Strategic Engineered Migration as a Weapon of War

KELLY M. GREENHILL

In recent years, it has been widely argued that a new and different armament – i.e., the refugee as weapon – has entered the world's arsenals. But just how new and different is this weapon? Can it only be used in wartime? And just how successful has been its exploitation? Using a combination of statistical data and case study analysis, this article tackles these questions and provides a detailed examination of the instrumental manipulation of population movements as political and military weapons of war. In addition to 'mapping the terrain' of the issue by providing a comprehensive typology of the most common means by – and desired ends for – which displaced persons have been used as political and military weapons since the end of the Cold War, the author also provides a portrait of the identities of the kinds of actors most likely to engage in this kind of exploitation. She also proposes an explanation for what motivates them to resort – and apparently increasingly so – to the use of this unconventional policy tool, despite the reputational and potential retributive costs of doing so.

In the aftermath of the mass expulsion of Kosovar Albanians in the spring of 1999, Harvard Law School's Professor Martha Minow declared, 'the nature of warfare has changed; now the refugees are the war'. She was not alone in declaring that a new and different armament – namely, the refugee as weapon – had entered the world's arsenals. But just how novel is this weapon? Who employs it? And, since the late Serbian leader Slobodan Milosevic eventually capitulated, should this unconventional weapon be dismissed as a dud?

Contrary to conventional wisdom, the fact is the instrumental exploitation of population outflows is neither a new nor a particularly unusual phenomenon. Rather, such exploitation has had a long, influential, and often successful history, one that includes both wartime and peacetime use, by both state and non-state actors. Furthermore, despite a widespread belief that the majority of outflows are simply the unintended consequences of man-made or natural humanitarian disasters – for example, wars, floods, famines – in reality most are created as the direct result of political decisions taken by sovereign states, often for specific political and/or military ends.1

In the last decade alone, we have witnessed their use in wartime in multiple locations and in numerous ways. They have been pressed into service as soldiers

Civil Wars, Vol.10, No.1 (March 2008), pp.6–21 ISSN 1369-8249 print/ISSN 1743-968X online DOI: 10.1080/13698240701835425 q 2008 Taylor & Francis (in the African Great Lakes region, for example), deployed as human shields (such as in Afghanistan and Iraq), and used to create logistical logjams (in Kosovo and Sudan, for instance). They have likewise been employed as propaganda tools to elicit international sympathy and support (including by all parties to the conflicts in Bosnia, to name just one example).

Ironically, however, while such manipulation is not particularly unusual, neither is it particularly well understood. While an appreciation of the fact that displaced people can be used tactically as 'refugee warriors' and 'human shields' has been growing in currency in recent years, this kind of tactical use is but one piece of the puzzle. Arguably, the more interesting and still under-appreciated piece surrounds the strategic manipulation of flows themselves, something that can be accomplished through means as obvious as the employment of massive military force, or as subtle as the judicious use of financial inducements. In the 1990s the world witnessed orchestrated population outflows in regions as diverse and far-flung as the Balkans, the African Great Lakes Region, the Caribbean, and Southeast Asia. Still, the manipulation of population movements as operational and strategic means to political and military ends remains poorly understood2 – so much so, in fact, that the idea that this kind of manipulation even exists is sometimes resisted.3 This is an unfortunate trend, because evidence suggests that such manipulation and exploitation is growing both easier and more frequent over time, while simultaneously remaining, for reasons I shall elucidate below, something of a self-hiding phenomenon.4

This article aims to address this lack of understanding in several distinct ways. First, it presents the first comprehensive taxonomy of the underappreciated phenomenon of strategic engineered migration.

Second, drawing upon data from this author's own database of these strategicallydriven out-migrations, it offers some observations on its general prevalence and the efficacy of one of its sub-variants; namely, the coercive variant – that is, the class of cross-border migrations designed to influence the political or economic behavior of potential host states and other state-level actors farther afield.

Finally, it presents a set of testable propositions about the nature of the actors who seek to employ this kind of unconventional weapon and an explanation as to what motivates them to resort to it, in the face of the potential reputational and retributive costs of doing so.

CHARACTERIZING AND UNDERSTANDING STRATEGIC ENGINEERED MIGRATION

Strategic engineered migration refers to those in- or out-migrations that are deliberately induced or manipulated by state or non-state actors, in ways designed to augment, reduce, or change the composition of the population residing within a particular territory, for political or military ends. Because the focus is on strategicdriven population movements, externalities-driven population movements – that is, those inadvertently generated as a consequence of other policies (e.g., construction

of the Three Gorges Dam in China), or of conflict (e.g., the Belgian and French refugees who fled the German offensive in World War I) – are excluded as are outflows that result from policies of neglect (e.g., the famine in Ethiopia in the early 1980s). The instruments employed in producing engineered migrations run the gamut from threats and the use of military force, through the promise of inducements and financial incentives, to the simple opening of normally sealed borders.

Building on earlier work by Myron Weiner and Michael Teitelbaum, one can distinguish between four distinct, but non-mutually exclusive, forms of strategic engineered migration, all of which may be utilized in wartime, by both state and non-state actors:5

• . Dispossessive the class of events, including both in and out-migrations, in which the principal objective is the appropriation of the territory and/or property of another group or groups, and/or the elimination of this group or groups as a threat to the ethno-political or economic dominance of the perpetrators; this includes what is commonly known as ethnic cleansing;
• . Exportive those displacements undertaken either to fortify a domestic political position –by expelling political dissidents and other domestic adversaries – or to discomfit or destabilize foreign government(s);
• . Militarized those displacements conducted, usually during active conflict, to gain military advantage against an adversary – namely, via the disruption or destruction of an opponent's command and control, logistics, or movement capabilities – or to enhance one's own force structure, via the acquisition of additional (sometimes reluctant) manpower and/or resources; and
• . Coercive the class of events in which (real or threatened) outflows are used, as a foreign policy tool, to induce (or prevent) changes in political behavior and/or to extract side-payments from the target(s); coercive use includes the propagandistic use of outflows (which are often generated by others) for their own benefit.

