Skip to main content

The dubious joys of standing up militias and building partner capacity: Lessons from Afghanistan and Mexico for prosecuting security policy through proxies

Vanda Felbab-Brown


After a decade and half of being involved in large-scale military operations in the Middle East and South Asia, the United States is increasingly looking for ways to prosecute its security interests with minimal involvement. The number and complexity of military conflicts burning in the Middle East, Africa, and Asia make large-scale military involvement prohibitive and potentially disastrous. Nor are desirable outcomes easy to achieve even when the United States is willing to devote large blood and treasure resources to such interventions. Yet allowing for large territories around the world to disintegrate into brutal conflicts or be dominated by terrorist and militant groups threatening the United States and its assets and allies may also be highly detrimental to U.S. interests. Increasingly, the United States has there embraced a strategy of building up the capacities of partners in such conflicts—preferably, those of national governments. But when national forces prove unable or unwilling to defeat the threat, such as in Iraq, Syria, Libya, Nigeria, and Somalia, the United States increasingly faces decisions whether to support (or establish) local militia and other irregular forces. Indeed, militias are back in vogue as a tool of U.S. security policy.

Building partner capacity and standing up militias may at times turn out to be the least bad of available policy options, and chosen as a last resort. The two approaches are not identical: Building up the official forces of a country allows for greater levels of accountability of the host nation to the donor and of the forces to their population than standing up irregular forces. Building up official partner capacity is thus clearly a preferable policy than embracing militias.

Nonetheless, both policies need to be adopted with a clear understanding of how limited their contribution will be to the prosecution of U.S. interests. Despite the seductiveness of the belief that in situations of limited U.S. engagement and resource transfers others will do for the United States what it is not willing to do for itself, a robust alignment of U.S. interests with those of its presumed partners will often be lacking. Instead, after a momentary intersection of interests, the United States could find its partners unreliable, having divergent interests, and prosecuting policies directly contradictory to those of the United States. Even when militias seem to be defeating the enemy du jour, they will have a tendency to go rogue and themselves become the source of long-term and deep-seated drivers of conflict. Partner governments may well pocket U.S. assistance and have a tendency to conduct policies to satisfy U.S. objectives only minimally while they use the U.S. resources for their own political objectives, including perhaps marginalizing or eliminating political rivals. Nor will the United States find it easy to extricate itself from such indirect interventions. Consequently, when relying on such proxies, U.S. long-term engagement, monitoring, and rollback capacities will often be necessary for securing U.S. interests.

This article explores the security and political effects of militia forces in Afghanistan and Mexico. From these case studies, it draws lessons for how to engage with militia forces more effectively and presents broader implications for U.S. foreign policy. In Afghanistan, building up militia forces has long been an important element of U.S security policy. In Mexico, anti-crime militias emerged spontaneously. Eventually, the behavior and visibility of the militias forced the government of Mexico to react to them, but the United States has not been involved with them in any way. Along many dimensions, the two cases are very different, and that is one reason why I chose such a compare-and-contrast pairing. Mexico is a middle-power country with a relatively strong economy, even though the state has been historically weak or absent in large areas, including those where militias are currently strong. In contrast, for several decades, Afghanistan has been a failed state or hovered on the verge of failure. For decades, Afghanistan has been caught up in insurgencies, civil wars, and foreign interventions. Only over the past decade has Mexico experienced highly violent crime. In Afghanistan, the United States and the Afghan government at least assumed that they could control some of the militias, such as those they actively recruited, including the Afghan Local Police, and steer some of the others that had existed, metamorphosed, and sometimes outright metastasized in Afghanistan since the late 1970s. At the same time, the Afghan government, the United Nations, and the United States have sought to disarm and dismantle other militia forces that came to be seen as particularly problematic.[1] In Mexico, the anti-crime self-defense forces emerged spontaneously without a direct and formal state effort to recruit them and without any nominal state control. Historically, of course, the Mexican government and military often recruited militias to fight insurgencies such as in Guerrero and Chiapas.

But it is striking that despite the differences between the two countries, consistent patterns of militia behavior emerge. Among the most important is the fact that no matter what the original motivations and justification for militia formation, militias have a strong tendency to escape control by their overseers and engage in problematic and abusive behavior. Even when militias spontaneously emerge in response to abuse that local communities find intolerable, the militias will have a strong tendency to deteriorate to such behavior themselves. The scale of such misdeeds often negates their previous usefulness, and the militias become a profound threat to order and rule of law and a new driver of conflict themselves.

Rarely do local communities or official state structures have the capacity to keep militias in check. But the less effort the national government puts into developing official mechanisms of control, restraint, and rollback, the worse the predation and deleterious effects the militias will have on stability and the long-term legitimacy of local political dispensations. Although militias might be local, their effects are not: they have profound and complex implications for political rivalries and balances of power throughout the political, militancy, and criminal systems. When the United States decides to sponsor militias, it needs to immediately build into its policies implementable ways to disband them once the U.S. objectives have been accomplished. Otherwise, it will have to live with the fact that the militias will likely become a major problem for U.S. interests. The United States thus must look beyond immediate pressing security imperatives and expand its policy horizons toward the long term.

Building Partner Capacity: A New Term for Good Old Internal Defense

Disavowing nation-building abroad (though really meaning state-building abroad) and determined to focus on nation-building at home,[2] the U.S. administration of President Barack Obama has strongly focused on “building partner capacity” as a way of securing U.S. interests in contested or “ungoverned”[3] spaces without dedicating large U.S. personnel and financial resources to such efforts.[4] In his January 2014 State of the Union Speech (SOTU), President Obama explicitly embraced building up partner capacity and quick in-and-out military engagement, such as by drones, as core elements of U.S. security policy.[5] The examples he picked—Somalia, Yemen, and Mali—have featured such limited operations. Like the Casper Weinberger-Colin Powell doctrine of the 1980s and early 1990s,[6] the January 2014 SOTU called for deploying the U.S. military only when the most vital national interests of the United States are at stake. But unlike the Weinberger-Powell doctrine, which emphasized the deployment of overwhelming U.S. force with no half-way measures, Obama underscored that U.S. military engagements would be surgical and constrained. The president rejected open-ended commitments and large-scale military interventions.

The key assumptions of U.S. external assistance for such internal defense policies is that there is a sufficiently strong overlap between U.S. interests and those of the selected external partners to design and conduct anti-militancy efforts in a way that is consistent with U.S. objectives, and that such an overlap of interests can be sustained for a sufficient period. In other words, through internal defense assistance, the United States hopes to motivate external actors to deliver U.S. objectives without the United States being extensively sucked into difficult internal conflicts abroad and without having to sacrifice U.S. blood and treasure on a large scale.

Building partner capacity and assisting in internal defense are concepts much broader than simply standing up militia or proxy forces. Indeed, ideally, internal defense or partner-capacity-building efforts would focus on state military and police forces only and be fully integrated into an official strategy of the host country to undertake security sector reform (SSR).[7] The ideal conditions and design for such an undertaking include a post-conflict setting, without the presence of low-intensity conflict still simmering or increasing, and with U.S. security-aid programming focused on creating a legitimate monopoly of force for the state. Such an assistance program would be a comprehensive systematic effort easily lasting a decade or more.

Yet the United States today faces security and foreign policy pressures that strongly push against such ideal conceptualization and implementation. Indeed, in an increasing number of significant conflicts around the world, such as in Syria, Libya, Somalia, and Nigeria, the United States is being drawn in to deliver security assistance after violent conflict has started, the weakness of the state has been critically exposed, and the legitimacy of the state is strongly contested. Battling for immediate survival, U.S. partners have limited (and possibly no) interest to engage in such as an externally-desired comprehensive and accountable reform of their security sector. Nor does the United States necessarily have the patience for a systematic long-term security-sector reform effort. Rather, building partner capacity is frequently a manifestation of a very limited U.S. (and Western) willingness to become involved in the conflict, and then  only in order to put out the most immediate security fire—whether by quick short-term in-and-out intervention with one’s forces or by providing military advisors.[8] Moreover, with national security forces often unable and sometimes unwilling to respond effectively against dangerous militant groups in the country of concern, the United States consequently often provides assistance to non-state militia forces, including in countries where it is also trying to beef up and reform official state forces, such as in Iraq and Afghanistan.

Militias to the Rescue in Afghanistan: 13th Time the Charm?

In Afghanistan, the embrace of militias and other irregular proxies has been a key feature of the war since 2001. Indeed, the 2001 U.S. intervention essentially consisted of a limited U.S. force deployment and support for anti-Taliban militias and warlords, such as those who formed the Northern Alliance of principally non-Pashtun anti-Taliban opposition. Although with the assistance of U.S. special operations forces (SOF) the forces, generously labeled Afghan Military Forces, did depose the Taliban, it did so only in a perfunctory way, allowing key Al Qaeda and Taliban leadership structures to escape into Pakistan. There the Taliban could over time rebuild itself and ignite a very potent insurgency in Afghanistan, fundamentally threatening the stability of the country and the state-building effort. Moreover, the governance that the United States and Northern Alliance helped usher in was pervaded by such abuses of power and ethnic and tribal marginalization that the brutal Taliban insurgency succeeded in deeply entrenching itself in the country.[9] The U.S. effort to thwart the Soviet invasion of Afghanistan in 1979, and the overall goal of weakening the Soviet Union in the Cold War strategic competition, was at the time also conducted through building up “militias”—namely, the anti-Soviet mujahideen forces. In turn, when in the late 1980s the Soviets realized they were not winning in Afghanistan, they too resorted to raising tribal militias, and used them as part of their exit strategy.[10] The U.S. and Pakistan-sponsored mujahideen ultimately prevailed and helped drive out the Soviets. But the collapse of the Afghan national military to which some of the mujahideen militias were recruited after the departure of the Soviets, and the subsequent infighting among the various militia forces and collapsed army units, and the resulting chaos and localized abusive governance plunged the country into a bloody civil war in the 1990s and gave rise to the Taliban.[11]

Early in the post-2001 era in Afghanistan, some attention and energy was devoted to attempting to demobilize irregular clan and warlord militias and particularly to collect their heavy arms such as tanks. But the effort soon fizzled out. Instead, as the Afghan National Army (ANA) and Afghan National Police (ANP) struggled against the growing Taliban insurgency, despite more and more intensive NATO training and resource transfers, the United States increasingly turned to building up Afghan militia forces to help defeat the Taliban.

