Jump directly to page content

Why People Stay

Decision-making in Situations of Forced Displacement and Options for Humanitarian Aid and Development Cooperation

SWP Research Paper 2023/RP 15, 05.12.2023, 30 Pages


Research Areas

Nadine Biehler is an Associate in the Global Issues Research Division at SWP.
This paper was written as part of the research project “Stra­tegic Refugee and Migration Policy”, funded by the German Federal Ministry for Economic Cooperation and Development (BMZ).

  • The proportion of affected populations who flee violent conflict is much smaller than is widely assumed. Many decide to remain in the conflict zones. They are often referred to as stayees.

  • Three groups can be identified. Some people stay voluntarily. Others do so involuntarily, for example because they lack the resources to flee or because violent actors restrict their freedom of movement. Another group acquiesce to their immobility.

  • Little is known about stayees, their needs and the reasons for their im­mobility. But several factors relevant to their decision-making can be identified. These include type of conflict, type of violence and personal situation.

  • Whether they remain voluntarily or involuntarily, stayees employ sur­vival strategies including collaboration, neutrality, protest and resistance.

  • Knowledge about stayees and their survival strategies is important for humanitarian aid and development actors. Only if they are well informed can they align their activities with actual needs and provide meaningful support to people living in and with violent conflicts.

  • It is therefore essential to consider the entire spectrum of (im)mobility and to understand this expanded perspective as a positive – without neglecting the forcibly displaced. The agency of civilians in violent conflicts needs to be recognised and they must be protected from abuse and exploitation by aid workers (do-no-harm principle). Finally, stayees must be systematically included in all post-conflict initiatives supporting vol­untary return and reintegration.

Table of contents

1 Issues and Recommendations

2 Why Are Stayees Overlooked?

2.1 “Mobility bias”: Concentrating on (internationally) mobile groups

2.2 Distinguishing stayees from refugees and IDPs

2.3 Involuntary immobility, immobilisation and “trapped populations”

3 Decision-making in Situations of Forced Displacement

3.1 Cost-benefit models

3.1.1 The functionalistic approach

3.1.2 Limitations of the model

3.2 Macro-level: Type of conflict

3.2.1 Location and geographical extent of conflict

3.2.2 Conventional versus irregular warfare

3.2.3 Foreign involvement vs. local conflict

3.3 Meso-level: Type of violence

3.3.1 Intensity of violence

3.3.2 Selectivity

3.3.3 Intention of violence

3.4 Micro-level: Personal situation and social markers

3.4.1 Sex and age

3.4.2 Income, employment, wealth, education

3.4.3 Access to social networks

3.4.4 Personal and ideological preferences and personality traits

3.5 Interim conclusion

4 Strategies for Navigating Armed Conflicts: Exit, Voice, Loyalty

4.1 Support for and collaboration with violent actors

4.2 Neutrality

4.3 Avoidance by changing patterns of mobility

4.4 Dialogue, protest and resistance

4.5 Self-defence

4.6 Interim conclusion

5 Recommendations for Humanitarian Aid and Development Cooperation

5.1 Consider the entire mobility spectrum and understand the (im)mobility per­spec­tive as added value

5.2 Acknowledge civilian agency

5.3 Do No Harm: Protecting from the helpers

5.4 A comprehensive perspective on return and reintegration

6 Conclusion

7 Abbreviations

Issues and Recommendations

Issues around refugees and forced displacement are high on the political agenda in Germany and the EU. One sign of this is the enormous media interest in migration routes to Europe. There is also increased funding for migration and refugee studies in Ger­many. Asylum and migration policy is hotly debated across Europe, and the potential for political mobili­sation is significant. The debate often divides into two highly polarised camps, with one side demanding solidarity for people forced to flee for their lives, the other regarding refugee movements as an economically driven security risk.

For all the controversy, the two sides share one thing in common: they concentrate on people who are on the move, internationally or within their own country. This “mobility bias” means that people who remain where they live despite war and violence tend to be overlooked, even though they usually represent the majority. One consequence of this is a lack of research into the motives and behaviour of these “stayees”.

The present study investigates why people stay in situations of violent conflict, what factors influence their decisions, and what alternative strategies they pursue. A number of recommendations for development cooperation and humanitarian aid are derived.

The contribution begins by defining stayees and distinguishing them from refugees and internally displaced persons (IDPs). The “aspirations-capabilities framework” from migration research is used to in­ves­tigate the decision-making processes. Three categories of stayee are identified: voluntarily immobile, acqui­es­cently immobile, and involuntarily immobile or immobilised (“trapped populations”).

The literature on forced displacement usually assumes that people in situations of violent conflicts possess agency and options and make rational choices to maximise their safety (by fleeing) – even if the options are generally limited. The present study ap­plies this functionalistic cost-benefit approach to deci­sions to stay. Given the lack of relevant research on stayees, factors that mitigate towards staying rather than fleeing are derived from the research on forced displacement, distinguishing type of conflict, type of violence and personal situation and social markers (macro, meso and micro levels). This offers a basis for explaining decision-making processes (although not to predict the migration decisions of individuals).

The present study supplements the research on forced displacement by supplying insights into the survival strategies employed by those who stay rather than flee. It analyses how people exploit the possibil­ities available to them to secure their survival, in par­ticular how they relate to violent actors and seek to influence them. Civilians in situations of violent con­flict are exposed to great danger and suffering; they undergo traumatic experiences and require protec­tion and (international) support. Humanitarian and development actors need to be aware of the options available to stayees and their strategies for coping with violent conflicts. If they are to provide effective support their interventions must be designed accordingly.