Dispossessive Engineered Migration

Although the term 'ethnic cleansing' did not become a household word until the mass expulsions associated with the Balkan wars of the 1990s, engineered migration designed to acquire territory or property and/or to alter the ethno–religious– political balance within said territory has had a long and sordid history, one that is often associated with acts of genocide against the group(s) displaced.6 Since biblical times, it has been carried out not only by states and their surrogates, but also by opposition groups and rebel movements, warlords and others who aspire to power or control over people or territory.7 Of the four variants, dispossessive is the most commonly recognized and arguably the most commonly occurring. Nevertheless, few incidents of dispossessive engineered migration appear to be driven solely by the desire for territorial acquisition, but rather tend to be tied to other political and/or military goals. Prominent recent examples include its use, since 2003, in the Darfur region of Sudan and throughout the most conflict-ridden zones of Iraq.

Exportive Engineered Migration

At least since the heyday of the Athenian empire, expulsion has been used as a way of dealing with domestic dissidents and class enemies. Sometimes these expellees have been individuals, but they have frequently been entire groups or social classes – for example, the Sri Lankan Sinhalese government's deportation of its Tamil tea estate workers. Exportive engineered migration has been particularly common in the aftermath of revolutions, as revolutionary regimes often view large-scale expulsions as a way to quickly transform the country's social structure, as was seen, for instance, in Cambodia in the mid-1970s.

Exportive engineered migration has also allegedly been used against foreign governments, either simply to embarrass them or, more often, in attempts to undermine them. Though finding definitive evidence of intent is often problematic, in some sense whether those generating the outflows actually intended to politically destabilize their neighbors matters less than the fact that those on the receiving end perceived that this was the generators' intent and that these perceptions conditioned their responses – which in the case of India in the early 1970s, for example, led to a full-scale war and the secession of East Pakistan.8

Militarized Engineered Migration

Militarized displacements are those conducted either to gain military advantage, through disruption or destruction of an opponent's command and control, logistics, or movement capabilities, or to enhance one's own force structure, either by acquiring additional manpower and resources or by 'draining the sea' in which insurgents swim through the use of 'regroupment camps' or 'strategic hamlets'.9 Militarized displacements may be generated by both state and non-state actors and for the benefit of military operations both at home and abroad. For instance, during their occupation of Afghanistan, the Soviets provoked refugee flows into Iran and Pakistan to deprive the guerrillas of bases of support and – in an ultimately failed attempt to create a cordon sanitaire along the Pakistani border – engaged in widespread carpet-bombing of the border areas.10

Likewise, in the early 1980s, Sandinista attempts to quash guerrilla resistance in Nicaragua through the use of regroupment backfired, generating not only international outrage, but also a growth in the number of local inhabitants eager to join the insurgency.11 This same unintended consequence has transpired in numerous counterinsurgency contexts, including as a result of the use of regroupment by the Portuguese in Angola, the French – and later, the US – in Indochina, the British in Kenya, and Burundian government forces, among myriad others.12

Another, more opportunistic, type of militarized engineered migration surrounds the exploitation of outflows generated by others to serve the military purposes and national security interests of receiving states. For instance, in the 1980s the Thais played host to a quarter of a million fleeing Cambodians, while also using them as a

human buffer zone to protect Thailand from the ongoing conflict within Cambodia.13 Similarly, in an ultimately unsuccessful attempt to quash an anti-governmental insurgency in the eastern part of his country, former Zairean President Mobutu Sese Seko armed many of the Hutu refugees and genocidaires who fled to Zaire from Rwanda in the wake of the 1994 genocide.14

Coercive Engineered Migration

A final type of engineered migration is the coercive variant, in which (real or threatened) outflows are used to induce (or prevent) changes in political behavior – that is, to compel or deter – and/or to extract economic side-payments from a target state or states. In practice, it bears a striking resemblance to more traditional forms of coercion and may be used both as an instrument of deterrence and of compellence, as well as for exercises in both coercion and counter-coercion –sometimes simultaneously. Operationally, it most closely resembles what Robert Pape refers to as 'coercion by punishment' – which functions by raising the costs or risks to a target's civilian populations – as opposed to 'coercion by denial' – which relies on the use of military force to prevent the attainment of political objectives or territorial ambitions.15

Circumstantial evidence suggests that most perpetrators prefer to employ what, in traditional coercion, are usually referred to as risk strategies, which is to say they prefer to inflict costs at a gradually increasing rate, threatening bigger punishment later for non-compliance. Because the punishment is not inflicted all at once, the coercer 'may interrupt operations temporarily in order to provide time for reflection or negotiation or to reward the target state for concessions, thus encouraging minor demonstrations or willingness to accommodate the [perpetrator's] demands as well as major concessions'.16

Risk strategies are likely preferred for two reasons. First, too large an outflow may itself destabilize a regime. Second, at least recently, massive outflows have increased the probability of military interventions undertaken to stop or reverse them. For example, evidence suggests Yugoslav President Slobodan Milosevic recognized that a massive outflow could only hurt him in the period leading up to the North Atlantic Treaty Organization's 1999 bombing campaign.17 Thus perpetrators generally prefer minimal or limited outflows to massive outflows. However, frequently perpetrators do not have the luxury of pursuing their preferred choice and, as often as not, find themselves generating larger outflows than they (strictly speaking) view as desirable.

