Afghanistan Trip Report V: The Afghan Local Police: “It’s Local, So It Must Be Good” – Or Is It?

Vanda Felbab-Brown

Editor’s Note: The following article is one of a series of reports based on Vanda Felbab-Brown’s fieldwork in Afghanistan in April 2012. Here she analyzes the Afghan Local Police and other self-defense forces programs. Read also her recent report on the progress of Afghan security forces in “

Firefight in Kabul

,” on governance problems in Afghanistan in “

The Road to Jalalabad

,” on the U.S. troop withdrawal in “

The U.S.-Afghanistan Strategic Partnership

“, and on ISAF’s logistical challenges and the complex political realities in northern Afghanistan in “

Crossing the Salang


Among the most controversial aspects of the transition strategy in Afghanistan are various efforts to stand up self-defense forces around the country. These Afghan “militias” are supposed to increase security in areas where Afghan National Army (ANA), Afghan National Police (ANP), and ISAF presence are highly limited. With ISAF denying that the various programs amount to a militia effort (calling the units  everything else but militias and insisting that they are based on Afghan traditions, such as arbakai), the most visible version of these efforts right now is the Afghan Local Police (ALP). The ALP currently numbers around 13,000 members and is slated to increase to at least 30,000 by the end of 2014.

The purpose of the ALP is to extend at least a modicum of security to communities where Afghan National Security Forces (ANSF) are unlikely to be deployed for a long time. In those communities, the ALP is relied on to weaken the Taliban, Hezb-i-Islami, and the Haqqanis by either hiring their soldiers away for the ALP or having the ALP fight them, and to generate intelligence for ISAF.

U.S. military officials claim to be thrilled with the program. In conversations with me (during my recent research in Afghanistan in April 2012 and previous research trips)[1], they describe the program in glowing terms and portray it as tremendous success. They report that the ALP is enthusiastically embraced by local communities and effective in fighting the Taliban—often characterizing it as a “game changer.” As a U.S. Special Operations Forces officer explained to me: “All politics is local. The ALP’s local, so it must be good. After all, that’s what counterinsurgency theory teaches us.”

In contrast to the rosy U.S. military’s portrait of the ALP, many Afghans tend to fear the ALP and other self-defense forces programs and have negative views of such militia efforts. Their experience with militias and arbakai is of these forces turning on local communities, extorting them and predating on them, engaging in the theft of land and goods and even murder, and brutalizing rival ethnic communities. Occasionally, some communities have positive views of the ALP and other arbakai and even volunteer to stand them up. But often, the Afghans who most enthusiastically embrace such programs are the ALP commanders themselves, or the powerbrokers who try to sell their unofficial militias to the formal self-defense programs for hefty payoffs.

The Tenth Time Around, We’ll Get It Right

The ALP efforts, and other concurrent versions of the self-defense programs, are nothing new; raising self-defense forces or inducing various tribes to do one’s fighting has been repeated in Afghanistan’s history many times. The Soviets in the 1980s resorted to raising tribal militias when they realized that they were not winning in Afghanistan, and used the militias as part of their exit strategy. Indeed, many Afghans associate the current militia programs with the Soviets’ defeat and see it as yet another signal that the United States is preparing to leave without a stable order in place.

Since 2002, various versions of the self-defense programs have been tried to stabilize pesky or troubled villages — including the Afghan Auxiliary Police, the Afghan Public Protection Program, and the Community Defense Initiative, also known as Local Defense Initiative groups in some areas. In some of these efforts, the self-defense forces receive a salary. In others, they are not supposed to be paid; but many of them insist on some sort of payment, so the non-payment rule is often adjusted.

Not all of the militias have been stood up under the supervision and blessing of the U.S. military. The National Directorate of Security (NDS), the Afghan intelligence agency, has for years been standing up or encouraging 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. Some of these separate programs have been at times folded into the ALP, others persist outside of the control and even access of ISAF. Particularly under the leadership of Bismullah Mohammadi, a prominent Tajik commander, the notoriously corrupt and ethnically-factionalized Ministry of Interior (MoI) has also sought to legitimize and formalize its own favorite militias. The competition among these various institutional and ethnic factions raises concerns about the intensification of predatory behavior by the self-defense forces. Many other militias are simply self-generated, or have long fought for a local powerbroker, simply carrying their weapons less ostentatiously after 2002.