International efforts to improve the coordination of humanitarian aid and development cooperation in such situations have been under way for some time now. Humanitarian aid and development cooperation frequently operate in parallel in countries with pro­tracted violent conflicts, fragility and insecurity, but still often without sufficient coordination. The hu­mani­tarian-development-peace nexus (HDP nexus) has an important contribution to make here.

In the German context, too, supporting civilian populations that remain in situations of violent con­flicts is one of the responsibilities of humanitarian aid – and these days also of development cooperation. Humanitarian aid still tends to dominate, and its approaches and instruments are designed for exactly such situations. Nevertheless, development cooperation plays a growing role today, increasingly employ­ing conflict-sensitive and peace-building approaches. Germany’s transitional development assistance oper­ates as a bridge between humanitarian aid and devel­opment cooperation, also in situations of protracted conflict.

The recommendations laid out at the end of the study are addressed to German and international humanitarian and development actors. The author argues that the entire mobility spectrum in violent conflicts needs to be taken into consideration in order to expand the existing focus on refugees into a broader (im)mobility perspective. In particular, awareness of the different forms of immobility can help to ensure that the needs of the civilian population are better served. That includes respecting and supporting (im)mobility decisions and providing assistance on the basis of vulnerability. In order to do so, external actors need to understand why civilians remain in situations of violent conflict and what survival strat­egies they employ.

The scope for humanitarian and development actors to contribute to physical safety in armed con­flict is often limited – but they can and must protect civilians from violence, abuse and exploitation by their own staff. This is especially pertinent in light of the feminist policies recently adopted by the German Foreign Office and Ministry for Economic Cooperation and Development (BMZ).

The question of preventing violence is also salient in connection with promotion of return and reinte­gra­tion in the Global South. Conflicts between re­turnees and stayees over land, employment and other re­sources can hinder reconstruction. External actors need to be aware of this and adopt conflict-sensitive approaches.

Why Are Stayees Overlooked?

People flee from persecution, war, violence and human rights violations, often experiencing great suffering and injustice.1 At the end of 2022 UNHCR counted 108.4 million forcibly displaced people world­wide,2 a new historic record. Forced displacement rightly receives significant political attention. Nevertheless, those who choose to remain, the so-called “stayees”, tend to be overlooked.3

“Mobility bias”: Concentrating on (internationally) mobile groups

Migration research focusses – practically by defini­tion – on mobile population groups. This phenom­enon is referred to as “mobility bias”4 in academic research.5 The same concentration on the mobile is also found in forced migration and refugee studies:6 those who stay are largely ignored.

Stayees currently receive much less attention than refugees and IDPs.

The reasons for this are primarily political. Countries and communities that host significant numbers of displaced persons incur considerable economic and political costs. Their security situation may deteriorate if the presence of refugees and IDPs contributes to violent conflicts.7 While the forced displacement of recent decades has occurred above all within and between developing countries – where most of those affected still live – more recent movements (in par­ticular since 2015/16) have brought the global refugee emergency closer to the European Union. So the cur­rent interest in refugees and IDPs is driven in part by the concern of European states to avoid further major arrivals, which they fear could have negative conse­quences including overstretching state services, popu­list instrumentalisation and social conflict.

International efforts to address forced displacement have been stepped up since 2015 in response to these developments. At the global level the New York Dec­la­ration of 20168 led to the Global Compact on Refu­gees in 2018.9 The same is true of Germany, with visible results in the shape of a realignment of Ger­man development cooperation to address forced dis­placement, for example through a new funding in­stru­ment, the special initiative on forced displacement (Sonderinitiative Flucht).10

There is also a second, more positive motivation driving the engagement. Situations of forced dis­placement are of great concern from the development perspective: the numbers affected by protracted dis­placement situations have been increasing and many of those affected experience great poverty and suf­fer­ing.11 Many displacement situations remain unresolv­ed as there are no durable solutions available, leaving millions of people facing a life of precarity and un­certainty. This also endangers the prospects of achiev­ing the Sustainable Development Goals (SDGs). Nor is there any doubt that people escaping armed conflict and persecution require humanitarian aid. The inter­national community does provide significant support in protracted displacement situations, for example for refugee camps, host communities, IDPs and especially vulnerable groups.

As a result stayees tend to be sidelined both in political discourse and in development cooperation and humanitarian aid.12 Little is known about them, their needs and their motives.

Distinguishing stayees from refugees and IDPs

Migration research (in contrast to refugee studies) does offer some starting points on the question of immobility. Carling lays out an “aspirations ability approach” to involuntary immobility.13 De Haas14 and Schewel15 develop this into an “aspirations-capabil­ities framework”, distinguishing between the wish to migrate and the ability or opportunity to do so, and including the question of immobility.

According to Carling the intention to migrate is influenced both by the macro level environment (the social, economic and political context) and by indi­vidual wishes at the micro level. Those who pos­sess both the intention and the means become (inter­national) migrants.16 Immobility on the other hand, according to Schewel, means either that the ability to migrate is lacking (for example because of political and legal restrictions, lack of funds or lack of social capital) or that immobility is preferred or at least acquiesced to – whether as an explicit wish or on account of internal constraints and/or social roles.17 Schewel defines immobility as “spatial continuity in an individual’s center of gravity over a period of time”.18

The same framework can be applied to forced dis­placement.19 The wish to escape the threat of war and persecution corresponds to “aspiration to migrate” in the figure, while the factors affecting the possibility to do so map to “capability to migrate”.