These Afghan “militias” are primarily supposed to increase security in areas where the presence of the ANA and ANP (and previously NATO’s counterinsurgency and stabilization deployment in Afghanistan, the International Assistance Security Force or ISAF)[12] is limited. At times, the objectives of these attempts “to stabilize Afghan villages” have also included extending the writ of the legitimate Afghan state. Nonetheless, the primary goal of these efforts has been to weaken the Taliban military capacity in rural areas and thus “increase security.”[13] The most extensive and visible iteration of the various militia stabilization efforts has been the Afghan Local Police (ALP). But there were many other iterations, including Afghan National Auxiliary Police (ANUP), the Afghan Public Protection Program (APPP), and the Local Defense Initiative (LDI), also known in some areas as the Community Defense Initiative (CDI).[14]

The U.S. Central Intelligence Agency (CIA) also created its own separate and special paramilitary forces, known as Counterterrorism Pursuit Teams. Highly secret and relatively unknown, the 3,000-member force was tasked with generating intelligence and conducting raids and combat operations against the Taliban and Al Qaeda in Afghanistan, but was not supposed to engage in lethal action when crossing into Pakistan.[15] Although supposedly tasked with a more narrowly defined counterterrorism, rather than broader counterinsurgency, mission, the existence of the Counterterrorism Pursuit Teams, and at times their reportedly rough behavior toward local populations, have had important counterinsurgency and governance implications.[16]

Not all of the militias were stood up under the supervision and blessing of the U.S. military or the CIA. The National Directorate of Security (NDS) (the Afghan intelligence agency) has for years encouraged the establishment of separate self-defense units, especially in areas where the Tajik-dominated NDS fears the presence of too many Pashtun self-defense groups.[17] Particularly under the former leadership of Bismullah Khan Mohammadi, the notoriously corrupt and ethnically factionalized Ministry of Interior (MoI) also sought to legitimize and formalize its own favorite: non-Pashtun militias in northern Afghanistan. At one point, the former Minister of Interior, for example, would allot at least 1,125 positions to Kunduz to establish a network of loyal non-Pashtun commanders.[18] In fact, the former minister tried to get ISAF to legitimize and support these non-Pashtun militias in exchange for sponsoring and expanding the creation of Afghan Local Police units in the south and east of Afghanistan. When ISAF was reluctant to agree to the loose practices in the north, Mohammadi would halt the disbursement of ALP salaries elsewhere in the country.[19] Such behavior is not surprising: Early in the post-2001 establishment of the Ministry, the MoI became the favorite dumping ground and legitimizing mechanism for militias associated particularly with Shura-ye Nazar faction (i.e., the military faction around Ahmed Shah Mahsoud) of the political party of Jamiat-i-Islami (consisting mostly Tajiks, Turkmen, and Uzbeks) and of other mostly non-Pashtun powerbrokers who were not able to place their militias into the newly forming ANA.[20] Powerbrokers associated with another northern political party, controlled by the Uzbek leader and current Vice-President Gen. Abdul Rashid Dostum, also obtained major influence over the ALP in the north.

Even at the height of the push to create the ALP in 2013, the United States and ISAF maintained that the various community-based self-defense programs were not militias, calling the local units everything but militias. They insisted that these self-defense programs are based on Afghan traditions, such as arbakai (tribal militias).[21] Although community-based self-defense is organic to Afghanistan and has a basis in previous Afghan militia versions, ISAF’s embrace of this concept was also stimulated by the success of the Anbar Awakening in Iraq, in which Sunni militias, along with the U.S. military surge, were important for beating down Al Qaeda in Iraq and decreasing the intensity of Iraq’s civil war in 2011.[22] By September 2014, the ALP numbered over 30,000, located at perhaps ninety-nine sites throughout Afghanistan.[23] ALP’s cost of $120 million per year (and about a quarter per militiaman of the cost of the ANA soldier)[24] has been footed by the United States After a significant reduction in NATO’s military role and substantial deterioration in security in the country, the ALP Directorate in Kabul has been requesting that the size of the ALP be increased to 45,000, with the United States fully funding the force and extending its operation from the current end date of September 2018 to at least 2024.[25]

Irrespective of the relatively modest effects on security of the ALP compared to the ANA, ANP, and the Taliban, the ALP’s and other militias’ troublesome impact on governance in Afghanistan is potentially large. Indeed, in contrast to the U.S. military’s rosy assessments of the ALP in 2012 and 2013,[26] many Afghans tend to fear the ALP and other self-defense force programs and often have negative views of such militia efforts. Local communities have frequently experienced militias and arbakai turning on them, extorting them, stealing land and goods, and even engaging in murder, kidnapping,  and brutalizing rival ethnic groups. “They come to our village and ask for food and money. And if we don’t give it to them, they take it anyway,” a resident of Baghlan told me in 2012.  “How can we resist the men with guns?” he asked.[27] And a Kandahari businessman feared that the proliferation of such militias would usher in “the chaos and war of the 1990s. The Soviets tried it, too, when they were defeated,” he argued, “and look where it brought Afghanistan. People just use the ALP as a justification to arm up their tribes. Everyone is afraid.”[28] Former President Hamid Karzai, fearing both the manipulation of the self-defense forces to locally empower his rivals and in general to further weaken his power, was highly reluctant to agree to the creation of the ALP and its other versions.[29] At the same time, many of the key powerbrokers on whom he relied, including his Vice-President Mohammad Fahim and key Shura-ye Nazar commanders, were enthusiastic and eager to use the ALP as a mechanism to shore up and legitimize their power vis-à-vis their Pahstun rivals. As the ALP recruitment drive coincided with the Afghan 2009 presidential elections, Karzai ultimately came to see its political usefulness himself.[30]

Of course, historically, local communities have also been able to materially benefit from their own arbakai, using them to strengthen their capacity for self-defense or to aggressively expand their lands. Today, too, some communities have positive views of the ALP and other arbakai and even volunteer to stand them up. “Before the ALP was created in our village,” a Pashtun resident in Baghlan told me in 2012, “no one could travel on the road, even in broad daylight. Even the ANA would not dare come to our village. Now it’s much better.”[31] Although he was not himself a member of the ALP, some of his relatives were. Indeed, often the Afghans who most enthusiastically embrace such programs are the ALP commanders themselves, or the power brokers who tried to sell their unofficial militias to ISAF and the Afghan government programs for hefty payoffs.

As the various self-defense forces frequently alter or reinforce local balances of power among ethnic groups or patronage networks and insert themselves into local conflicts, the Ministry of Interior, the National Directorate of Security, and the Independent Directorate of Local Governance, as well as local government officials, have often fought intense turf battles over who controls these forces. Many other militias are simply self-generated or have long fought for a local power broker, and simply carry their weapons less ostentatiously since 2002.[32]

In short, to the extent that institutional biases and ethnic competition and political rivalries, rather than a careful evaluation of local dynamics, drive the creation of such self-defense and paramilitary forces, supervision and control are bound to loosen significantly. Such dynamics in turn raise troubling concerns about the intensification of predatory behavior by the self-defense forces themselves.

The oversight mechanisms and controls over the militias have often been insufficient. Like many of the previous programs, the ALP, too, was mostly supervised and trained by U.S. Special Operations Forces. The original plan was that the U.S. SOFs would do all the training and programming, but as the program expanded amidst U.S. troop drawdowns, U.S. conventional forces also increasingly came to be used for implementation, even though the SOFs are better suited for such a mission.[33]

The U.S. trainers embedded with the ALP in the village or an area where the ALP operated. Embedding could mean a variety of arrangements: under optimal circumstances, exemplified by the ALP program in Arghandab (an effort that built on the previous experience with the Local Defense Initiative in the district), SOF units lived in the village for six weeks, developed a knowledge of the community’s composition and political actors, fully participated in the vetting of the recruits, and subsequently maintained close supervision of the ALP units.[34] In other cases, however, such as in parts of Kunduz, the SOF monitoring amounted only to visiting the village once a week.[35] Yet particularly Kunduz was a place where very close and nonstop monitoring of the ALP would be crucial.  At the time of the ALP formation, and still today, Kunduz was not only a vortex of local ethnic rivalries and persisting violent conflict, but also of hidden (and not so hidden) powers. One notorious powerbroker in Kunduz, Mir Alam, had previously successfully manipulated DDR efforts to establish himself as the dominant force in the province, and was equally able to manipulate the ALP formation.[36]

U.S. training of the ALP recruits mostly consisted of teaching them how to handle small firearms, deliver first aid, and communicate with the SOFs. With the dramatic reduction of U.S. forces in Afghanistan starting in 2013, and with possibly only a minimal presence of 1,000 military personnel for the protection of the U.S. embassy after 2016, the U.S. supervisory teams planned to hand control over to Afghan special operations forces. But the latter frequently did not materialize. Increasingly, many Afghan ALP and other militias are on their own and on the loose.

During the formal ALP creation between 2012 and 2013, those recruited for the force were to be vouched for by three maliks (tribal elders) or a village shura or both. The maliks or shuras were relied on to determine that the ALP recruits would not secretly work for the Taliban or other antigovernment elements and turn on their U.S. advisers, or abuse the local community. Frequently, however, a powerbroker would control the village elders, dictating his preferences through the elders in a way that escaped the outsiders’ scrutiny. At other times, the village elders would have no problem vouching for militia members as long as they only extorted a rival village.[37]

A second control mechanism of the ALP program, as in the case of the Afghan Public Protection Program, was that the district police chief was to supervise the ALP units. But in Afghanistan the post of the district police chief has often been entangled in some of the worst and most consistent corruption, intense politicization, and ethnic factionalism. Appointments of district police chiefs are rarely based on the high moral character and outstanding professional qualifications of the individual; even the absence of a criminal record is often asking too much (or not enough). Instead, the position is, at times, bought by those who can afford to pay for it (expecting to collect hefty “taxes” from the local population), or most frequently, the appointment is negotiated between Kabul and local powerbrokers, with the goal of satisfying their competing demands.[38] Consequently, the credibility and quality of the police chief’s supervision of the Afghan Local Police units are often poor.

Recruitment and vetting became especially problematic when the ALP and other militia recruits were former members of two of Afghanistan’s most prominent insurgencies—the Taliban or Hezb-i-Islami. Indeed, the ALP was at times meant to function also as a mechanism to provide employment for reintegrated insurgents and to keep them on a payroll.[39] Many ex-insurgent groups that have gone through the reintegration programs have asked to keep their arms and be given a checkpoint, and the ALP and other militia efforts have provided a means to satisfy their demands.[40]

The ALP would thus—somewhat paradoxically—mix a trifecta of goals: weakening the Taliban and Hezb-i forces through proxies, serving as a DDR, or more precisely reintegration only, program for the Taliban and Hezb-i forces who could be flipped, and keeping order. Not always did these three separate goals result in a happy merger under the ALP umbrella. Indeed, this “fox-in-the-henhouse” approach may buy reductions in violence, but it hardly guarantees that a community feels safer or is not abused by the ALP, or that the community in any way more closely embraces the Afghan government. Moreover, many of the flipped Taliban and Hezb-I ALP members and units, and indeed, many other ALP units not originating in the anti-government forces would be rather open about their loyalty being very thin and a function of the payroll. In interviews with me in Baghlan in 2012, ex-Taliban and Hezb-i ALP members stated rather openly that should they no longer be paid by the United States or the Afghan government, they would join the insurgency again. A similar finding was confirmed in more recent interviews conducted by The International Crisis Group.[41]

The greatest weakness of the ALP effort, and of the many previous and concurrent versions of standing up irregular forces in Afghanistan, was that there were no firmly established mechanisms for disarming an ALP unit which started misbehaving, such as extorting or otherwise abusing its own or rival communities. A formalized and diligently implemented procedure could greatly reduce the ALP’s propensity toward abuse and increase the program’s legitimacy.

