The United Nations (UN) and its peacekeeping arm have a serious public relations problem. Media reports on UN peacekeeping tend to highlight the negative aspects of missions, including deaths that occur in the field, sexual abuse by personnel and often stunning organizational failures. Boot (Reference Boot2000), for example, has pointed out that the UN has an ‘almost unrelieved record of failure in its peacekeeping mission’. Autesserre (Reference Autesserre2019) has declared that ‘UN peacekeepers too often fail to meet their most basic objectives’. And headlines such as ‘The UN's Peacekeeping Nightmare in Africa’ are common (Oladipo Reference Oladipo2017).
Politicians have used these assessments to help justify deep cuts in UN peacekeeping funding. Since entering office, President Trump has sought to slash the US contribution to UN peacekeeping by 40 per cent, pledging to provide less than half the $2.5 billion it delivered in 2016 (Williams Reference Williams2018). According to a December 2016 tweet by Trump, ‘the United Nations…is just a club for people to get together, talk and have a good time. So sad!’ (Reuters 2017).
Reality, however, tells a different story. Over the last twenty years, numerous empirical studies have examined the role of third-party peacekeeping in reducing violence around the world. The data overwhelmingly reveal that peacekeeping, especially UN peacekeeping, is surprisingly effective. Using different datasets and statistical models, leveraging different time periods and measuring peacekeeping in somewhat different ways, dozens of researchers at different universities, with diverse funding streams and different preferences, have all found that peacekeeping has a large, positive and statistically significant effect on reducing violence of all sorts. Despite the very real problems associated with UN peacekeeping, it is remarkably effective at bringing peace.
This review article has three goals. The first is to summarize the results of past empirical research to move the debate beyond the question of whether peacekeeping works to the more pressing questions of how, when and why it works. The second goal is to reveal the limitations of current quantitative studies in order to identify areas in which scholars can make big, new contributions to the field. The final goal is to propose a new research agenda that is heavily evaluative – one that informs policy makers about the specific practices, mission compositions, and mandates that work, and also identifies the local, regional, and international conditions that amplify or diminish peacekeeping's success. This type of research could help reduce the costs of peacekeeping operations (PKOs), eliminate some of the negative consequences of interventions and potentially save even more lives.
What the Current Evidence Reveals
Numerous large-n statistical studies have explored the relationship between third-party peacekeeping and different forms of violence (Dorussen Reference Dorussen2014; Gizelis, Dorussen and Petrova Reference Gizelis, Dorussen and Petrova2016); Di Salvatore and Ruggeri Reference Di Salvatore2017; Sandler Reference Sandler2017). What is most striking about these studies is the consistency of their findings. Almost all of them find that peacekeeping is highly effective at preventing violence before it begins, reducing violence in the midst of war and preventing violence from recurring once it has ended. All else equal, countries and regions that receive peacekeeping missions experience less armed conflict, fewer civilian and combatant deaths, fewer mass killings, longer periods of post-conflict peace and fewer repeat wars than those that do not receive peacekeepers. This relationship – between peacekeeping and lower levels of violence – is so consistent across different large-n analyses that it has become one of the strongest findings in the international relations literature to date.
The power of peacekeeping is all the more striking given that the UN tends to intervene in the toughest cases. Multiple scholarly studies have found that the UN Security Council tends to send peacekeepers to countries with more violence, particularly bad governments and ongoing conflict (Beardsley and Schmidt Reference Beardsley and Schmidt2012; Costalli Reference Costalli2013; de Jonge Oudraat Reference De Jonge Oudraat and Brown1996; Fortna Reference Fortna2004a; Fortna Reference Fortna2008a; Gilligan and Sergenti Reference Gilligan and Sergenti2008; Gilligan and Stedman Reference Gilligan and Stedman2003; Hegre, Hultman and Nygard Reference Hegre, Hultman and Nygard2019 Ruggeri, Dorussen, and Gizelis Reference Ruggeri, Dorussen and Gizelis2017). The most recent study by Ruggeri, Dorussen and Gizelis (Reference Ruggeri, Dorussen and Gizelis2018) found that UN peacekeepers are deployed to places experiencing active conflict. Using geographically and temporally disaggregated data on UN peacekeepers' deployment in eight African countries between 1989 and 2006, they find that peacekeepers tend to be deployed to the dangerous frontlines of a conflict. Their research did reveal some problems with deployment. Peacekeepers tended to arrive late and be sent to conflict zones near urban areas, ignoring violence further afield. But the fact that peacekeepers often go to the most demanding places (even if late and disproportionately urban) suggests that academic studies have probably underestimated the effectiveness of peacekeeping at reducing violence.
These findings do not imply that peacekeeping works all the time, or as efficiently and successfully as it could. There are many well-known cases such as Bosnia and Rwanda where UN peacekeeping failed. There are also additional cases such as those in South Sudan and Malí where peacekeeping missions are not progressing well (Autesserre Reference Autesserre2010; Autesserre Reference Autesserre2014; Day Reference Day2019; Howard Reference Howard2008; Van der Lijn Reference Van der Lijn2019).Footnote 1 Moreover, we understand that the data and analyses of even the best quantitative studies are imperfect. Analysis of peacekeeping and violence is constrained by the messy and sometimes unreliable nature of data on death rates, the potentially untenable assumptions made by scholars in employing statistical models in cross-national and cross-conflict analysis, and the hard constraints (both practical and ethical) on scholars' ability to randomize treatments in ways that make airtight causal inferences possible. The findings in the quantitative literature can also appear mixed or contradictory on some topics, when in fact the disparity is due to different research designs and empirical strategies. This can be confusing and can also distract from the bigger and more important picture, which is that peacekeeping works amazingly well given its many challenges.