The performance and outcomes of these various militia programs has been highly varied, but often they are cumulatively negative, and problems abound. The experience with militias in Nangarhar illustrates some of the complexities of such efforts, and the limits to how easily they can be controlled.  In 2010, the Sepoy tribe of the Achin region of Nangarhar province spontaneously decided to raise an anti-Taliban militia. Enthusiastically embraced by U.S. military command in Afghanistan and promised financial sponsorship, the tribes were incorporated into the Community Defense Initiative. Quickly, however, the Sepoy militias turned on a rival tribe – the Alisherkhel. Using U.S.-provided weapons and claiming to have U.S. backing, they “reclaimed” land disputed between the two tribes and triggered violence in large portions of the province. After a series of negotiations and efforts to reduce hostilities, a delegation of NDS, ANP, ANA, and U.S. SoFs was sent in 2011 to the area to disarm the two tribes. In the fighting that ensued, eighteen Sepoy were killed, but still only the Alisherkhel agreed to disarm, incorrectly believing that the Sepoy would hand over their weapons too. The land dispute is not resolved, though a ceasefire was extended. The experience of Achin notwithstanding, the ALP is being stood up in Nangharhar: it is already in existence in the Goshta district, but plans are under way to stand it up also in Achin and other areas.

Experience with militias in northern Afghanistan reveals other problems and complications. Three year ago in the provinces of Kunduz and Baghlan, the U.S. military created the so-called pro-government militias (PGMs) (obviously running out of ideas for names and not getting the message up there in the northern Afghanistan that these programs are not militias). Later on, a decision was made to convert the PGMs into the ALP. Many were, but because of limits on the ALP size in the area, some PGMs were left out and told to hand over their weapons and go home. Instead, they went back to their checkpoints keeping their arms – no one has dared disarm them. They are still there, running out their former PGM salary, and, according to Baghlanis from the area, are getting angrier and angrier at being left out of the Afghan government-ISAF payroll. The Baghlan residents I spoke with overwhelmingly expected that these former PGMs would resort to highway robbery and extortion when their remaining money finally fully ran out.

Weak Control Mechanisms

When compared with the other self-defense programs, the ALP has by far the strongest oversight mechanisms, and the U.S. military officials are quick to note that the ALP program is far more sophisticated and far better than the Soviet militia program. Even so, the oversight mechanisms and controls are hardly sufficient.

The ALP is supervised and trained by U.S. Special Operations Forces who are to embed with the ALP in the village or area where the ALP operates. Embedding may imply a variety of things – from living in the village for six weeks to visiting the village once a week. Training mostly consists of teaching the recruits how to handle small firearms (which they either have and know already how to handle, or are issued), medical training, and communicating with the SoFs.

Those recruited are to be vouched for by three maliks (elders) and/or a village shura (council). The maliks or shura are to determine that the ALP recruits will not secretly work for the Taliban or other anti-government elements, turn on their U.S. advisors, and abuse the local community. This control mechanism is believed to be adequate since the maliks are assumed to know the men they are recommending. The problem with this elder control mechanism is that, not infrequently, a powerbroker controls the village elders, dictating his preferences in a way that escapes international scrutiny. At other times, the village elders have no problem vouching for the militia members as long as they only extort a rival village. Finally, outsiders, such as the SoFs, often have difficulties assessing how credible an elder a person is. If they live in a village for a period of time, they may well develop a good sense of the power distribution in the locality and level of acceptance purported elders have in the community.  But when they first arrive to identify the maliks who are then to identify the militiamen, any three older men with beards and turbans (or pakuls) may walk up to them and claim to be the three wise elders of the village.

The final control mechanism in the ALP program is that the district police chief is to supervise the ALP units. The problem with this mechanism is that the post of district police chief has often been associated with some of the greatest and most consistent corruption in Afghanistan. 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 a bridge too far. 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, more frequently, negotiated between Kabul and local powerbrokers, with the goal of satisfying their competing demands. Thus, the police chief’s quality of supervision of the ALP is often poor. In some areas, the problems with MoI’s supervision, such as desire to cleanse ethnic rivals from the ALP, has led the U.S. Special Operations Forces to operate via other programs, such as the PGMs to guarantee that the self-defense forces were in fact being paid and not manipulated by the MoI.