Source: Based on Kerilyn Schewel, “Understanding Immobility: Moving beyond the Mobility Bias in Migration Studies”, International Migration Review 54, no. 2 (2020): 328–55 (331), doi: 10.1177/0197918319831952.

In light of the above, we distinguish three groups of stayee: Those who choose to stay (voluntary im­mo­bility), those who would like to leave but cannot (in­voluntary immobility), and a third category who are neither willing nor able to leave but tolerate their immobility (“acquiescent immobility”).20 The latter attitude is frequently found among older people,21 as well as resource-poor groups living at a relatively safe distance from any fighting who therefore lack both the motivation and the ability to leave.

It is important to note that while such categorisations are useful for understanding immobility, they are also a great simplification. None of the combi­nations of intention and ability are purely binary in reality, and are better conceived as a spectrum on which a person’s position may change over time.22 There are also interactions between the two axes, for example when a person’s migration opportunities are so constrained, for example by poverty, that the possibility of fleeing does not even occur to them. Nevertheless, the model’s matrix of mobility aspi­ra­tions and capabilities is expedient for understanding and discussing the issues.

Involuntary immobility, immobilisation and “trapped populations”

The spectrum of involuntary immobility and im­mo­bilisation is broad. It includes those living in camps under access restrictions and/or curfews. Pro­tracted refugee situations can be tantamount to im­mo­bility,23 where individuals cannot return to their country of origin but have no possibility of local integration or resettlement in a third country.24

The term immobilisation is frequently applied where freedom of movement is intentionally re­stricted as part of the strategy of armed actors (rather than resulting from more or less random circumstances). One example of this occurred in 2020 in Colombia, when the Ejército de Liberación Nacional (ELN) guerrilla organisation enforced curfews using landmines and threats of violence. The United Nations described this form of immobilisation as “confine­ment” and estimated that more than 70,000 people had been affected. The tactic offers several advantages for the violent actors. The confined population is easier to control and can be used as a human shield, and the region and its resources can be more easily exploited.25

Immobility can also occur as a side-effect of war and violence. The Mozambican civil war of 1977 to 1992 provides a good example. Here fighting over con­trol of territory and borders prevented traditional labour migration and made it extremely hard for those affected to make a living.26

Criminal violence can have the same effect. In Haiti in 2022 fighting between gangs27 left thousands of residents of the capital Port-au-Prince cut off from food, drinking water and medical care.28

The Russian invasion of Ukraine in February 2022 also created a broad spectrum of involuntary immo­bil­ity. While the situation of civilians trapped in besieged cities like Mariupol in summer 2022 was widely re­ported, the selective travel ban for Ukrainian males aged between 18 and 60 has attracted less atten­tion. This group represents a special case of im­mo­bilisa­tion: they are unable to leave the country and their freedom of movement within it is also restricted.29

Whether involuntary immobility is enforced by vio­lent actors or a product of circumstances, it is usually associated with heightened vulnerability, not least because those affected lack the possibility, resources and/or ability to flee from danger.30 Yet these individuals are frequently overlooked.31

The diversity of forms and causes of involuntary immobility is reflected in the terminology.

The diversity of forms and causes of involuntary immobility is reflected in the terminology. As well as the widely used “involuntary immobility” and “immobilisation”, “confinement” (see above) and “trapped population”32 are also found.33

It is difficult to distinguish conceptually between voluntary, involuntary and acquiescent immobility. As with voluntary and involuntary migration, the differences are better understood as a matter of degree rather than hard category. Every form of im­mo­bility is associated with a certain level of compulsion and choice.34

The lack of knowledge about stayees often makes it difficult to differentiate in practice between volun­tary, acquiescent and involuntary immobility. That would require knowledge about their (im)mobility capabilities and aspirations that is rarely available.

Decision-making in Situations of Forced Displacement

We intuitively understand flight from armed conflict as the obvious and almost inevitable response, essen­tially the norm. Research confirms that violence is generally the immediate trigger of forced displacement;35 or, put differently, displacement is a response to violence,36 often in fact the only possibility for the civilian population to protect themselves and save their lives.37 Media reporting also tends to portray the forcibly displaced as traumatised victims who have no choice but to flee. That contrasts with the empirical observation that many people actually stay rather than leave,38 even in cases where others are fleeing.39 In fact it is not uncommon for a majority to decide to stay.40 It is rarely acknowledged that people choose consciously and freely to remain in conflict zones, and too little is known about the circumstances and factors that influence their decisions.41 This research gap is astonishing,42 given that stayees generally out­number refugees and IDPs. The fact that so many remain in conflict zones despite danger and hardship is plainly in need of explanation.

Cost-benefit models

The “aspirations-capabilities framework” (see above) offers a basis for distinguishing stayees from other, mobile groups and a theoretical framework for ana­lysing mobility and immobility in contexts of violent conflict. It permits the external circumstances and individual wishes to be considered, along with their interactions. However, the lack of research on stayees in violent conflicts and the sparsity of information on their (im-)mobility aspirations and capabilities means that the framework and all its nuances can not be ap­plied here. Instead we must resort to more simplified functionalistic explanatory models.43 Additional research on staying is plainly needed, but hard to realise because of the dangers associated with work­ing in active conflicts zones.