Two factors tend to impede the successful employment of risk strategies: the migrants themselves and the nature of risk strategies. As Thomas Schelling argued, 'the ideal compellent action would be one that, once initiated, causes minimal harm if compliance is forthcoming and great harm if compliance is not forthcoming'.18 The problem is that mass migrations are often 'a gift that keeps on giving'. Once an outflow has been initiated, perpetrators often lose some degree of control over it. The fact that generators can lose control is not surprising. Those who conduct ethnic cleansing may be irregulars or even simply 'bands of thugs' who lack discipline and may even pursue their own self-serving strategies, which may be to the detriment of the perpetrators' own strategies.19 Likewise, those displaced have their own agendas. Once outside the sending state, they are frequently capable of autonomous actions that are not necessarily compatible with the goals of the perpetrators, and which can undermine stable equilibria and torpedo potential deals between those generating outflows and their targets.

Also, as in traditional coercion, risk strategies are often viewed as incredible. 'Instead of being convinced of the perpetrator's resolve to inflict maximum damage if demands are not met, the opponent is more likely to be convinced that the coercer will never escalate far above current restrained levels.'20 For reasons discussed more fully below, targets often fail to take seriously perpetrators' threats until a massive outflow has been initiated. Thus, the kind of graduated approach upon which risk strategies are predicated often fails.

At the same time, coercive engineered migration also differs from traditional coercion in some crucial ways. Whereas in traditional coercion, potential assailants tend to be deterred from even attempting coercion unless they possess superior military capabilities that can protect them from the victim's retaliation, in cases of coercive engineered migration perpetrators are frequently undeterred by their target's military superiority, because retaliation by the target is only rarely a politically feasible option. This is due to the fact that – also unlike traditional coercion – targets generally value the issues at hand less than do the coercers, who tend to be highly dissatisfied with the status quo and more highly resolved than are their targets.21 For instance, as disconcerting as West German leaders found the periodic inflows of large numbers of Eastern bloc refugees, neither they nor their NATO allies were ever going to be willing to risk starting World War III by taking retaliatory military action against East Germany. Perpetrator dissatisfaction often stems from the fact that even before any outflow has occurred, targets and/or others are often themselves engaged in trying to coerce the perpetrators, via the use of sanctions, embargoes, or diplomatic pressure. Also, unlike in traditional military coercion, coercion by punishment does work and works frequently, despite the fact that the balance of material capabilities most often favors the target.22

Although the aims and specific tactics employed vary, in the majority of cases these three types share a common and significant feature: namely, asymmetric coercion of the strong by the relatively weak,23 through the employment of techniques that may render the power and influence differentials between them significantly less meaningful than is the norm; namely, by using the strengths of the powerful against them. To be clear, coercive engineered migration also may be employed by strong actors against each other and against weaker actors. In those cases where strong actors have been the perpetrators, the goal appears to be the achievement of political goals at an even lower cost than they could be achieved through military means.

An appreciation for all four of the aforementioned variants is critical to understanding the broader phenomenon, and its use in wartime, because the variants are not mutually exclusive – namely, many outflows comprise multiple motivations and objectives. For instance, during the 1992–95 Bosnian war, all four types of engineered migration – dispossessive, exportive, militarized, and coercive – were

employed, and they were employed by all three sides – that is, the Bosniaks, the Croats, and the Serbs—but to varying degrees and with significantly disparate levels of success.24 Likewise, during the Vietnam War, the North Vietnamese (NVN) employed exportive, militarized, and coercive engineered migration against the South, the US, and its own people, while the US and the South Vietnamese engaged in militarized engineered migration against NVN and Vietcong forces via its use of resettlement campaigns and strategic hamlets, and in coercive engineered migration via their attempts early in the war to incite North Vietnamese to flee to areas controlled by the South.25 Because of the existence of overlapping motivations and objectives, any analysis of one type of engineered migration must take into account the possible impact of the others when evaluating the success or failure of any given attempt to use refugees strategically.26

Recognizing the existence of overlapping motivations and objectives is likewise critical because a failure to do so may undermine the potential resolution of crises at the time they transpire. Consider, for instance, an example in the context of the coercive variant. Put in strategic interaction terms, during a crisis overlapping objectives may lead to perpetrators and targets holding mutually exclusive assumptions. A powerful target assumes there will be no challenge to it, as such a challenge would appear to be irrational, and thus the target fails to make immediate (and/or credible) deterrent threats or to resolve the underlying dispute with concessions that would be sufficient to reassure the perpetrator. Thus neither effective deterrence nor reassurance is provided to dissatisfied would-be perpetrators. Instead the two end up in a situation often referred to as 'strategic ambivalence' – that is, a situation in which a defender's policy does not provide an unequivocal statement of intent to protect the status quo or provide concessions that can satisfy a challenger's ambitions.27 For instance, during the 1998–99 Kosovo crisis, evidence suggests that NATO's focus on the dispossessive component of Serbian behavior may have blinded the alliance to Milosevic's attempt to signal his intent to use Kosovar Albanian refugees coercively.

Moreover, one should also keep in mind that overlapping motivations and objectives may also exacerbate the situation on the ground for the true victims in these crises, the refugees themselves. Consider, for instance, that the Thais' successful use of coercive engineered migration against the US in the late 1970s – namely, via threats to close border camps housing Indochinese refugees –then precipitated the subsequent use of these self-same refugees as a military buffer zone in the years that followed.28

THE EXISTENCE AND PREVALENCE OF STRATEGIC ENGINEERED MIGRATION

The format of this journal precludes graphical representation of the prevalence of strategic engineered migration. Data gathered by this author indicates, however, that – contrary to conventional wisdom – strategic engineered migration is both real and relatively common. Identification of well over 100 documented cases in the last half-century belies the claims alluded to at the outset that the instrumental use of cross-border population movements (i.e., the refugee as weapon) is either new or novel.29 It likewise suggests that the conventional wisdom that the majority of population outflows are simply the unintended consequences of human and natural disasters should be considered suspect at best.