Furthermore, although the ALP likely is more closely monitored and perhaps controlled than other militia programs in Afghanistan, local Afghans rarely have the ability to distinguish among the various self-defense groups and tend to just call them all arbakai. Attributing abuses to a particular group may be especially difficult for the local population. The ALP is issued their own separate uniforms or at least arm bands, but as one Afghan civil society organizer in Nangharhar told me: “If they go to houses to demand money, they take off the uniform. They’re not stupid.” At other times, of course, various armed and criminal bands in Afghanistan do put on others’ uniforms to stop cars and extort money.

The greatest weakness in the ALP effort, and its many predecessors and concurrent programs, is that there are no established mechanisms for disarming an ALP unit that has gone rogue and predates on its own or rival communities. A formalized and diligently carried-out procedure could greatly reduce the ALP’s propensity toward abuse and increase the program’s legitimacy. When in Baghlan, for example, the ANP arrested twenty-two ALP militias for human rights violations and extortion, and they were sentenced to jail for six to seven years, the level of abuses by the ALP overall was perceived to have declined by the wider community. (The only problem was that the some members of the community believed that behind the guise of clamping down on abuses lay ethnic purging of the ALP in the area.)

A Great Variation in Outcomes: The Local Context Does Matter

The U.S. Special Operations Forces officer in charge of the ALP in an area who insisted to me that the ALP is all about local politics was of course correct. The problem is that the local context in Afghanistan is often very nasty. The structure, composition, history, and insider-outsider relations all significantly influence how well-behaved a local self-defense unit will be.

If a community is homogenous and isolated and subject to outside Taliban extortion and abuse, it may well enthusiastically welcome the creation of the ALP and the ALP may significantly improve security and the life of the community. Thus, in parts of Kandahar, for example, some communities volunteered for the ALP even before U.S. SOFs or the MoI arrived in the village to recruit the ALP.

Although such communities abused by Taliban outsiders sometimes could generate a force on their own to fight the Taliban, the benefit of the ALP structure is that it can relieve some of the logistical problems that an independently operating self-defense group may have — the ALP units are given small arms and ammunition. The logistical support is hardly perfect: When in Wardak, for example, the Afghan Population Protection forces, one of the ALP’s other versions, ran out of ammunition due to MoI’s lack of support, 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.”[2]

The ALP also receives not only supplies, but, critically, the back-up of U.S. Special Operations Forces during a firefight. Since in many parts of Afghanistan, tribal structures have been critically weakened and communities ravaged by war, their ability to fight to the Taliban can be very limited.  A SOF backup can be a life-saver for the self-defense forces if such groups come under overwhelming pressure from the Taliban. In Baghlan, for example, an ALP commander reported that the Taliban dared not attack them beyond minor harassment because the local Taliban groups knew that the ALP would call the SoFs if they came under serious Taliban attack. So the Taliban would only occasional plant IEDs or shoot at the ALP checkpoint. Of course, to the extent that the deterrent effect comes predominantly from the SOFs backup, the disturbing question arises as to what will happen with the improved security in the area and the state of the ALP when one day the U.S. Special Operations Forces are not available in the area as a back-up.

If a community is systematically disfranchised from power in an area — for example, Ghilzai Pashtuns do not have representation in the local district government and police and army forces, such as in Uruzgan — establishing ALP units in such a community empowers it. This empowerment is against the local Afghan government rather than the Taliban, but the community nonetheless is likely to greatly appreciate such a village-stabilization program. (The U.S. Special Operations Forces thus became the heroes of Baghlan Pashtuns groups from which they stood up ALP units, when during a firefight between the Andarabi-Tajik- dominated ANP in Baghlan and the Pashtun ALP units, they took the side of the ALP. The local Andarabi-Tajik dominated ALP had not been thrilled about how the ALP units were disturbing the Andarabis’ hold on power in the province.)

Under the best of circumstances, the ALP extends security against anti-government forces, such as the Taliban, to communities previously left to suffer, opens up roads to villages previously-deemed too dangerous to travel and hence boosts economic activity in the area, and perhaps even reduces local crime, extortion, and land theft.

Difficulties and complexities in many forms, however, tend to arise quickly when a community or an area is not homogenous, and when the Taliban or Hezb-i-Islami or other anti-government elements are not merely thuggish outsiders in the area.