The functionalistic approach

The dominant decision theory models used to explain decision-making44 in forced displacement situations assert that people possess agency and options even amidst violent conflict, and even if they are severely constrained by violent actors or freedom of choice extends no further than being able to choose the lesser evil.45 This approach presumes that civilians are rational actors who weigh the costs and benefits of dif­ferent courses of action, and that their cost-benefit calculations will lead them to decide to leave in the event of mortal danger.46 The same decision theory approach can also be applied to stayees and their deci­sions. The assumption here is that personal safety takes priority47 and that people decide to leave if the costs of doing so appear smaller than the costs of stay­ing.48 This is different from the neoclassical expla­nations of voluntary migration, which are based on maximisation of economic utility.49

The present study brings together factors and vari­ables identified in quantitative and qualitative re­search as influencing decisions to stay or leave, and examines and contextualises their applicability to stayees, in order to better understand why stayees decide to stay.

The factors that influence these decisions include type of conflict. Is this a conflict between states or a civil war? What type of warfare is involved? The type of violence also plays a role.50 How is it exercised; is the intensity high or low? Is violence targeted against individuals or is it indiscriminate? Is displacement the objective? Finally, the existence of stayees sug­gests that even a functionalistic cost-benefit model produces different individual outcomes. If only some of the residents of a conflict zone flee, we can con­clude that personal factors such as age, gender and membership of political and/or social groups influ­ence decision-making.51 Using Raleigh’s categories of levels,52 we can distinguish macro level (type of conflict), meso level (type of violence) and micro level (personal situation and social markers).53

Limitations of the model

While the functionalistic model permits analytical investigation of stayees’ decisions, it is subject to a number of limitations in practice.

First of all, the cost-benefit model assumes that those affected are able to assess the repercussions of their decision to stay or leave for their future safety.54 That presupposes adequate information, which is not always available.55 Although people in conflict situa­tions have strong incentives to observe their surround­ings very closely in order to make well-informed decisions,56 important information may be hard or impossible to obtain, for example where disinformation is disseminated. Psychological responses to violence represent a second aspect that plays a role in decision-making but is inadequately reflected in rational choice approaches. They may include freez­ing in fear, passivity and/or obedience, even where a different reaction would appear “more sensible”.57

The assumption that personal safety will be pri­oritised over all else is another oversimplification. It tends obscure other reasons to stay, for example where personal convictions lead to disobedience and resistance or a person stays in order to care for a rela­tive. The role of political, ideological and “patriotic”58 motives also needs to be considered.59

The cost-benefit model also assumes that decisions are made by individuals in isolation. In fact they must be understood in their social context,60 and in the con­text of the decision-making systems of families and households.61 Echoing the phenomenon of risk diversification in voluntary migration,62 it is observed that some household members remain in contexts of conflict while others leave,63 for example in order to secure an additional source of income elsewhere while protecting property at home from looting. A model based on individual decisions can only pay limited consideration to such factors.

The distinction between staying and leaving is not always clear cut in reality.

The distinction between staying and leaving in the model is not always so clear cut in reality. Dealing with instability and uncertainty is simply part of the daily routine for many people affected by armed con­flicts, where every escalation represents a deterioration that they have to cope with. Their response must not necessarily be immediate flight.64 However, this does not preclude mobility – in the sense of tem­po­rary or intermittent change of residence – as a sur­vival tactic. In other words, the distinction between stayees, IDPs and refugees does not necessarily accu­rately reflect their lived reality.

Nor does the model permit us to weigh the influence of individual factors on decisions to stay or leave.65 This applies for example to economic factors such as growing impoverishment in conflict situations.66 At the same time researchers agree that vio­lent conflict is the principal trigger of flight.67 The forms of violence and their consequences for safety are therefore assumed to be central to the decision to stay or leave. While the individual factors of income, wealth and education do feature in the present study, the potential influence of conflict-related economic decline does not.

Despite these shortcomings the cost-benefit model does offer a good starting point for understanding stayees’ decisions. What the model cannot do is pre­dict whether and when an individual will stay or leave.

Macro-level: Type of conflict

The type of conflict is an important factor influencing the danger to the civilian population, the size of migratory movements and ultimately also the num­ber of stayees.68 The specific local manifestation of conflict and violence also influences decision-making and survival strategies.69 For example the effects of conventional war between states on the civilian popu­lation differ from those of civil war and other kinds of armed conflict. Even if research is not always un­am­biguous, a number of observations can be made.

Location and geographical extent of conflict

Violent conflicts cause people to flee primarily where the fighting occurs. So citizens of countries involved in conflict are not necessarily all stayees; the classifi­cation applies only to those who remain in countries where fighting is occurring.70

The example of the Russian invasion of Ukraine in February 2022 illuminates this aspect. The number of people fleeing Russia has been comparably small and their departure was not triggered directly by the fight­ing. Those who have left have been individuals who opposed the war or objected to increasing restrictions on their civil rights and liberties.71 After Russian Presi­dent Vladimir Putin announced a partial mobili­sation at the end of September 2022 several hundred thousand men left to avoid being mobilised.72 Despite the dangers experienced by political dissidents in Russia, the civilian population’s risk of experiencing war-related violence is small because there has to date been no significant fighting on Russian territory. As such neither they nor the men who remain in Russia despite the partial mobilisation are regarded as stayees – unlike those who remain in Ukraine.