But local powerbrokers often prevented disciplinary actions, including full disbanding, from taking place. A case in Kunduz provides a vivid example. In revenge for a Taliban assassination of a militiaman, a pro-government militia associated with Mir Alam and led by commanders Qaderak and Faizak, massacred eleven Pashtun boys and men in September 2012. Qaderak had previously attacked the Pashtun village at least three times because it had refused to pay him “taxes.” Although the militia was according to U.S. officials not part of the ALP,[42] the Pashtun community nonetheless labeled Qaderak and Faizak’s militias as Afghan Local Police units and organized protests to demand the dismantling of the militias, also blaming the Americans for the massacre and seeing them as responsible for the reaction and sponsorship of the militias. Nonetheless, Alam and Vice-President Fahim exercised their influenced and prevented meaning prosecution of the perpetrators and their backers.[43] Qaderak and Faizak continued operating their militias until the Taliban killed Qaderak in August 2014.[44]

Effectively correcting, punishing, and disbanding unruly and abusive units have mostly been a rare sporadic occurrence, rather than a systematic policy. In Wardak, for example, the U.S. military took action against a particularly troublesome ALP unit, reorganizing it and sending away its out-of-district members, with the apparent outcome being a reduction of ALP predation.[45] However, the lesson did not seem to stick: When in 2014 an elder from Wardak complained in Kabul about ALP abuses, he not only failed to achieve redress of his grievances, but was subsequently brutally killed by the very ALP members.[46]

In Baghlan the regular police arrested twenty-two ALP members for human rights violations and extortion, and the offenders were sentenced to jail with six- or seven-year prison terms. As a result, the community perceived a decline in the level of abuses perpetrated by the ALP.[47] A possible—and major—glitch in the cleanup effort was alleged by some of the Baghlan residents I interviewed, who claimed that the Afghan National Police’s clampdown on the abuses of the ALP was really a subterfuge by the local Afghan National Police contingent—heavily dominated by Tajiks—to reduce the number of its ethnic rivals—the Pashtuns—in the local police. In fact, the reason the Pashtun community embraced the ALP was that it saw it as a mechanism to redress their marginalization vis-à-vis the Tajiks and their dominated and discriminatory ANP.[48]

At the end of December 2011, upon the urging of President Karzai, ISAF moved to dismantle another little-known irregular police force under its supervision, the Critical Infrastructure Police (CIP). The outfit operated in northern Afghanistan, at least in the provinces of Balkh, Kunduz, Jowzjan, and Faryab, and perhaps also in Sar-i-Pul. Formerly unpaid, the CIP operated as an abusive militia of untrained self-armed men under the control of local power brokers, extorting money from local communities. These militia members numbering in the low thousands—were then rolled into an ISAF irregular forces program and paid between $150 and $200 a month from an American discretionary fund.[49] Under the supervision of ISAF’s Regional Command-North (the command structure responsible for northern Afghanistan), the CIP officially operated for at least three months before President Karzai heard about it.[50] Fearing that the effort reinforced the power of rival power brokers in the north, such as Atta Mohammad Noor, Karzai ordered the program to be shut down. But the disbandment efforts were deeply troubled: CIP strongmen refused to disarm in Faryab, for example, and retained influence throughout the north, such as ex-CIP commander Nabi Gechi in Kunduz who is reputed to illegal tax even weddings.[51]

The performance of the ALP and other self-defense forces around the country has been highly uneven. Under opportune circumstances, such as in areas of homogenous communities or communities without the presence of ethnic, tribal, or other rivalries and threatened by non-local Taliban, the results have at times been positive. Such ALP units and other local self-defense forces could increase security, and the local population might welcome them. In heterogeneous communities torn by various rifts and divisions, however, the creation and embrace of such self-defense forces have often greatly complicated the security environments that local populations face. And even in the supposed success cases, the long-term impact of this program on governance is deeply worrisome.

Communities abused by Taliban outsiders can generate a force on their own to fight the Taliban. And sometimes they do. The negative outcome of such self-arming, subsequently blessed with an ALP label by the U.S. military or the Afghan government, is described later in this discussion, in the context of an intensifying tribal conflict in Nangarhar. However, in other instances, the outcome of such self-arming may be more positive—at least in the extremely narrow sense of expelling the Taliban.

In Andar district in Ghazni province—an area riven by tensions between Hazaras and Pashtuns and where the Taliban has been strong for several years—members of several villages rose up against the Taliban in the summer of 2012.[52] Incensed by Afghan government corruption, people in the conservative district at first welcomed the Taliban. But as Taliban outsiders from Pakistan increasingly tried to impose an ever more restrictive rule, including shutting down a boys’ school and closing several bazaars, men in one village took up arms against the Taliban. The tribal uprising quickly spread to fifty villages representing 4,000 people and fielded 250 armed men who fought some tough skirmishes against the Taliban. Yet that did not mean that the villagers were aligning themselves any closer with the Afghan government, which they continued to resent. Instead, the Ghazni communities in revolt were adamant that they wanted nothing to do with the Afghan government, even rejecting government and U.S. funding and arms, unlike most communities that decided to rise up against the Taliban and then wanted to be anointed as ALP and receive U.S. support.[53] Yet concerns quickly emerged that the Taliban’s insurgent rivals also fighting against the Afghan government—Hezb-i-Islami—were trying to take over the Ghazni rebellion, by raising funds and providing logistics for the rebellion. Hezb-i-Islami has long fought the Taliban over turf in Ghazni. Other Afghan power brokers, including Asadullah Khalid, then chief of security for southern Afghanistan, now director of the National Directorate of Security, who comes from Ghazni, and Faizanullah Faizan, former Hezb-i commander, an Andar native, and, like Khalid, also a former governor of the province, were also actively seeking to control and take over the rebellion, such as by providing it with their own money outside of official government channels.[54] Khalid attempted to integrate the rebellion into his patronage and procurement networks and even brought in outside commanders to take over the leadership of the rebellion, much to the chagrin of the local rebels.[55] Eventually, many of the ALP units came to accept U.S. funding, ammunition, and even armed vehicles—and now bemoan that they lost this U.S. support as U.S. forces withdrew from the province.[56]

With the departure of U.S. forces from the province, the energy of the anti-Taliban self-defense rebellion rapidly withered, while the Taliban made a renewed push in the province. Insecurity for the population significantly increased, including from the ALP units that were supposed to protect them. The absence of supervision by and support from U.S. forces meant that the militias quickly turned to predation and abuse, and neither local community structures nor official government structures, such as district provincial police chiefs, were unable (or unwilling) to restrain them. Unpaid and highly vulnerable to the Taliban’s brutal retribution, ALP units of commanders such as Rahimullah or Abdullah (the latter notorious for extrajudicial killings) came to extort money from communities, gravely beating or even killing those who refused to pay or whom they accused of being pro-Taliban.

Although the Andar rebels originally refused U.S. support, outside backup and logistical support have often been key for the success of the militia forces and for dissuading them from predation on local communities. Yet such outside support has often been critically lacking. In the Afghan province of Wardak, for example, the Afghan Public Protection Program there—a predecessor of the ALP—had originally been issued one rifle and three magazines of ammunition per volunteer by the Afghan Ministry of Interior, as well as one vehicle for every twenty-five men and one radio for every ten men by the United States.[57] But when the APPP ran out of ammunition as a result of Ministry of Interior neglect and shed their uniforms to avoid being targeted by the Taliban, all the U.S. military supervisor of the program could offer was to encourage them “to put on a brave face and look like you have ammo.”[58]

The poorer the backup and the more uncertain and vulnerable the logistical chains, the higher the odds that the self-defense force would surrender to their enemy, such as the Taliban, or even defect to them. Yet the Afghan Ministry of Interior has routinely failed to deliver weapons and other supplies, and for reasons of political contestation and ethnic rivalries, has also at times denied salaries to some ALP units.[59] The ALP members I interviewed in Baghlan and Nangarhar[60] openly disclosed that if U.S. or Afghan forces stopped providing sufficient backup and protection for them from retaliation of their former insurgent comrades, rival tribels, or local powerbrokers aligned with the Afghan government but with whom they were in a hostile relationship, they would work out a new deal with the insurgencies, defect back to them, or at least stop functioning as an ALP unit. Indeed, the ALP units have been primary targets for the Taliban, with their casualty rate three to six times as high as that of the ANA or ANP.[61]

ALP presence in a community has an important political impact, which sometimes is positive. If the community has been systematically disfranchised from power in an area—for example, Ghilzai Pashtuns in Uruzgan do not have representation in local district government or the local police forces—establishing ALP units in such a community does empower it. This empowerment, however, can be used against the district Afghan government as much as against the Taliban. Instead of providing official rule of law or extending the Afghan state, ALP units may well provide protection from the rule of Kabul and local powerbrokers. In fact, this protection from former rule and its abuses may be the primary reason why a community would welcome the self-defense forces in the first place.

Under the best of circumstances, the ALP can increase security against antigovernment forces, such as the Taliban, in communities previously left to suffer; open up roads to villages previously deemed too dangerous for travel and hence boost economic activity in the area; and even reduce local crime, extortion, and land theft. ISAF frequently highlighted the case of Arghandab, a relatively homogeneous Kandahar district without major rifts and conflicts where about 80 percent of the population belongs to the Alikozai tribe, as the exemplar of ALP effectiveness.[62] Overall, the ALP units in Arghandab purportedly achieved excellent results, were effective at fighting the Taliban, and made local communities happy with increased security.[63] Kandahar’s strongman and the provincial police chief, Gen. Abdul Raziq, himself accused of widespread torture, extrajudicial killings, and various other crimes and corruption,[64] has apparently been able to strictly enforce discipline over the ALP and prevent predation on local communities while encouraging the units to engage in unrestrained killing of the Taliban.[65] He also supplemented the ALP pay (officially 60% of the ANP salary) from his own funds.

Difficulties and complexities in many forms, however, tend to arise quickly when a community or an area is not homogeneous, where rifts and conflicts exist, and when the Taliban, Hezb-i-Islami, or other antigovernment elements are not simply thuggish outsiders. In very heterogeneous, polarized, and fractured communities, the establishment of ALP units often exacerbates security dilemmas among the communities and triggers an armament spiral.

Experience with militias in northern Afghanistan reveals such multiple problems and complications. Often, they have engaged in ethnic and tribal infighting and in predation on local communities. In the northern province of Kunduz and western province of Farah, for example, after the ALP defeated the Taliban in their own villages, they started extorting the communities and demanding taxes for themselves.[66] In Faryab, they were accused of looting, raping, and keeping a torture chamber with snakes.[67]

At other times, the various ALP units, when drawn from rival communities or supported by rival warlords, have turned on each other instead of fighting the Taliban. In Faryab, ALP units associated with Junbish engaged in fighting ALP units associated with Jamiat in the fall of 2014, as tense rivalry over the formation of the new national government in Kabul inflamed tensions among the powerbrokers behind the ALP units.[68] Another notorious case of such infighting occurred in Uruzgan in 2010, a place particularly troubled by tribal discrimination and power broker rivalries as well as by severe underdevelopment.[69] Another bad incident with the self-defense forces in Uruzgan took place in August 2012, when a Hazara militia commander, Shujayee, massacred at least nine Pashtuns in a village in revenge for the deaths of two fellow Hazaras. The outraged Pashtun communities believed him to be a member of the U.S.-sponsored local ALP—a claim denied by the United States military as well as some Afghan officials in Uruzgan—but there seemed to be widespread confusion among local officials and the population as to whether Shujayee belonged to the ALP or not. [70] The Pashtun community in response organized anti-government and anti-U.S. protests that the Taliban was able to exploit to mobilize within the community. Far from extending security to the community, the self-defense militias—whether ALP or not—increased insecurity, Taliban mobilization among the Pashtuns, and ethnic tensions.

