of Afghanistan.

With regard to the negotiated settlements, there is just over 1 percent war recurrence, thus making any inference from such data unreasonable. In the same way, less than 1 percent of war recurrence out of total terminations suggests: does the war recurrence factor deserve that much attention at all? There is no compelling reason for policy makers to discourage peace agreements and support one side to push a war to its “natural end”. The nature of the data does not provide sufficient scope to be able to accept or discard Wagner’s or Downes’ hypotheses, either. Wagner argues that rebels use negotiations to re-gather strength and re-wage wars, and Downes claims that negotiated settlements are not the right solutions to ethnic or identity wars.

Recurrence Cases in Toft’s Data Set

In the following, the recurred cases of the data are qualitatively analyzed. The cases that are qualified as recurring in the given data set are a victory case of Burundi, and two negotiated settlement cases of Philippines (Moro Rebellion) and Angola. The stalemate or ceasefire case recurrence is not in the scope of this article, hence it is omitted. According to Toft’s data, Philippines and Burundi fall into identity war category, while Angola does not (Toft, 2010c). UCDP categorizes Angola case as a war for seeking central governance (Kreutz, 2010).


In Angola, after the independence war against Portugal, three rival forces that emerged engulfed Angola in violence, turning the state into a Cold War proxy war situation. Among the three groups – FNLA (National Liberation Front of Angola), MPLA (People’s Movement for the Liberation of Angola), and UNITA (National Union for Total Independence of Angola) – UNITA would appear as an incorrigible faction, not accepting defeat and breaching ceasefire agreements. During the Cold War, UNITA re-strengthened itself with international supply, and in the post-Cold War period, it relied on the diamond industry. Although UNITA was induced to sign the ceasefire of 1994, owing to government victories (Toft, 2010b), which gave it a power-sharing in the government, it re-waged armed conflict in 1998.

Wagner’s hypothesis that defeat would leave rebels so weak that they could hardly re-gather the strength to re-wage war does not hold true in this case. Because, taking into account the fact that UNITA was almost completely destroyed at one point, which was then revived by US support into a solid warring side (Ray, 2007; also see “Former Rebels in Angola Shun Unity Meeting” in New York Times, 1992/11/22), enlightens the fact that relapse into war is contingent on the international structure (for the effect of international structure on civil wars, see also Hartzell et al., 2001) and interest, rather than on the sabotaged conditions of warriors, as Wagner claims (Licklider, 1995). In the context of Cold War, UNITA had enjoyed a significant support from the Reagan administration (Ray, 2007).Though the Cold War-type of international interest was irrelevant when UNITA resumed war in 1998, Congo-Brazzaville and Zaire first, and then Burkina Faso and Togo, helped resupply the diamond rich UNITA.

As regards the peace agreement, apparently the reason UNITA reverted back to armed struggle against MPLA government was mainly MPLA’s ill-intentions, disputed elections, etc. However, the character of the agreement and the international community’s position, which failed to induce the spoiler leader Jonas Savimbi vis-à-vis his diamond supplies, were arguably a major motive for which Savimbi embraced war rather than peace (Stedman, 2000).

A strong message UNITA warfare carries for the peacemakers is that, where there regional or international interest and involvement exists, or in the cases of cross-border conflicts, taking civil wars only as an internal war and supporting one side till defeat will perhaps never resolve the problem. Because some incorrigible groups will always be willing to exploit outsiders’ interest and wage wars again and again; whereas it could be inferred that creating a framework by negotiated settlements would convince outsiders that there remained no such groups anymore whom they could mobilize for their “interest”. In 2002, Savimbi was killed in a skirmish with government forces and, subsequently, the army and rebels signed a ceasefire to end the conflict (for the details in UNITA warfare, see Prendergast, 1999).


In the Philippines, the southern islands of Mindanao and Sulu have faced long-standing battles between the Christian Philippino government and the minority Muslim insurgents. President Ferdinand Marcos’ martial law (in 1972) ignited counteroffensives of the Moro rebellion for full-independence of the region Bangsamoro, “the Moro Homeland”, from the Philippines (Bacani, 2005). Though the MNLF (Moro National Liberation Front) accepted peace settlements with the government on increased scope of autonomy in the regions in the Jakarta Peace Agreement in 1996 (Toft, 2010b), the breakaway faction MILF (Moro Islamic Liberation Front) stood as another incorrigible force who would take full-independence as non-negotiable. The agreements between the government and the MILF did not endure. The government resorted to full-fledged attacks to uproot the insurgency and, consequently, the insurgents took a turn into terrorist outfits, increasing the potentiality of transforming the war from a secessionist war into a religious war (Bacani, 2005). In 2000, the all-out offensive from the government not only managed to derail the peace process (Bacani, 2005), but also managed to have MILF declare Jihad against the Philippines.

