Military integration and intelligence capacity: informational effects of incorporating former rebels

ABSTRACT Military integration seeks to improve counterinsurgency and peacebuilding outcomes by incorporating former rebels into preexisting or new state security forces during or after civil wars. While peacebuilders continue to promote military integration, there is mixed evidence about its effectiveness and the mechanisms through which it affects counterinsurgency and peace duration. One underexplored mechanism is the effect of military integration on intelligence capacity and the information available to security forces. Information is key to successful counterinsurgency and peacebuilding efforts, and I argue that military integration of ex-rebels can improve intelligence capacity by providing gains in knowledge of human and physical geography, access to preexisting social networks and informants, and knowledge of the relative effectiveness of government and rebel tactics. I illustrate these improvements with evidence from conflicts across time and space and brief case narratives from the Philippines, Uganda, and Rwanda. I conclude by discussing policy implications, cases of unsuccessful integration and negative effects on intelligence, and questions for future research on the intelligence aspect of military integration.

toward sustaining peace (Krebs and Licklider 2015). Existing evidence is mixed and it is unclear if the mechanisms through which military integration is posited to contribute to achieving and maintaining peace are generalizable across conflict settings or hold independent causal power (Licklider 2014b). 2 Resolving this question is a major challenge of the military integration research agenda, but here I highlight one previously underexplored potential effect of military integration that might increase the likelihood of conflict resolution and longer-term peace: improved information and intelligence capacity in integrated forces.
I discuss intelligence and local knowledge's near-absence from the military integration literature, and I present a theoretical justification for how military integration can improve the intelligence capacity of integrated forces, potentially resulting in faster conflict resolution or a more sustained peace. I argue that military integration of ex-rebels can improve intelligence capacity through gains in three complementary factors: 1) knowledge of human and physical geography; 2) preexisting social ties and potential informants; and 3) knowledge of rebel tactics, government counterinsurgency tactics, and reasons for their success or failure.
I then examine the empirical record of how military integration has improved intelligence capacity in a variety of civil wars and post-civil war environments, with brief case studies of the Philippines, Uganda and Rwanda. I focus on cases in which security forces facing intelligence weaknesses in counterinsurgency and conflict resolution efforts were able to, at least in part, overcome them by integrating ex-rebels. This focus on successful cases is to demonstrate the plausibility of the mechanism, but in the discussion and conclusion I also examine potential negative cases where rather than yielding improved intelligence capacity, military integration contributed to information leakage or defection.
Military integration is often a component of internationally-brokered peace agreements, yet governments also decide on integration as their own strategy, independent of international facilitation or pressure. I focus on three of these cases of 'autonomous development' of military integration (Licklider 2014b) to demonstrate that integration can lead to improvements in intelligence capacity even in the absence of the external assistance or supervision of forces that might be seen as necessary to overcome commitment problems. Within these cases, there is variation across regime type and across the number of challengers faced by the governments.
The Philippines offers a case of adoption of military integration by a democraticallyelected government facing multiple simultaneous rebellions and splinters, and is also the case examined by the only study of which I am aware that explicitly discusses indepth the potential benefits of integration for intelligence and improved counterinsurgency efficacy (Santos 2010b). Uganda similarly saw the autonomous and repeated use by the current National Resistance Movement government of military integration across multiple insurgencies. However, this was a government that, in contrast to the Philippines, came to power as victorious rebels and established an authoritarian (and now competitive authoritarian) political system. Uganda is also a case with which I was familiar, and the case study draws on archival research conducted in 2016 at the Centre for Basic Research, the Parliamentary Library, and Makerere University. The Rwandan case was selected as a similar case to Uganda in that the current authoritarian Rwandan Patriotic Front (RPF) government came to power through rebel victory. The RPF government has faced one main rebellion against it, though, and Rwanda is a hard case for the embrace of military integration: the RPF came to power in an intensely violent civil war against government forces committing genocide, who then fled into exile and continued fighting, now as rebels.
From this brief analysis, I inductively derive hypotheses for testing in future studies, outlining directions for a research agenda on military integration's effects on intelligence capacity and through it on military effectiveness, conflict resolution, and peace duration. Finally, I discuss policy implications, and also information leakage as a serious potential downside of military integration for intelligence capacity.

Existing literature
The comparative literature on the implementation, effects, and effectiveness of military integration after civil wars is beginning to provide a broader base of evidence against which to evaluate military integration's role in peacebuilding and prevention of civil war recurrence (for reviews, see Knight 2009;Krebs and Licklider 2015;Licklider 2014a). Cross-national quantitative studies have examined factors affecting military integration's inclusion in peace agreements (Hartzell 2014;Hartzell and Hoddie 2007, ch.2); impacts of integration on the likelihood of peace agreement signing and/or implementation to end a civil war (Glassmyer and Sambanis 2008;Hartzell and Hoddie 2007, ch.3;Joshi and Quinn 2017); and whether partially or fully implemented agreements produce sustained peace or conflict recurrence (Berg 2020;Derouen, Lea, and Wallensteen 2009;Hartzell and Hoddie 2007, ch.4;Hoddie and Hartzell 2003;Jarstad and Nilsson 2008;Joshi and Quinn 2017;Mattes and Savun 2009;Ottmann and Vüllers 2015;Quinn, Joshi, and Melander 2019;Toft 2009). Military integration is suggested to be especially helpful as a strategy in civil wars in which a government is facing multiple rebel groups, since ending conflict in one dyad can allow a government to concentrate on combatting remaining rebels Quinn, Joshi, and Melander 2019), now with increased forces, 3 though the peace process itself may lead the integrated group to splinter, with some members joining the state security forces and others returning to the field as insurgents in a fragment of the original group or by joining with other rebel organizations (Johnson 2020(Johnson , 2021Quinn, Joshi, and Melander 2019;Reiter 2016;Stedman 1997).