While the effects of establishing Afghan Local Police units are highly contingent on local contexts, cumulatively the ALP phenomenon transcends the local context and can, through contagion, as it were, generate a widespread and complex security predicament for the whole country. Regardless of the order that ALP units not travel or operate outside of their villages (and, of course, they violate that rule), their reputations do travel among widespread communities. What happens in one Afghan village does not stay in that Afghan village. Instead, rival communities, observing that their antagonists are being armed or retaining their arms, seek to arm themselves as well.

Indeed, both the establishment of militias, such as the ALP, or their demobilization profoundly alter local and national balances of power. Hence militia formation, legitimization, or disarmament are extremely politically sensitive. How these supposedly local and decentralized dynamics intensely reverberate in Kabul and at the national level is also evident in the above analysis of MoI’s factional power-plays with the ALP and other militias.

Even when security improves as a result of the creation of a local ALP outfit, the robustness of that improvement may be far less than meets the eye. Sometimes security in an area improves simply because a community typically hedges its bets and pays part of its income, including what it gets through the ALP salary payments, to the Taliban.[71]

One of the biggest problems with the Afghan Local Police, and other militias, is how to disband them. The U.S. military in Afghanistan always emphasized that the ALP was a “temporary” solution. But no firm plans were developed, let alone implemented, to retire the ALP. Thus even when the need for a long-term approach was to some extent recognized in the ALP policy formulation, it assumed unrealistic sustainability of policy, inconsistent with U.S. domestic political realities. Although the hope was to eventually roll the ALP and other militias into the Afghan National Police, merely establishing supervision, let alone full control, by either Afghan special operations forces and the Ministry of Defense or by the Ministry of Interior, has been a massive challenge and mostly an unfulfilled dream.

Instead, as the 2014 and 2015 fighting in Afghanistan, such as in Kunduz, got very tough, the Afghan local and national government and security forces rushed in to recruit as many anti-Taliban militias as possible. More than 1,000 militia members were recruited in the short span of a few weeks,[72] without any of the vetting, logistical support, and supervision that the ALP effort at least aspired to have. Even prior to the latest recruitment drive, Kunduz was believed to have between 3,000 and 10,000 armed militiamen associated with various powerbrokers and political parties, and not counting the ALP.[73]

Far from stabilizing Kunduz, militias, including many of the ALP brand, have dangerously contributed to the cauldron of ethnic and tribal rivalries, instability, and political exclusion and favoritism that Kunduz has been for years. Instead of the militias succeeding in expelling the Taliban from Kunduz, it is this insecurity, augmented and perpetuated by the presence of militias, that has consistently pulled the Taliban into the province and enabled the insurgency to persist there. And yet throughout Afghanistan, the ALP Directorate now seeks to expand the program by 50% of its current strength to 45,000 and have the United States pay for at least these militias for another decade. The drive is already on to raise at least 5,000 militiamen in seven provinces.[74]

Although insurgency and counterinsurgency dynamics and violent conflict among communities have overshadowed other criminal misbehavior of militias in Afghanistan, many also engage in ordinary banditry, extortion along roads, and drug smuggling. The ALP units have not been immune to these criminal temptations. The lure of drug smuggling is as strong to them as it is to all other actors in Afghanistan, including members of the Afghan government, powerbrokers on all sides of the conflict, the Taliban, and local populations.[75] This is hardly surprising as drugs easily constitute a third of the Afghan economy, a portion that is likely to increase as Afghanistan experiences a likely prolonged recession.[76] In fact, many a local powerbroker smuggling drugs, such as in Shindand, Herat,[77] as appropriated the ALP label to cloak his drug ventures. In Mexico, of course, the recent formation of militia forces has been very much about drugs.

Mexico: Self-Defense Vigilantes or New Crime Gangs Under a Cloak?

Amidst intense and shifting criminal violence, which since 2006 has resulted in the death of between 80,000 and 100,000 people in Mexico,[78] the country’s mountainous center stands out. Although the intensity of homicides has been smaller there than the peak homicide rates in some of the northern cities, such as Ciudad Juárez, Tijuana, and Monterrey,[79] the central states of Michoácan and Guerrero are nonetheless very violent. For decades, they have been some of Mexico’s most prominent locales of illegal cultivation of poppy and production of heroin, both of which have been greatly expanding since 2013 in response to a growing demand in the United States for illegal opiates. Large parts of their territories, including the so-called Tierra Caliente, have historically experienced minimal state presence; and the underdeveloped Guerrero in particular has been one of Mexico’s most lawless states, pervaded by insurgents, criminals, rogue politicians, and militant unions. And Guerrero and Michoacán have also featured some of the most iconic episodes of Mexico’s crime wars, including the killing of students in Iguala, Guerrero in September 2014[80] and the mass killing of presumed members of the Jalisco New Generation Cartel in Ecuandureo, Michoacán in May 2015.[81] Moreover, those two central states have seen the most visible expansion of “anti-crime” militias, capturing the attention of Mexico’s public and ultimately also the Peña Nieto administration.

Militias, whether genuine self-defense forces or private security forces of powerful Mexican politicians, have a centuries-old history in Mexico. Even in the post-WWII period, many a municipal police force in Mexico essentially functioned like a personal (and often abusive) militia force of the district mayor.[82] And many municipal police forces in Mexico are deeply penetrated and often outright controlled by organized crime. So are many municipal governments, particularly in places like Guerrero and Michoacán. Historically, of course, the Mexican government and military often recruited militias to fight insurgencies, such as in Guerrero and Chiapas. Adding to the militia context are officially-sanctioned militias of indigenous communities—defined as indigenous community police force and indigenous justice system—which have been permitted under Mexico’s constitution for several decades.

But over the past several years, the self-defense forces that emerged in response to the extortion and violence of criminal groups in Michoacán and Guerrero came to symbolize the weakness of the central state in providing public safety. As in the different context of Afghanistan, the militias seemed to alleviate violence in the initial period, but soon became predatory and abusive themselves. Nonetheless, the Mexican state, much stronger than the Afghan state, has been better able to respond to their emergence than the government in Afghanistan has been toward Afghan militias.

The home state of former Mexican president Felipe Calderón, Michocán was an early focus of his administration in response to the rapid growth of the violent criminal cartel La Familia Michoacana (LFM). In 2006, LFM was one of Mexico’s most vicious drug trafficking groups, and its authority was expanding over large parts of the state, particularly in La Tierra Caliente. It engaged in brutal violence, visible on the streets of Michoácan.[83] It launched an aggressive extortion campaign that targeted even major businesses in the state, such as avocado growers and logging companies. Not even businesses operating in the state capital of Morelia were immune. By 2009, LFM reportedly had influence over (or extorted anyway) perhaps as many as 180,000 sale outlets in gasoline stations, truck shops, street markets, movie theaters, and other businesses. Its daily earnings were reported (likely highly exaggerated) to be USD 1.9 million.[84]

La Familia’s control over some communities was such that it would monitor the entrance of anyone into a town or village, permitting or denying their entrance and exit, sometimes extorting the person for money. Mixing religion and rituals under a cultish cloak, it also established “courts” and “dispute resolution” procedures for residents of areas under its influence. Indeed, some residents of Michoácan’s Tierra Caliente as well as Morelia told in me in spring 2011 that they actively preferred such courts of La Familia to the formal state justice.[85] Others were just terrified, believing that the group had halcones (lookouts and informants) everywhere; had deeply penetrated mayor’s offices, municipal councils, and local police forces; and could strike anyone.[86] But La Familia also had to battle other criminal groups for turf, including the super-violent and expanding Zetas as well as smaller rivals, such as the Millenio Cartel. Along with government actions against La Familia, it was ultimately a splinter group—Los Caballeros Templarios (like La Familia, embracing cultish religiosity and referring to a saintly purifying mission with which to cloak their criminal endeavors)—which hastened its demise.

During the Calderón administration, Michoacán became one of the first areas where the Mexican military was deployed to combat criminal groups. Like elsewhere in Mexico, one of the military’s key missions was to back up, and in some circum­stances completely replace, Michoacán’s municipal police forces which typically were undertrained, under-re­sourced, deeply corrupt, and completely overwhelmed by organized crime.

Equally important, the new military policing strategy, consisting of high-value targeting and searches at fixed checkpoints, failed to restore or, perhaps more precisely, expand state authority and control. Nonetheless, the high-value targeting strategy was capturing many of LFM’s top leaders; and in the spring of 2011, Los Piños (the seat of the Mexican president) declared LFM dismantled.

Within weeks, however, a new criminal group—Los Caballeros Templarios—emerged and took over the illegal and informal markets in Michoacán that La Familia used to run. Although portraying themselves as a self-defense force to protect Michoacán residents and purge the area of organized crime, Los Templarios soon came to behave like the evil they purported to seek to ostracize: Even more aggressively than LFM, they extorted legal, informal, and illegal businesses. In addition to kidnapping relatives of rich businessmen,[87] they, too, demanded extortion fees from avocado farmers and logging companies, and expanded the extortion racket into iron ore extraction and shipping through Michoacán’s principal port and economic hub, Lázaro Cárdenas. In fact, in March 2014, the Mexican government’s special envoy for restoring rule of law in Michoacán, Alfredo Castillo, claimed that Los Templarios made more of their money from extorting the iron ore extraction, processing, and transshipment operations than from drug smuggling or other extortion.[88] Regardless of whether this assessment of the cartel’s financial portfolio is accurate, the Templarios, exploiting their strong territorial presence and a fearsome reputation, succeeded in turning themselves into a multifaceted mafia with fingers in many an illegal racket in the state and widespread extortion.

By the spring of 2014, Los Templarios were the area’s most feared authority. Despite their purported emergence in reaction to the abuses and excesses of La Familia Michoacana, the Templarios also overreached in their demands for extortion fees and obedience and triggered a backlash. Going beyond mere popular resentment, anti-Templarios militias began forming in Michoacán’s countryside even before the influence of the Templarios peaked.

Anti-crime self-defense forces, such as in Michoacán’s Cherán, began emerging as early as 2011, but the Calderón administration did not pay much attention to them. But their spread to other areas, visibility, and increasingly questionable behavior continued to grow through 2013. By then, the militias were arresting people whom they accused of working for the Templarios and other criminal groups, and held their own court trials and meted out sentences. They were particularly active in Michoacán’s towns of Tepalcatepec, Buena Vista, and La Ruana, where they gathered whatever weapons they could find and seized controlled of police stations. By the time the self-defense forces came to beat up, expel, and detain not just municipal police officers, but also soldiers,[89] the administration of Calderón’s successor, President Enrique Peña Nieto, could not remain placid about their growth. But even detentions of militia members who were engaged in the worst excesses, such as kidnappings of police personnel, did not appear to deter them.[90]

The militias also grew in the neighboring state of Guerrero, one of the most violent areas in Mexico during the Peña Nieto administration so far, with 73.2 homicides per 100,000 in 2013, compared to the national average of 29.3 per 100,000 that year.[91] Although its homicide rate decreased in 2014, Guerrero remained the second most violent state in Mexico.[92] A plethora of small, fragmented, unstable, and highly violent criminal gangs emerged in the state in the wake of the federal government’s high-value-targeting interdiction policy against the once dominant Beltrán Leyva Cartel. Like in Michoacán, the Jalisco New Generation Cartel from neighboring Jalisco has also been encroaching on their territory, triggering violent battles.