Instead, somewhere in the world, at almost any given time since the signing of the United Nations (UN) Refugee Convention in 1951, population displacements have been used strategically, and often by multiple actors simultaneously. The frequency of this phenomenon (at . ¼ 1.82 incidences per year) is significantly lower than the number of annual interstate territorial changes (4.82/year) and is dwarfed by the frequency of militarized interstate disputes (20.06/year). At the same time, its frequency is (at its lower level) comparable with the number of extended immediate deterrence crises (0.58/year) and intrastate wars (0.68/year); thus the conventional wisdom is simply wrong.

At the same time, because of the chaos that often surrounds mass outflows and the aforementioned overlapping motivations problem, it is not surprising that the prevalence of this kind of manipulation is opaque. This opacity is exacerbated – at least in the case of the coercive variant – because few actors will wish to appear to be weaklings and some perpetrators may not wish to be viewed as bullies, those involved may go to some trouble to keep attempts at coercion from becoming transparent. Despite these obstacles, however, and for reasons of statecraft and humanitarianism, it behooves us not only to better understand the nature and prevalence of this kind of exploitation, but also why some actors decide to employ this unconventional weapon.

HYPOTHESES ABOUT THE MOST LIKELY PERPETRATORS OF STRATEGIC ENGINEERED MIGRATION

To aid in our understanding of why some actors would resort to the manipulation and exploitation of the displaced as a policy tool, below I propose some hypotheses about the nature of perpetrators of strategic engineered migration and an explanation as to appeal of this unconventional policy instrument. These hypotheses have been inferred by building on earlier work from a variety of literatures, including asymmetric and guerrilla warfare, bargaining and negotiation, international law, and democratic peace theory.

It is a widely accepted axiom in international politics that, paraphrasing Thucydides, 'the strong do what they can; the weak endure what they must'. Yet as Dwight D. Eisenhower also famously noted, 'It's not the size of the dog in the fight. It's the size of the fight in the dog.' In other words, with a well-chosen strategy and the right set of circumstances, highly motivated and resolute, albeit weak, actors can triumph over more powerful ones.30 Thus for such actors, a resort to unconventional, asymmetric methods such as the manipulation of population movements may appear both eminently rational and very attractive. Hence, I hypothesize that those who actively generate refugee flows are most likely to be weak, illegitimate, or semilegitimate actors, who lack effective recourse to more conventional methods of

influence.31 I further suggest that they will be drawn to this instrument for five distinct, and wholly rational, cost-benefit driven reasons.

Asymmetric Leverage

A variety of scholars who have done research on the negotiating strategies of weak actors have found that they often view crisis generation – refugee or otherwise – as a necessary precursor to negotiations with their more powerful counterparts.32 This is likely a relic of the fact that in the absence of crisis-generating incentives, powerful states tend to be reluctant to yield concessions, particularly to weaker challengers, for two distinct reasons.

One, they tend to doubt the credibility of weaker actors' threats. This is due to the fact that powerful actors frequently cannot fathom the idea that their weaker counterparts would initiate a crisis or conflict they seem destined to lose, based on relative capabilities. This tendency may be further exacerbated by the fact that targets may also underestimate the magnitude of the threats facing weak actor(s) when the issues at stake seem so small to them, thus leading them to further discount the probability of crisis initiation.33 Moreover, because the majority of targets would not themselves initiate refugee flows, they tend to find perpetrators' threats to do so incredible.

Two, powerful targets may feel that concessions to weaker adversaries will be viewed as a sign of weakness by others – allies and adversaries alike. Although the validity of such concerns is questionable, history nevertheless provides ample evidence that powerful states sometimes view tough posturing, even on issues of small consequence, as essential for the purposes of signaling their intentions to, and maintaining their reputation with, both friends and adversaries.34

Yet weak actors have a tried-and-true strategy for overcoming powerful actors' reluctance to negotiate and leveling the playing field: the generation of crises. Crisis generation represents one of the few areas in which weak, illegitimate actors may possess relative strength vis-a-vis other more powerful target states, and certainly – in the case of migration crises, at least – also vis-a-vis their even weaker domestic victims. After creating crises, weak actors can then offer to make them disappear in exchange for financial or political pay-offs. In short, crisis generation may permit weak actors to increase their leverage relative to their aggregate power.35 As Scott Snyder has noted: 'this kind of crisis diplomacy has proven an effective way for weak states to force [their] way to the top of the negotiating agenda'.36 Indeed, negotiators report a recognizable pattern of 'drama and catastrophe'.37 And in the face of catastrophe, an overlapping bargaining space may develop rapidly where before there was none.

Put another way, crisis generation acts as a kind of force multiplier for weak states, thereby enhancing their credibility and improving their coercive capabilities. In a discussion of more traditional methods of coercion, Alexander George has argued that the key to successful coercive diplomacy is finding a way of injecting the message of the challenger's (read perpetrator's) threats into the adversary's calculations and leading him to comply with the demand(s) made. Whether he will succeed is contingent upon the significance of the demand made, the determination of the opponent and the extent to which the opponent feels that the threatened punishment is both sufficiently credible and potent to cause him to comply.38

However, because the traditional means of influence at the disposal of weak states generally do not favor their success, a resort to unconventional methods – such as the generation of migration and refugee crises – can enhance the 'potency' of their threats for several reasons. For one thing, via the generation of massive outflows refugee generators may inflict a punishment upon targets disproportionate to the costs of compliance. Although targets may be reluctant to concede ex ante, in the grand scheme of things the demands being made by weak actors are often nominal compared with the costs of managing a sustained, large-scale outflow. In addition, because in-kind retaliation is rarely an option for targets – and alternate responses may also be problematic39 – crisis generators may achieve a kind of escalation dominance over potential targets.