Comparative case studies have explored military integration's role in peace agreement implementation and peace duration, offering mixed evidence on when integration improves the chances of conflict resolution and sustaining peace (Burgess 2008;Gaub 2011;Krebs and Licklider 2015;Møller and Cawthra 2007;Wilén 2016). Single-case studies have detailed the development and implementation of military integration policies and their relative effectiveness in building peace and constructing more capable militaries (Baaz and Verweijen 2013;Licklider 2014b;Neads 2020;Samii 2013;Santos 2010b;Warner 2013).
Synthesizing the literature's lessons, Licklider (2014a, 4-5) suggests that military integration could prevent civil war recurrence and sustain peace-both avoiding integrated rebel groups rebelling again and reducing the severity or preventing the emergence of other conflicts-through three mechanisms: 1) the presence within the integrated military of individuals or units from different factions increasing the sense of security for factions' fighters or for their popular constituencies; 2) employing and providing financial security to demobilized fighters who might otherwise use or offer their skills as 'specialists in violence' to return to war; and 3) offering security to society overall, increasing the likelihood of socioeconomic and political development that can improve state capabilities and reduce incentives to rebel.
Rarely, however, do the works cited above mention intelligence as an aspect or potential benefit of military integration. Krebs and Licklider (2015, 100) only mention intelligence once, discussing some militaries' desire to segregate units to isolate newly integrated former enemies from sensitive intelligence roles (see also Simonsen 2007, 579). The chapters in Licklider's (2014b) edited volume rarely mention intelligence, and never examine intelligence as a factor significantly affected by military integration. 'Information' is primarily discussed as a factor in the commitment problems inherent in reaching and implementing a settlement (Hartzell 2014;Hoddie and Hartzell 2003;Verweijen 2014, 154). Other studies similarly do not address the potential effects of military integration on intelligence capacity (Burgess 2008;Gaub 2011;Møller and Cawthra 2007;Samii 2013;Warner 2013;Wilén 2016). Driscoll (2012, 121) mentions that integration 'brings soldiers with knowledge of the terrain and tactics,' but does not explore this point further. Baaz and Verweijen (2013) discuss integration's effects on intelligence capacity, though only in a negative case, where integrees were leaking intelligence. The primary exception to this limited discussion of intelligence as a potential benefit, Santos's (2010b) study of the integration of the Moro National Liberation Front into the national security forces of the Philippines, is discussed in detail below, 4 but I now turn to the mechanisms through which military integration may improve intelligence capacity.

Knowledge and networks: intelligence benefits of military integration
Integration of former rebels into the national military offers potential intelligence benefits in three ways. First, rebels bring with them knowledge of the human and physical geography of their areas of operations. Second, and relatedly, rebels have preexisting social networks and access to informants within their areas of operations. Third, rebels possess knowledge of what government tactics succeeded and failed against them and how their own tactics of rebellion worked, offering insights for countering other rebel groups. 5 These factors are important as rebellions tend to arise in similar peripheral areas of low state capacity and rough terrain (e.g. Fearon and Laitin 2003;Hendrix 2011), often among populations that have spawned past insurgencies (Daly 2012), and governments frequently face multiple rebel groups at the same time (Bakke, Cunningham, and Seymour 2012;Christia 2012;Cunningham 2006;Quinn, Joshi, and Melander 2019). Military integration of ex-rebels may occur while fighting is ongoing against a rebel group or after a peace agreement is signed.
If rebels have been fighting an 'irregular war' using guerrilla warfare, the dominant 'technology of rebellion' in civil wars since World War II (Kalyvas and Balcells 2010), 6 then they have often developed close relations with the population in their primary areas of operations. Guerrillas should move among the population 'like fish in water,' to paraphrase Mao Tse-Tung. Unless a group can rely on resource wealth or external patronage (Weinstein 2007) or is interested solely in central state power and wealth (Day and Reno 2014;Mkandawire 2002), maintaining good relations with civilians is crucial to ensuring a nascent insurgency is not snuffed out and gains viability. Rebel groups thus form networks of civilian informants and collaborators who can provide information, supplies, shelter, and other assistance to protect and sustain the group, and try to ensure civilians do not pass information about the rebels to the state (on the importance of information in civil wars, see Kalyvas 2006;Kilcullen 2010;Lewis 2012Lewis , 2017. The balance of social ties and sentiment is especially likely to disproportionately favour rebels when pro-government or intervening forces have been mechanized (cf. Lyall and Wilson 2009) or are otherwise less embedded on the ground, inhibiting the development of trust and deeper social networks, which are vital in intelligence gathering. Nepal's more conventional military forces were seen as distant from the population and lacking trust from the 'masses,' something likely to be improved with the integration of Maoist former guerrilla forces after the 2006 peace agreement (Nayak 2009, 739). Foreign interveners, as outsiders, are at an even greater informational disadvantage, making integration especially valuable for improving intelligence. A US Special Forces officer in Afghanistan's Mohmand Valley fighting against the Islamic State in 2018 described with distaste a notoriously violent former Taliban fighter-turned-police chief, then acknowledged his usefulness: 'he had a lot of contacts in the area and we needed his influence' (Quilty 2018).
