It has been argued that the deployment of gendarmerie forces can help establish public order and security in international interventions. However, little is known empirically about gendarmeries in multinational operations. This article examines the Italian Carabinieri in post-war Bosnia-Herzegovina, post-war Kosovo, and war-time Afghanistan. It shows that in these missions, the Carabinieri were able and willing to carry out a variety of activities, including crowd management, arrests, and the training of host state police, and that the gendarmeries flexibly adapted to the respective mission context. However, the article also points at limitations of gendarmerie operations and identifies knowledge gaps relating to the activities and the effects of such operations. The debate on, and practice of, militarized police intervention in war-torn countries would benefit from the systematic collection of data and information. Unfortunately, such collection faces significant challenges.
* Dr. habil. Cornelius Friesendorf is a senior researcher at the Institute for Peace Research and Security Policy (IFSH), University of Hamburg. Previous positions include working as a senior advisor for an EU police reform support project in Myanmar and teaching International Relations at Goethe University Frankfurt. His work examines security sector reform, counter-insurgency, organized crime, and privatized security. Publications include How Western Soldiers Fight: Organizational Routines in Multinational Missions (Cambridge: Cambridge University Press, 2018).
Many war-torn countries with an international security presence experience public security gaps.2 International civilian police take a long time to deploy and may not have the means to deal with well-organized and well-armed spoilers of peace. International military forces are trained and equipped for high-intensity combat and therefore hard-pressed to carry out tasks such as crowd management. Local armed forces are often predatory. Public security gaps can have disastrous consequences. For example, in Iraq the US failed to plan for public security following the invasion of the country in 2003, and the ensuing chaos fueled the insurgency.3
One way of preventing or filling public security gaps is to deploy in-between forces able to carry out tasks for which neither combat troops nor civilian police services are suitable. Some authors have praised gendarmeries for their versatility and pleaded for their increased deployment.4 In their Mediterranean variant, gendarmeries are police forces with military status. The deployment of gendarmeries as part of multinational operations has mostly occurred in the context of NATO and European Union (EU) missions. But the trend to deploy militarized police forces is visible in other contexts, too. By 2011, half of the police officers deployed in United Nations (UN) peacekeeping operations were Formed Police Units (FPUs), i.e. cohesive units used primarily for the establishment of public security.5
While there are rife debates about public order problems and ‘policekeeping’,6 little is known empirically about gendarmeries in multinational missions. This article contributes to a better understanding of this phenomenon, by discussing gendarmerie activities and identifying knowledge gaps that need to be addressed before calling the deployment of gendarmeries necessary or, indeed, successful. In doing so, the article also contributes to bridging academic fault lines. Military sociologists tend to study regular military forces,7 not gendarmeries. Police scholars have expanded their traditional focus on domestic policing, now also studying police in peace operations.8 However, authors examine civilian police rather than gendarmeries.9
The focus here is on the Italian Arma dei Carabinieri, which has provided niche capabilities in between regular military forces and civilian police agencies in numerous post-Cold War foreign missions. The Carabinieri were chosen because they are a highly versatile organization and can therefore be expected to make valuable contributions to multinational missions, and because they have played a more prominent role abroad than other gendarmerie forces such as the French Gendarmerie Nationale of the Spanish Guardia Civil. The article discusses three missions: post-war Bosnia-Herzegovina (in the following referred to as Bosnia), post-war Kosovo, and war-time Afghanistan.
Focusing on the operational and tactical levels, the analysis reveals that in these missions, the Carabinieri carried out a variety of functions and that they tailored levels of militarization to the level and types of threats in mission areas. It also reveals knowledge gaps regarding the type, number, and sites of operations. If data and information on output is sketchy, this is even more the case with outcomes, such as the number of persons arrested or weapons confiscated, although available evidence suggests that some outcomes are underwhelming. If assessing output and outcome is difficult, we know even less about impact, in particular the extent to which the Carabinieri improved the safety of local residents.
The first part of the article discusses police interventions in general and gendarmeries in particular. The subsequent sections examine, respectively, Bosnia, Kosovo, and Afghanistan, discussing Carabinieri activities and, tentatively, their effects. A conclusion suggests inroads for future research on gendarmerie operations and points at methodological and practical challenges of collecting data and information.
The analysis is based on secondary literature, official documents, and media reports. Moreover, the account relies on personal observation as well as informal background talks and formal meetings with Carabinieri and representatives of other agencies in the Balkans (between 2006 and 2008) and in Afghanistan (between 2009 and 2013).10
Militarized Police Interventions and Gendarmeries
Many contemporary wars feature a variety of state and non-state armed groups and large-scale human rights violations.11 International security forces deployed to such conflict zones face a variety of problems. Depending on their mandates, they may have to secure ceasefires, disarm former combatants, secure elections, arrest criminals, and contain riots. These tasks do not clearly fall into the purview of military combat forces or civilian police services.
Attempts to close resulting public security gaps may blur military and police roles.12 Soldiers are policized, by engaging in crowd management or handling criminal evidence. Police are militarized, not only in war-torn and authoritarian countries but also in stable democracies. Most countries have Special Weapons and Tactics (SWAT) teams for dealing with bank robbers or hostage takers, as well as cohesive units for crowd management.13
Peter Kraska regards police as militarized (or, to use his term, paramilitary) when police have specific material, cultural, organizational, and operational characteristics.14 Operationally, militarized police agencies and units carry out tasks requiring ‘robust’ capabilities beyond the means available to individual civilian police officers, such as SWAT missions or crowd management. In terms of organizational indicators, militarized police operate in larger units than civilian police (who often patrol in pairs of two), and under centralized command structures. Material indicators of militarized policing include weapons such as automatic rifles, as well as body armor. In terms of culture, militarized police use martial language and hold beliefs akin to those of soldiers.