In what follows we present overwhelming evidence from more than two dozen studies that peacekeepers: (1) reduce civilian and military deaths, (2) prevent the spread of violence, (3) help belligerents achieve peace and (4) help countries maintain the post-conflict peace.Footnote 2 We also provide an overview of the general problems related to peacekeeping that the qualitative literature explores.
Reducing Civilian and Combatant Killing in Civil War
Since 1999, all multidimensional UN peacekeeping operations have a mandate to protect civilians (Howard and Dayal Reference Howard and Dayal2018). Do peacekeepers protect by reducing violence against civilians and combatants in the midst of war? To date, more than a dozen quantitative studies have examined the effect of peacekeeping and outside interventions in protecting civilians, and the answer is yes (Costalli Reference Costalli2013; Carnegie and Mikulaschek Reference Carnegie and Mikulaschek2018; Di Salvatore Reference Di Salvatore2018; Fjelde, Hultman and Nilsson Reference Fjelde, Hultman and Nilsson2018; Hultman, Kathman and Shannon Reference Hultman, Kathman and Shannon2013; Kathman and Wood Reference Kathman and Wood2011; Kathman and Wood Reference Kathman and Wood2016; Bove and Ruggeri Reference Bove and Ruggeri2016; Reference Bove and Ruggeri2018; Kirschner & Miller Reference Kirschner and Miller2019; Melander Reference Melander2009; Phayal Reference Phayal2019; Phayal and Prins Reference Phayal and Prins2019).Footnote 3 These studies examine somewhat distinct sets of cases and time periods, and measure key variables slightly differently. They also often use different research designs. Nonetheless, all find that peacekeeping or similar forces are associated with lower levels of violence against citizens.Footnote 4
We start with a study of all intrastate armed conflicts in sub-Saharan Africa from 1991–2008 by Hultman, Kathman and Shannon (Reference Hultman, Kathman and Shannon2013). They find a significant inverse relationship between UN peacekeeping and violence against civilians: the greater the number of UN peacekeepers committed to protecting noncombatants in these conflicts, the fewer civilians are killed.Footnote 5 The effect is substantial. Their analyses show that, on average, deploying several thousand UN military personnel and several hundred UN police dramatically reduces the number of civilians killed in a conflict.Footnote 6
Why do UN military personnel and police have these beneficial effects? Hultman, Kathman and Shannon theorize that military personnel and police save civilian lives by placing themselves between combatants, reducing direct hostilities and decreasing the incentives to target civilians as a tactic of war. They also suggest that robust peacekeeping forces can impose physical barriers between combatants and civilian targets, making violence a less attractive way for armed factions to extract support or resources from local citizens.
Carnegie and Mikulaschek (Reference Carnegie and Mikulaschek2020) also find that peacekeeping is strongly related to less violence against civilians when one looks at all UN peacekeeping missions deployed to ongoing conflicts between 1989 and 2010. What is particularly noteworthy about this study is that it deals with difficult endogeneity and selection bias issues by using an instrumental variable approach – in this case, variation in Security Council membership. They find that the presence of UN peacekeeping personnel (troops, police and military observers) reduces violence against civilians in civil wars, but only if that violence is perpetrated by rebels and not the government. This finding suggests that peacekeeping may constrain the bad behavior of rebels more than that of governments – a pattern that makes sense given that governments must consent to such missions. Phayal and Prins (Reference Phayal and Prins2019) ask a similar question, using original geocoded data of UN deployments in four key African conflicts (Sudan/Darfur, South Sudan, Democratic Republic of the Congo and Cote d'Ivoire). They find that peacekeepers react quickly to violence against civilians but are slower to act when governments perpetrate such violence.
Di Salvatore (Reference Di Salvatore2018) confirms that UN troops reduce civilian deaths. However, she finds that peacekeepers become less effective as the power asymmetries between armed ethnic groups grow. In examining the UN mission in Sierra Leone from 1997 to 2001, she finds that violence against civilians increased when one ethnic group was very strong and the other was quite weak.
A number of scholars look more closely at the composition of missions and whether it helps or hurts civilians. In a study of UN peacekeeping missions in Africa between 1991 and 2008, Bove and Ruggeri (Reference Bove and Ruggeri2016, Reference Bove and Ruggeri2018) find that peacekeepers recruited from an ethnically and linguistically diverse range of countries and those who are culturally closer to local populations are more successful at reducing violence against civilians. They hypothesize that this type of diversity allows peacekeepers to communicate more effectively with the local population, helping to deter misconduct. This suggests that there may be an ideal composition of UN peacekeeping missions, which includes a combination of ethnolinguistically heterogeneous troops, police and observers who are also culturally similar to local citizens.
Other scholars explore additional questions related to peacekeepers’ characteristics. Haass and Ansorg (Reference Haass and Ansorg2018), for example, use monthly data on the make-up of peace operations and find that when one looks at similarly sized peacekeeping missions, those with a large share of troops from countries with higher-quality militaries (defined as well-trained and equipped soldiers with significant political support) are better able to protect civilians in conflicts.Footnote 7
Do peacekeepers have a similar effect after a war has ended? Kathman and Wood (Reference Kathman and Wood2016) evaluate the role of UN PKOs in protecting civilians after conflict has ended. In a study of twenty-six African countries between 1992 and 2010, they find that fewer civilians are killed if the UN is present, and that this is inversely related to the number of UN military troops. Once again, the size of the peacekeeping force appears to matter. Mirroring studies by Hultman, Kathman and Shannon (Reference Hultman, Kathman and Shannon2013), they find that the opposite is true of UN observers: the greater the number of observers, the greater the number of civilians killed. Why is this the case? The authors argue that observers may signal that the UN is ‘not deeply invested’ in resolving the conflict, perhaps encouraging bad behavior. The bottom line, however, is clear: when UN missions are outfitted with more forces, fewer citizens are targeted and killed.