In many cases, rebel group members share identities with the civilian population in their main area of operations, identities that government forces may lack. Shared identities can deepen cultural knowledge, social ties, norms of reciprocity, and trust (e.g. Habyarimana et al. 2009). Civilians may, for instance, be more willing during wartime to inform against insurgents to security forces that share their ethnic identity (Lyall, Shiraito, and Imai 2015). Shared identity, as Lyall (2010, 15) notes, provides 'the ability to access information-rich local networks … closed to outsiders,' and in counterinsurgency 'coethnic soldiers face lower levels of uncertainty over who the insurgents are,' permitting more precise 'persuasion or coercion toward' individuals and more credible threats.
At higher levels of politics and central command, military officers and government officials cannot 'properly comprehend the mindset of those involved in insurgency … or the mindset of those who simply reside in an area subject to counterinsurgency' without local social knowledge, argues Souleimanov (2015, 92), making it 'extremely difficult' to anticipate rebel behaviour. Where former or potential rivals are included in the government in interethnic power-sharing agreements, it can lead to better knowledge of events and satisfaction or discontent in the included actors' areas of influence or among their ethnic groups, giving warning signs about nascent insurgencies and helping preserve peace in those constituencies (Roessler 2016).
Beyond the human geography of areas of operations, rebels possess knowledge of the physical geography often surpassing that of government forces. As noted above, insurgencies often begin in 'rough terrain' areas like mountains, forests, swamps, and deserts, where contemporary and historical state presence is limited. Rebels use their territorial knowledge to evade state forces, organize clandestine supply routes, and survive even in harsh conditions, knowledge equally useful to counterinsurgents (see also Brandt 2020, 30-31).
Over the course of a rebellion, rebels also develop organizational knowledge on the effectiveness of the tactics and practices of their own group and their government opponents. Writing about state relationships with paramilitary 'proxies, ' Biberman (2018, 758) argues that ex-rebels make ideal counterinsurgents, because they are 'skilled fighter[s] with insider knowledge of the insurgency and its logistics network.' Military integration can assimilate this knowledge and lessons learned, offering opportunities to improve counterinsurgency strategies and tactics and better understand domestic processes of rebel group formation, fighting, and persistence. Counterinsurgency learning and adaptation by government militaries and intelligence agencies tends to be slow and sporadic (Gentry 2010;Jackson 2008). Successful military integration, however, can offer a shortcut toward developing better counterinsurgency practices, both by incorporating skills and knowledge about rebellion, and also by learning from integrees about state weaknesses and errors that rebels exploited, which government forces can then try to address.
The effects of integration on intelligence capacity likely vary based on three factors: commander buy-in, ex-rebel buy-in, and effective monitoring. Commanders must be receptive to ex-rebels' knowledge and skills to benefit from them, a problem identified by Samii (2014) in Burundi, where officers retained disdain for integrees. Ex-rebels must also feel well-integrated into their new force and the government needs effective monitoring in place. Otherwise, intelligence can flow in the other direction, to ex-rebels' erstwhile comrades, as occurred in the Democratic Republic of Congo (DRC) (Baaz and Verweijen 2013). Solving these commitment problems and achieving trust between the state and ex-rebels are crucial to the success of integration more generally (Krebs and Licklider 2015;Licklider 2014b) and to gaining potential intelligence benefits. This requires socialization to break down enmities (e.g. Jowell 2014; Neads 2020), and also efforts to ensure that ex-rebels are treated fairly and offered aid and opportunities to overcome inequities in education (e.g. Hall 2014; P. A. Martin 2018). Increasing trust over time can increase confidence in the quality of intelligence provided by ex-rebels, though this may also work in the other direction, with vital intelligence tips helping build trust.
The degrees of integration and monitoring likely vary by different institutional arrangements. Ex-rebels can be integrated into military units or brought onto the government side in separate paramilitary units. Second, if they are integrated into preexisting forces, ex-rebels can either be in fully mixed units with government loyalists or under the same overall command, but in separate subunits. 7

Comparative evidence
Ideally, a study of the intelligence effects of military integration could provide systematic data and test for a causal linkage between the presence or absence of the three factors proposed to improve intelligence capacity through integration (rebel geographic and cultural knowledge, social ties, and tactical and strategic knowledge). Yet these types of knowledge and networks are very difficult to code and systematically analyze across a large set of cases, since assessing them requires in-depth interviews with military officials, ex-rebel integrees, and community members; ethnographic observation; and/ or access to sensitive internal military and government documents. Difficulties of variable operationalization and systematic comparison are problems throughout the military integration literature (Krebs and Licklider 2015, 93-95), especially given that multiple, interacting mechanisms (Licklider 2014b) can lead from military integration to effects on conflict outcomes or preservation of peace. I can, however, present anecdotal evidence for the presence and effectiveness of the three modes of intelligence improvement across a spatially and temporally diverse set of civil wars, and more detailed evidence from the cases of the Philippines, Uganda, and Rwanda. This evidence establishes the importance of considering intelligence capacity in assessing the benefits and drawbacks of military integration, showing how it affected counterinsurgency operations or efforts to prevent conflict after security forces integrated former rebels.