Applying these characteristics to specific police units, a SWAT team, for example, is militarized because it carries out tasks likely to involve the use of force, comprises significant number of officers with military-grade weapons, and uses military-style language and has a military-style look. By contrast, a Criminal Investigative Department (CID) has less militarized characteristics.
The public and scholarly debate on militarized policing is rife. Kraska points at the negative consequences of this trend in the US, showing that militarized police fighting the ‘war on drugs’ often use disproportionate force. Other scholars second his observation about the negative effects of police militarization.15 However, some observers point at the advantages of making police more robust. P.A. Jim Waddington argues that specially trained and equipped crowd management units are less likely to use disproportionate force than civilian police who, in case of riots, must improvise.16
Militarized policing is not only a domestic but also an international and transnational phenomenon. Foreign actors are involved in militarized policing in missions where international actors have executive law enforcement authority, such as in ‘international protectorates’ (Kosovo and Timor Leste are the primary cases), as well as in UN peacekeeping operations that include FPUs. More typically, foreign agencies assist militarized police units of host states, through the delivery of training and equipment.
Security gaps in Haiti, Bosnia, Kosovo, Iraq, and elsewhere led observers to call for militarized policing capabilities as part of internationally-supported stabilization efforts.17 Other authors point at the limitations and risks of international engagement in militarized policing.18 Citizens who suffered from paramilitary forces during war may fear militarized police after war. Also, international efforts to reform the security sector of fragile states include efforts to demilitarize the police. The Organization for Economic Cooperation and Development (OECD) points at the ‘sense of elitism’ marking militarized police forces.19 To the extent that international actors deploy militarized police or train and equip such host state forces, they risk sending confusing signals to host states and societies and to undermine democratization efforts.
Militarized policing is thus not a panacea to stabilization. Alice Hills argues that militarized police forces may carry out niche functions in international missions but that they also have limitations, e.g. when facing organized military resistance.20 David Bayley and Robert Perito plead for a division of labor: police organizations should include both civilian and robust units.21 The OECD, although critical of militarized policing, acknowledges that militarized police might deliver security in high-threat environments.22 Indeed, research suggests that when violence is rife, the primary concern of local residents is security, whoever may offer it – even brutal and corrupt police may find public support if they offer security.23
Hence, there is no agreement over how police militarization can contribute to stabilization and the protection of people in conflict and post-conflict zones. For David Bayley, democratic police agencies are responsive to the public, subject to the rule of law, protective of human rights, and transparent.24 Police organizations that include militarized units do not necessarily violate these principles as long as effective oversight institutions exist.25 Without proper oversight, any armed force can threaten human rights, as Western security assistance to unaccountable and repressive forces during and after the Cold War shows.26
In fact, more democracy may mean more militarized policing. Democracies are wary about deploying regular military forces for domestic law enforcement, given the risk of disproportionate force. Instead, they create SWAT teams or crowd management units. In Northern Ireland in the mid-1970s, the Royal Ulster Constabulary (RUC) replaced the British Army as the primary guarantor of public security. West Germany created an elite police force following a hostage situation during the 1972 summer Olympic games. The United States has seen a proliferation of militarized police units since the 1980s.27
The Italian Carabinieri
The Italian Arma dei Carabinieri is a force that can cover almost the entire spectrum between civilian policing and military combat. This includes carrying out tasks for which neither military combat soldiers nor individual civilian police are the best choice. Applying Kraska’s characteristics to the case of the Carabinieri highlights the hybrid nature of this organization.
In terms of their operational characteristics, the Carabinieri carry out a broad range of functions. These include typical civilian police activities such as patrolling, regulating traffic, and investigating crime. Another routine activity, though involving more ‘robust’ units, is crowd management. Towards the military end of the spectrum, the Carabinieri comprise SWAT teams and military support units, in particular military police and the paratroopers of the 1st Carabinieri regiment Tuscania. Although versatile, the organization cannot remedy all situations; thus, few Carabinieri (the paratroopers) can engage in high-intensity combat against conventional military enemies.
The Carabinieri’s hybrid character has been forged over the last 200 years. The force was created in 1814 and based on the model of France’s Gendarmerie Nationale, in line with a trend in Europe at the time to create centralized and militarized police forces. All major Italian military missions, at home and abroad, saw the involvement of Carabinieri.28 Examples are the Crimean War, military action during the unification of Italy, and the occupation of Abyssinia and Yugoslavia under Mussolini. During the Cold War, the Carabinieri focused on domestic security tasks, some of which resembled low-intensity warfare. This was the case with counter-crime operations in Italy’s restive South as well as the fight against separatists in South Tyrolia.29 The Carabinieri also routinely carried out crowd management, often using heavy-handed means that exceeded those used by police in other West European states.30 After the Cold War, the Carabinieri participated in a variety of multinational missions, especially in the Balkans, due to the region’s proximity to Italy and links between organized crime groups in the Balkans and in Italy. The Carabinieri were also active further away, especially in Iraq and Afghanistan.31
With regard to its organizational characteristics, the Carabinieri is a police organization with military status, constituting the fourth branch of the Italian armed forces. The force covers the entire Italian territory and is under the central command of a headquarters in Rome; its members have military ranks. The defense ministry and interior ministry share control over the Carabinieri, with the former offering training in military skills and the latter in policing skills. Its hybrid legal status is one of the Carabinieri’s main assets in multinational missions: it allows placing the Carabinieri under either civilian or military command, thus increasing the flexibility of strategic and operational planning.