Melander (Reference Melander2009) looks at how peacekeeping conducted by any state, global or regional organization – not just the UN – influenced the risk of mass killings. This study includes a longer time period (1947 to 2004) and a larger number of civil wars than Hultman, Kathman and Shannon (Reference Hultman, Kathman and Shannon2013). Melander finds that all forms of peacekeeping that included military and/or civilian personnel decrease the risk of mass killings during civil wars.Footnote 8 In a study of a much broader category of military interventions by third parties between 1946 and 2005, Kathman and Wood (Reference Kathman and Wood2011) find that outside military intervention designed to aid the government leads to an increase in government-led genocide and mass killing.Footnote 9 The same was not true of neutral intervention, which tends to exacerbate genocidal violence in the short term but diminish it over time. This work suggests that military intervention may be able to reduce mass killings if the mission is impartial and is willing to remain involved over a longer period of time.
What about combatant deaths? Peacekeepers, particularly UN peacekeepers, also appear to reduce the number of soldiers killed during a civil war. In a study of all intrastate armed conflicts in Africa between 1992 and 2011, Hultman, Kathman and Shannon (Reference Hultman, Kathman and Shannon2014) find that increasing the number of UN armed military troops significantly reduces violence on the battlefield. When more troops are deployed, fewer soldiers are killed in combat. This relationship is confirmed by Beardsley, Cunningham and White (Reference Beardsley, Cunningham and White2018), who find that a greater number of peacekeeping troops in active African conflicts between 1989 and 2008 substantially reduces violence.Footnote 10 The same is not true of police and observers, neither of which was found to reduce the intensity of fighting in a statistically significant way.
In sum, the quantitative literature on peacekeeping to date has consistently identified a positive and statistically significant relationship between the presence of UN peacekeepers and lower rates of civilian and combatant deaths in civil wars. It has also found that a larger presence of peacekeepers is particularly efficacious. When one looks at large patterns over long periods of time, the evidence is clear: the presence of peacekeepers in a country saves lives.
Preventing the Spread of Violence
A second challenge associated with civil wars relates to the spread of violence: such conflicts can move from one region of a country to another (Cunningham, Gleditsch and Salehyan Reference Cunningham, Gleditsch and Salehyan2009). They can also move across international borders, creating a ‘contagion effect’ (Braithwaite Reference Braithwaite2010; Buhaug and Gleditsch Reference Buhaug and Gleditsch2008; Salehyan Reference Salehyan2009). Three quantitative studies assess whether peacekeepers can prevent or reduce these types of expansion. All find that they do.
First, Beardsley and Gleditsch (Reference Beardsley and Gleditsch2015) explore whether peacekeepers can prevent violent conflict from spreading within a country during a civil war. Using geo-referenced conflict polygons between 1990 and 2010, the authors find that peacekeeping missions that deploy at least 1,000 military personnel can limit the spread of violence geographically.Footnote 11 This is especially true as the number of deployed military troops increases: larger missions are the more likely to contain violence.Footnote 12
Beardsley and Gleditsch (Reference Beardsley and Gleditsch2015) theorize that peacekeeping missions limit the geographic movement of violence in three ways: (1) peacekeepers make it more difficult for armed units from any side to move about the country undetected, (2) they establish protected areas against rebel attacks and (3) they decrease the willingness of government forces to use heavy-handed tactics to clear areas of rebels and their supporters.
A second, more recent investigation finds that peacekeeping can reduce local conflict (thus also reducing the possibility of a widening war) even when the overall war does not stop. Using spatially disaggregated UN peacekeeper deployment data for eight African countries between 1989 and 2006, Ruggeri, Dorussen and Gizelis (Reference Ruggeri, Dorussen and Gizelis2018) find that UN peacekeeping decreases the probability that subnational conflict (defined as conflict between a government and rebel forces in a particular locality) will continue for another year. This is true whether the UN deploys a small or large number of peacekeepers, although the authors find that the greater the number of peacekeepers sent, the more effective they are.
In a final study, Beardsley (Reference Beardsley2011) looks specifically at the problem of spillover across international borders. Using event history analysis of all states in the international system from 1946–2005, Beardsley finds that peacekeepers (defined as military personnel sent to a foreign state by the UN, a regional security organization or a group of states) deployed to a neighboring country can prevent conflict from dispersing. If a civil war broke out and peacekeepers were stationed in a neighboring state, the risk of spillover did not significantly increase. However, if no peacekeepers were sent, the expected risk of armed conflict increased by over 70 per cent. Beardsley posits that peacekeepers have this effect because they are able to reduce violence in two ways. First, peacekeepers make it more difficult for transnational insurgencies to move across borders. Secondly, they inhibit the ability of external patrons to supply nascent rebel groups, making it more difficult for them to continue to fight.
Taken together, these studies suggest that sizeable peacekeeping missions (those with at least 1,000 troops) play an important role in reducing the geographic spread of conflict, both across borders and within countries experiencing violence. In what follows, we explore whether they also help combatants reach and implement comprehensive peace agreements.
Reaching and Implementing Comprehensive Peace Agreements
What role, if any, do peacekeepers play in convincing combatants to sign and implement negotiated solutions to civil wars? Once again, the positive effect of external peacekeepers is striking. At least four major empirical studies have examined the role of peacekeepers on signing and/or implementing peace agreements. All find the same pattern: outside peacekeepers stationed in a country during the post-treaty period are key to the successful implementation of peace settlements.Footnote 13
The first two studies examined all civil wars initiated between 1940 and 1992. Walter (Reference Walter1997, Reference Walter2002) finds that combatants are significantly more likely to sign a negotiated settlement if a third party is willing to help verify or enforce the terms.Footnote 14 When a third party steps in, negotiated settlements almost always bring at least five years of peace. When a third party does not, combatants almost inevitably return to war. Combatants appear to be looking down the road and factoring in the presence of peacekeepers when deciding whether to sign an agreement or continue to fight.