In Guerrero, the provenance and control of the militias seems even murkier than in Michoacán. Some of the self-defense militias appeared to be permeated by organized crime groups, such as the Jalisco New Generation Cartel.[93] In fact, a response of the cartels to the self-defense group formation has been to label their own hitmen as self-defense groups and to attempt to penetrate and subvert the existing self-defense groups. At the same time, the militia forces in Guerrero have also been intricately intermeshing with the so-called “community police forces,” legally permitted under Mexico’s constitution and allowed to carry firearms, which operate mainly in indigenous communities. In the spring of 2013, there were 45 such community police groups in 14 of Mexico’s 32 states.[94] In Guerrero’s municipality of Ocotito, for example, the local self-start-up militia force appeared to have the assistance of the Union of the People and Organization of the State of Guerrero (Unión de los Pueblos y Organizaciones del Estado de Guerrero: UPOEG) community police force.[95]

Moreover, an extensive whispering campaign emerged in both Guerrero and Michoacán that the militias might also be taking justice into their own hands more aggressively—such as by killing those they conceived of being their opponents. At minimum, they would trot around with machine guns, expel or arrest municipal police officers they saw as incompetent or corrupt, and block roads, using their own discretion to determine who could go in and out.

The original reactions of high officials of the Peña Nieto administration were to denounce the militias. The president, for example, pointedly stated: “[W]hatever the denominations of these groups, the practice they have of taking justice into their own hands [is] outside the law, and my government will combat it.”[96] But at the same time, state officials in Michoacán continued hinting that the militia existence could be tolerated. In Guerrero, the contradictions between state and federal-level authorities and among state responses were even more pronounced: On the one hand, the state was providing the self-defense forces with funds, uniforms, and communications equipment, while on the other hand, it was arresting at least some militia members. In the spring of February 2014, as one of Guerrero’s militia groups seized villages on the outskirts of the state capital, Chilpancingo, Mexico City dispatched further military battalions and federal police units to stop them from moving into the city itself.

As the process unfolded, even federal level officials learned that doing away with the militias was not easy. Negotiating with the militias to effect their disarmament proved especially difficult, as militia members emphasized that they would be subject to retaliation and could only disarm after the criminal gangs, including the key leaders of the Templarios, were arrested. But forcibly dismantling the militias could set off a bloody and problematic fight between them and the federal government, in which assistance from local and state authorities could not necessarily be counted on. After all, the militias’ own narrative claimed that they were merely defending themselves and their families and communities against the brutality of the crime groups because the state had failed to do so, which indeed was often the case.[97]

The increased deployment of Mexico’s military into Guerrero and Michoacán, which President Peña Nieto boosted there by fifty percent at the beginning of 2013, did not slow the formation, spread, and audacity of the militia forces. By the end of 2013, 47 out of Michoacán’s 113 municipalities experienced their presence. In the neighboring state of Guerrero, they operated in more than half of the state’s 81 municipalities in the spring of 2014.[98] Areas that were key Templarios hotbeds in Michoacán, such as Apatzingán, experienced dramatic fire-arm battles between the Templarios and the self-defense forces. Elsewhere, the self-defense forces set up checkpoints. In January 2014, the self-defense forces took over the municipal building in Parácuaro and blocked off entry points to the town, digging in for a battle with the Templarios, until the Federal Police negotiated its own entry. The militias also seized control of a nearby town, La Huerta. In some parts of Michoacán, the Federal Police began operating joint checkpoints with the self-defense forces. Membership in the militias swelled to the thousands, by some reports to as many as 20,000,[99] though no reliable counts were conducted, and the militias had an incentive to exaggerate their strength. To accommodate to the militias’ insistence that they could only stop their vigilantism if the government arrested key leaders of the Templarios, the government launched a dragnet in Michoacán and over several months captured key Templario leaders.

When a prominent Templario leader known as “El Tío” was arrested in January 2014, Mexico’s Interior Minister Miguel Ángel Osorio Chong announced that the government negotiated a deal with the groups to absorb them into a new state security entity known as the “Rural Defense Corps.” The deal specified that the corps would be temporary and required that the militia leaders would provide the government with a registry of their members. Putting a time limit on the existence of the militias was a highly appropriate provision since dismantling any unofficial and extralegal forces and vigilantes, however motivated, always needs to be the position of a state adhering to the rule of law.

Even so, there were good reasons to doubt the desirability of the arrangement. The fact that the government was not able to prevent and dismantle the militias in the first place, and was essentially left to make a deal with them, was glaring evidence of the weakness of the state in the rural center areas of Mexico. The deal also created a bad precedent, signaling that if one wanted to get on the payroll of the state and take the law and its enforcement into one’s own hands (or cloak one’s extortion and other crimes with legitimacy), one only had to set up a self-defense militia. More immediately, there were good reasons to be skeptical about the accuracy of the member registry handed over to the state by the militia leaders and the ability of the state to do its independent re-vetting of the militia members. Moreover, it was not obvious just how committed the militias were to the deal: A key militia leader, Dr. José Manuel Mireles, was not at the signing; and another militia group from the Ruana area was not only absent, but occupied the government building in the Peribán municipality that very same day. In Guerrero, the militias rejected a similar deal to be folded into an official rural defense force, claiming they did not believe Mexico’s federal government was truly motivated to combat the criminal groups.[100]

But however problematic, the deal to form the Rural Defense Corps was clearly better than the previous policy of just allowing the militias to run loose and act without restraint. While not desirable, the Rural Defense Corps concept at that moment was likely the least bad option the government had available. It was only a matter of time before the unsupervised militias would start engaging in predation on local communities, designating as a criminal anyone who crossed them, arrogating “justice” to themselves, and further damaging the already poor bonds between the state and the population. And it was not too far-fetched to imagine that they might be tempted to take over some illicit markets.[101]

Indeed, such problematic developments surrounding the militias and their speedy descent into going rogue was exposed just a few weeks after the deal was signed. By the middle of March 2014, Mexican authorities had to arrest one of the top militia leaders, Hipolito Mora, indicting him for the murders of two members of a rival militia faction in Buenavista Tomatlán. Government authorities also detained 28 other vigilante members, accusing them of stealing and appropriating the property of alleged Templario members, such as ranches, land, and horses, while demanding money from local citizens for returning their property stolen by the Templarios.[102] Announcing the arrests, Mexican authorities implied that they would no longer tolerate the militias, now that the government had developed independent intelligence networks to go after the Templarios. In April 2014, an additional 100 militia members were arrested on charges that they were in fact criminals (some belonging to the Templarios) merely posing as self-defense forces.[103] The militias, including those of other factions, such as the Tepacaltepec group led by Mireles, claimed that the government was unjustly prosecuting them while failing to deliver on its part of the negotiated deal, and that Mexico’s government still could not cope with security in Central Mexico without help from the militias.[104] Another vigilante spokesman, Estanislao Beltrán, admitted that some bad elements, including criminals, might have infiltrated the militias, but that the militias would clean their own ranks themselves and continue operating, though preferably under a government hat.

Thus in April 2014, the federal government declared that the self-defense groups agreed to disarm by May 10—but the deadline was missed and the militias showed little interest in obeying the basic deal struck in January 2014. At the same time, José Manuel Mireles declared that the self-defense groups under his influence would now work with federal forces in cities like Morelia, Uruapan, and Lázaro Cárdenas to take down all remaining members of the Templarios, including middle-level managers, thus changing the terms of the deal and parameters of the disarmament of his militias. He also stated that as part of a new deal with the government, the federal authorities agreed to release many of the arrested self-defense group members.[105]

Overall, the government deal with the militias starting breaking down almost as soon as the ink on the paper had dried. Some militias joined the Rural Defense Corps, receiving guns, uniforms, and salaries from the government, while others continued to drag their feet. For the rest of 2014, the Mexican government kept negotiating with the various militia factions, arresting leaders and members of some, only to release them later. Nonetheless, by December 2014, most of the major militia factions in Michoacán, including those of Hipolito Mora and his rival Luis Antonio Torres, known as “El Americano,” were nominally folded into the Rural Defense Corps.

But their nominal presence in the state-sanctioned outfit did not guarantee that the state had adequate control over the behavior of the militias. In the middle of December 2014, Mora’s and Torres’s factions engaged in a bloody shootout with each other in the town of La Ruana, leaving 11 people dead, including Mora’s son. Although Mora and Torres handed themselves over to state authorities, and later were indicted with homicide and kidnapping charges, violence among their factions and between those factions and a new offshoot of Los Templarios, Las Viagras, continued into January 2015. Official military and federal police forces also began responding with greater violence toward the militias, including in a notorious incident after one of the militia forces tried to seize the town hall of the city of Apatzingán.[106]

Both in Michoacán and Guerrero, violence and the rise of the militias also had an effect on Mexico’s midterm elections held in June 2015. In Michoacán, the leader of one militia faction, Enrique Hernandez, was assassinated in March as he tried to campaign on the ticket of the left-leaning Movement for National Regeneration, or Morena, party.[107] He had earlier spent three months in jail, but was released for a lack of evidence.

Meanwhile in Guerrero, the federal state did not manage to even sign a deal with the militia, let alone enforce it. Various militia groups, whether genuinely indigenous police forces or fronts for local criminal gangs, continued to arrest and detain soldiers and government officials, and homicide and extortion rates remained high.[108] The election campaign in 2015 took place amidst bitter memories of the Iguala massacres and widely-assumed state complicity, widespread intimidation by rival militias and organized crime groups, disappearances, and assassinations of local government officials and political candidates.

One of the most dramatic incidents involving Guerrero’s self-defense forces took place in early May 2015 in the town of Chilapa. Although small in size, Chipala is strategically-located on the foothills of a major poppy growing area and a major logistical hub for the drug trade since it is the place with the only gas station in miles. Following an assassination of a local political candidate in April 2015, 300 civilians armed with rifles, machetes, and sticks, followed by pickup trucks with men sporting high-caliber weapons, seized the town. Although the Mexican military and federal and municipal police were present, they failed to act against the self-proclaimed self-defense group. Whether out of intimidation, indifference, complicity, or on orders from higher up, the military and police stood by as for several days the militias controlled the town, set up checkpoints, and detained people. At least 11 of those detained (and perhaps as many as 30) have not been seen since. Townspeople believed that the self-defense force, which after several days left on its own accord, was actually the criminal gang Los Ardillos, fighting over the important heroin-turf with another gang, Los Rojos.[109] Regardless of whether the armed invasion was by a self-defense force that ran amok or the self-defense label was appropriated by an organized crime group, its effect on the community was the very opposite of increasing security.

In some ways, the willingness of the government to act against the militias, including to arrest and prosecute some, has been more encouraging than its other anti-crime policies.[110] The original plan of folding them into the Rural Defense Corps also was the least bad option; however, the government has failed to effectively enforce it with the militias. In Guerrero, the government has not even been able to force them to sign such a deal. In both Michoacán and Guerrero, many of militias have become important sources of conflict and abuse, hardly acting as a stabilizing force. Indeed, the Mexican government needs to retain the resolve to monitor the militias diligently; prosecute those who engage in criminal acts, such as extortion and murders; and use any opportunity it can to roll them back and dismantle them—even if such efforts have not been going well so far. The militia option might seem seductive in the short term at a moment of crisis, but it spells long-term problems for security, rule of law, and state legitimacy, as much in Mexico as in Colombia or Afghanistan. To the extent that Mexico’s struggle against criminality is not merely about reshuffling who has control and power in the criminal market, but about a broader extension and deepening of the rule of law and accountability in Mexico, any official endorsement of the militias fundamentally contradicts that project.