Predation on host or neighboring communities can easily develop, serious abuses of human rights can take place, and the ALP may in fact undermine the security of the local community. In Kunduz or northern Farah, for example, after the ALP defeated the Taliban in their villages, they started extorting the communities and demanding taxes for themselves. Many instances of abuse by the ALP have been reported from Takhar where the ALP unit would identify its personal enemies as the Taliban for ISAF to get. Sometimes, the various ALP units, when drawn from rival communities or supported by rival warlords, turn on each other, instead of fighting the Taliban. One notorious case of such infighting took place in Uruzgan in 2010.

At other times, the negative effect on human security and perception of safety is more subtle. The establishment of an ALP unit may attract the Taliban to start attacking the unit and the community, expanding the insecurity bubble. The U.S. military considers it a sign of the ALP’s effectiveness that it draws the Taliban fire, but from the perspective of the local community, security may be considerably worse than before the creation of the local ALP outfit.

In very heterogeneous, polarized, and fractured communities, the establishment of ALP units frequently critically augments the security dilemma among the communities and triggers an armament spiral among them. Baghlan provides a prime example. The ALP units there have been drawn predominantly from the Pashtun minority in order to drain the swamp of the Taliban and Hezb-i-Islami. The Pashtuns in Baghlan have felt deeply marginalized and disfranchised after 2002, especially as many government positions, including in the ANP in the area, have been dominated by former Northern Alliance members. Many Pashtun communities there have considered their villages to be extorted by the Andarabi-Tajiks-dominated ANP and power structures. The Taliban and Hezb-i-Islami, as a result, easily mobilized support among the Pashtun communities, and establishing the ALP among the same groups was seen as mechanism to reduce the Taliban strength. But rival Tajik and Uzbek communities and powerbrokers have in turn felt extremely threatened by the arming of their rivals (often one and the same as the Taliban and Hezb-i-Islami) and began arming their own ethnically-based militias and lobbying the local MoI to recognize their militias as the ALP and not the Pashtuns. One such commander, who went by the nom-de-guerre Afghankush (which implies “Pashtun Killer”)[3] and had been renown for fighting “the Taliban” for several years before the ALP was stood up and for keeping the Baghlan-Kunduz road flowing, demanded that his Uzbek-militia be recognized as part of the ALP. When he was rejected, he simply announced that his men already were the ALP and were just not yet being paid. Several other warlords in the north adopted that approach, often with the backing of their governors and chiefs of police. Many Tajik and Uzbek communities and powerbrokers in the north have felt betrayed and treated unfairly as a result of the ALP program because they believed that the ALP program rewarded Taliban-supporting villages, whereas those who have been fighting the Taliban for two decades were being sidelined.

In other words, although the effects of establishing the ALP are highly contingent on local contexts, cumulatively the ALP phenomenon transcends the local context and can easily set off a widespread security dilemma within Afghanistan. Even though the ALP are physically not to travel and operate outside of their villages (of course, they violate the rule), their reputation travels among communities. What happens in one Afghan village does not stay in that Afghan village. Instead, rival communities, observing that their antagonists are being armed, seek to do the same.

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. The local ALP reaches a modus vivendi with the Taliban and the Taliban reduces its attacks. Indeed, such hedging is typical of Afghan history, with local warlords, khans, and tribes siding and making peace with those they sense would prevail in a conflict, and easily breaking deals if the situation on the battlefield changes. Sometimes, such as in Logar in 2010, the accommodation between the militias and the Taliban even results in temporary improvements in security in the area and the community welcomes it. But the reduction in violence often exists only at the mercy of the Taliban and the deal collapses when the Taliban chooses to renege on it or when the external payoffs dry up and can no longer be divided among the various warring parties.

In some cases, such as in Kunduz and Baghlan, creating the ALP merely means putting the Taliban and Hezb-i-Islami on the ISAF/Afghan government payroll and paying them not to fight. Buying off one’s enemies is of course another time-honored tradition in Afghanistan, and at times groups align themselves either on the basis of assessing on which side of the conflict there is more money to be made, rather than on the basis of any hardened ideological preferences or deep-seated communal rifts.  But paying your enemies not to fight you can turn out to be a short-lasting solution. What happens if the ALP money suddenly dries up? An ALP commander (“former” Hezb-i-Islamic leader) answered the question for me: “It’s all about money. Now the Afghan government pays us. If the opposition starts paying us more, we switch to them.”