The Moro rebellion case showed that even heavy offensives would not truncate the insurgency, given that the rebels resorted to guerilla tactic and built nexus with terrorist organizations, which only made the conflict more intractable. Eventually MILF softened its claim from full-independence to autonomy; however, though the government set up a semi-autonomous region in 1996, war continued as the agreement did not address the grievances of southern Muslims. After several episodes of failed ceasefires, in 2012, MILF and the government signed a framework peace plan that would set up autonomous Muslim regions (see “Philippines and Muslim rebels sign key peace plan” at, 2012/10/15). However, the extremist group of Abu Sayyaf, that is affiliated to al-Qaeda, remained unreconciled. In the Moro rebellion case, the taking-one-side-till-the-enemies-are-finished argument remains weak, as the government’s all-out offense that enjoyed international support only made matters worse. Similarly, for the international community, taking the side of rebels was out of the question, given their terrorist character. Although Downes recommends partition as the best solution for such territorial autonomy war (Downes, 2004), perhaps due to the fear of creating precedent, no supporters seem likely to follow such policies. Therefore, it seems there is no alternative to negotiating a settlement.


At the time when Belgium’s domination of Burundi came to an end, the Tutsi minority group had the political predominance. Tutsis’ political domination over a huge Hutu majority was the major grievance of Hutus and the underlying cause of the deep animosity between these two groups. Since immediately after the independence, Hutus revolted against the Tutsi minority rule, and Burundi became a consistent battleground. The long-standing animosity between Hutu and Tutsi caused tit-for-tat-looking attacks. Tutsi government and the Tutsi army, in particular, succeeded to keep the Hutu rebels at bay for years by committing large-scale massacres of Hutu population.

In 1991, when Tutsi government started political liberalizations, a Hutu opposition party, PALIPEHUTU (Party for the Liberation of Hutu People-Burundi), waged terrorist-type attacks against the government. The government again succeeded to suppress that war (Toft, 2010b). However, the war was about to recur soon, and would last for years as one of the most intractable wars of the Great Lake regions of Africa. In 1994, the genocides of Tutsis by Hutus took place. One of the major causes of the intractability of Burundi conflict is not only the ethnic or identity-based animosity among Hutu and Tutsi groups, but the transnational character of the conflict, which encompasses mainly Rwanda, and also DR Congo and Tanzania. The list of countries involved also includes Zaire and Zimbabwe. John Prendergast and David Smock’s report suggests that “conflicts in the Congo, Rwanda, and Burundi are linked inextricably through cross-border insurgencies, cross-border ethnic linkages, and cross-border economic ties” (Prendergast and Smock, 1999). In several episodes of wars in the 1990s, several Hutu groups joined the agreements, but again, the alliance of PALIPEHUTU-FNL (Hutu Forces for National Liberation-Burundi) would remain incorrigible and continued to wage war. Other rebel groups FDD (Forces for the Defense of Democracy) and FNL operated from Congo and Tanzania to launch attacks on targets in Burundi (Prendergast and Smock, 1999).

Wagner’s claim that rebels intend to re-strengthen themselves during the talks do not hold in Burundi’s case because, while several rebel groups joined the peace formula and remained in the process, only one group denied peace. Downes’s claim that identity conflict is better resolved by victory as opposed to negotiated settlement does not hold, either, given that Burundi’s case here is a victory case, but still the war recurred. And Downes refers to either government victory, which would consolidate the status-quo, or rebel victory, which would create independent states. By contrast, Toft’s claim seems to hold because she argues for rebel victory, not government victory. However, given the fact that Burundi is part of another cross-border conflict, it has to be compared with studies that examine whether rebel victory will sustain where neighbors are “re-fueling” either warring sides of state. The studies discussed in this article do not explain this possibility.

Findings and Conclusions

Despite the fact that the shift in foreign policy after the Cold War encouraged to halt wars on the negotiating table (Hartzell, 1999), the evidence shows there is only one more case of negotiated settlement compared to victories – 17 and 16, respectively. This fact raises the policy question that, given the huge human cost and destruction owing to continuation of wars, why are there still almost equal cases of victories? Similarly, the nature of the data set does not provide enough cases to be able to convincingly test the hypotheses related to civil war terminations (particularly in the period of post-Cold War which this article examines). Toft’s criteria that wars having 1,000 battle deaths average per year are included in the dataset misses several other civil war terminations – “low level conflicts” are not included (see Kreutz, 2010). This fact itself gives an optimistic prospect that there are not many wars going on which cause 1,000 battle deaths in a year. In the same way, less than 1 percent of recurrence out of all civil war terminations – just over 1 percent in negotiated settlements – demands caution while claiming that negotiated settlements cause recurrence of war significantly more than victory cases do.