Fundamentally the larger the area and the number of cities affected by fighting the more forced displace­ment is to be expected. In fact those two factors influ­ence the scope of forced displacement more strongly than the intensity of the fighting.73 The larger the area the larger the number of people affected, who usually divide into a (smaller) group who leave and a (larger) group who stay.74 If the conflict-affected area is small the decision to stay may be comparatively easy for those who live in the same country but far from the fighting. In fact if the fighting is distant enough the question of fleeing may not even arise; these individuals fall under the category of immobility-acquiescent stayees. But if larger areas are affected – or a person’s place of residence – the decision to leave may appear more obvious.

Conventional versus irregular warfare

Conflicts conducted with irregular tactics, such as guerrilla warfare (in contrast to conventional wars fought by organised and identifiable armies), present additional risks to the civilian population. Civilians may be drawn directly into the fighting for example where combatants shelter among the population or rely on their support. This may make the civilian popu­lation a target, especially where it becomes hard to distinguish between fighters and civilians.75

Irregular warfare can turn civilians into targets and increase the civilian casualty toll. Under this logic the civilian population is attacked to harm the non-state armed groups who depend on the civilian population for support.76

Finally, in so-called “new wars” the civilian popu­lation is deliberately attacked and human rights are violated on a large scale.77

It follows that conflicts fought by irregular means involve a greater likelihood that civilians will become targets of violence, whether intentionally or uninten­tionally. In conventional conflicts, where the civilian population is not directly involved in the fighting, this is less likely to be the case and staying is more plausible. Given the predominant forms of contempo­rary armed conflict, this is not possible in most cases.

Foreign involvement vs. local conflict

As Schmeidl observes, foreign involvement in a con­flict tends to be associated with escalation and greater numbers of refugees.78 One example of this is the long civil war in Somalia and the exodus from Moga­dishu in 2007/08. Here Ethiopia’s intervention in the conflict was one of the reasons why so many residents ultimately decided to flee the city even though they had remained, in some cases for decades, through po­litical instability and more or less intense fighting.79

Raleigh’s quantitative research finds few cases where inter-communal violence is the direct cause of forced displacement. But she does find that this type of violence leads to the impoverishment of rural com­munities and consequently to greater migration to the cities. This form of chronic low-intensity conflict is endemic in nomadic pastoral societies in parts of Somalia, Ethiopia, Kenya, Uganda, and the northern Sahel region.80

The implications for decision-making on staying or leaving are as follows: a civilian in a situation of con­ventional conflict, where identifiable armed actors fight in a defined geographical region, is at less risk than one who is dealing with violent actors demand­ing support and thus making the civilian population a military target. The risk to the civilian population also increases where third states intervene militarily. On the other hand, armed conflicts conducted by clearly identifiable forces and irregular conflicts where civilian support is not relevant are associated with less risk to stayees.

Meso-level: Type of violence

Alongside the type of conflict, the strategy and tactics of the parties and the forms of violence they use also influence civilians’ decisions to stay or leave. One fun­damental tenet is that violence is employed inten­tionally as a means to an end,81 and is not haphazard or an end in itself. In other words, in an armed con­flict violence is used against the civilian population (or parts thereof) to gain strategic advantage.82 Here it should be noted that simply threatening violence may already have the desired effect.83

In relation to type of violence, three factors contrib­ute to decisions to stay or leave: intensity, selec­tivity and intention of violence.

Intensity of violence

The intensity of violence varies within affected coun­tries; its level is not uniform. Using the ACLED data­base (Armed Conflict Location and Event Data), Raleigh and colleagues find that: “The average per­centage of area covered by civil war … is approximately 48%, but the average amount of territory with repeated fighting is considerably smaller at 15%.”84 This creates zones of different intensity of conflict and violence, which in turn lead to correspondingly larger or smaller movements of people. The most dangerous places are the front lines and contested territories. In less affected regions most of the popu­lation stays.85

Intensity of violence is less important for decision-making than the geographical extent of the fighting,86 but it does still influence behaviour. The number of people fleeing increases with intensity of violence, but not in a linear fashion.87 Bohra-Mishra and Massey observe this phenomenon in their research on the Maoist insurgency in Nepal (1996–2006), where num­bers fleeing increased sharply when a particular intensity of violence was exceeded. Low and medium levels of violence were associated with comparatively low levels of forced displacement. Their explanation for this was that mobility in armed conflicts is asso­ciated with risks: roadblocks and curfews, general insecurity, looting, robbery and assault. Civilians can escape these dangers by staying close to home and limiting their radius of movement. Only when vio­lence exceeds a critical threshold do the immediate dangers outweigh the risks of fleeing.88

Lindley documents this effect in the Somali capital Mogadishu in 2007/08, during an escalation of the decades-long civil war. Until that point the level of violence had been relatively low and not led to a mass exodus. From the population’s perspective it was possible to stay as long as one remained vigilant and avoided particular places. But as the violence in­creased in intensity it also became less predictable.89 Urban fighting and arbitrary shelling of residential areas90 caused increasing numbers of civilian casual­ties and made it impossible to move around the city.91 About two-thirds of Mogadishu’s residents had left by the end of 2008, mainly due to the intensity of the violence.92

A conflict’s geography may also produce threshold effects. There is generally less fighting and violence in regions that are controlled by a single group. In such cases the risks associated with leaving will often appear greater than the dangers of staying. Where territory is actively contested the fighting becomes unpredictable for the civilian population and the heightened risk can motivate people to leave.93

Staying may be the safer option both where the level of violence is low and where it is very high.

Interestingly staying may be the safer option both where the level of violence is low and where it is very high. In the first case because the danger is small, in the second because the risks of leaving are too high. So for the civilian population the preference to stay is connected to the geographical extent of the conflict and the changing shape of the front lines. During intense fighting people stay in place for safety. If one side emerges victorious the level of violence falls again, and staying appears less risky than leaving.