Finally, despite the fact that warnings tend to precede them, for reasons articulated above, refugee outflows – or at least the scope of their potential consequences – tend to take targets by surprise. For instance, in the aftermath of the 1999 Kosovo crisis, the forewarned German Foreign Minister Joschka Fischer later said 'he regretted not having taken Milosevic seriously' when the Yugoslav president said he could empty Kosovo in a week.40

Expected Marginal Costs versus Expected Benefits I – The Paradox of Liberalism There exists an ever-growing belief, in both academic and policy circles, that liberal democracies possess particular characteristics that make them (and their leaders) behave differently than other regime types. Those in liberal democratic states see themselves as distinct and frankly superior to those of other regime types. As Michael Doyle has put it, 'because non-liberal governments are in a state of aggression with their own people, their foreign relations become for liberal governments deeply suspect. In short, fellow liberals benefit from a presumption of amity; non-liberals suffer from a presumption of enmity.'41 Moreover, authoritarian states are expected to 'aggress against others if given the power and the opportunity'.42

Because contemporarily illegitimate and/or illiberal actors will already be viewed with suspicion and contempt by the most powerful members of the international community at large, it is logical that they will also believe they have little left to lose by abrogating the moral norms associated with the generation of population outflows. As Louis Henkin notes:

for any nation, the cost and advantage of law observance or violation must be seen largely in the context of its foreign policy as a whole... Nations generally desire a reputation for "principled behavior, for propriety and respectability.... They do not wish to be accused or criticized. They "know that violation will bring protest, will require reply, explanation, and justification.43

However, for actors that are already isolated and outside the 'club', the cost of violating international laws may be regarded as low relative to the costs of compliance. In short, states and/or non-state actors that are already viewed as internationally illegitimate (e.g., 'rogue' and 'pariah' states) have significantly less reason to 'fear the moral wrath of the "international community" than states with reputations to protect'.44 As Susan Woodward has put it in reflecting on the Bosnian case:

Those parties who already had widespread international support were more vulnerable to international opinion but less likely to be exposed, whereas those who were most accused of such atrocities and on whom media attention focused– the Bosnian Serbs in the case of Bosnia-Herzegovina – were far less susceptible because they had little international support to lose or to try to maintain.45

In short, non-democratic, and hence, 'illegitimate' states and non-state actors face a double-whammy: they are frequently neither strong enough to impel others to take them seriously under normal conditions, nor are they likely to be trusted to negotiate in good faith. Hence, the reputational barriers to resorting to normsviolating tactics (e.g., the use of refugees as weapons) are lower, while the bargaining advantages of doing so are far greater – at least in part because specific reciprocity (threatening a counter-crisis) is not an available option for most targets.

Expected Marginal Costs versus Expected Benefits II – The Exploitation of Liberalism

At the same time, however, sometimes the virtues of liberalism can be exploited in the context of the refugee weapon. As has been argued elsewhere, rebels trying to overthrow their governments, and secessionist groups that stand little chance of winning their independence alone, may act in ways designed to generate international political, economic and/or military support for their causes.46 The key to engendering such support is twofold: first, generate visible international 'moral outrage' on one's behalf and second, be viewed as a victimized group. But victims only become 'victims' when their oppressors have been identified and 'labeled as tyrants' by outsiders.47 It has long been recognized that conventionally superior forces could be made to act in ways that entailed great political costs for the governments they served. Hence, the most efficient way for insurgent groups to garner the requisite sympathy for themselves and derision for their government is to provoke attacks upon themselves by said governments, attacks that frequently lead to the generation of refugee flows.48

Such flows not only provoke outrage and evoke sympathy, but also generate fear within potential recipient states, and thus may precipitate international action on the part of the international community, or, more likely, of a 'coalition of the willing', that is, the exposed). For instance, reports have surfaced that Bosnian forces did not defend the 'safe area' of Gorazde, in the expectation that the consequences would lead to increased NATO involvement in the war for Bosnia-Herzegovina.49 Likewise, there is ample evidence to suggest that the Kosovar Albanians behaved in a manner calculated to bring down the wrath of the Serbian government upon them, because they had credible reason to believe NATO would intervene on their behalf. As the fact of the 1999 war for Kosovo demonstrates, this turned out to be a wise gamble.

However, such gambles do not always pay off. For instance, during the Algerian bid for independence, the rebels suffered greatly, at least in the short run, from their miscalculation that the West would intervene in the face of the French brutality they brought upon themselves. Although some French actions generated an international stir (and, in the case of the 1958 bombing of the refugee camp at Sakiet-Sidi-Youssef in Tunisia, generated a ministerial crisis) the international community was not prepared to intervene on behalf of the guerrillas, even if they were refugees.50 In other words, while by 1999, Yugoslav President Slobodan Milosevic was acknowledged to be an international 'tyrant', the leaders of the French Fourth Republic in the 1950s were not, and that made all the difference.51

Operational Advantages

Strategic engineered migration can also be a relatively cost-effective policy tool. Compared with more conventional military operations, the generation of outflows is usually relatively cheap, particularly as the number of troops required is frequently small, and the manpower necessary to effect successful population displacement need not be highly trained or well-equipped. In fact, the use of regular troops is often not even necessary; it can also be done with paramilitary 'shock troops' and even bands of thugs, as the recent wars in the Balkans and the conflict in Darfur demonstrate.52 The inducement of engineered migration does not rely on direct combat, but instead on the expectations associated with the demonstrative capacity of the violence that can be brought to bear. Sometimes no force need be used at all. Little more than the fear of future violence may be sufficient to cause people to flee.53

Conversely, waging a successful campaign for the hearts and minds of a local population is much costlier and time intensive. In the case of national control, it may also necessitate ceding some amount of power, which engineered migration does not. Moreover, history illustrates that engineered migration can be a very effective way to consolidate control over territory. Consider, for instance, Mao's conquest of China, Israel's successful territorial consolidation during the 1948–49 war of independence, and King Hassan II of Morocco's 1975 'Green March' into the disputed territories of Spanish Sahara – an operation that required no troops at all. In short, engineered migration may offer relatively weak actors substantial rewards for a relatively small investment, just as long as it does not precipitate an international response.54