Colonial militaries commonly integrated former rebels, recognizing integration's intelligence benefits. The British in 1930s Palestine relied on ex-rebels and local militias who had sometimes opposed British authority in order to combat Arab revolts (Hughes 2016). In 1950s Kenya, the British found that the 'pseudogangs' comprised of former rebels 'proved to be the most potent weapon tracking down the Mau Mau' (Anderson 2005, 241-43;285 in Lyall 2010, 17), drawing on their local and tactical knowledge. In Malaya, one British commander 'began deploying surrendered rebels to evaluate his own companies, a practice that led to significant operational innovations designed to throw off the enemy and exploit its weaknesses;' British forces increasingly used 'surrendered enemy personnel' (SEP) to take advantage of their knowledge of local society and of the structures and motivations of the rebels, including creating 'an interrogation centre staffed entirely by former rebels who were adept at convincing recently captured or surrendered cadres to speak' (Ucko 2010, 24-26). French counterinsurgency in Algeria was more successful after creating the Mobile Security Groups (G.M.S.), units of Algerian Muslims, many of them ex-rebels 'who possessed intimate knowledge of insurgent tactics … and proved invaluable to counterinsurgency operations' (Clark 2006, 9).
In post-colonial contexts, foreign interveners have also sought to integrate former rebel forces to overcome the 'invader's dilemma' (Tamm 2020) and gain local knowledge and legitimacy, in addition to military capacity. Interveners have then pushed for these local forces' integration into host state militaries. The United States in Iraq recruited exinsurgents and militias in Anbar province to act as scouts and provide intelligence, and then formalized the recruits into the 'Sons of Iraq,' paramilitary forces that provided crucial local knowledge, intelligence, and fighting power in combatting Al Qaeda in Iraq-though efforts to integrate these fighters into the Iraqi Security Forces have been bumpy (see e.g. Rabasa et al. 2011, ch.6). In Syria, Russian forces in Daraa governorate have helped build the new Fifth Assault Corps in the Syrian Army, including creating the 8th Brigade, a force comprised primarily of ex-rebels, which has helped local control and stability thanks to ex-rebels ties to their communities, along with improving fighting capability against continuing rebellions (Al-Jabassini 2019). Relations with the Syrian government, however, have been prickly, with Russia wanting to retain influence over the 8th Brigade and the Syrian Army unwilling to recognize it (Al-Jabassini 2020), meaning any intelligence benefits are unlikely to last beyond a Russian withdrawal.
Russian forces in Syria were building on a pattern from home. In Chechnya, the local social knowledge of Chechen units comprised largely of ex-rebels proved crucial to the effectiveness of Russia's brutally successful counterinsurgency campaign, and these Chechen units outperformed their Russian counterparts (Biberman 2019;Lyall 2010;Šmíd and Mareš 2015;Souleimanov 2015). Souleimanov (2015, 91) argues for the 'crucial importance of cultural knowledge … of patterns of social organization, persisting value systems, and other related phenomena in the relative success of the eradication of the Chechnya-based insurgency.' Ex-rebel Chechen units' local knowledge and social networks 'dramatically increased the amount and quality of intelligence coming from villagers' (Souleimanov 2015, 107). Lyall (2010, 15) notes that the common tactic of kidnapping 'is an information-intensive form of abuse since it requires not only accurate knowledge of the targeted individual and his family, but also his whereabouts,' and Chechen units, but not Russian ones, had the necessary social networks to identify and then coerce individuals. For former rebels, it is also possible that 'their past status as insurgents enables [them] to identify, convert, or, if necessary, kill remaining fighters and their supporters with greater efficacy than Russian soldiers' (Lyall 2010, 16).
Elsewhere, in Northern Mali, the government has found success in counterinsurgency and maintaining local order by forming special Tuareg security units tied to the armed forces, incorporating many ex-rebels 'who know the terrain and guerrilla tactics much better than the Malian regular army' (Rabasa et al. 2011, 154). Nigeria's military has similarly integrated local militias in the northeast into the Civilian Joint Task Force (CJTF), using local knowledge of communities and terrain to successfully combat Boko Haram (Agbiboa 2018), with CJTF leaders recruiting demobilized and reintegrated Boko Haram fighters to incorporate their skills and knowledge (ICG 2021, 11). Integration of Maoist ex-rebels in Nepal was argued to increase military capabilities through rebel tactical and social knowledge, since the military 'has been fighting a conventional war and the [rebels] are trained in guerrilla war' (Nayak 2009, 739).

Military integration and intelligence in the Philippines
Uniquely in the military integration literature, Santos's (2010b) case study of the integration of Moro National Liberation Front (MNLF) fighters by the Philippines' national security forces highlights in detail the positive intelligence effects of the integration process, and further literature on the Philippines case supports his findings.