The Carabinieri also have militarized material characteristics.32 Gear includes not only handcuffs, batons, and side arms; units providing military support also have access to machine guns, rocket-propelled grenades, armored vehicles, and helicopters, and command and control is centralized. The appearance of Carabinieri carrying out SWAT or infantry support missions is martial; this is less the case for units carrying out civilian policing duties, such as criminal investigators who wear plain cloths. Crowd management units, equipped with batons and shields, are in between.
Culturally, this organization comprises a variety of subcultures. A crime forensics expert of the force is likely to employ different professional language and to hold different beliefs than a member of a SWAT team, a paratrooper, or a military police officer. But intra-organizational cultural diversity notwithstanding, all Carabinieri share an esprit de corps.
Believing in its professionalism, the Carabinieri organization is protective of its autonomy over operations and tactics. In fact, even Mussolini struggled to completely control the force.33 In its post-Cold War foreign missions, the Carabinieri were largely able to avoid political ‘micromanagement’, also because Italian politicians as well as the public pay little attention to foreign and security policy.34 The Italian government tends to content itself with authorizing the deployment of Carabinieri abroad, and thus to placate the demand of allies, in particular the US, for robust policing capabilities. Carabinieri missions have been politically expedient because while being inexpensive they enhanced Italy’s international standing.35
Hence, the Carabinieri can be classified as a militarized police force, in terms of their operational, organizational, material, and cultural characteristics. The following sections show that these characteristics, and the ability to adapt to the local context, informed the Carabinieri’s operations and tactics in Bosnia, Kosovo, and Afghanistan. The analysis focuses on the most important activities that the Carabinieri carried out in these missions. In the Balkans, these were crowd management, the arrest of suspect criminals, operations against organized crime, and the training of host state law enforcement agencies. In Afghanistan, the main Carabinieri activity was preparing Afghan special police for supporting counter-insurgency. Thus, given the different types of missions (peace support operations in the Balkans and counter-insurgency in Afghanistan), the aim is not to compare the same types of activities across mission areas but to show that the Carabinieri adapted to the respective mission context.
The Dayton agreement of 14 December 1995 formally ended the Bosnian war. It also created a multinational military mission (the Implementation Force [IFOR], which in late 1996 was replaced by the Stabilization Force [SFOR]) and a civilian police mission, the International Police Task Force (IPTF). IFOR/SFOR were placed under the command of NATO and based on a robust Chapter VII mandate of the United Nations (UN) Security Council.36 The IPTF was a UN mission, and its officers were unarmed and did not have law enforcement powers.
The Dayton agreement left tasks undefined that fell into military-police grey zones. IFOR, and during its early rotations also SFOR, were unwilling to arrest suspect war criminals wanted by the International Criminal Tribunal for the Former Yugoslavia (ICTY), to dismantle illegal check points, to disarm paramilitary Bosnian police forces (especially in the Republika Srpska), and to stop rioters. Regular military soldiers were generally not trained for such tasks, and sending states worried about own casualties.37 The IPTF had neither the mandate nor the capacity to deal with war criminals and battle-hardened former fighters. With spoilers of peace moving around freely and with displaced persons trying to return to their former homes facing threats, post-war Bosnia experienced a public security gap.38
Over time, it became evident that public insecurity prevented sustainable peace. Troop-contributing states understood that they had to do more to help implement civilian aspects of the Dayton agreement such as the return of displaced persons. To further this objective, states started looking for forces with niche capabilities. This was especially the case for the US government, given Pentagon concerns of involving US soldiers in police-like tasks.39 The US government thus pushed for the creation of a Multinational Specialized Unit (MSU), which would operate under SFOR command across Bosnia and undertake tasks such as crowd management.40
The MSU became operational in mid-1998, with most of its 300 gendarmerie officers from the Carabinieri, who also developed the unit’s doctrine.41 Italy was keen to police the Balkans, given the negative impact on Italian security of Balkan smuggling and trafficking in people and goods.
However, most SFOR commanders were not familiar with gendarmeries and therefore primarily employed the Carabinieri as a reserve force in case of riots (instead of utilizing them for preventing riots such as by talking to demonstrators or by observing situations in plain clothes), or for tasks such as static guard duty.42 On several occasions, riots broke out that might have been prevented had SFOR leaders employed the Carabinieri more systematically.43 Also, the frequent rotation of SFOR commanders and units meant that the Carabinieri had to explain the role of gendarmeries over and over again.