In the third study, Kathman and Benson (Reference Kathman and Benson2019) also find that combatants are able to end their war sooner in a negotiated settlement when the UN commits more armed soldiers to a mission. Using monthly data on all minor and major armed conflicts between 1992 and 2004, they find that the greater the number of UN troops, the shorter the time to a negotiated solution.
Finally, Lundgren (Reference Lundgren2016) finds that peacekeepers can play an important role in resolving civil wars even before they arrive. In a study of all mediation attempts by international organizations between 1975 and 2004, he finds that combatants are much more likely to sign a peace agreement to end their war if they expect peacekeeping and monitoring forces to be sent. This suggests that even the prospect of peacekeepers can help end a civil war.
These findings are significant for two reasons. First, they reveal the crucial role that third-party interventions can play in resolving the difficult strategic challenges of post-conflict treaty drafting and implementation. Secondly, they demonstrate that the credible promise of peacekeepers can encourage combatants to sign settlements even before they are put in place. The knowledge that a third party, including the UN, is willing and available to help with implementation appears to motivate combatants to consider a negotiated outcome to their war.Footnote 15
Maintaining Peace in the Aftermath of War
Quantitative studies have carefully examined whether peacekeeping prevents the resumption of civil war. Civil wars have a surprisingly high recidivism rate (Collier and Sambanis Reference Collier and Sambanis2002). Of the 108 countries that experienced some form of civil war between 1946–2017, only 27 per cent avoided a subsequent return to war.Footnote 16 How effective is peacekeeping in preventing a war from returning?
Nine large quantitative studies have attempted to determine the role of peacekeeping in maintaining the peace once civil war has ended (Collier, Hoeffler and Söderbom Reference Collier, Hoeffler and Söderbom2008; Doyle and Sambanis Reference Doyle and Sambanis2000; Doyle and Sambanis Reference Doyle and Sambanis2006; Fortna Reference Fortna2004a; Fortna Reference Fortna2008a; Gilligan and Sergenti Reference Gilligan and Sergenti2008; Hartzell, Hoddie and Rothchild Reference Hartzell, Hoddie and Rothchild2001; Sambanis Reference Sambanis2008; Vivalt Reference Vivalt2015). Additional studies assessed the role of peacekeeping in preventing renewed interstate war (Fortna Reference Fortna2003a; Fortna Reference Fortna2004b; Fortna Reference Fortna2004c). All find a strong relationship between peacekeeping and longer post-war periods of peace.
Hartzell, Hoddie and Rothchild (Reference Hartzell, Hoddie and Rothchild2001) examine the relationship between third-party enforcement of a peace agreement and at least five years of peace after the agreement is signed.Footnote 17 In a study of all intrastate conflicts between 1945 and 1998, they find that peace agreements lasted significantly longer if a third-party enforcer was present to separate or protect former combatants from each other. These assurances could come from third-party states or from regional or international organizations.Footnote 18
In a study of all civil wars from 1944–1997, Doyle and Sambanis (Reference Doyle and Sambanis2000; Reference Doyle and Sambanis2006) find that multilateral UN PKOs (defined as those that include extensive civilian functions, economic reconstruction, institutional reform and election oversight) increase the likelihood that peace lasts 2–5 years after a war ends.Footnote 19 For Doyle and Sambanis, this type of peacekeeping was effective because it signaled international interest in ending the conflict and offered badly needed aid and technical expertise to the parties. Both of these factors are believed to boost domestic capacities and reduce hostility.
Sambanis (Reference Sambanis2008) looks more closely at the effect of different types of UN peacekeeping on peace (observer missions, traditional peacekeeping, multidimensional peacekeeping and enforcement missions). He finds robust results for the hypothesis that UN missions have a positive effect on peacebuilding outcomes, and the longer peacekeepers stayed in a country, the more likely the peace was to last.
Collier, Hoeffler and Söderbom (Reference Collier, Hoeffler and Söderbom2008) find a similar result. In a study of all post-conflict peace periods between 1960 and 2002, they conclude that the more money the UN spent on post-conflict peacekeeping per year, the longer the post-war peace lasted. The effect was large: doubling expenditures reduced the risk of renewed war from 40 to 31 per cent.
Fortna (Reference Fortna2008a) analyzes whether different types of peacekeeping (broadly defined as observer, traditional, multidimensional and enforcement) extend the duration of peace and reduce the risk of another civil war.Footnote 20 In a study of post-conflict countries between 1989 and 1999, she finds that each type of peacekeeping mission significantly reduces the risk of war.Footnote 21 Fortna conservatively estimates that peacekeeping after the Cold War reduced the risk of repeat war by over half. Her less conservative estimates point to an even more dramatic effect, with the risk of war dropping by 75–85 per cent in cases where peacekeepers are deployed.Footnote 22
Fortna also finds that Chapter VI consent-based missions (observation, traditional and multidimensional missions) are as effective as the larger and more robust Chapter VII enforcement missions at maintaining post-conflict peace. This is quite different from the findings noted above about missions designed to protect civilians, where bigger and stronger missions appear to be more effective. These results suggest that the size of a mission and its mandate may matter more if its goal is to protect civilians in the midst of war rather than to prevent war from restarting once it has stopped. They are consistent with qualitative evidence in Fortna (Reference Fortna and Sisk2008b) and Howard (2019) and quantitative evidence in Matanock and Lichtenheld (Reference Matanock and Lichtenheld2017) that political and economic inducements are more important than military force for explaining how peacekeeping works to maintain peace.