MNLF fighters were integrated into the Philippines' military and police forces following decades of conflict over the ethnopolitical demands of the Muslim Moro people in the Mindanao region of the Christian-majority country. Cooperation between MNLF units and the military began even before the peace process was finalized, as the government began 'to exploit the MNLF intelligence network and use its firepower' to help respond to and resolve 'critical incidents' including high-profile kidnappings and hostage-taking by the extremist Abu Sayyaf Group and MNLF dissidents (Lara 2012, 104). The military eventually integrated 5,750 people from the MNLF, while 1,750 joined the police (Caramés Boada 2009, 11), 8 though not every integree was an ex-combatant, since some MNLF fighters had their relatives join in their place (Hall 2014). Santos (2010b, 163) states that 'independent appraisals' found the 7,000 integrated former MNLF combatants 'served as a vital and trusted link between the security forces and the community.' The military put processes in place to try to ensure cultural sensitivity to Muslim ex-rebels (Hall 2014;Santos 2010b), and even set up a special office to handle issues between lowerranking integrees and senior officers (Santos 2010b, 171-72), aiming to build trust with the integrated fighters.
Integrees' community ties and cultural knowledge proved highly valued and useful to military officers as they continued to battle separatist forces in the region. Unit leaders 'expressed confidence in the intelligence information provided by MNLF integrees, describing it as generally highly reliable,' with positive outcomes like weapons seizures, while 'Because of their familiarity with the area, people, and language, integrees are often sent on surveillance missions against enemy targets.' Meanwhile, 'the MNLF integrees could be relied upon to liaise with community and local leaders. Their familiarity with the local people and language improved public relations for the battalion' (Santos 2010b, 173). Lara (2012, 102) similarly remarks that integration helped 'cement … legitimacy at the local level,' building trust. In another assessment, 'a top army general said the biggest contribution [of integration] was the build-up of the military intelligence database from integrees … ' with police in Mindanao 'during the first two years after MNLF integration … [having] confiscated more illegal firearms, received more rebels returning to the fold of the law, and enhanced its public relations with the community' (Santos 2010b, 174). Hall (2014, 103, 106) offers support for these findings, emphasizing how the integrated Muslim ex-MNLF forces' cultural knowledge 'engendered culturally sensitive military regulations and policies,' with integration of Muslim fighters 'expected to enhance its effectiveness in operations in Muslim communities. To the army, the Muslim recruits were critical to the strategy of winning hearts and minds.' In this case, integration was also serving as a symbol of national unity and building trust (Licklider 2014a), helping ensure Muslim communities were not alienated by the military and thus might be more willing to act as intelligence sources.
Other sources likewise confirm that military integration improved the government's intelligence capacity and that this contributed to conflict resolution, counterinsurgency, and the preservation of peace. Local commanders claimed integrated former MNLF fighters were 'more effective than those from the state army in conducting operations in Mindanao' (Laitin 2014, 234). MNLF cooperation and integration helped ensure state success in 'the identification [and monitoring] of key leaders of the Abu Sayyaf … and the ring-fencing of their camps,' as well as 'destroying other armed groups such as the Al Khobar and Pentagon gangs' (Lara 2012, 104-5), ultimately reducing the range of threats faced by the Philippines security forces.
Military integration of the MNLF provided new fighters for the state, increasing its coercive capacity, but given that one of the state's main opponents in Mindanao, the Moro Islamic Liberation Front (MILF), had initially emerged as a splinter from the MNLF, MNLF integration also provided a crucial opportunity for state forces 'to deepen their knowledge of the MILF's capabilities, and to deal with the eruption of clan violence that often escalates into wider conflict' (Lara 2012, 109). This helped the military combat an existing rebellion and better address the conflicts that might metastasize into insurgencies in the future. The intelligence provided by integrees also helped the government combat 'Abu Sayyaf and regional extremist groups such as the Jemaah Islamiya,' justifying United States military aid that further bolstered capabilities (Lara 2012, 109).
The MNLF never fully demobilized, and several factions were not interested in participating in the peace process or cooperating with the state (Hall 2014;Lara 2012). Even so, the relative success of integrating ex-MNLF fighters without full implementation of the peace agreement led the Philippines government to adopt military integration more broadly, for instance integrating into the military 246 members of the Cordillera People's Liberation Army after years of a stalled peace process (ICG 2013;Santos 2010a, 145-46). 9 Even hardline, mano dura President Rodrigo Duterte has been open to integrating former rebels to benefit from their local knowledge In 2017, he called on not only former MNLF, but also on members of the more recalcitrant MILF to join with the government in combatting Islamic State-affiliated rebels in Marawi, Mindanao (Xinhua 2017), with MILF's 'local intelligence and fire support' greatly aiding the government's cause (Brandt 2020, 104).

Military integration and intelligence in Uganda and Rwanda
Two militaries that emerged from rebel victories in civil wars have also pursued military integration, finding it helped increase intelligence capacity. The militaries of the contemporary Ugandan and Rwandan regimes, the National Resistance Army (NRA) in Uganda and Rwandan Patriotic Army (RPA) 10 in Rwanda have integrated both soldiers from the militaries of the governments they toppled and ex-rebels who fought in subsequent insurgencies. Victorious rebels can draw on their own experiences opposing the state as they conduct counterinsurgencies, but many groups have only been active in certain regions or among certain segments of the population. Victorious rebels also tend to reorganize their militaries along conventional lines, conforming to international norms and losing touch with irregular tactics and skills that often prove more useful in counterinsurgency contexts (see e.g. Farrell 1998).