Over time, SFOR learned to better utilize the force and also developed doctrine that foresaw that regular SFOR troops would secure an outer area (the so-called green box) and the MSU operate in the inner area (the so-called blue box). To be sure, problems did not go away. For instance, on 6 April 2001, Bosnian Croat mobs wounded eleven Carabinieri in Herzegovina when UN auditors tried to confiscate documents from branches of a bank used by Bosnian Croat hardliners for illicit transactions. French SFOR troops had failed to secure the green box in which the Carabinieri were operating, and SFOR had deployed too few Carabinieri.44 This disaster was a catalyst for a better utilization of the Carabinieri. Shortly after, in July 2001 the MSU protected a commemoration of the 1995 massacre in Srebrenica, with SFOR crediting the force for deterring violence.45
In addition to these public order functions, the Carabinieri were active in another field: the arrest of suspect criminals. Arrests would generally be made by SWAT-units of the Carabinieri that used military command structures, military-grade weapons, and body armor. Some of the persons arrested included persons suspected of being involved in organized crime.46 The MSU also helped to arrest suspect war criminals. For example, in spring 2004, SFOR, with MSU participation, tried to arrest Radovan Karadžić in Pale.47 Moreover, the Carabinieri, after 9/11, supported the US ‘war on terror’ in Bosnia. On 25 September 2001, Carabinieri seized two Bosnians working for a Saudi charity organization in Sarajevo and subsequently questioned one of the suspects before transferring him to the main US military base near Tuzla.48 Over the following months, the MSU participated in more operations that lead to the arrest of approximately 20 individuals suspected of planning attacks on US sites in Bosnia.49
In 2004, a European Union military mission (EUFOR Althea) replaced SFOR, and the MSU became an Integrated Police Unit (IPU). The remit of the unit remained the same, and Italy became the lead nation of the IPU. The new mission comprised two mobile elements.50 The first were platoons undertaking patrolling, crowd and riot control, and surveillance. The second mobile element comprised specialized units put together on a need-basis, such as SWAT teams. Regular EUFOR troops backed up the IPU through means including helicopters.
A main task of the IPU SWAT teams was the arrest of suspect war criminals. The IPU’s proactivity earned them the praise from observers who criticized the passive stance of regular military forces.51 But the Carabinieri’s most important function was to train Bosnian police forces in skills including crowd management, arrests, and surveillance. The main partner of the Carabinieri was the State Investigation and Protection Agency (SIPA), the only centralized police force in an otherwise highly decentralized state.52 The Carabinieri supported SWAT teams within SIPA that were tasked, among other functions, with arresting suspect war criminals.53
Effects of Carabinieri Activities
The few accessible works that assess the Carabinieri’s role in Bosnia paint a positive picture. A Pentagon report states that the MSU resolved 261 out of 263 situations ‘without the use of force through a combination of deterrence, dissuasion, and negotiation’, yet without providing further evidence on specific operations.54 One could also argue that the Carabinieri contributed to security in Bosnia by supporting the transfer of wanted persons to the ICTY and that they enabled Bosnian police to undertake tasks such as the arrest of human traffickers.55 Not least because of Carabinieri training, SIPA, during the first half of 2008, was able to arrest around 80 persons, including suspect war criminals.56
However, claiming that the Carabinieri made major contributions to security in Bosnia would be problematic. The exact role of the Carabinieri in many operations is shrouded in secrecy. It is unclear, for example, to what extent the Carabinieri themselves arrested suspect war criminals or only provided back-up to other forces; how many persons were arrested in operations that included the Carabinieri; and how they used force during these operations.
Moreover, various factors stymied the effectiveness of the Carabinieri. These include the frequent failure of SFOR commanders to understand the role of gendarmerie forces, as well as the fragmented policing structure of Bosnia, which undermined cooperation among police agencies both within and across the entities.
Also, some of the outcomes of Carabinieri operations were underwhelming. For instance, in 2005, the Carabinieri-led IPU documented that it had seized, among other goods, ‘12.71 cubic metres of wood without proper documentation’, a narcotising spray, ten grams of marihuana, a toy gun and small amounts of smuggled cigarettes.57 Such seizures may reflect a lack of actionable intelligence on the part of the Carabinieri or the fact that Bosnia was not a haven of organized crime.58
Last, claims of success gloss over the fact that deploying gendarmeries can have negative unintended consequences in the absence of strong accountability and oversight mechanisms. The Carabinieri’s involvement in the US ‘war on terror’, for example, supported US policies that were problematic from a human rights point of view. One of the persons arrested by the Carabinieri in the above-mentioned 2001 operation in Sarajevo was reportedly tortured by US personnel after his arrest.59
During the second half of the 1990s, the situation in the Serbian province of Kosovo deteriorated. Yugoslav troops and irregular forces committed violence against ethnic Albanians (Kosovars), and the rebel Kosovo Liberation Army (KLA) fought back. International actors tried to broker a peace deal but failed. On 24 March 1999, NATO started to bomb Yugoslavia, to compel Yugoslav president Slobodan Milosevic to end violence committed against Kosovars. After 78 days of war, Milosevic agreed to negotiate a peace agreement. The ‘Kumanovo’ agreement of 9 June foresaw a timeframe for the withdrawal of Yugoslav forces from Kosovo and the disarmament of the KLA. In parallel, the UN Security Council agreed on a governance structure for Kosovo that effectively created an international protectorate.60 An international military presence was to ensure compliance with the peace agreement; this Kosovo Force (KFOR) was placed under NATO command. On the civilian side, the United Nations Interim Administration Mission in Kosovo (UNMIK) became responsible for policing the province.
On 12 June, troops of the NATO-led Kosovo Force (KFOR) entered Kosovo. In contrast to IFOR and SFOR in Bosnia, KFOR could not abrogate its responsibility for establishing public order. After all, UNSC Resolution 1244 mandated KFOR to ensure ‘public safety and order until the international civil presence can take responsibility for this task’.61 But KFOR was hard-pressed to undertake public order functions since it consisted primarily of combat troops. Regarding civilian agencies, it took the UN a long time to deploy police officers to Kosovo.62 Yugoslav police forces had withdrawn from Kosovo, and the KLA was ill-suited for policing the province, given its ethnic bias. Hence, a public security gap opened up, threatening Serbs and other minorities in a situation where the tables had turned.