Another factor affecting decisions to stay or leave is the question of how selectively the conflict parties use violence and above all whether it is targeted against individuals on account of their behaviour94 or ethnicity,95 or simply indiscriminate. Those aspects decide whether an individual has a realistic chance of avoiding violence.

Violent actors have various reasons for using violence selectively. It can be an effective means of intimidation to force civilians to cooperate and to elicit information, especially for forces whose control of territory is weak, leaving them lacking the re­sources that stronger groups can use to incentivise coopera­tion.96 So where armed actors are weak, civilians must expect an elevated level of violence. Civilians who remain in uncontested areas under the control of strong armed groups face comparatively less danger.

Stronger violent actors whose control of territory is firm frequently possess comprehensive information about the civilian population. They are able to use violence selectively, identifying specific individuals as targets. This enables them to intimidate the population and shift the balance of power in their favour.97 Killing local leaders and putting corpses on display serve to intimidate the population and break resist­ance.98 The use of violence can also change over the course of a conflict, for example if violent actors come under military pressure and are less able to target their violence. Civilians who lived in the Iraqi city of Mosul in 2014–2017 reported that the violence of the fighters of the so-called Islamic State became increasingly indiscriminate as time went on, in par­ticular as they faced loss of control of the city.99

Conflict parties also use selective violence to govern and administer, rather than ruling captured territories by force alone. In that case sanctioning crime and enforcing rules is a strategy to persuade the population to take a positive view of the violent actors.100 This also reduces the dangers of living in a conflict zone, and civilians’ willingness to remain may increase if it appears possible to reduce the danger by adapting one’s behaviour (for example obeying rules).101

These observations on the selective use of violence show that staying is safer than leaving if one side has clearly gained the upper hand, if violence is used selec­tively, and for persons who do not belong to a group targeted on the basis of behaviour, background or (perceived) attributes. As soon as one of these con­ditions is absent the cost-benefit calculation changes and the incentives to leave increase.

Intention of violence

Finally, intentionality of violence is an important factor. Although leaving is frequently regarded as a consequence or side-effect of fighting, forced dis­place­ment can also be a deliberate strategy in various types of conflict. The purpose of deliberate expulsion of civil­ians is to harm the enemy and/or to eliminate specific unwanted population groups.102

Displacement is sometimes a deliberate strategy.

Especially in civil wars, where conflict parties often rely on civilian support, expulsion of civilian popu­lations can be a severe blow to the other side.103 In the civil war in Aceh, Indonesia, the army conducted expulsions in 1999–2002 to isolate the separatist Free Aceh Movement (Gerakan Aceh Merdeka, GAM) from its civilian base.104 The Ugandan government forced IDPs to live in camps during its war against the Lord’s Resistance Army (LRA) in the mid-1990s and again in 2004 in northern Uganda. The objective was to deny the rebel organisations access to labour.105 Forced depopulation of regions and cities by shelling or bombing can also be part of a military strategy, as for example practiced by the Russian armed forces in the Chechen capital Grozny during a conflict that lasted from 1994 to 2009.106

Deliberate forced displacement also occurs in cases of genocide, for example in Rwanda in 1994, when almost one-third of the population had to leave their homes.107 This form of forced displacement, where an entire group is forced to flee on the basis of ethnicity (frequently accompanied by mass atrocities), is often cynically referred to as “cleansing” as for example in the Bosnian war in 1992–1995.108

But supporters of the opposing side have also been expelled when territory changed hands in political conflicts. In the civil wars in Spain (1936–1939) and Colombia (since 1964) graffiti and leaflets were used to inform specific parts of the civilian population that they should leave.109

Regardless of the motives of the violent actors, expulsion campaigns increase the risk of remaining. That applies in particular to those whose group mem­bership puts them at risk of being attacked or ex­pelled. On the other hand the risks involved in fleeing are unlikely to be worthwhile for those who are not tar­geted for expulsion, or who identify with the armed actors controlling the area they live in.

Micro-level: Personal situation and social markers

The risks of staying or leaving vary depending on the individual characteristics of the civilians affected by violence.110 Where violence is directed against the civilian population, specific individuals and groups are at greater risk of being attacked by armed groups, for example on grounds of their (group) identity or political affiliation, and therefore have a stronger incentive to flee.111 It follows that others will be able to stay specifically on account of their personal char­acteristics. In this context some characteristics are especially salient. In the event of a violent conflict, certain identities will be more important than others. Regardless whether identities are ascribed or self-selected, any characteristic that places an individual on one side of the conflict will be of relevance to the violent actors.112 Individuals and households have to assess their specific level of danger, and decide whether they can improve their safety by staying or leaving. Those who stay believe that the risk of experiencing violence is smaller and perceive a lower level of per­sonal risk, for example because their specific profile is not targeted.113 The group who feel personally safe enough to stay also includes those who possess influ­ence over violent actors and those who may even stand to profit from the conflict for example by ex­pand­ing their income, wealth or political power.