Uganda
When the NRA and its political wing, the National Resistance Movement (NRM), came to power in Uganda in 1986, multiple other rebel groups and soldiers from the toppled Obote and Okello regime militaries remained armed and active. The NRA responded by trying to integrate many of these groups, a strategy they have continued to pursue over three and a half decades of NRM rule, across multiple amnesties for enemy fighters and battles against sixteen rebel groups and dozens more small armed groups and dissident organizations (Day 2011;Kiyaga-Nsubuga 1999;Lewis 2012Lewis , 2017Ondoga Ori Amaza 1998).
Already, while fighting as rebels, the NRA sought to ally with and integrate other rebel forces. The NRA's leader, current Ugandan President Yoweri Museveni, had initially been leading a smaller guerrilla force, the People's Resistance Army, with roots in western regions and the Banyankole ethnic group, but merged with the primarily-Baganda Uganda Freedom Fighters. This enabled the merged NRA to gain legitimacy, support, and local geographical and social knowledge among the Baganda, the country's largest ethnic group, who live in the central region surrounding the capital, where fighting was concentrated (Kasfir 2000(Kasfir , 63, 2005Twaddle 1988, 318-19). The Baganda began to see the NRA as more inclusive and responsive to their interests, increasing willingness to supply information (Kasfir 2005;Tidemand 1994).
While expanding operations to mountainous western Uganda, the NRA worked to integrate the forces of the Rwenzururu kingdom, who had previously fought against the national government and thus knew the terrain and government operations in the region. After the NRA takeover in 1986, Rwenzururu forces split and the NRA subsequently had difficulties rooting out Zairean rebels who had crossed over the border, lacking intelligence to distinguish between rebels and the local population (Syahuka-Muhindo 1993, 314-15).
Once in power, the new NRM regime almost immediately faced rebellions, and the erosion of state capacity under previous regimes, economic collapse, and their opponents' irregular warfare strategies led the NRA to emphasize integration and prioritize human intelligence. As Minister of State for Internal Affairs Tom Butime stated in a 1991 parliamentary debate, 'the Security Forces cannot get information from a satellite, they cannot get information from newspapers about where the rebels are, neither can they get information from telephone. Information can only be given by the people in that particular area to the forces' (Republic of Uganda 1991, 10). The NRA used integration of ex-rebels into the military or local militias to improve this flow of information and to remove competitors from the battlefield, beginning in late 1986 with the integration of the Uganda National Resistance Front, which helped maintain peace in West Nile, as former rebel leaders 'were working actively to ensure that their people supported the new regime' (Brett 1995, 145; see also Twaddle 1988, 325).
In 1986 and 1987, while fighting against the Ugandan People's Democratic Army (UPDA), 'as the NRA moved into [the northern region of] Acholiland and UPDA members began to change sides, the NRA used ethnic Acholi members to gain the trust of these anti-NRA fighters, encouraging them to form relationships to help integrate new converts into the NRA' (Bell 2016, 508). Locals in Acholiland later requested that more integrated ex-UPDA forces be brought back and deployed in the region to combat Alice Lakwena and Joseph Kony's Holy Spirit Movement (HSM) rebels, 'as they know the Holy Spirit better' (Lamwaka 1988), having operated in similar areas and at times allied with the HSM, and so were able to use this knowledge of guerrilla operations to improve counterinsurgency effectiveness and further build local communities' trust.
In Teso in eastern Uganda, the Uganda People's Army (UPA) rebellion quickly grew from 1989-1990, and the NRA, with relatively limited experience and ties in the region, had difficulty responding. By the middle of 1991, however, as more UPA fighters began to surrender, 'often with promises to be absorbed into the NRA military or government, the counterinsurgents gained valuable information about the UPA that enabled more military gains' (Lewis 2012, 145), relying on these ex-rebels knowledge of the geography and population and their former comrades' operations.
Former rebels' knowledge and networks have proven useful for counterinsurgency when they have taken on political, rather than military, roles as well. 11 Lewis (2012, 241-42) found that many former rebel leaders integrated into local and regional governments can 'draw on their network of former footsoldiers in their rebellion in order to keep appraised [sic] of potential threats emerging in their home area.' This high level of integration of former enemies has been a key component of the Ugandan government's military and intelligence apparatus and ex-rebel leaders' local knowledge has enabled it to deter and detect rebellions, 'nipping them in the bud' or fighting relatively effective counterinsurgencies (Brett 1995;Day 2011;Lewis 2020).
Integration was also crucial in making the military better reflect national demographics (Brett 1995, 147-48), improving trust-at least outside of the north (see e.g. Branch 2005)-as a symbol of national unity. Military socialization and political education programmes, both within the military and for civilians, helped ensure integrees' and their constituencies' loyalty to the state, built commanders' trust in them, and also created systems of monitoring within the armed forces and in communities (Bell 2016;Kabwegyere 2000;Lewis 2012Lewis , 2020.