Some of the Carabinieri deployed to Kosovo served as military police to regular Italian troops in the West of Kosovo. Carabinieri also participated in the UNMIK Police mission. However, the organization’s main role was in KFOR: the Carabinieri led a Multinational Specialized Unit (MSU), which was modeled on the MSU in Bosnia after a US request to deploy gendarmerie forces to Kosovo.63 Most of the around 300 MSU members in Kosovo were Carabinieri. The MSU was directly placed under the commander of KFOR (COM KFOR) and had a broad mandate: it was to patrol anywhere in Kosovo, maintain public order, gather information, and support counter-terrorism.64
As in Bosnia, the Carabinieri in Kosovo felt they possessed the necessary skills for operating in civil-military grey areas. For example, regular US troops in Multinational Brigade East struggled to carry out police-like functions, and there were too few US Military Police officers. The MSU deployed 15–20 members to that sector, assisting with crowd management, the collection of information on drugs and weapons smuggling, and the detention of smugglers.65
Carabinieri also protected minorities. For example, as soon as the MSU became operational, the Carabinieri shielded Serb Orthodox monasteries from attack by radical ethnic Albanians.66 Several years later, on 17–18 March 2004, mayhem broke out across Kosovo: riots triggered by claims that Serbs had caused the deaths of two Kosovar children left 19 people dead, over 900 injured, over 700 homes of Serbs, Ashkali, and Roma damaged or destroyed, over 30 Serbian churches and monasteries damaged and destroyed, and around 4,500 persons displaced.67 KFOR (as well as UNMIK) first failed to prevent the riots and then to stop them, due to weak (a) intelligence, (b) mobility across sectors, and (c) crowd management capacities, as well as the timidity of some national units (such as the Germans in Prizren).68
The MSU’s role during the riots was limited because the force was relatively small and could not be everywhere at the same time. Also, KFOR did not issue adequate orders to the MSU, given deficient coordination between KFOR and UNMIK as well as deficient KFOR command and control.69 But within these limitations, the Carabinieri prevented worse from happening at the sites where they were present. Around 15 Carabinieri, together with around 30 regular Italian soldiers, protected the monastery of Deçani in the West of Kosovo, which is a main site of Serb Orthodoxy. The Italians placed armored vehicles in front of the monastery and fired warning shots to deter a 500-strong crowd throwing incendiary devices from overrunning the monastery.70 At other trouble spots in Western Kosovo, Carabinieri serving under the Italian army parachute regiment Folgore evacuated Serb civilians by helicopter. Italian troops evacuated 32 Serbs in Bijelo Polje and 56 Serbs in Bica. In Djakova, Carabinieri rescued an Italian police officer from a Kosovar mob, dispersing the crowd through warning shots.71 In Grabac, Italian soldiers forced Serbs unwilling to abandon their houses to leave in the face of a mob around 2,000 strong.72 After the riots, NATO requested more Carabinieri, leading Italy to deploy 80 members of the force from Bosnia to Kosovo.73
Over the following years, the Carabinieri remained active. In 2007 and 2008, for example, members of the force patrolled the streets of the capital city of Pristina. These patrols showed how Carabinieri adapted their characteristics to civilian policing when threat levels were low: officers patrolled in small units, wore non-camouflage uniform and soft hats, carried their weapons in a non-threatening manner, and interacted with local residents as much as possible.74
Counter-crime activities were another area where the Carabinieri felt they had a competitive edge. Early on during the KFOR mission, the MSU confiscated drugs and weapons, and also raided brothels where victims of sex trafficking were thought to be held.75 One Deputy COM KFOR frequently requested the MSU to observe or arrest suspect war criminals and persons involved in organized crime.76 In February 2003, the MSU assisted British special military forces when these arrested three Kosovars who the ICTY had indicted for war crimes.77 Several years later, the Carabinieri supported a British-led Intelligence, Surveillance, and Reconnaissance (ISR) unit by providing human intelligence. The unit was to locate and enable the arrest of persons suspected of serious offenses such as organized crime; the arrests were made by KFOR, UNMIK Police, and as of spring 2008 the European Union Rule of Law Mission EULEX, which succeeded UNMIK Police.78 Low-level offenders, too, became Carabinieri targets. Members of the force frequently raided private homes to search for illegally held weapons. They also set up vehicle checkpoints to search for contraband and wanted persons. At one checkpoint between Pristina and Gracanica in late 2008, the Carabinieri, over a period of 1,5 hours, stopped a dozen cars.79 Depending on the type of task and the risks involved, the Carabinieri adapted their characteristics, such as by changing the size of units and their weaponry.
The Carabinieri also threw their weight behind the Kosovo Police. Sometimes, the MSU would hand over detained persons to Kosovo police officers so that these would make the arrest.80 At other times, the MSU would provide back-up to SWAT teams of the Kosovo police (and of UNMIK police).81 The Carabinieri also trained the Kosovo police in skills including crowd management, weapons handling, search, VIP protection, and the gathering of criminal intelligence.82
Effects of Carabinieri Activities
However, as in the case of Bosnia, a lack of data makes it difficult to evaluate the performance of the Carabinieri. Little is publicly known, for example, about their involvement in crowd management. The reports cited above describing the performance of Italian troops during the March 2004 riots do not clearly distinguish between regular troops and the Carabinieri. As a consequence, the counterfactual claim that without the presence of the Carabinieri more Serb sites would have been destroyed and more Serbs harmed by rioters must be made cautiously.