Straightforward Economic Benefits

Finally, strategic engineered migration can be economically quite remunerative. Those in the sending state can appropriate the assets of the displaced, as happened after Idi Amin's expulsion of Ugandan Asians in 1972, for instance. Likewise, in cases when only some members of a group, class or population are expelled or have fled, the sending state may benefit from direct or indirect access to future remittances from those have fled to those who remain. Finally, sometimes generators (and/or those

carrying out the operations) profit handsomely from the acquisition of bribes from would-be refugees. For instance, in Vietnam, where it is widely-acknowledged that in the late 1970s the military extracted 'departure taxes' from many of those it pushed out, it has been estimated that, during 1978 alone, Hanoi's receipts from the refugee traffic totaled \$US 115 million, or 2.5 percent of the total estimated gross national product.55

CONCLUSIONS

As stated at the outset, the intent of this paper was to provide a framework for understanding how, why, and how frequently the creation and exploitation of crossborder population movements are utilized as an instrument of statecraft and, especially, as a weapon of war. The article introduced four distinct varieties of this kind of exploitation, that is, dispossessive, exportive, militarized, and coercive, and presented data about its relative prevalence in the late twentieth to early twentiethfirst centuries. It also offered a portrait of the kind of actors who would employ it and why. It was hypothesized that weak and illegitimate actors favor this tactic, because they face fewer behavioral constraints and possess more material incentives to engage in strategic engineered migration than do more powerful and internationally legitimate actors. Thus, while morally reprehensible, it is hardly surprising that such actors should sometimes resort to this unconventional policy tool.

Long before 9/11 galvanized a new preoccupation with border security, issues surrounding refugees and illegal migrants had in many countries transmuted from a matter of 'low politics' to a matter of 'high politics', involving a shift in the definition of national security threats and in the practice of security policy. In such an era, and one in which actors as globally diverse as Libya's Muammar Gaddafi, Belarus's Aleksandr Lukashenko, and North Korea's Kim Jong-Il, are able to actively or tacitly exploit the fears associated with massive cross-border population movements – a better understanding and appreciation for both the prevalence of this phenomenon and the motivations of actors desperate enough to risk the dangers of instigating migration crises is critical for policymakers and practitioners alike.56 Crafting appropriate policies to confront these dangers and proactive responses to such threats demands no less, especially since, in the post-Cold War era, evidence suggests both the attractiveness of this weapon and frequency of its use appear to be growing.

NOTES

• 1. Gil Loescher, 'Introduction,' in Gil Loescher and Laila Monahan (eds.), Refugees and International Relations (Oxford: OUP 1989) p.8. Loescher is one of the few scholars and practitioners who have recognized and written about this fact; others include Charles Keely, Michael Teitelbaum, and Myron Weiner.
• 2. There are a few exceptions to this general rule. See, for instance, Gil Loescher, Refugee Movements and International Security, Adelphi Paper 268 (London: International Institute for Strategic Studies 1992); and Barry R. Posen, 'Military Responses to Refugee Disasters', International Security 21/1 (Summer 1996) pp.72–111.
• 3. Such resistance is particularly surprising, given that some perpetrators of refugee-driven coercion have explicitly acknowledged their role in such manipulation. See, for instance, Frank Johnson, 'East Germans' refugee ploy upsets the West', The Times, 26 July 1986, Issue 62519; Barbara Demick, '58 N. Korean Defectors Held; Authorities capture groups bound for South Korea and Japan at boat terminal in China', Los Angeles Times, 20 Jan. 2003, Part 1, p.3; and Steven Greenhouse, 'Aristide Condemns Clinton's Haiti Policy as Racist', New York Times, 22 April 1994.
• 4. See, for instance, 'Growing Global Migration and Its Implications for the United States', National Intelligence Estimate 2001-02D (March 2001), p.3; and Kelly M. Greenhill, Constructed Calamities: Engineered Migration and the Coercive Power of Unnatural Disasters (under review).
• 5. Both Weiner and Teitelbaum's own typologies focused exclusively on state-level manipulation of flows – as opposed to state and non-state actor manipulation – and neither considered the strategic use of refugee flows for military ends. See, for instance, Michael S. Teitelbaum, 'Immigration, Refugees, and Foreign Policy', International Organization 38/3 (Summer 1984) pp.429–50; and Myron Weiner, 'Bad Neighbors, Bad Neighborhoods', International Security 21/1 (Summer 1996) pp.5–42.
• 6. See, for instance, Heather Rae, States, Identities, and the Homogenisation of Peoples (Cambridge, UK: CUP 2001).
• 7. Nicholas Van Hear, New Diasporas: The Mass Exodus, Dispersal and Regrouping of Migrant Communities (Seattle: Univ. of Washington Press 1998). See also Andrew Bell-Fialkoff, Ethnic Cleansing (Basingstoke, UK: Palgrave Macmillan 1996) and – for a recent example – see Milcho Manchevski, 'NATO gave us this ethnic cleansing – The Macedonian war is a fight about borders, not human rights', The Guardian, 15 Aug. 2001, p.12.
• 8. See again Hasan Zaheer, The Separation of East Pakistan: The Rise and Realization of Bengali Muslim Nationalism (Oxford: OUP 1994); and Nicholas J. Wheeler, Saving Strangers: Humanitarian Intervention in International Society (Oxford: OUP 2001) Chapter 2:' India as Rescuer? Order versus Justice in the Bangladesh War of 1971'.
• 9. For an examination of the efficacy of this tactic, see Kelly M. Greenhill, 'Draining the Sea, or Feeding the Fire?: The Use of Population Relocation in Counterinsurgency Operations' (under review).
• 10. While the displacements deprived the guerrillas of some sources of food, shelter, and protection, it also compounded the general condemnation of the Soviet invasion, as well as enhanced Western sympathy for the Afghan refugees. See Marek Sliwinski, 'Afghanistan: The Decimation of a People', Orbis 33/1 (1989) pp.39–56.
• 11. Digital National Security Archive, US Dept. of Defense, JCS, 'Defense Intelligence Agency (DIA), Managua, Routine Report to DIA, Washington DC, Subject: Nicaragua: Refugee Exodus', 22 Aug. 1983.
• 12. See again Greenhill, 'Draining the Sea' (note 9).
• 13. Clark Neher, Southeast Asia in the New International Era (Boulder, CO: Westview Press 1994).
• 14. See, for instance, Van Hear, New Diasporas (note 7) pp.236–7. The Israelis were likewise unsuccessful in their attempt to crush the Lebanese guerrillas through the use of militarized migration in spring 1996. See, for instance, Human Rights Watch, 'Operation Grapes of Wrath', Israel/Lebanon 9/8 (Sept. 1997).
• 15. Robert Pape, Bombing to Win: Air Power and Coercion in War (Ithaca, NY: Cornell UP 1996) p.13.
• 16. Ibid. p.19.
• 17. See, for instance, William Hayden, 'The Kosovo Conflict and Forced Migration: The Strategic Use of Displacement and the Obstacles to International Protection', Journal of Humanitarian Assistance; posted 14 Feb. 1999.
• 18. Thomas C. Schelling, Arms and Influence (Cambridge, MA: Harvard UP 1960) p.89.
• 19. As one Yugoslav journalist put it regarding the 1999 offensive in Kosovo: 'there were differences between the police and the army. The police were in favour of expulsions because they could steal money from people. The intelligence guys were against it because they said it was bad for us.' Quoted in Tim Judah, Kosovo: War and Revenge (New Haven, CT: Yale UP 2000) pp.241–2. See also John Mueller, 'The Banality of Ethnic War', International Security 25/1 (Summer 2000) pp.42–70.
• 20. Pape, Bombing to Win (note 15) p.28.
• 21. Ibid. pp.6–7. This makes sense in that perpetrators are often fighting for their very political survival, whereas for targets the issues at hand tend to be of more limited importance.
• 22. Also, for reasons associated with the credibility of the perpetrators ex ante, contrary to traditional coercion, coercive threats tend to be more, not less, credible than deterrent threats. See Pape, Bombing to Win (note 15) pp.6–7.