Ominous Future

Herein lies one of the biggest problems with the ALP. The U.S. military readily agrees that the ALP is a “temporary” solution. No one, however, knows as yet, how and when it will be retired. What happens after 2014 when even regular Afghan forces of the ANA and ANP are to be reduced by as much as 130,000 because the Afghan state will not be able to afford to pay for the current size of the ANSF? What will happen if a NDS or ANP very strongly dominated by one ethnic group after 2014 decides not to pay the ALP of rival ethnic groups? If the ALP officially persists, will Afghan Special Operations Forces be able to exercise even the current inadequate level of supervision over the ALP? How likely are such units to disarm if they told to so after 2014? If the ALP units are simply told to go home after 2014, they may well offer themselves up to the highest bidder, the warring party that is most likely to prevail in an area, or turn to crime.

Under even the best circumstances — if after 2014 the ANA and ANP can effectively take over security in Afghanistan and prevent the country from disintegrating into an ethnically-factionalized civil war of which the Taliban will be one of many factions — the (former) ALP will still represent a huge challenge for improving governance. Its predatory tendencies will be hard to control and in very polarized communities, the presence of (former) ALP may be the trigger of local conflict. If a source of local crime, the ALP will undermine perceptions of public safety and rule of law more broadly and legitimize actors such as the Taliban who deliver “order and justice.” Thus even when the program is performing its function of beating down the Taliban in a particular locale at a particular time, it still leaves behind armed men who can and often do challenge the already weak central government and engage in predation of local communities.

Under the worst circumstances, the (former) ALP militias will be one of the actors in a post-2014 civil war. Already many of the prominent former commanders are digging up their weapons and resurrecting their former militias. The ALP will be just one of many warring “self-defense” groups. If one takes this dire view of Afghanistan as its most likely future, one can see several reasons to continue and even augment the ALP effort. One is that yes, the militias are not a good development, but they are happening anyway, with or without ISAF’s and Afghan government’s sponsorship, so why not get into the spirit of the time and exploit them to U.S. advantage? Especially, if the civil war is believed to be coming, the concerns about ALP’s negative effects on governance and security in Afghanistan can be discarded, and instead the ALP can be seen to maximize the number and strength of groups that may oppose the Taliban and hence reduce the Taliban’s post-2014 power and territorial control. Putting many of the Taliban on the payroll may also extend the time before a civil war arrives after 2014.

Yet if one believes that a civil war is not unavoidable, that there is still a chance to stabilize Afghanistan sufficiently to avert a full-blown civil war and prevent extensive Taliban territorial control, then eagerly listening to the ALP siren song is not the way to go. Whatever limited tactical gains the ALP may bring are likely to be offset by the long-term conflict triggers it also carries within itself and the long-term negative impact on the already poor quality of governance in Afghanistan. And without improving governance in Afghanistan, it is hard to imagine how the current dispensation in Afghanistan can be sustained after 2014 even with improvements in the ANSF. Thus, if one still believes in a reasonably stable Afghanistan after 2014, then building up the ALP further is not the way to go. Instead, credible and robust mechanisms should develop right away to roll back rogue ALP units already in existence. Now is also the time to start developing a serious program of how to disarm and demobilize the ALP at the end of 2014. The United States and the international community should commit themselves to carry out that disarmament and to establish a program to divert the former ALP from predation. Such a program will likely only be credible if other militias, whether under the aegis of the United States or belonging to Afghan warlords, are also incorporated in it. For if they are not, the security dilemma will be triggered by them, just as it is currently triggered by the ALP, and the ALP units may shed their uniforms but not put down their weapons. The United States will do Afghanistan a great disservice if it rushes to stand up as many ALP units as possible before the end of 2014 and then hands them over to the Afghan government to worry about.


[1] During my recent and previous research in Afghanistan, I conducted interviews about the ALP and other self-defense programs in Afghanistan with ISAF and U.S. military officials at different levels of the chain of command in Kabul and in various regional commands, Afghan government officials at all levels of the government, Afghan Police officers, members of the Afghan Local Police, maliks, Afghan civil society organizers, businessmen, as well as ordinary Afghans such as shopkeepers. All interviews were conducted under the no-attribution rule.

I would like to thank Philipp Rotmann of the Global Public Policy Institute for his invaluable input into this article.

[2] David Axe, “Fourth Time the Charm for NATO’s Afghan Militia Plan?” World Politics Review, July 21, 2010. On the positive side, the Wardak APPP militias showed considerable restraint in getting involved in local tribal disputes, such as between the Hazaras and the Kuchis over grazing lands, and stayed out.

[3] Afghankush was assassinated by a suicide bomber in 2011.