We charged one thousand persons with violations, of which five hundred were arrested [by UNMIK police]. We seized thirty-five thousand items, including mortars, bullets, grenades, and other explosives. We seized over a hundred tons of contraband cigarettes. We freed fifty young women who had been forced into prostitution. We identified thirty-five criminal groups and compiled criminal records on ten thousand individuals.83
Moreover, a variety of conditions stymied the effectiveness of the Carabinieri. These included the difficulties of regular KFOR units to engage in police-like tasks, the climate of violence and revenge that prevailed in the province following NATO’s air war, and the sophistication of Kosovo’s shadow networks.84
Weak coordination and turf battles did not help either. KFOR and UNMIK struggled to coordinate. Different Italian agencies were not on the same page either. An Italian police officer from the Guardia di Finanza (Financial Guard) criticized the Carabinieri for not sharing information and intelligence.85 Another observer accused the Carabinieri of acting independently and of being ‘reluctant to submit to the law as it is enforced by the UN, thereby undermining cohesion and the promotion of one rule of law applicable to all.’86 The presence of gendarmeries also added another layer of difficulty to intra-organizational cooperation. The MSU was unpopular among KFOR sector commanders who tended to resent the fact that the MSU were mandated to conduct Kosovo-wide operations (upon the request of COM KFOR). They worried that Carabinieri activities such as weapons searches of private homes would reduce public support for KFOR.87 In one case from late 2008, a Multinational Task Force commander vetoed an operation by the Carabinieri, also because he thought that the outcomes of Carabinieri searches (the confiscation of single Kalashnikovs or old hunting rifles rather than major arms caches, he said) did not justify antagonizing the local population.88 Indeed, there is no evidence that the outcomes of Carabinieri counter-crime operations following the immediate post-war period were significant. Asked about important confiscations during 2008, members of the force listed six new pistols.89
To be sure, not only the Carabinieri but all international actors struggled to investigate crime, not least due to a lack of actionable intelligence.90 Bishop Artemije of the Serbian Orthodox Church claimed that during the first post-war year, over 1,000 Serbs were killed in Kosovo, around 1,200 were kidnapped or disappeared, and over 10,000 Serb homes and 80 churches were destroyed.91
Caution in praising gendarmeries in Kosovo also ignores possible unintended consequences. Providing training and intelligence to the Kosovo police created capacity but at the same time was problematic because the Kosovo police included few Serbs and tended to be biased against minorities. In one Serbian enclave in southern Kosovo, an official of the Organization for Security and Co-operation in Europe (OSCE) claimed that local Serbs feared special police units of the Kosovo police since these were exclusively Kosovar and dominated by former KLA fighters.92
In contrast to the peace support operations in the Balkans, in Afghanistan the Carabinieri operated in an armed conflict. In supporting the Afghan National Police (ANP), which was their main task, the Carabinieri adapted their characteristics to the growing insurgency: they taught tactical military skills, used military-grade weapons, and employed militarized language.
Weak policing was a main obstacle for stabilizing post-Taliban Afghanistan. Troops of the US-led Operation Enduring Freedom (OEF) focused on hunting remnants of the Taliban and their al Qaeda allies. The International Security Assistance Force (ISAF) was confined to Kabul until 2003, when NATO took command of the mission. There were no statutory Afghan security forces to speak of, only militias answering to local and national strongmen. Decades of war had eroded traditional police and justice mechanisms, leaving Afghans at the mercy of warlords and predatory government forces. This growing public insecurity benefited the Taliban, who re-infiltrated Afghanistan from safe havens in Pakistan and started an insurgency.
International actors increasingly came to see a functioning ANP as pivotal to stabilization and counterinsurgency. In early 2002, Germany became lead nation for police reform but its contributions were piece-meal. As of 2003, the US State Department began to fill the street-level policing gap, by hiring contractors to train ANP officers. In 2005, the Pentagon de facto took control over police-building, investing vast amounts of money and tasking regular soldiers to train Afghan police, primarily in tactical military skills.93 Such training could not alleviate perennial problems of the ANP such as human rights abuses, corruption, illiteracy, and drug consumption. For many Afghans, cops were robbers.94
To support police-building, the US turned towards the Carabinieri. Some Carabinieri had taken part in the OEF mission while others served in a military policing capacity for regular Italian troops in Western Afghanistan.95 In early 2008, the US asked Italy to send Carabinieri as ANP trainers and mentors to Afghanistan. Against the backdrop of the Balkans missions, US diplomats thought that the Italians had ‘mission-critical skills’.96 In Kabul, the Carabinieri were placed under the NATO Training Mission Afghanistan (NTM-A), which NATO had mandated to strengthen the ANP ‘at the district level and below to increase operational capabilities and survivability. The ANP needs more robust training and mentoring, as well as equipment.’97
The main recipient of Carabinieri training was the Afghan National Civil Order Police (ANCOP), which operated across Afghanistan (in contrast to regular ANP units, which were confined to districts), had a planned size of 5,000 members, and initially was to focus on crowd management.98 In comparison to regular police officers who received 6–8 weeks of training, ANCOP officers received 16 weeks of training; the outfit was also better equipped than the regular ANP.