The quantitative evidence, however, strikingly shows that that there is zero recurrence in rebel victory. It could be said that Toft’s hypothesis that rebel victory causes a lasting peace is verified. However, an important policy question arises from there: could the international community support rebels even if they had a brutal character and were fighting to impose their ideology? How could such rebels pave the way for democracy and economic prosperity? Similarly, the qualitative analyses – Angola and Burundi – signal towards a striking fact that, while civil wars are basically considered internal wars amongst internal groups, cross-border linkages, neighbors’ interests, and geo-politics of the region have apparently the decisive role in waging or ending wars.

Therefore, this article argues that combining both internal and external dimensions of a state having civil wars can better explain the phenomenon in question (see also Gleditsch, 2007). For instance, civil war studies can be complemented by applying the realist theory of relative gain. This theory assumes that one state sees its gain not only in its own progress, but in undermining the progress of another state. The reason for selecting realist theory is because most civil war studies are basically realist in nature. They see warring groups’ behavior through self-interest, power, and deceits.

All in all, given the data size and the character of civil wars in the post-Cold War period, the theories built upon the large number of cases covering several decades of Cold War, or even periods before that, seem to be less relevant. Although statistical analyses need as many cases as possible to make good inferences, conflicting conclusions are coming out from them. The point here is to  not to talk about the same methodological debate whether quantitative method/large-N studies or qualitative method/case study is better, but to simply point out the fact that the quantitative approach somewhat should take different world orders into account to better explain contemporary phenomena and prepare a data set accordingly.


Bacani, B. R., 2005. Special Report, The Mindanao Peace Talk: Another Opportunity to Resolve the Moro Conflict in the Philippines. United States Institute of Peace, accessed 08 August 2014.

Downes, A.B., 2004. The Problem with Negotiated Settlements to Ethnic Civil Wars. Security Studies, 13(4), pp.230–279.

Gleditsch, K.S., 2007. Transnational Dimensions of Civil War, Journal of Peace Research, Vol. 44, No.3:293-309.

Hartzell, C., 1999. Explaining the Stability of Negotiated Settlements to Intrastate Wars, The Journal of Conflict Resolution, Vol 43, No. 1: 3-22.

Hartzell, C., Hoddie, M., Rotchild, D., 2001. Stabilizing the Peace After Civil War: An Investigation of Some key Variables, International Organization, Vol. 55, No. 1 : 183-208.

Kreutz, J., 2010. How and When Armed Conflicts End: Introducing the UCDP Conflict Termination Dataset, Journal of Peace Research, Vol. 47, No. 2: 243-250.

Licklider, R., 1995. The Consequences of Negotiated Settlement in Civil Wars 1945 -1993. American Political Science Review, 89(3), pp.681–690.

Prendergast, J., 1999. Angola’s Deadly War: Dealing with Savimbi’s Hell on Earth, United States Institute of Peace, accessed 08 August 2014.

Prendergast, J, and Smock, D.,1999. Post Genocidal Reconstruction: Building Peace in Rwanda and Burundi, United States Institute of Peace, accessed 08 August 2014.

Ray, J.L., 2007. American Foreign Policy and Political Ambition. Washington DC: CQ.

Stedman, S.J., 2002. Ending Civil Wars: The Implementation of Peace Agreements J. S.

Cousens, Elizabeth M., Rotchild, D., Stedman, ed., Colorado: Lynne Rienner.

Stedman, S.J., 2000. Spoiler Problems in Peace Processess, in International Conflict resolution after the Cold War, Stern, P.C., Daniel, D., eds., Committee on International Conflict Resolution.National Research Council.

Toft, M.D., 2010 a. Ending civil wars: A case for Rebel Victory? International Security, 34(4), pp.7–36.

Toft, M.D., 2010 b. Data Appendix for Monica Duffy Toftaccessed 08 August 2014.

Toft, M. D., 2010 c. Securing the Peace:The durable settlements of civil wars. Princeton, New Jersey: Princeton University Press.

Zartman, W.I. ed., 2007. Peacemaking in International Conflict: Methods and Techniques, Washington DC: United States Institute of Peace.

Further Reading on E-International Relations

Please Consider Donating

Before you download your free e-book, please consider donating to support open access publishing.

E-IR is an independent non-profit publisher run by an all volunteer team. Your donations allow us to invest in new open access titles and pay our bandwidth bills to ensure we keep our existing titles free to view. Any amount, in any currency, is appreciated. Many thanks!

Donations are voluntary and not required to download the e-book - your link to download is below.


Get our weekly email