Rwanda
Rwanda has also been identified as a highly successful case of military integration and counterinsurgency, despite integration having occurred in a difficult post-genocide, post-rebel victory environment of state collapse and ethnic polarization (Burgess 2014;Mills 2008;Wilén 2016). The Rwanda Patriotic Army (RPA), the Rwanda Patriotic Front's (RPF) military wing, was organized by ethnic Tutsi Rwandan refugees living in Uganda. The core leadership had, in fact, served in the Ugandan NRA, both as rebels and in the new state military. RPA leaders deserted the NRA and invaded Rwanda from Uganda in 1990, eventually seizing power and ending the ongoing genocide in 1994. Throughout the fight as rebels, the RPA worked to integrate different Tutsi diaspora populations from around the region, gradually also gaining recruits from within Rwanda (Jowell 2014, 280). Around the RPA victory in 1994, large portions of the Hutu-dominated Rwandan Armed Forces (FAR), the deposed regime's military, fled to the eastern DRC, and with other Hutu refugees and former génocidaires began an insurgent campaign against the new RPF government. The RPA had already begun integrating former enemy personnel during the civil war, and continued integrating former government and militia forces after taking power (Burgess 2014, 91-94;Jowell 2014;Rehder 2008, 20-22;. While they rarely held territory within Rwanda, the insurgent forces, generally still referred to as the FAR or ex-FAR, enjoyed strong support in northwestern Rwanda, as 'the majority of the insurgents came from the northwest; they knew the terrain and people; and their extended families and friends provided protection,' food and supplies, and information, as well as joining the insurgents to participate in attacks (Orth 2001, 90). Former government fighters integrated into the RPA had by 1997 been included in counterinsurgency campaigns (Burgess 2014, 95).
The RPA had difficulties penetrating insurgent networks and effectively countering the insurgents in the northwest, however, until they began a wider programme of integrating ex-FAR insurgents into the new military in 1997-1998. This integration quickly helped the RPA turn the tide against the insurgents in the northwest, 'as the people realized that the army was not the enemy' and 'the former insurgents knew all the 'tricks' of the EX-FAR, including terrain particulars' (Orth 2001, 95), demonstrating benefits in terms of local knowledge, rebel tactical knowledge, and social ties.
The RPA had an intensive ingando socialization process for ex-rebels and the units they were joining, designed to break down barriers between government loyalists and exrebels and to forge a new collective military identity, seeking to avoid commitment problems and generate trust during integration (Burgess 2014;Jowell 2014;. The military and its intelligence officers sought to cultivate ties to the population by having integrees return to their home communities after ingando, training, and joint deployments to 'break down myths and social barriers and to encourage inter-communal trust' (Jowell 2014, 283). 12 Overall, 10,500 former FAR officers and soldiers were integrated into the RPA from 1995-1997, while 39,200 ex-FAR soldiers, former insurgents, and Hutu militia members were integrated between 1998and 2002(Burgess 2014Rusagara 2011, 114-15). Hutu ex-rebels were even put into command positions, sending 'a significant signal to those below' and increasing trust among integrees and Hutu constituencies (Jowell 2014, 283). 13 The integrated military proved effective in combat both domestically and in a transnational campaign against Hutu insurgents and government forces in the DRC (Burgess 2014, 98-99;Rusagara 2011, 116). Integration's success in building unity and trust among former enemies gave the government greater social knowledge and legitimacy, and led communities in the northwest to be more willing to work with the military and the government and provide information to it-though this legitimacy and trust has likely eroded as the government has become more authoritarian.

Hypotheses for further study
Based on the literature review and the evidence presented above, I inductively identify five hypotheses that can be tested as part of a research agenda on the intelligence effects of military integration. I divide these into three categories: organizational, geographic, and political.
Organizationally, I hypothesize that intelligence benefits will be higher and last longer when ex-rebels are integrated into mixed units, rather than kept in separate units, be they subunits within the state military or separate security forces. Separate units may offer built-in unit cohesion for integrees and a greater sense of personal security, which is important for morale, but they decrease opportunities that exist in mixed units to develop ties and loyalty to integrees' enemies-turned-comrades and the state, and also limit ex-rebels' communities seeing them operating equally with state forces, factors that were key to successful integration in Rwanda (Burgess 2014;Jowell 2014;. Separate units can make monitoring more difficult for commanders and increase the potential for defection, though they may be preferred by commanders when trust in integrees is low (Krebs and Licklider 2015, 100;Simonsen 2007, 579). The ability to benefit from integrees' knowledge and networks in either organizational structure will likely also depend on integrees' degree of socialization into state forces (see Jowell 2014; Neads 2020), and on unit and force commanders' respect for ex-rebels (see Hall 2012;Samii 2014).
In terms of geography, in line with the insurgency and counterinsurgency literature (e.g. Fearon and Laitin 2003;Kilcullen 2010;Lyall 2010;Lyall, Shiraito, and Imai 2015), I hypothesize that intelligence benefits will be greater when ex-rebels are operating in terrain with which they are familiar and if they are operating among populations with whom they share a salient identity. Politically, I hypothesize that intelligence benefits will vary based on perceptions of the degree to which the peace agreement or other integration arrangement is being upheld and implemented (cf. Joshi and Quinn 2017). If exrebels believe terms are being violated, they and their communities will be less willing to provide information, while if government commanders believe terms are being violated, they may be suspicious and dismissive of the information they receive from integrated exrebels. The MNLF case suggests that intelligence benefits may still accrue even without full implementation of the peace agreement and incomplete integration of forces (Lara 2012;Santos 2010b), but over time, failures of implementation and compliance are likely to erode cooperation.