Training began in September 2008 in Herat province, involving around 60 Carabinieri trainers. That program ended in 2012, after around 11,000 ANP had been trained in over 90 courses.99 The US praised the Carabinieri’s role, with US General David Petraeus going so far as to call the Carabinieri ‘the Michael Jordan of police trainers’.100 As of December 2009, the Italians added two more programs, one also in Herat province and the other in Kabul.101 The peak was reached in 2011 when around 130 Carabinieri were training ANCOP.102 In summer 2013, the force ended its training in Afghanistan.
As the insurgency worsened, the Carabinieri helped morphing ANCOP into a light infantry. Operationally, they emphasized tactical military skills. This supported the strategy of the US and the Afghan government to make ANCOP a core pillar of efforts to hold districts cleared from insurgents – as happened for example in early 2010 in the town of Marjah in the southern province of Helmand where ANCOP followed a major offensive by ISAF and the Afghan Army.103 By 2011, the primary task of ANCOP was to hold contested districts, although the force also participated in clearing operations, that is military offenses to drive insurgents out of areas.104 In 2013, a US general saw ANCOP ‘more in an army’ than a police role.105
These operational dynamics also changed the material and cultural elements of Carabinieri training programs. In preparing ANCOP for combat, the Italians divided the training into three phases. During the first, the Carabinieri taught ANCOP basic tactical movements. This was followed by lessons on complex movements, the setting up of checkpoints, and searches. Last, ANCOP was taught skills such as how to patrol high-threat areas in small units. During these phases, the Carabinieri showed their trainees, among other skills, how to use firearms while moving, how to handle improvised explosive devices (IEDs), and how to evacuate casualties.106 The Carabinieri would still teach crowd management, yet techniques differed from those taught in the Balkans. For example, the Italians showed ANCOP how to respond to demonstrators’ using live rounds.107
This operational outlook had implications for the material characteristics of trainers and trainees: preparing for high-intensity violence involved not only the handling of regular police weapons such as pistols but also the use of military-grade weapons such as machine guns and rocket-propelled grenades. Carabinieri justified such militarization by arguing that the ANP was involved in a growing insurgency.108 Indeed, the ANP suffered high numbers of casualties – around 75 officers killed per week in mid-2013.109
Praising their own approach was often accompanied by criticism of others. Carabinieri were especially critical of the private security company DynCorp, which the US Department of State, in 2003, had contracted to carry out police training. One Carabiniere said that DynCorp had failed to teach ANCOP how to shoot while wearing body armor.110
To institutionalize training, the Carabinieri developed ANCOP curricula (working jointly with French gendarmes). These curricula employed terms including enemy, camouflage, combat, and contact.111 When German police reformers, in April 2011, presented a draft document for a train-the-trainer program involving non-commissioned ANP officers, the Carabinieri were concerned that the curriculum did not include training in the use of machine guns. The Germans were trying to teach civilian policing as much as possible, in line with their own characteristics; for them, machine guns were taboo.112
Effects of Carabinieri Activities
In the light of institutional incentives to secure legitimacy and funding, it is not surprising that the Carabinieri evaluated their own contribution to the stabilization of Afghanistan in a positive light. The US military, which dominated over other international actors, concurred, given that the US welcomed any support for building the capacities of Afghan forces they could receive, and because the presence of gendarmeries allowed US soldiers to focus on combat. Other international actors, largely working in silos in Afghanistan, pursued their own distinct approaches.
As is the case for the Balkans, claims that gendarmeries in Afghanistan were successful cannot be substantiated due to a lack of evidence. ANCOP was better paid and trained than the regular ANP, yet it is unclear whether this made the force less prone to corruption and violence. One report claims that Afghans preferred ANCOP to the regular ANP.113 But other sources mention ANCOP failures such as a lack of interaction with the local population, human rights violations, and drug abuse.114 NATO and the US military undertook efforts to assess the performance of Afghan security forces. According to a US watchdog institution, the assessments were overly optimistic and methodologically flawed.115 Moreover, such assessments do not show whether security assistance changes the behavior of former trainees. As noted above, the Carabinieri provide numbers of trainers and persons trained. But we do not know to what extent such output and outcome translated into impact, that is if training reduced ANCOP’s involvement in violence and corruption after the completion of training. The Carabinieri were not collecting such information, and neither was the Afghan Ministry of Interior.
It is safe to argue, though, that a host of conditions made it difficult for the Carabinieri to support effective and accountable policing. Some of these conditions relate to the overall context of the Afghanistan war, including foreign support to the insurgency, a misbalance between international military and economic aid, a lack of contextual knowledge on the part of international actors, and the deployment of too few troops to this vast country.116 Other problems relate to ANCOP more specifically. One issue was literacy: although formally only literate recruits were accepted, in practice many were illiterate. Another issue was language. Few Carabinieri spoke English fluently, few translators spoke Italian, and few ANCOP officers spoke English. If proper translation from Italian or English into an Afghan language was provided, it was mainly in Dari, disadvantaging ANCOP members who spoke Pashtu only.117 While from this perspective it was fortunate that only few members spoke Pashtu, the small number of Pashtu speakers turned into a problem once trained units were deployed to Pashtu-speaking areas since it limited ANCOP’s ability to interact with local residents, which was a precondition for winning their trust. Another factor undermining the development and utilization of gendarmerie capabilities was that ANCOP was often employed to perform military operations that exceeded their capabilities.118 Flawed operational and tactical planning contributed to high rates of attrition within ANCOP.119
The Carabinieri training was part of an overall policy of NATO and the US to quickly increase the size of the Afghan security forces so that international troops could be drawn down. This policy de facto prioritized quantity over quality. Concomitantly, international capacity building was largely reduced to train-and-equip, with little attention paid to the creation of internal and external accountability mechanisms for Afghan forces. Under these conditions, the training offered by the Carabinieri was, at best, ineffective. At worst, it inadvertently exacerbated the violence in Afghanistan, by sending poorly trained and equipped troops into a war where many of the troops would lose live and limb without a proper strategy for peace, and where some of the trained troops risked violating human rights with impunity.