Key Conclusions and Implications for U.S. Foreign Policy

Although the cases of Afghanistan and Mexico are very different in many structural dimensions, such as state weakness and economic strength; although one country has been at war for decades whereas the other has faced very violent criminality for only over one decade; and although there are some important differences in the origins and pervasiveness of militias, some consistent striking take-aways emerge.

From a policy perspective, the most salient findings are the following:

  • In Afghanistan and Mexico, militias seemed to have the least proclivity toward abuse of local and rival communities when they emerged spontaneously from the local community, faced a particularly abusive external force (whether outside Taliban or outside criminal groups), and if major rifts and conflicts were absent from the community of the militia’s origin.
  • In communities contested along ethnic, tribal, or indigenous versus non-indigenous lines, or with intense conflict over land and legal and illegal resources, the rise of militias intensified local conflicts and resulted in predation on communities.
  • In the absence of effective supervision by and support from strong official forces, such as powerful domestic or outside military or police forces, militias in both Afghanistan and Mexico quickly turned to predation and abuse, no matter what their original motivations and self-justification. Although U.S. special operations forces in Afghanistan sometimes developed effective understanding and control, they failed to create such Afghan supervisory structures (whether from the Ministry of Interior or Afghan special operations forces). With the departure of U.S. special operations forces, the tendency of the militias toward predation and rogue behavior increased significantly.
  • In both Afghanistan and Mexico, local community structures have often been unable (or unwilling) to restrain the behavior of the militias.
  • In Afghanistan, not surprisingly, given the weakness of the Afghan state, official government structures, such as district provincial police chiefs, have been equally unable (or unwilling) to restrain the militias. The Mexican federal government has made more of an effort to regularize the militias, including by folding them into official, if ad hoc and presumably temporarily-created, police structures. The government also set limits on what kind of activity the militias could engage in and established some vetting procedures of members. But it has been unable to fully implement and enforce these formal rules. Nonetheless, the Mexican government has been at least willing to indict and arrest militia leaders for the most notorious abuses perpetrated by their units, such as murders, kidnapping, and extortion.
  • No matter what their origins and motivations, the rise of militias profoundly changes local balances of power. Consequently, both local and outside actors seek to appropriate the militias or establish rival ones. In Mexico, even when the militias rose to oppose the brutality and extortion of criminal groups, cartels sought to take them over or establish rival “militias.” In Afghanistan, rival tribes and powerbrokers reacted in the same way.
  • In both countries, such competition over control and establishment of militias was also present in official government structures: In Afghanistan, the Ministries of Interior and Defense, as well as the intelligence services, competed in controlling their militias and sought to subvert the militias of the rival ministry. Similar conflicts regularly occurred between district and provincial police chiefs and rival ministries. In Mexico, municipal and state government officials often had militia policies directly contradictory to those of the federal government.
  • Although the formation of militias may have originated as a local matter, the security and political effects the militias had did not remain contained within a small locality or a village. The balances of power they affected were much broader. So were the contagion effects they set off, diminishing perceptions of state strength.

The experience of building up militia forces in Afghanistan, as well as of their rise in Mexico, should induce the United States to seriously reexamine whether conducting its security policies through such proxies is a reliable and wise strategy. Even though the United States may well be—appropriately—unwilling to engage in large-scale military interventions, it should not delude itself into believing that interventions via proxies are easy, quick in-and-out engagements. The proxies will often require a long-term management and deep involvement by the United States so that they deliver the objectives the United States seeks and do not go rogue. Nor is the United States the only power that should have recently learned such a lesson: France’s inability to extricate itself from the presumably quick in-and-out intervention in Mali and find or establish adequate partners to whom to hand over a sufficiently stable “post”-intervention order is a potent example.

Indeed, whether in building partner capacity of official government forces or establishing militias against the enemies of U.S. interests, the United States needs to assume the very opposite of what has come to underpin such policies—namely, that the provision of security assistance will produce a robust alignment of interests. Rather, it is very likely that a divergence of interests between the United States and U.S. partners and proxies will rapidly emerge, and that the proxies will prosecute their own interests, and not sufficiently those of the United States.

Just like the Northern Alliance and other warlords on whom the United States relied in the 2001 Afghanistan intervention, the presumed partners, whether in government or out, will often pocket the money (and weapons) and run. They will service their own interests, not those of the United States. They will often undertake the absolute minimum action necessary to keep U.S. resources flowing while striking their own deals and accommodations, hedging their bets, or targeting their own enemies, not those of the United States. “Partner” action might thus often come with side-effects that the United States will find highly undesirable, such as blatantly eliminating one’s political enemies with U.S. weapons, and at times directly subverting U.S. objectives.

Moreover, various actors in the recipient country will likely have different political interests and objectives and will seek to manipulate security assistance to serve them, competing with and undermining rival ministries or rival militias. Even policy execution consistent with U.S. objectives and design by one official interlocutor may be subverted by other powerful official actors in the system later (or even at the same time).

Instead of a robust and lasting alignment of security interests and joint understandings of how to build stable and legitimate political order, the United States and its presumed partner might often enjoy only a momentary intersection of interests and understandings. Such divergence of interests is particularly likely to arise during intense and fluid conflict situations; but it may well characterize the relationship of the two countries even during times of peace. Thus presumed U.S. allies—whether official, such as government forces, or unofficial, such as non-state militias—may well undertake systematic actions that the United States considers problematic and potentially directly contrary to U.S. national interests.

Does this imply that the United States should not engage in building up capacities of presumed partner governments and standing up militia forces? The question cannot be answered in the abstract. While such a determination needs to be made on a case by case basis, one of the critical policy evaluation criteria needs to be an in-depth consideration of when and how these presumed partners will defect from U.S. objectives, and where the presumed momentary intersection of interests ends. Policy evaluations need to include considerations of when and how recipients of U.S. security assistance will undertake actions that will contradict, or at least are inconsistent with, U.S. interests.

Crucially, policy evaluations and choices need to consider what kind of restraint and rollback capacity of these presumed partners—whether state or militias – the United States will have at the point of interest divergence, and what kind of actions the United States could take to mitigate the negative impacts on U.S. interests and values.

Particularly in the case of nonstate irregular forces, such as militias, the presumption for U.S. policy should be not to stand them up. Supporting them against more dangerous and vicious U.S. enemies should be the last resort since the odds are very high that they will turn rogue and predatory and ultimately contradict U.S. interests.

To the extent that the United States does provide security assistance to a new “partner,” it should primarily seek to engage formal military and police forces as they can presumably be more accountable to the United States—as well as, and no less importantly – to local populations. Such form of security assistance is also more easily consistent with future stabilization and state-building efforts than supporting non-state militias.

Nonetheless, if supporting militia forces is the least bad option, such as because formal military and police forces are totally corrupt, collapsed, or antagonistic to the United States, the United States needs from the very beginning to build into the effort a consideration of how to end and roll them back. Merely defunding the militias is not enough: indeed, cutting them off from funds without dismantling them is only likely to encourage them to engage in predation on local communities or sign up with U.S. enemies willing to pay them. Ideally, any militia policy program would include developed and implementable policies of disbandment when needed. Such plans need to be based on realistic assessment of policy sustainability and include worst-case scenario rollback options, understanding that political will to persist in a direct or even indirect military engagement is likely evaporate much faster than many a military plan assumes.

Ideally, when embarking on building up partner capacity or standing up militias, the United States would retain the option of being engaged in the country for a long time and involved in its political processes. Nuanced intelligence and a broad understanding of the multiple political impacts the militias and the rollback processes have are equally required. If such mechanisms of restraint are elusive (which will often be the case in many situations of intense conflict with dangerous terrorist groups and miserable governance by official state actors), the United States needs to realize that its embrace of militias will merely defer immediate threats to later and long-term problems, even compound them. Our friendly militias today will likely end up a threat to our interests in a matter of time.

Assessments of the chances of success thus need to be much broader than merely eliminating a particular terrorist group. They also need to include judgement of whether a sufficiently stable, sustainable, and legitimate order and governance will ensue or whether supporting “partners” merely perpetuates structural causes of instability.

Upholding human-rights restrictions in deciding to whom U.S. security assistance can be provided, such as embodied in the Leahy Amendment, is important. It is crucial not only because of U.S. values and humanitarian interests. It is also important because such assessments of eligibility are a good indication of whether U.S. assistance will be effective or end up supporting an actor whose rapacious, predatory behavior ultimately fuels a cycle of instability and keeps the United States stuck in a mire—even if via a proxy.

Setting up diligent, effective, and lasting monitoring mechanisms, informed by input from as broad a set of local interlocutors, as possible needs to be a required element of any U.S. effort to build partner military capacity or stand up proxy forces.

When the United States determines that building up militia forces or problematic government proxies cannot be avoided even while sufficient rollback mechanisms are lacking, the United States must be willing to renege on any deal with them when they start contradicting U.S. interests. That may involve letting them sink politically and militarily. For reasons of honor and military creed, such a policy assumption might be difficult for the U.S. military to adopt. Indeed, such a conditional policy entails many problems, including setting problematic precedents and dissuading some actors from even temporarily cooperating with the United States. But otherwise, the United States risks being sucked into long-term and potentially unresolvable conflicts with partners who are allies one day and enemies the next.

Building up partner capacity of irregular militias as well as official government forces can only be successful if the efforts are based on detailed as well as broad intelligence. The kind of intelligence that is needed is not merely tactical intelligence for a particular advisory support effort of special operations forces. The intelligence that is necessary is much deeper and broader intelligence that comprehends cultural and institutional contexts and the nuances of historical and political rivalries, such as who did what to whom forty years before the current eruption of crisis. Indeed, the less the United States is directly engaged, and the more it relies on someone else to carry the water of its interests, the shrewder it needs to be about how to deal with the interests and pressures of its presumed partners.

Indeed, overall the United States need to recognize that providing security assistance, building up partner capacity, or standing up militias are not merely technical endeavors or tactical military tools. They are all profoundly political undertakings, with potentially large impacts on power distribution and political arrangements in recipient countries. All such efforts by the United States thus need to be linked to a larger U.S. political strategy of how to create sustainable structures of stability and order, backed up by sufficient and sustainable resources and legitimacy.

Finally, in all such security assistance endeavors, the United States needs to comprehend its limited ability to calibrate and leverage the situation and its assistance.

But to base national security policy on such an understanding does not imply that the United States is impotent. It means rather that the United States needs to use its leverage to demand policy accountability from recipients of its security assistance.