Peacekeepers have also been quite effective at maintaining peace after interstate wars. In a study of all such wars between 1947 and 1997, Fortna (Reference Fortna2003b, Reference Fortna2004b) finds that stronger peace agreements (defined as those with more formal mechanisms, such as the deployment of peacekeepers, a formalized agreement, third-party guarantees and dispute settlement procedures) lead to longer periods of post-war peace after a ceasefire agreement is signed. A related study of the effect of peacekeeping in interstate wars over the same period finds that peacekeeping ‘helps even the most deadly of adversaries to avoid war’ (Fortna Reference Fortna2004c, 481).
How reliable are these findings about peacekeeping and a durable peace? As discussed earlier, one potential criticism is that peacekeepers deploy to places where peace is likely to emerge with or without peacekeepers. However, studies that take this endogeneity issue seriously continue to find that the results hold up. Using matching techniques, Gilligan and Sergenti (Reference Gilligan and Sergenti2008) confirm that UN peacekeeping missions significantly lengthen the post-war peace. Examining civil wars that took place between January 1988 and December 2003, they find that peace lasts longer in countries that experienced a UN intervention than in those that did not.Footnote 23 Vivalt (Reference Vivalt2015) uses an instrumental variable approach to help identify the causal effect of peacekeeping on ending episodes of conflict and maintaining peace. She exploits the fact that membership on the Security Council rotates, and that peacekeeping is less likely in countries holding a seat on the council at the time. Her use of a more rigorous causal identification strategy confirms that peacekeeping significantly increases the duration of peace. This provides additional confidence that the relationship between peacekeeping and post-conflict peace is not spurious.
The over-riding conclusion from all these studies is that peacekeeping can play an enormous role in reducing and preventing violence.Footnote 24 When one compares cases with peacekeeping to those without, measuring outcomes in different ways, and controlling for a wide variety of variables, a powerful and consistent picture emerges. Peacekeepers can contain the spread of conflict within and across borders, reduce civilian and military deaths, make settlements more likely and reduce civil war recurrence. In the end, the question is not whether peacekeeping works, but how the world can enable PKOs – especially those led by the UN – to do an even better job.
Mixed and Contradictory Findings
The findings from the qualitative literature tend to showcase peacekeeping's problems and tend to be more pessimistic in tone than quantitative studies.Footnote 25 The qualitative literature has two general strands – one that engages directly with theoretical debates and another that is more policy oriented. Here we address the former, focusing on book-length studies that elucidate five sets of challenges that peacekeeping must address: (1) the difficulty target societies face in adapting to the new political and economic institutions introduced via peacekeeping, (2) the challenges of local ownership, (3) the tradeoffs involved with attempting to gain both internal and external legitimacy, (4) the organizational pathologies inherent in a multinational organization and (5) the difficulties associated with using force. We examine each in turn.
Paris (Reference Paris2004) wrote an early and influential critique of peacekeeping that examines fourteen large, multidimensional PKOs. In each case, the UN sought to promote rapid political liberalization and economic marketization. Paris concludes that quick, externally driven political and economic engineering on such a large scale tends to create new tensions and conflict. Some of Paris' observations have been prescient. For example, some post-conflict countries have massive crime and corruption problems (Aoi, de Coning and Thakur Reference Aoi, de Coning and Thakur2007; Cheng and Zaum Reference Cheng and Zaum2011). However, fifteen years after publication, Paris' direst predictions have not come to pass.Footnote 26 Of the fourteen cases, the UN has fulfilled its complex mandates in eleven, and none has returned to full-scale civil war (Howard Reference Howard2019a).Footnote 27
Secondly, issues of ‘local ownership’ have plagued peace operations from their earliest days. Pouligny's research (Reference Pouligny1999, Reference Pouligny2005, Reference Pouligny2006) has revealed how external peacebuilders often do not understand the local languages, customs and contexts in which they operate. This can lead to misunderstandings and misinterpretations that reduce the helpfulness of peacekeeping. The solution, she argues, is to include more local agents in the peacebuilding process. Similarly, Autesserre (Reference Autesserre2010, Reference Autesserre2014) attributes peacebuilding failures in the Congo and elsewhere to a peacebuilding culture that privileges national-level debates and state building over local-level issues and peacebuilding (see also von Billerbeck Reference von Billerbeck2017). These criticisms have convinced the UN and other peacebuilders to incorporate greater local involvement in peacebuilding processes over time. It is too early to tell if this has enhanced UN effectiveness or not.
Other qualitative studies view internal and external legitimacy as the central dilemma in peacekeeping and state building. Coleman (Reference Coleman2007) discusses an inherent tradeoff between multinational operations that tend to enjoy international legitimacy and local operations that have greater internal legitimacy. Zaum (Reference Zaum2007) similarly points out the inherent paradox of state building: PKOs often compromise local sovereignty in order to establish a sovereign state (see also Whalan Reference Whalan2013).Footnote 28 Relatedly, Barma (Reference Barma2017) illustrates how efforts to establish legitimate and effective political economies in Cambodia, East Timor and Afghanistan have empowered post-conflict elites, unintentionally forging neo-patrimonial order rather than legitimate local democracy.
A fourth strand in the qualitative literature examines organizational pathologies and organizational learning. Barnett and Finnemore (Reference Barnett and Finnemore2004) contend that in Rwanda, a bureaucratic and rules-obsessed UN Secretariat failed to recommend intervention to halt the genocide. Howard's (Reference Howard2008) qualitative comparison of all completed complex peacekeeping missions finds that some of these problems can be corrected through ground-level organizational learning. However, Campbell (Reference Campbell2018) reveals that innovation at the local level often entails breaking with higher-level organizational practices.