23. Though weak vis-a`-vis their targets, perpetrators are generally strong relative to their victims.

• 24. See Susan L. Woodward, Balkan Tragedy: Chaos and Dissolution after the Cold War (Washington DC: Brookings Institution 1995); Nik Gowing, 'Real-time TV Coverage from War: Does it Make or Break Government Policy?' in Bosnia by Television (London: British Film Institute 1996); and James Gow and James Tilsley, 'The Strategic Imperative for Media Management,' in Bosnia by Television, p.103.
• 25. On the use of militarized migration by the US, see, for instance, LBJ Library, NSF Vietnam CO File, Folder: Vietnam NODIS, Vol.3(B), 10/56–6/66, 'Department of State Telegram to the President', 15 Sept. 1965. On encouraging the flight of North Vietnamese, see, for instance, George Herring, America's Longest War: The United States and Vietnam 1950–1975, 3rd edition (New York: McGraw-Hill 1996); and LBJ Library, NSF Vietnam CO File, Folder: Vietnam NODIS, Vol. 3(B), 10/56–6/66, 'Department of State Airgram, CINCPAC for POLAD, From US Embassy, Saigon on the Joint GVN Security Council-US Mission Council Meeting', 25 March 1965, p.4.
• 26. At the same time, it must be noted that such overlaps can make disentangling primary and secondary motivations profoundly difficult.
• 27. See, for instance, T.V. Paul, Asymmetric Conflicts: War Initiation By Weaker Powers (Cambridge UP 1994) p. 150.
• 28. See again Neher, Southeast Asia in the New International Era (note 13); and Fiona Terry, Condemned to Repeat: The Paradox of Humanitarian Intervention (Ithaca, NY: Cornell UP 2001).
• 29. For details, see Greenhill, Constructed Calamities (note 4) Chapter 2.
• 30. See also Ivan Arreguin-Toft, 'How the Weak Win Wars', International Security 26/1 (Summer 2001) p.105; and Andrew Mack, 'Why Big Nations Lose Small Wars: The Politics of Asymmetric Conflict', World Politics 27/2 (Jan. 1975) pp.175–200.
• 31. Exceptions to this general rule tend to occur in the context of civil wars e.g., Turkey's use of dispossessive, exportive, and militarized migration against its Kurdish population – and counterinsurgency campaigns – e.g., the militarized use of refugees by the French in the French– Algerian War and by the US in the Vietnam War. Between 1990 and 1999, the Turkish Army burned, leveled or forcibly evacuated more than 3,000 Kurdish villages. There are large swaths of territory in southeastern Turkey depopulated at the village level. By forcing the Kurds out of the rural areas, the government effectively cut off all logistical support to the Kurdish guerrillas which has been an effective military strategy. Frelick interview in 'Refugees as Weapons of War', a program produced by America's Defense Monitor, 17 Oct. 1999.
• 32. See Scott Snyder, Negotiating on the Edge: North Korean Negotiating Behavior (Washington DC: US Institute of Peace 1999). See also Hans Binnendijk, How Nations Negotiate (Washington DC: National Defense Univ. 1987); and Paul, Asymmetric Conflicts (note 27).
• 33. Paul, Asymmetric Conflicts (note 27) p.17.
• 34. Ibid.
• 35. Mark Habeeb, Power and Tactics in International Negotiation: How Weak Nations Bargain with Strong Nations (Baltimore, MD: The Johns Hopkins UP 1988).
• 36. Snyder, Negotiating on the Edge (note 32) p.69.
• 37. Ibid. p.43.
• 38. Alexander George, cited in Lawrence Freedman (ed.), Strategic Coercion: Concepts and Cases (Oxford UP 1998) p.17.
• 39. For instance, launching a war to counter outflows may be an option in certain circumstances, but often the expected costs associated with escalation to that level far exceed the expected costs of conceding to perpetrators' demands in whole or in part. Likewise, if the perpetrator is already internationally isolated, the methods short of war that powerful states may employ in response may be slow-acting (e.g., sanctions) and thus inappropriate as a method of counter-coercion during a crisis. See Daniel Byman and Matthew Waxman, The Dynamics of Coercion (Oxford: OUP 2002).
• 40. Lara Marlowe, 'War and peace revisited', Irish Times, 25 March 2000, p.68.
• 41. Michael W. Doyle, 'Liberalism and World Politics', American Political Science Review 80/4 (Dec. 1986) pp. 1151–69.
• 42. Bruce Russett, 'Why Democratic Peace?' in Michael E. Brown et al. (eds.), Debating the Democratic Peace (Cambridge, MA: MIT Press 1996) p.93.