[1] The disarmament, demobilization, and reintegration process (DDR) in Afghanistan has been largely ineffective, despite its multiple incarnations. The various DDR efforts have often been manipulated by influential Afghan powerbrokers seeking to disarm their rivals while they kept their own militias, and often contradicted by U.S. counterterrorism and NATO’s insurgency policies. For a first-rate detailed examination of the many DDR efforts in Afghanistan and their shortcomings, see Deedee Derksen, “The Politics of Disarmament and Rearmament in Afghanistan,” United States Institute of Peace (USIP), May 20, 2015,
[2] Office of the Press Secretary, The White House, “Weekly Address: A New Chapter in Afghanistan,” May 5, 2012,
[3] Few spaces and domains are truly ungoverned; most are merely governed by non-state entities, including possibly insurgent and terrorist groups or criminal organizations. See, for example, Anne L. Clunan and Harold Trinkunas, “Conceptualizing Ungoverned Spaces: Territorial Statehood, Contested Authority, Softened Sovereignty,” in Anne L. Clunan and Harold Trinkunas, eds., Ungoverned Spaces: Alternatives to State Authority in an Era of Softened Sovereignty (Stanford: Stanford University Press, 2010); and Vanda Felbab-Brown, “Rules and Regulations in Ungoverned Spaces: Illicit Economies, Criminals, and Belligerents,” in Anne Clunan and Harold Trinkunas, eds., Ungoverned Spaces: Alternatives to State Authority in an Era of Softened Sovereignty (Stanford: Stanford University Press, 2010).
[4] For a review of U.S. government concepts and programming, see Government Accountability Office (GAO), “Building Partner Capacity: State and DOD Need to Define Time Frames to Guide and Track Global Security Contingency Projects,” November 2014,
[5] Office of the Press Secretary, the White House, “President Barack Obama’s State of the Union Address,” January 28, 2014,
[6] For details on the Weinberger-Powell doctrine, see Seyom Brown, Faces of Power: Constancy and Change in United States Foreign Policy: From Truman to Obama (New York: Columbia University Press, 2015): 414-16.
[7] For a review of key analytical concepts, see Christopher Paul, Colin P. Clarke, Beth Grill, Stephanie Young, Jennifer D. P. Moroney, Joe Hogler, and Christine Leah, “What Works Best When Building Partner Capacity and Under What Circumstances?” RAND, 2013,
[8] What military advisors actually do on the ground can include gathering and providing intelligence, directing artillery and air support fire, training local units, directing their operations, as well as monitoring their performance and the use of U.S. assistance.
[9] For details, see Vanda Felbab-Brown, Aspiration and Ambivalence: Strategies and Realities of Counterinsurgency and State-building in Afghanistan (Washington, DC: The Brookings Institution Press, 2013).
[10] See Geraint Hughes, “The Soviet-Afghan War, 1978–1989: An Overview,” Defense Studies 8, no. 3 (2008): 326–50.
[11] See, for example, Barnett Rubin, Fragmentation of Afghanistan (Yale University Press, 1995); Ahmed Rashid, Taliban (Yale University Press, 2001); Anthony Davis, “How the Taliban Became a Fighting Force,” in William Maley, ed., Fundamentalism Reborn: Afghanistan and the Taliban (New York University Press, 1998): 43-64; Ahmed Rashid, “Pakistan and the Taliban,” in William Maley, ed., Fundamentalism Reborn: Afghanistan and the Taliban (New York University Press, 1998): 72-84; Steve Coll, Ghost Wars (New York: Penguin Press, 2004); and Bruce Riedel, What We Won: America’s Secret War in Afghanistan, 1979 to 1989 (Washington, DC: The Brookings Institution Press, 2014).
[12] Since January 2015, NATO’s previous ISAF mission has changed to a more limited training and advising mission known as Operation Resolute Support. For details, see Vanda Felbab-Brown, “Blood and Hope in Afghanistan: A June 2015 Update,” The Brookings Institution, May 26, 2015,
[13] For two comprehensive and insightful studies of Afghan militias and their impacts, see Jonathan Goodhand and Aziz Hakimi, “Counterinsurgency, Local Militias, and State-building in Afghanistan,” United States Institute of Peace (USIP), December 2014; and The International Crisis Group (ICG), “The Future of the Afghan Local Police,” Asia Report No. 268, June 4, 2015,
[14] For an overview of these various efforts, see David Axe, “War Is Boring: Fourth Time the Charm for NATO’s Afghan Militia Plan?” World Politics Review, July 21, 2010.
[15] For details, see Craig Whitlock and Greg Miller, “U.S. Covert Paramilitary Presence in Afghanistan Much Larger than Thought,” Washington Post, September 22, 2010. See also, Bob Woodward, Obama’s Wars (New York: Simon and Schuster, 2010).
[16] Whitlock and Miller, “U.S. Covert Paramilitary Presence.”
[17] Interviews with the National Directorate of Security and Afghan National Police (ANP) officials, Kabul and Baghlan, Afghanistan, September 2010 and April 2012.
[18] Philipp Münch, “Local Afghan Power Structures and the International Military Intervention: A Review of Developments in Badakhshan and Kunduz Provinces,” Afghan Analyst Network,
[19] Interviews with ISAF, U.S. Embassy, United Nations Assistance Mission in Afghanistan (UNAMA), and Afghan government officials, Kabul, fall 2010 and spring 2012.
[20] Derksen: 7-11.
[21] For a background on the historic arbakai, see Mohammed Osman Tariq, “Tribal Security System (Arbakai) in Southeast Afghanistan,” Occasional Paper 7 (London School of Economics, Crisis States Research Center, December 2008),
[22] Interviews with high-level ISAF officials and members of the U.S. Special Operations Forces responsible for the ALP, Kabul, Kandahar, Herat, and Balkh, fall 2010 and April 2014. For details on the Anbar Awakening and its relationship to the U.S. military surge in Iraq, see Stephen Biddle, “Lessons Learned in Afghanistan and Iraq,” in U.S. Policy in Afghanistan and Iraq: Lessons and Legacies, edited by Seyom Brown and Robert Scales (Boulder, Colo.: Lynne Rienner, 2012), pp. 89–98. See also Marc Lynch, “Explaining the Awakening: Engagement, Publicity, and the Transformation of Iraqi Sunni Political Attitudes,” Security Studies 20, no. 1 (2011): 36–72; John McCary, “The Anbar Awakening: An Alliance of Incentives,” Washington Quarterly 32, no. 1 (2009): 43–59.
[23] See, Inspector General, Department of Defense, Assessment of U.S. Government and Coalition Efforts to Develop the Afghan Local Police, Report DODIG-2012-109, July 9, 2012,
[24] ICG, “The Future of the Afghan Local Police”: 1.
[25] Ibid.
[26] For an academic military assessment of a similar positive view, see, for example, Daniel Green, “Retaking a District Center: A Case Study in the Application of Village Stability Operations,” Military Review 95(2) March/April 2015: 118-124. For a more critical assessment, see Jon Strandquist, “Local Defense Forces and Counterinsurgency in Afghanistan: Learning from the CIA’s Village Defense Program in South Vietnam,” Small Wars and Insurgencies 26(1), 2015: 90-113.
[27] Interview with a street vendor in Pul-e-Khourmi who lived in a nearby village, Baghlan, September 2010. This section overall is based on interviews with Afghan officials at all levels of the government, ANP officers, ALP members, maliks, Afghan civil society organizers, businessmen, as well as ordinary Afghans such as shopkeepers, Kandahar, Helmand, Uruzgan, Nangarhar, Baghlan, and Zabul, spring 2009, fall 2010, and April 2012.  For a recent, highly visible case of such abuse attributed to the ALP, see Alissa Rubin, “Rape Case, in Public, Cites Abuse by Armed Groups in Afghanistan,” New York Times, June 1, 2012.
[28] Interview, Kandahar, fall 2010. For a comprehensive documentation of recent cases of abuse by irregular pro-government forces in Afghanistan (both the ALP and others), see Human Rights Watch, Just Don’t Call It a Militia: Impunity, Militias, and the “Afghan Local Police” (New York, 2011),
[29] Apparently, General Stanley McChrystal and General David Petraeus had to spend considerable time persuading Karzai to concede to the efforts. Interviews with ISAF officials, Kandahar, spring 2009, and Kabul, fall 2010. See also Matthew Rosenberg and Alissa Rubin, “Afghanistan to Disband Irregular Police Force Set Up under NATO,” New York Times, December 26, 2011; David Cloud, “U.S. Plans to Beef Up Rural Afghan Forces,” Los Angeles Times, August 18, 2012.
[30] Derksen: 35.
[31] Interview in Baghlan, April 2012.
[32] For contradictions within the effort to disarm illegal armed groups (DIAG), see Robin Edward-Poulton, DIAG Evaluation  – Disbandment of Illegal Armed Groups in Afghanistan, A Project of the United Nations Development Programme and Afghanistan’s New Beginnings Programme (ANBP) (Oxford: EPES Mandala Consulting, April 2009),; Michael Vinay Bhatia and Robert Muggah, “The Politics of Demobilization in Afghanistan,” in Security and Post-Conflict Reconstruction: Dealing with Fighters in the Aftermath of War, edited by Robert Muggah (New York: Routledge, 2009): 126–64.
[33] Cited in Inspector General, Assessment of U.S. Government and Coalition Efforts: ii.
[34] See David Flynn, “Extreme Partnership in Afghanistan: Arghandab District, Kandahar Province, 2010–2011,” Military Review, March-April 2012: 27–35.
[35] Interviews with ANP officer in charge of the local ALP, Baghlan, and with political advisors to ISAF, Regional Command-North, fall 2010 and spring 2012.
[36] Derksen: 37-38.
[37] “Afghanistan: Fears over Child Recruitment, Abuse by Pro-government Militias,” IRIN (UN Office for the Coordination of Humanitarian Affairs news service), January 20, 2011.
[38] See, for example, Barnett Rubin, “Saving Afghanistan,” Foreign Affairs 86, no. 1 (2007): 57–78.
[39] Interviews with a SOF officer, Herat, April 2012.
[40] For details see, Human Rights Watch, Just Don’t Call It a Militia: 97–98.
[41] ICG, “The Future of the ALP”: 3.
[42] Author’s interviews with officers of U.S. Special Operations Forces responsible for the ALP, Kabul, July 2013.
[43] Derksen: 38.
[44] Problems of restraining, disciplining, and disbanding the ALP in other cases an areas, such as Parwan and Langham, are documented in ICG, “The Future of the ALP”: 11.
[45] I am grateful to Ronald Neumann, former U.S. ambassador to Afghanistan, for the information about the Wardak ALP.
[46] ICG, “The Future of the ALP”: 10.
[47] Interviews with ALP members, ANP officers, civil society representatives, and local residents, Baghlan, April 2012.
[48] Author’s interviews with Pashtun leaders and ALP commanders, Baghlan, September 2012.
[49] Interviews with former and current political advisers in ISAF’s Regional Command-North, Balkh, April 2012. See also Rosenberg and Rubin, “Afghanistan to Disband.”
[50] Rosenberg and Rubin, “Afghanistan to Disband.”
[51] ICG, “The Future of the ALP”: 5.
[52] For further details on this uprising, see “The Worm Turns,” The Economist, August 18, 2012.
[53] David Young, “An ‘Afghan Summer’ of Revolt,” Foreign, September 12, 2012,