The few existing empirical analyses of gendarmerie operations in multinational missions stress that this type of force is vital for avoiding public security gaps in international interventions.120 This article started from the observation that the activities and effects of gendarmerie missions in multinational peace and stabilization operations are under-researched. In focusing on the Italian Carabinieri as a militarized police force, the article discussed gendarmerie activities and possible effects.
The analysis shows that gendarmeries, as paramilitarized police forces, carry out a variety of tasks that military combat forces and civilian police services struggle with in war-torn countries, including crowd management, counter-crime operations, and arrests. The analysis also underlined the flexibility of gendarmeries whereby they adapt their characteristics to the mission context. Thus, in the Balkans the Carabinieri characteristics were more civilian and in Afghanistan more military.
Moreover, preliminary evidence suggests that the Carabinieri produced considerable output. They were keen to demonstrate their competitive advantages and went on patrol, made arrests, engaged in searches, and trained host state forces as much as possible. However, translating organizational activity into changes on the ground was difficult. The outcomes of some gendarmerie operations, such as with regard to amounts of weapons or drugs seized, were underwhelming, despite a high operational tempo. Problems can be found on the international level (such as deficient coordination between gendarmeries and regular military forces in the Balkans, and the relatively small size of gendarmerie contingents, which limits their geographical scope and number of activities); on the interface between international and host state actors (such as problems of translation in Afghanistan); and within war-torn countries (such as weak oversight of armed forces). If scope conditions for effective and accountable stabilization such as sufficient troop numbers, coordination, and host state legitimacy are not given, the deployment of gendarmeries may be ineffective. Worse, their deployment may produce negative unintended consequences, such as empowering unaccountable host state forces.
There are many inroads for future research but also challenges. Regarding output, we need to better understand the who, where, when, and how of militarized police interventions. Such analysis would reveal that multinational missions are not homogenous but that their component organizations have varying characteristics. Unfortunately, at least from a research perspective, much data and information is classified. While the Carabinieri tend to grant access to researchers in mission areas, they are reluctant to disclose tactical procedures, the number, types, and sites of operations, or policy documents such as manuals; the force, as any other security organization, also does not readily share lessons-learned studies with institutional outsiders.
Analyzing the outcomes of gendarmerie operations is even more challenging. Outcomes need to be assessed against a baseline; without knowing, for example, how large the illicit economy in a specific area is, statistics on the amount of seized goods have little meaning. Any such data collection necessarily includes much speculation, even in stable countries. To increase the reliability of findings, analyses should rely on multiple sources, not only gendarmeries themselves who, after all, have a vested interest to inflate their achievements. Triangulation has limitations though because other institutions may not possess data and information on gendarmerie operations and merely comment on these operations in line with their own vested interests and organizational cultures.
Measuring impact is the most challenging. Surveys could reveal how local residents perceive gendarmeries. But most people do not distinguish between different foreign security forces. Qualitative case studies relying on process tracing and counterfactual analysis could reveal mechanisms of how gendarmeries affected changes on the ground. Such studies would help assess the validity of claims that without gendarmeries, more weapons and drugs would have been smuggled in and out of Bosnia or Kosovo, that the damage during the March 2004 riots in Kosovo would have been worse, or that Afghan forces would have been more violent. Such studies also require a significant amount of data, including knowledge of decision-making by the gendarmeries’ adversaries. Case studies should honestly point at empirical and explanatory black boxes.
Case studies discussing impact should also be sensitive to multicausality and non-linearity. Gendarmerie operations are only one factor among many others influencing decision-making by criminal groups or insurgents, the performance of gendarmerie allies, or overall levels of crime and violence. One could compare, for example, compliance with human rights norms of a host state unit that received gendarmerie training with one that did not. This would require controlling for context factors influencing compliance, such as orders from superiors. Context factors thus qualify claims of success. In Kosovo in March 2004, Carabinieri were not the only ones trying to prevent a mob from marching on the monastery of Deçani; Kosovar leaders used their influence to persuade the crowd from taking such action.121 The broader the impact to be measured, the more problematic the attribution of causality; levels of crime and violence are influenced by a web of normative, demographic, and economic factors.122
Impact assessments should also specify the beneficiaries and take into account complex effects. Operations that reduce crime with a transnational dimension may increase the security of sending states but damage the livelihood of people depending on small-scale illicit activities; operations that evict insurgents from an area might expose local residents to state forces that are even more predatory than the insurgents.
Knowledge of the activities and effects of gendarmerie operations is too inconclusive yet to support the claim that the deployment of gendarmeries and other hybrid forces is vital for stabilizing war-torn countries. We need systematic evaluations of militarized policing interventions in order to substantiate debates on such interventions and to improve practices.