A final set of qualitative studies examines the thorny issue of UN military force. Although many scholars and policy makers would like peacekeeping missions to become more ‘robust’, peacekeepers, by design, have neither the capacity nor the resolve to use compellent force (Howard Reference Howard2019a). Peacekeeping is different from other forms of military intervention because of its three doctrinal rules of consent of the parties, impartiality and the non-use of force (except in self-defense, and more recently in defense of the mandate). These rules mean that peacekeepers differ significantly from state military forces (Howard Reference Howard2019b). For instance, peacekeepers in all UN missions hail from dozens of different countries. They do not train together before they deploy. They do not speak each other's languages, or share interoperable matériel. Paddon-Rhoads (Reference Paddon-Rhoads2016) illustrates how contingents of different nationalities interpreted the mandate to use force differently in the Congo. Most importantly, UN peacekeeping forces do not employ a standard military chain of command and control: a force commander from one country cannot order a battalion from another to neutralize a rebel group unless the government of the battalion's sending country agrees (Findlay Reference Findlay2002; Howard Reference Howard2019a). Nevertheless, more often than not, peacekeepers succeed at accomplishing their goals. Howard (Reference Howard2019a) argues that peacekeepers exercise power successfully not through the use of compellent military force, but rather through financial and institutional inducements, and non-material forms of persuasion. Howard traces causal mechanisms in both positive and negative cases, in contrast with much of the qualitative literature, which is concerned with failure alone.
There is a reason why the qualitative literature tends to be negative about peacekeeping while the quantitative literature is not: the two literatures ask different questions and analyze different sets of cases. Qualitative scholars often explicitly select cases of peacekeeping failure so as to examine and identify what went wrong. Quantitative scholars, however, tend to examine the full range of cases, including failures, successes, and cases where peacekeepers were not deployed in order to determine larger patterns over time and across cases. Both are necessary to fully understand the phenomenon.
Disaggregating and Expanding the Data on Peacekeeping
The quantitative literature could improve in two areas. The first is related to data. Most recent quantitative studies rely heavily on macro-level (country and yearly) data that are too blunt to offer detailed policy recommendations to practitioners. We still have almost no systematic evidence to help inform when peacekeepers should arrive in a particular conflict, what exactly they should do once they arrive, how long they need to stay, or how many personnel, and what types of personnel, are needed to have the greatest chance at success given conditions on the ground. If the UN Secretary-General were to ask scholars how best to design a mission – anywhere – all scholars could offer are generalities: ‘go early’, ‘leave late’ and ‘the bigger the force and mandate, the better’. From a policy perspective, this is not very helpful.
Disaggregating the data based on the different contexts in which peacekeepers operate and the different types of missions they undertake – which the qualitative literature is already doing – would help reveal what exactly works in specific environments (Gizelis, Dorussen and Petrova Reference Gizelis, Dorussen and Petrova2016). We have some evidence that peace is harder to generate and maintain when war is ongoing, when the government is winning, when contraband resources fuel the fight, and in identity-based or secessionist conflicts (Fortna Reference Fortna2008a, 18–46; Howard Reference Howard2008; Costalli Reference Costalli2013). But this is still quite vague. PKOs are not monolithic institutions operating in consistent environments. It almost certainly matters whether peacekeepers operate in regions that are resource rich or poor, rural or urban, friendly or hostile, Muslim or Christian, clan based or tribal. Quantitative studies need to take into account the many ways in which local actors and conditions might matter (Gizelis, Dorussen and Petrova Reference Gizelis, Dorussen and Petrova2016).
Quantitative scholars are now starting to disaggregate data in ways that facilitate more granular analysis of different types of missions, mandates, contexts and time periods (Costalli Reference Costalli2013; Di Salvatore Reference Di Salvatore2017; Hultman, Kathman and Shannon Reference Hultman, Kathman and Shannon2013, Reference Hultman, Kathman and Shannon2014; Ruggeri, Dorussen and Gizelis 2016; Ruggeri, Gizelis and Dorussen Reference Ruggeri, Gizelis and Dorussen2013). Costalli (Reference Costalli2013) was the first to analyze the effectiveness of UN peacekeeping using local-level data. In a study on violence and peacekeeping in the Bosnian conflict between 1992 and 1995, he found that the United National Protections Force (UNPROFOR) was unable to reduce violence in the ongoing war because it did not have a sufficiently clear mandate and enough manpower to stop the violence or influence the conflict dynamics. This is a big step forward in our understanding of peacekeeping, but much more work needs to be done.
Scholars have also begun to expand the types of outcomes they study to include the full range of problems peacekeepers are asked to address (Dorussen Reference Dorussen2014). This includes offering humanitarian assistance, promoting human rights, establishing or re-establishing the rule of law, instituting free elections, and supporting economic development and social justice.Footnote 29 This research, however, is still in its infancy and the findings inconclusive. For example, we still do not know whether peacekeepers enable good governance after a civil war. Research by Fearon (Reference Fearon2010) and Walter (Reference Walter2010, Reference Walter2015) found that the presence of peacekeepers in a country does not appear to improve the quality of governance. However, Blair (Reference Blair2019, Reference Blair2020) finds that it does, at least when the data are disaggregated down to the mission level. He demonstrates that the UN Mission in Liberia (UNMIL) – a peacekeeping force of over 16,000 military personnel, civilians and police officers, the goal of which was to monitor a ceasefire agreement – significantly improved governance in that country. Mvukiyehe (Reference Mvukiyehe2018) also examines UNMIL and finds that citizens were more likely to participate in national politics if they had interacted with this mission, suggesting that positive peacekeeping experiences may strengthen the rule of law (see also Bara Reference Bara2020). Further research is needed to tease out the precise relationship between peacekeepers and good governance.