Sex and age

Sex114 and age are decisive factors for civilian safety. (Young) men are more likely to be the targets of lethal atrocities while women tend to experience non-lethal forms of violence. Men of fighting age are usually regarded as potential fighters and thus legitimate tar­gets, unlike women, children and older people.115 How­ever, the heightened risk of rape and sexualised violence for women in situations of conflict can also expedite a decision to leave.116 Young adults, male and female alike, are at greatest risk of violence or forced recruitment and therefore most likely to flee.117

Research into the insurgency of the Communist Party of Nepal (Maoist Centre) (1996–2006) demonstrates that the likelihood of deciding to flee decreases with age,118 possibly because older people are less likely to be perceived as a potential threat by violent actors and are therefore exposed to less risk, or be­cause age-related frailty prevents them from leaving. It is known from migration research that older people are less mobile and less willing to migrate than younger people; longer residence leads to stronger attachment and less propensity to emigrate.119 In combination with the aforementioned smaller risk of experiencing violence, this effect may also play a role in contexts of conflict.

In the Sri Lankan civil war, when the separatist Liberation Tigers of Tamil Eelam (LTTE) fought for an independent Tamil state (1983–2009), it was the women and older people who took agricultural pro­duce to the markets because it was easier for them to pass the army checkpoints. Young men quickly fell under suspicion of belonging to the rebels and were therefore at risk of arrest. They also had to reckon with being forcibly recruited into the LTTE. For those reasons they avoided checkpoints wherever possible.120

Nobody is safe in an armed conflict, but the inten­sity of danger varies. Individuals have to weigh up what their gender and age mean for their safety, in the specific conflict and in light of the strategies of the violent actors, and whether they are therefore able to stay or should better flee.

Income, employment, wealth, education

The findings on factors such as income, education, wealth and education in relation to staying or leaving appear at first glance contradictory.

Refugees from conflict zones generally exhibit comparably high levels of education and wealth. People with low incomes and no savings are frequently simply unable to afford to flee and therefore invol­un­tarily immobile.121 Because they are often dependent on agriculture they have to return at intervals to tend their crops, or even to stay for the duration of the farming season.122 During the Sri Lankan civil war it tended to be the wealthier Tamil and Singhalese farm-owners who left the conflict zone.123 An­other possible reason for the wealthier to flee is that their wealth can make them targets.124 Alongside income and wealth, particular kinds of employment can also be associated with higher risk. In Mosul for example, senior government officials were targeted by the so-called Islamic State when it controlled the Iraqi city from 2014–2017.125 In Mogadishu journalists were at particular risk from 2007, when they were threatened and attacked by all the warring parties.126

The example of Mogadishu also shows how relative wealth can make it possible to stay. Those who pos­sessed connections and economic resources on account of belonging to a powerful clan were not only able to stay in the city through the decades of conflict but in some cases even profited financially. Only when it became harder to do business from the end of 2006 did many of them find themselves forced to flee after all – after they had stuck it out for more than ten years.127

Several studies on the conflict in Nepal from 1996–2006 contradict the above findings in different ways and suggest that the question is complex.128 While their findings are only partially comparable, they conclude that people whose assets are less trans­portable and whose skills are harder to apply else­where are more likely to decide to stay.129 Change over time and categories that overlap in reality also offer possible explanations for the disparate findings.

Altogether the empirical findings show that economic factors such as income and wealth are not static; they change over time and in relation to con­text and conflict dynamics. Here again the details are crucial: a stable income can create an incentive to leave at an early stage – or to stay, potentially pur­chasing security.

Access to social networks

Membership of social groups, political parties, ethnic groups and clans can also influence decisions to stay or leave, regardless of whether these are self-selected political or ascribed ethnic categories. Membership of a group can offer protection, participation in deci­sion-making processes, and economic influence.130

For example ethnic Javanese fled during the sepa­ratist insurgency in Aceh, Indonesia. Their ethnicity made them targets but their networks outside Aceh also helped them to escape.131 In the Sri Lankan civil war, the more affluent Tamils and Singhalese in the conflict zone used ethnic or family contacts in other parts of the country or abroad to get out of the con­flict zone.132

Social networks facilitate emigration, but may also enable individuals to remain.

It is known from migration research that social net­works facilitate emigration.133 Such networks need not necessarily be tied to ethnicity and can also be social in nature.134 Networks outside the conflict zone can also facilitate flight by supplying information about potential destinations, contacts and resources.

Although conflict alters local networks, they may still possess the potential to enable individuals to stay in safety.135 Membership of social networks is there­fore a context-dependent factor in decisions to stay or leave. If a person recognisably belongs to the oppos­ing side, for example on the basis of their language or religion, and has contacts in the diaspora they can improve their security by leaving. But a person who is seen as belonging to the same ethnic group as the dominant armed actors may even be able to profit economically from staying.

Personal and ideological preferences and personality traits

Personal and ideological preferences and personality traits are hard to pin down. They include feelings of place attachment and patriotism, as described by some of those who stayed through the siege of the Bosnian capital Sarajevo during the 1992–1995 civil war.136 One reason why it is so hard to distinguish in practice between toleration of mobility restrictions and re­main­ing voluntarily is that these may be retrospective narratives whose function is to rationalise choices after the fact.137

There is another reason for staying in place after a change of power that should not be underestimated, namely, when the residents welcome the new regime. In Mosul in 2014 parts of the civilian population wel­comed the so-called Islamic State because they reject­ed the Iraqi government or felt that the Islamic State would make a better job of running the city.138

Personal preferences, such as ideological or prag­matic approval of new rulers, can improve civilians’ secu­rity. Openly demonstrated loyalty can (some­times) make staying safer. The situation is similar with patriotism: if it leads to – in their eyes credible – support for the violent actors it can contribute to im­proving security. But again, it does not automatically enhance the safety of the individual.