Going into a case study or field research with these hypotheses in mind, it may be more feasible to conduct interviews or find documents that allow for hypothesis testing. Many of the different factors proposed to affect integration and intelligence benefits are likely to vary over time, so tracing particular integration efforts temporally can generate a more fine-grained understanding of what issues matter when, and how the intelligence impacts of integration affect counterinsurgency and peacebuilding. Detailed declassified military information may be necessary to test for an independent causal effect of intelligence gains from military integration on counterinsurgency effectiveness or overall peacebuilding in recent cases, but it might also be possible to test these propositions using qualitative or quantified historical data from conflict archives (see Balcells and Sullivan 2018).

Conclusion
This article has examined a gap in the existing literature on military integration: the potential intelligence benefits of military integration for counterinsurgency operations and preventing future rebel group formation and development. I have highlighted three key ways that integration can improve intelligence capacity: former rebels' knowledge of human and physical geography, social networks and informants, and knowledge of the effectiveness of rebel and counterinsurgent strategies and tactics. Based on evidence from a variety of historical, political, and geographic contexts, it appears highly plausible that improved intelligence may interact with other factors highlighted in the military integration literature to contribute to more successful counterinsurgency or peacebuilding outcomes. I have also identified a set of hypotheses to test this proposition and the mechanisms underlying it.
Ideally, intelligence capacity improved by military integration would be used for more selective targeting and to restrain and limit violence against civilians (Kalyvas 2006), 14 and would also inform nonviolent actions aimed at redressing anti-government grievances. Even defeated rebel groups may retain legitimacy in areas where they held influence, and integrating ex-rebels may be better from a human rights perspective and in costs and effectiveness of governance, instead of violently crushing the rebellion as occurred in Sri Lanka (Kubota 2017). At the least, debates and assessments regarding military integration should account for intelligence effects. Highlighting intelligence benefits of integration may also help convince reluctant state security elites to embrace integration and accept ex-rebels, avoiding the problems identified by Samii (2014) of military officers rejecting integrees.
Future research must also consider that the intelligence benefits of integration may not flow solely to the incumbent government: integrated ex-rebels can act as an 'early warning system' for their social or political constituencies, leaking information (Baaz and Verweijen 2013;Verweijen 2014) and potentially defecting when they learn of policies or strategic plans of which they disapprove. For instance, integrated ex-rebels can take signals from the integration process, promotion patterns within the military, and where integrated units are stationed, and can use these to judge government intentions toward not only their professional interests but also their broader constituencies. This was the case with the defection and return to rebellion of integrated former Anya-Nya fighters in Sudan (D. H. Johnson and Prunier 1993;LeRiche 2014). If there are not careful screening mechanisms, integration can also mean notorious human rights violators joining the military, as occurred in the DRC (Arnould 2016, 333), potentially making civilians less willing to share information. Investigating these and other cases of negative effects of military integration on intelligence capacity can help refine theory and highlight factors that may undermine integration, as well as potentially pointing to junctures or warning signs at which practitioners should rethink military integration. 15 Beyond my hypotheses, the potential intelligence benefits of military integration raise several questions for further investigation. How can potential intelligence benefits of integration be isolated empirically to test their hypothesized effects? How does improved intelligence complement or contradict other proposed mechanisms for military integration's contribution to peacebuilding or counterinsurgency success? How does taking intelligence capacity into account alter the calculus of bargaining models of civil war resolution incorporating military integration (Hartzell 2014, 17-18)? For example, if a government realizes it lacks intelligence capacity in a region or among a certain population, this might increase its likelihood of including or accepting integration as a component of a peace agreement. Finally, in all three cases examined in this article, military integration was adopted autonomously, without significant international pressure or supervision, and in Uganda and Rwanda it occurred after rebel victory. Future research can further investigate similarities or differences in the adoption of military integration and its intelligence effects when it is driven by international actors or when it takes place after government victory or a military stalemate and negotiated settlement.
This extension of the military integration research agenda will provide deeper and more precise theoretical and empirical knowledge to evaluate when, where, and why military integration is a better or worse policy choice. Focusing attention on the intelligence aspect of military integration also reveals further potential linkages between counterinsurgency and peacebuilding research and practice, and can provide an opening to try to convince even conservative government policy elites to consider military integration and reconciliation with former rebels and their civilian constituencies. Notes 1. In shorter samples, Ottmann and Vüllers (2015) also found that 41.4% of the 46 negotiated settlements from 1989-2006 included military integration provisions and Martin (2017) found such provisions in 54% of signed peace agreements from 1975-2011, while Berg (2020) found that from 1960-2012, 38% of years in peace spells after major civil wars feature military integration agreements, and military integration plans were fully implemented in 29% of peace spell-years. 2. Krebs and Licklider (2015, 102) summarize the hypothesized mechanisms leading from military integration to reduced conflict recurrence: 'by serving as a costly signal and thereby mitigating the security dilemma, by providing security to vulnerable populations, by employing former combatants, and by serving as a symbol of the united nation' and helping generate trust. 3. See also Ottmann (2020) on 'extra-agreement violence' by forces not part of peace agreements. 4. During revisions, I found Brandt's (2020, 30-31) dissertation, which also briefly discusses the potential for military integration to improve intelligence, and then draws on Santos (2010b)