[54] Ibid.
[55] Ibid.
[56] Joseph Goldstein, “After U.S. Exit, Rough Justice of Afghan Militias,” New York Times, March 18, 2015.
[57] Mathieu Lefèvre, “The Afghanistan Public Protection Program and the Local Defense Initiative,” in Snapshots of an Intervention: The Unlearned Lessons of Afghanistan’s Decade of Assistance (2001–2011) , edited by Martine van Bijlert and Sari Kouvo (Kabul: Afghanistan Analysts Network, 2012): 74–79.
[58] On the positive side, the Wardak Afghan Public Protection Program militias showed considerable restraint by avoiding involvement in local tribal disputes, such as those between the Hazaras and the Kuchis over grazing lands. See Axe, “War Is Boring.”
[59] Interviews with ISAF and UNAMA officials, Kabul, Kandahar, Herat, Balkh, and Baghlan, fall 2010 and spring 2012. See also: House Armed Services Committee, Subcommittee on Oversight and Investigations,  Hearing on Afghan National Security Forces and Security Lead Transition: The Assessment Process, Metrics and Efforts to Build Capacity, Testimony of Kenneth Moorefield, Deputy Inspector General for Special Plans and Office of Inspector General, Operations, Department of Defense, 112 Cong. 2 sess., July 24, 2012.
[60] Author’s interviews, Baghlan and Nangarhar, Fall 2012.
[61] ICG, “The Future of the Afghan Local Police”: 8.
[62] See Flynn, “Extreme Partnership in Afghanistan.”
[63] For a detailed description of their successes against the Taliban, and for a contrast with the problems that have plagued the ALP in the north, see Luke Mogelson, “Bad Guys vs. Worse Guys in Afghanistan,” New York Times, October 19, 2011.
[64] See, for example, Matthieu Aikins, “Our Man in Kandahar,” The Atlantic, November 2011; and Felbab-Brown, Aspiration and Ambivalence: Chapter 5.
[65] Author’s interviews with Kandahar elder, Kabul, summer 2013 and officers of U.S. special operations forces, Kabul, summer 2013.
[66] For other human rights abuses and acts of crime perpetrated by the ALP and other self-defense militias in other parts of Afghanistan, see Lynn Yoshikawa and Matt Pennington, “Afghan Local Police: When the Solution Becomes the Problem,” Foreign, October 27, 2011,
[67] ICG, “The Future of the ALP”: 8.
[68] Ibid.: 12.
[69] Interviews with ISAF officials, UNAMA officers, and local community representatives, Uruzgan, spring 2009, and with ISAF and U.S. Embassy officials and U.S. journalists, Kabul, Afghanistan, September 2010. For a more recent example of problems with ALP units in Uruzgan, see Rod Nordland, “Afghan Officer Sought in the Killing of 9 Colleagues,” New York Times, March 8, 2012.
[70] Alissa Rubin and Sangar Rahimi, “Afghan Officials Cite Revenge Killings in Latest Outbreak of Ethnic Hatred,” New York Times, August 3, 2012.
[71] For a nuanced discussion of how broader insecurity impinges on the determination of local communities to embrace the counterinsurgency effort and generate self-defense forces, see John Nagl, “A Better War in Afghanistan,” Statement to Senate Foreign Relations Committee, September 16, 2009,
[72] Feroz Sultani and Kay Johnson, “Stretched Afghan Army Falls Back on Militias to Help Defend Kunduz,” Reuters, May 3, 2015.
[73] Joseph Goldstein, “Police Force Is Studied for Ties to Taliban,” New York Times, February 9, 2015; and Gran Hewad, “Legal, Illegal: Militia Recruitment and (Failed) Disarmament in Kunduz,” Afghanistan Analyst Network, November 10, 2012,
[74] ICG, “The Future of the Afghan Local Police”: i.
[75] The emerging Islamic State in Afghanistan, so far located in three districts of Nangarhar, all major opium poppy areas, may be an exception. So far, it has not only prohibited the cultivation of poppy, but also refrained from taxing drug smugglers. See Shah Mahmoud Shinwari and Abubakar Siddique, “Pakistani Militants Lead IS Push Into Eastern Afghanistan,” Gandhara, June 30, 2016,
[76] For an overview of drug trends and policies in Afghanistan since 2001, see Vanda Felbab-Brown, “No Easy Exit: Drugs and Counternarcotics Policies in Afghanistan,” Improving Global Drug Policy: Comparative Perspectives and UNGASS 2016, The Brookings Institution, April 2015,–Afghanistan-final.pdf?la=en.
[77] Author’s interviews with members of U.S. special operations forces, Shindand, Herat, July 2013, and Herat government officials and powerbrokers, September 2014.
[78] Kimberly Heine, Cory Molzahn, and David Shirk, “Drug Violence in Mexico: Data and Analysis through 2014,” Justice in Mexico Project, April 2015,
[79] For background on violence in these cities and on the rise of criminal violence in Mexico, see Vanda Felbab-Brown, “The Violent Drug Market in Mexico and Lessons from Colombia,” Foreign Policy at Brookings, Policy Paper No. 12, March 2009,; Vanda Felbab-Brown, “Calderon’s Caldron: Lessons from Mexico’s Battle Against Organized Crime and Drug Trafficking in Tijuana, Ciudad Juarez, and Michoacán,” Latin America Initiative Paper Series, The Brookings Institution, September 2011,; Vanda Felbab-Brown, “Pena Nieto’s Piñata: The Promise and Pitfalls of Mexico’s New Security Policy,” Foreign Policy @ Brookings Paper Series, February 2013,; and Vanda Felbab-Brown, “Changing the Game or Dropping the Ball: Mexico’s Security Policy under Enrique Peña Nieto,” The Brookings Institution, Washington, DC, November 17, 2014,
[80] Nick Miroff, “Mass Kidnapping of Students in Iguala, Mexico, Brings Outrage and Protests,” Washington Post, October 11, 2014,
[81] “Many Questions in Mexico Cartel Battle That Killed 43,” Associated Press, May 23, 2015.
[82] For a further background on the history of self-defense forces and militias in Mexico and an analysis of their pros and cons in the current period, see International Crisis Group, “Justice at the Barrel of a Gun: Vigilante Militias in Mexico,” Latin America Briefing No. 29, Mexico City/Bogotá/Brussels, May 28, 2013, See also, George Grayson, “Threat Posed By Mounting Vigilantism In Mexico,” Strategic Studies Institute, September 2011; and Íñigo Guevara y Moyano, “Gendarmes, Rurales y Autodefensas,” El Excelsior, March 16, 2014.
[83] One notorious incidence of violence occurred in Michoacán’s capital, Morelia, on September 15, 2008, when a grenade was thrown into a crowd celebrating Mexico’s Independence Day. La Familia Michoacana was widely accused of the crime. It denied responsibility and accused the Zetas, which were later officially blamed for the incident.
[84] Francisco Gómez, “Piratería, el otro frente del narco,” El Universal, March 1, 2009.
[85] Author’s interviews in Michoacán, spring 2011. For a similar system of counterculture and of criminal groups providing not just employment, but also acquiring political capital and legitimacy and protection from local communities, see José Arturo Yañez Romero’s study of the Iztapalapa borough of Mexico City, where pirated and stolen goods are distributed – “Modelo para el Estudio de la Inseguridad, El Caso de Iztapalapa, Center for U.S.-Mexican Studies, University of California-San Diego, May 2005. For a similar study of the fluid legal, informal, and illegal markets in Guadalajara’s San Juan de Dios neighborhood, see José Carlos Aguiar, “Nuevos objetos en la agenda de seguridad pública: La ‘lucha contra la piratería’ en el Mercado se San Juan de Dios, Guadalajara,” in José Carlos Aguiar and María Eugenia Suárez de Garay, eds. Policía, seguridad y trasición política: Acercamiento al estado del México contemporáneo (Amsterdam: Centre for Latin American Studies and Documentation, 2008).
[86] Author’s interviews in Michoacán, spring 2011.
[87] Author’s interviews in Michoacán, March 2011. Regarding the iron ore extortion, see Dave Graham, “Chinese Ore Trade Fuels Port Clash with Mexican Drug Cartel,” Reuters, January 1, 2014,
[88] Cited in Eduardo Castillo, “Knights Templar Drug Cartel Counts Iron Ore as Main Income Source,” The Associated Press, March 17, 2014,
[89] Miguel García Tinoco, “Liberan a Militares Secuestrados en Michoacán,” El Excelsiór, March 1, 2013.
[90] Marguerite Cawley and James Bargent, “Following Arrest, Mexico Vigilantes Take Soldiers Hostage,” InsightCrime, March 13, 2013,
[91] James MacAuley, “Self-Defense Group’s Refusal to Disarm Increases Risk of Conflict with Military in Mexico’s Guerrero State,” IHS Global Insight, January 29, 2014.
[92] Heinle, Molzahn, and Shirk (April 2015): 17-8.
[93] Dudley Althaus, “Michoacán, Guerrero, and Mexico’s Mixed Responses to Vigilantes,” InSight Crime, February 13, 2014,
[94] Laurence Allan, “’Community Police’ Disarm Police Officers in Violence-Plagued Mexican State,” HIS Global Insight, May 10, 2013.
[95] Richard Fausset, “Mexico Under Siege: Guerrero State Sliding into Chaos,” Los Angeles Times, February 16, 2014.
[96] Georgina Olson, “Combatiremos defensa por propia mano: Enrique Peña Nieto,” Excélsior, April 10, 2013.
[97] Some scholars also embraced the formation of the militias – see, for example, Patricio Asfura-Heim and Ralph H. Espach, “The Rise of Mexico’s Self-Defense Forces,” Foreign Affairs, July/August 2013,
[98] Fausset, “Mexico Under Siege: Guerrero State Sliding into Chaos.”
[99] “Mexico to Draw Line on Vigilantes,” Associated Press, March 14, 2014.
[100] Fausset.
[101] For how many of the self-defense forces – rondas campesinas ­– created to combat the Shining Path ended up as important local drug trafficking entities, having often been formed out of the cocaleros to start with, see Vanda Felbab-Brown, Shooting Up, pp: 54-67.
[102] “Mexico to Draw Line on Vigilantes.”
[103] “Mexico Arrests 110 Posing as Vigilantes,” Agence France-Presse, April 22, 2014, and “Mexico Arrests 46 Criminals Posing as Vigilantes,” Associated Press, April 22, 2014.
[104] “Vigilantes Say Mexico Government Prosecuting Them,” Associated Press, March 16, 2014.
[105] Laura Castellanos Enviada, “Acuerdan disolución de las autodefensas en Michoacán,” El Universal, April 14, 2014; and Dalia Martínez, “Próximo, desarme de autodefensas: Castillo,” El Universal, April 4, 2014.
[106] Tracy Wilkinson, “Death Toll Rises in Mexico’s Roiling Michoacán State,” Los Angeles Times, January 12, 2015.
[107] Assassinations of politicians are not rare in Mexico, and hardly linked solely to organized crime or militias. Since 2008, at least 24 political candidates have been slain in Mexico.
[108] For a background on Guerrero and criminal violence there and its entanglements with the state’s political system, see Chris Kyle, “Violence and Insecurity in Guerrero,” Woodrow Wilson Center, January 2015,
[109] For details, see, Deborah Bonello, “Five Days of Terror in a Mexican Town,” Los Angeles Times, May 21, 2015; “13 People Missing in Southern Mexico after Vigilantes Withdrew From Violence-Wracked City,” Associated Press, May 19, 2015. For violence in Guerrero, see also Daniela Pastrana, “Drug Violence Leaves a String of Ghost Towns in Mexico,” InterPress Services, February 7, 2015.
[110] See, Felbab-Brown, “Dropping the Ball.”


Get daily updates from Brookings