Nor have macro-level studies been able to obtain a consensus on peacekeeping's effect on democratization, partly due to different operationalizations of democracy and democratic transition. Doyle and Sambanis (Reference Doyle and Sambanis2006) argue that peacekeeping helps to foster at least a minimal level of democracy through the terms of the settlement they enforce. Sambanis (Reference Sambanis2008) also finds that peacekeeping has a positive effect on political openness, especially in the short term. Pickering and Peceny (Reference Pickering and Peceny2006) and Joshi (Reference Joshi2013) also find empirical support for the notion that UN interventions foster transitions to democracy. Other studies, however, find that peacekeeping has a negligible, or no, effect on democracy. Gurses and Mason (Reference Gurses and Mason2008) find no significant effect of UN peacekeeping on democratization. Fortna (Reference Fortna and Sisk2008b) finds that peacekeeping has no clear effect on post-war democratization, and argues that the positive and negative effects cancel each other out.Footnote 30 In a more comprehensive study on the same theme, Fortna and Huang (Reference Fortna and Huang2012) also find that peacekeeping has no effect. Qualitative studies that analyze peacekeeping's effect on democratization also suggest mixed effects (Sotomayor Reference Sotomayor2014; Von Hippel Reference Von Hippel2000; Zanotti Reference Zanotti2011). These divergent results call out for additional work on this critical relationship.
Scholars have also begun to collect individual-level data on the effects of peacekeeping on average citizens. Gordon and Young (Reference Gordon and Young2017) use an original survey of a random sample of residents in Port-au-Prince, Haiti to determine if citizens are more likely to co-operate with peacekeepers if the intervening forces engage in positive activities such as security and relief activities rather than abuse. They find that they are. Mironova and Whitt (Reference Mironova and Whitt2017) and Kelmendi and Radin (Reference Kelmendi and Radin2016) use surveys to explore how peace operations affect local attitudes toward peace attempts. These micro-level data offer more nuanced insights into questions of legitimacy that are believed to be key to peacekeeping's success.
Three related issues concerning gender in peacekeeping remain under studied, and ripe for better data and analysis: (1) sexual abuse and exploitation (SEA) in peacekeeping, (2) the effects of gender equality efforts in peacekeeping, and (3) the effects of peacekeeping on sexual and gender-based violence (SGBV) in civil wars. Since the early 1990s, women in countries where peacekeepers are deployed have attested that peacekeepers bring both peace and rape and increased prostitution, and wish the UN ‘had done it better, had been smarter and more thoughtful’ (Whitworth Reference Whitworth2004). Nordas and Rustad (Reference Nordas and Rustad2013) offer the first quantitative study that compares variation in SEA levels by external forces, citing domestic factors such as low development and high sexual violence in the conflict as sources of variation. The findings, however, are limited because the data on SEA committed by external forces remain limited and poor. Rather than relying on mission reporting, Beber et al. (Reference Beber2017) draw on a randomized survey sample of Liberian women between the ages of 18 and 30, and find that more than half had engaged in transactional sex, mainly with UN personnel. The authors estimate that each additional peacekeeping battalion significantly increases the risk that a woman will engage in her first transactional sex. Looking more broadly at human trafficking across countries near peacekeeper deployments, Bell et al. (Reference Bell, Flynn and Martinez Machain2018) demonstrate that deploying more peacekeepers increases the risk of sex trafficking. In an attempt to offset the problems that accompany peacekeepers, Karim and Beardsley (Reference Karim and Beardsley2016) argue that the greater the number of women in peacekeeping, the less SEA committed by peacekeepers, but they acknowledge that more and better data are needed. Not all of the news is disheartening. Kirschner and Miller (Reference Kirschner and Miller2019) show that the higher the number of UN peacekeepers, the less SGBV is committed during conflict. Johansson and Hultman (Reference Johansson and Hultman2019) report similar findings, noting that greater numbers of UN police reduce the risk of sexual violence committed by rebels. Nevertheless, questions remain: what are the best means of reducing SEA committed by peacekeepers? Which peacekeepers are most effective at reducing SGBV in civil wars? Does simply adding more female peacekeepers solve the problems?
The next wave of quantitative studies will almost certainly continue to embrace more fine-grained data at the mission and local levels to help policy makers determine where, when and how peacekeepers are having the most beneficial effect. Disaggregated primary data are now available through the Norwegian Ministry of Foreign Affairs, from the Peacekeeping Operations Location and Event Dataset (PKOLED) developed at the Department of Government at the University of Essex. New data on different dimensions of peacekeeping are also increasingly available, including data on personnel contributions to missions (Kathman Reference Kathman2013), gender composition (Karim and Beardsley Reference Karim and Beardsley2016), mission leadership (Bove, Ruffa and Ruggeri Reference Bove, Ruffa and Ruggeri2019; Bove, Ruggeri and Zwetsloot Reference Bove, Ruggeri and Zwetsloot2017) and georeferenced event data on PKOs (Cil et al. Reference Cil2019; Dorussen and Ruggeri Reference Dorussen and Ruggeri2017).Footnote 31 These data will allow scholars to trace the varying effects that different types of missions have on a wide range of security and non-security outcomes.
Better Research Designs
There is one glaring gap in our knowledge about peacekeeping: there is no consensus on why it is so powerful. Quantitative scholars, qualitative scholars and practitioners offer several different (and often conflicting) explanations of why peacekeeping is correlated with peace.
The next big advance in our understanding of peacekeeping and peace will thus likely come from carefully designed policy evaluations. We believe the best way to do this is through more evaluation research – ideally conducted in a policy lab focused on peacekeeping. Our vision is for a team of scholars to create the peacekeeping equivalent of the Poverty Lab at MIT, or the Immigration Policy Lab at Stanford. Such a lab would enable groups of collaborating scholars to amass primary data on numerous elements of PKOs in order to determine the causal pathways linking peacekeeping to peace. This would not necessarily tell us why peacekeeping works, but it would clearly elucidate what specific policies work in different contexts.