• 43. Ibid.
• 44. Fiona Terry, Condemned to Repeat: The Paradox of Humanitarian Intervention (Ithaca, NY: Cornell UP 2001), p.34.
• 45. Woodward, Balkan Tragedy (note 24) p.322.
• 46. As Bill Frelick, then of the US Committee on Refugees argued, 'the KLA [Kosovo Liberation Army] which was not in a position to fight a straight out battle between standing armies, used their civilian population as part of its tactic to win international support and to really bring the international community as an ally in their struggle against the Serbs'. From Bill Frelick interview in 'Refugees as Weapons of War'; Kelly M. Greenhill, 'The Use of Refugees as Political and Military Weapons in the Kosovo Conflict,' in Raju G. C. Thomas (ed.), Yugoslavia Unraveled: Sovereignty, Self-Determination, and Intervention (Lanham, MD: Rowman & Littlefield 2003) pp. 205–42; Alan J. Kuperman, 'Tragic challenges and the moral hazard of humanitarian intervention: how and why ethnic groups provoke genocidal retaliation', PhD dissertation (MIT 2002); and Clifford Bob, 'Beyond Transparency: Visibility and Fit in the Internationalization of Internal Conflict,' in Bernard Finel and Kristen Lord (eds.), Power and Transparency in the Age of Transparency (Basingstoke, UK: Palgrave MacMillan 2002).
• 47. Rony Braumann, 'When Suffering Makes a Good Story', in Edward R. Girardet (ed.), Somalia, Rwanda and Beyond: The Role of the International Media in Wars and Humanitarian Crises (Dublin: Crosslines Global 1995) pp.135–48.
• 48. As one Kosovar Albanian demonstrator said in March 1998, a year before the start of the NATO bombing campaign, 'We are going to have to spill a lot more of our own blood before we can expect the outside world to risk getting heavily involved here... But I can't see any other way that we can hope to give better lives to our people. I don't believe Milosevic is ever going to do anything for us or give us any freedom to do it for ourselves'. Geoff Kitney, 'The Killing Fields Of Blackbirds,' Sydney Morning Herald, 21 March 1998, p.41.
• 49. See, for instance, Oliver Ramsbotham and Tom Woodhouse, Humanitarian Intervention in Contemporary Conflict: A Reconceptualization (Cambridge, UK: Polity 1996) p.186.
• 50. Jean-Christophe Rufin, Le piege humanitaire, suivi de Humanitaire et politique depuis la chute du Mur, revised edition (Paris: Jean-Claude Lattes 1993) pp.119–20. As Asprey put it in his seminal work on guerrilla warfare: 'In choosing force, however, the rebels also displayed an arrogance of ignorance... They underestimated both the umbilical cord linking Algeria to France in the minds of the great majority of public opinion in metropolitan France ... and naively failed to realize that France's allies, however disapproving, would neither interfere nor proffer advice until French public opinion had reconciled itself to the eventuality of Algerian independence. Robert Asprey, War in the Shadows: The Guerrilla in History (London: Macdonald and Jane's 1975) pp.982–3.
• 51. Nevertheless, the Algerians did in the end achieve their goal of independence, and it was largely because of the political costs the war inflicted on the French. So perhaps it was not such a failure after all, just a far more costly success.
• 52. See, for instance, Mueller, 'The Banality of Ethnic War' (note 19).
• 53. See, for instance, James Gow, 'Coercive Cadences: Yugoslav War', Chapter 11 in Freedman Strategic Coercion (note 38).
• 54. There is also the danger that expelled populations will return with a vengeance, as, for instance, did the Ugandan-based and Tutsi-dominated, Rwandan Patriotic Front (RPF), which wrested power from Hutu-dominated Rwandan regime in 1994. See, for instance, Gerard Prunier, The Rwanda Crisis: History of a Genocide (New York: Columbia UP 1995).
• 55. Guy Sacerdoti, 'How Hanoi Cashes In: Boat Organizers Tell of Taxes on Refugee Trade', Far East Economic Review, 15 June 1979, pp.23–6. See also Bernard Gwertzman, 'US Assails Vietnam for Refugee Policy', New York Times, 13 June 1979, p.A6.
• 56. See, for instance, 'Sunk', The Economist, 24 Aug. 2006; Volker ter Haseborg, 'Offering Asylum in Chernobyl's No Man's Land', Der Spiegel, 14 October 2005 (online edition); and 'Kim Jong-Il Goes Ballistic', The Economist, 6 July 2006.