In an ideal world, a Peacekeeping Policy Lab would be created at a major research unit and would involve scholars working on the subject from around the world. Data would be housed centrally and be publicly available. Academics are perfectly situated to provide this important evaluative role. Scholars at major universities around the world could conduct the frank and realistic assessments of what is most effective at promoting a particular outcome without fear of publicizing unpleasant findings (Gowan Reference Gowan2019).
Randomized controlled trials (RCTs) are one way to establish that a causal relationship exists even if they do not tell us how and why it exists. Peacekeeping troops, for example, might undertake pilot programs in randomly assigned villages in a given conflict area to allow researchers to study whether particular types of treatments are more successful than others at delivering order, stability, rule of law or peace. These trials could determine the exact effects of different mandates, force compositions, durations and timings of deployment, as well as locations, and outcomes.
On many key questions, however, RCTs, will not be feasible. These trials are difficult to implement, especially in conflict-affected environments. There are also numerous logistic and bureaucratic challenges associated with working with (and through) peacekeeping organizations. Finally, there are serious ethical issues to consider when offering to send peacekeepers to one village but not another, especially when there is evidence that peacekeepers save lives. Researchers will need to grapple with these challenges, which have also bedeviled RCTs in fields such as economic development. Nevertheless, when logistical and ethical considerations can be managed, the potential rewards of such trials could be high.
Meanwhile, additional types of experimental research are emerging that will help answer questions about the micro-level effects and causal processes of peacekeeping. Scholars are starting to engage in this research on their own in the form of randomized surveys (Myukiyehe and Samii Reference Mvukiyehe and Samii2012), field experiments and natural experiments. (Di Salvatore and Ruggeri Reference Di Salvatore and Ruggeri2017). Blair (Reference Blair2019), for example, finds in a survey experiment that exposure to UNMIL increased Liberian citizens' willingness to go to state authorities to resolve serious incidents of crime and violence rather than non-state authorities. He also finds that exposure to the UN mission increased non-state authorities' reliance on legal over illegal mechanisms of dispute resolution. These findings suggest that UN peacekeeping reduces violence in part by strengthening the rule of law in participating countries.
Quasi-experimental methods could also help us better understand when and how peacekeeping works. A working paper by Mvukiyehe and Samii (2018), for example, presents the results of such an experiment on sets of communities in post-war Liberia. These communities were similarly likely to receive peacekeeping bases, but only some actually did. They use communities’ proximity to a base to study the effects of peacekeepers' presence. Their study finds that peacekeeping missions did not succeed by increasing local security, as many theories have surmised, but rather through mechanisms such as signaling and deterrence.
Lab experiments are another possible route. In a series of laboratory experiments with ethnic Serbs and Albanians under different third-party intervention treatments and peacekeeping environments in Kosovo, Mironova and Whitt (Reference Mironova and Whitt2017) found that third-party enforcement was more effective at promoting norms of trust between ethnic Serbs and Albanians than monitoring alone or no intervention at all.
Finally, the UN Security Council likely engages – unintentionally – in numerous natural experiments as it assigns peacekeepers to certain places and not others, often independently of the expected impacts. Budget cuts, changes in the composition of the Security Council, and changes in leadership at the Department of Peace Operations have all led the UN to adjust missions in ways that are unrelated to events on the ground. These shifts might allow us to assess, for example, how the behavior of target populations changed when the mandate was adjusted. All of these studies are well suited, in complementary ways, to move us beyond macro-level assessments of the effectiveness of peacekeeping to a deeper understanding of the causal mechanisms at work.
Conclusion
Most people do not realize that peacekeeping is as successful as it is – hence the public relations problem discussed earlier. Its negative reputation stems in part from media coverage of UN missions that tends to spotlight failures and abuses rather than positive results. Most people know about the UN's fiasco in Rwanda but not about its achievements in Namibia. Newspaper accounts that focus on dramatic failures with little context make it easy for the public to conclude that the world should, perhaps, invest less in peacekeeping, not more.
The field of peacekeeping, however, is blessed with a mass of high-quality empirical studies that reveal just how critical peacekeeping is to peace. Almost every large-n study finds a strong, statistically significant relationship between peacekeeping and reduced bloodshed. These studies make it clear that without peacekeeping, we would see more violence around the world and costlier types of intervention would be necessary. A 2018 General Accounting Report commissioned by the US Congress found that a unilateral US deployment to the Central African Republic would cost ‘almost 8 times’ more than what the United States contributes to the current UN mission (Government Accounting Office 2018). UN peacekeeping is not only necessary to reduce and prevent violence; it is also a cost-effective way to do so.
We all know that the UN is slow to make decisions, often inefficient in its implementation, hamstrung by red tape, and vulnerable to corruption and scandal. These inefficiencies, however, reveal just how much more the UN could do if it were properly resourced and organized. Hegre, Hultman and Nygard (Reference Hegre, Hultman and Nygard2019), for example, estimated that the UN could have significantly reduced the violence in an additional four to five major conflicts between 2001 and 2013 if the international community had been willing to spend more on peacekeeping and provide existing missions with stronger mandates. Less peacekeeping will not make the world safer. It will only facilitate more violence.
Leaders around the world can play an instrumental role in aiding the UN and other peacekeeping organizations by refusing to cut their budgets. But academics can help as well by continuing the ground-breaking qualitative and quantitative work that is critical to assessing peacekeeping's role in a wealth of different local contexts. More importantly, they can begin to collaborate with practitioners to design studies that evaluate the most pressing policy questions. This will not only ensure the best research designs and analyses; it will also make sure that the findings will be carefully vetted and widely disseminated. If we think peacekeeping is good now, improved data and research designs will only make it better.
Acknowledgements
We would like to thank Filip Savatic and Shea Minter for research assistance, and the anonymous reviewers for helpful suggestions.