Reasons of this kind for staying, which do not fol­low the logic of maximising security and therefore appear “irrational”, are regularly flagged in qualitative studies and should not be neglected.

Personality traits such as risk tolerance/aversion also play a role. In north-eastern Nigeria, which is affected by the conflict with the Islamist group Boko Haram, it is reported that especially risk-averse people tend to flee, while the more risk-tolerant remain. That is the opposite of the situation with voluntary economically motivated migration.139

Interim conclusion

We can draw two obvious conclusions from these em­pirical findings. Firstly, that the cost-benefit model cannot adequately explain behaviour in situations of conflict; secondly, stayees tend to be risk-tolerant by nature, but justify their behaviour retrospectively in terms of ideological motives. Post-conflict compensa­tion claims for example are frequently justified in terms of suffering experienced. Accordingly, narra­tives emerge that both leavers and stayees use to situate their own behaviour in broader, often also ideological rationales. Both conclusions point to places where further research would be warranted.

Stayees have options; their agency should be acknowledged.

Applying the cost-benefit model to stayees shows that they possess (heterogeneous) options and should be taken seriously as actors. At the same time the levels and factors described supply pointers as to which aspects play a role in which conflict contexts.

The described levels and factors cannot always be clearly differentiated in reality, however. Nor are the factors influencing decisions all equally relevant in all contexts and for all affected individuals. Further­more, interactions exist between the three identified levels. For example personal characteristics and social markers are largely irrelevant in a conventional air war. Even if attacks on civilian targets occur in such situations, social markers tend not to be terribly sig­nificant and the civilian population is not (usually) directly involved in the fighting.

Because the factors discussed here are derived from forced migration studies, the existence of other fac­tors that are only relevant to stayees – and therefore do not feature in investigations of decisions to flee – cannot be excluded. Contradictory findings on eco­nomic factors suggest that the specific context is decisive.

For humanitarian aid and development actors the present study can do no more than offer an overview and an initial orientation. There is a need for further research to investigate the weighting and interaction of the factors. Absent such evidence, policy-makers in donor countries and humanitarian and development actors operating on the ground will have to understand the context and background as best they can and offer context-sensitive support on that basis.

Strategies for Navigating Armed Conflicts: Exit, Voice, Loyalty

Flight can be a strategy to ensure personal survival, but is plainly not the only one – as demonstrated by the sheer number of stayees and the ways they live and survive in conflict zones.

The strategies found in armed conflicts can be clas­sified along a spectrum inspired by Hirschman’s140 systematisation of the responses to decline exhibited by social entities such as firms, organisations and states. As well as “exit” (flight as the response rejected by stayees), between the extremes “voice” (resistance) and “loyalty” (collaboration), we also find attempts to remain neutral or to evade violent actors by modifying patterns of mobility.

Support for and collaboration with violent actors

There are cases where civilians freely support and collaborate with violent actors. In Aceh for example many civilians in the rebel territories expressed their support through graffiti and flags and by providing shelter and supplies to the fighters.141

Support for violent actors can take different forms and occur for different reasons. Collaboration in exchange for protection is one possibility, regardless whether this occurs voluntarily or under pressure.142 In the civil war in Uganda for example civilians co­operated with the Lord’s Resistance Army by oper­ating as informers, carrying out small tasks, or lever­aging family ties to LRA officers to obtain money in exchange for providing shelter, food and information. But in many cases the civilian population collaborated out of fear of torture and death.143

Life goes on in armed conflicts, albeit under altered circumstances.

Life goes on even in violent conflicts, albeit under altered circumstances. The civilian population has to find ways and means to cope with the uncertainty of the new situation.144 That does not necessarily mean lawlessness and chaos. Armed groups are aware of the importance of civilian cooperation and collaboration, and work to create a sense of order and security by clamping down on crime,145 establishing efficient civil­ian administrations complete with the requisite bureaucracies,146 and collecting taxes in order to build both institutions and legitimacy.147

Civilians are not always innocent victims. They may betray others to save themselves, or exploit the conflict to settle old scores.148 Some even manage to profit from the conflict, potentially making themselves complicit.149 Certain civilians succeed in deriv­ing economic benefits through connections to local politicians and military rulers. Ethnicity and socio-economic status play a role here, as does the geo­graphy of the front lines.150 More or less voluntary sexual and romantic relationships with commanders, officers and other influential males represent a relatively gender-specific collaboration strategy for women, both for protection and for economic gain.151

As described above the risk involved in collaboration or its refusal depends heavily on the violent actors and their strength.152


People living amidst armed conflict often pursue a strategy of neutrality.153 This may adopt various forms: adapting to the new rulers and their rules; or attempt­ing to draw as little attention as possible by withdraw­ing from political and economic activities; and where encounters are unavoidable underlining one’s own neutrality or if necessary “playing dumb”.154

Even if they feel ideologically drawn to the violent actors civilians have to weigh the benefits of collabo­ration against the risks. The prospect of retribution can make pay, loot and other material gains unattrac­tive – quite apart from the risk of being killed in the fighting. It must be assumed that order, security and functioning basic services are usually more important for the civilian population.155

Seen from outside, strategies of inconspicuousness and neutrality may appear easy to realise. But that is not necessarily the case. Deliberate non-participation can be associated with costs and risks more or less comparable with those of active participation in fighting.156

Collective neutrality strategies are complex and visible. Sometimes villages or neighbourhoods suc­ceed in preventing violence and attacks by declaring themselves neutral, preventing recruitment and negotiating with violent actors. In her case studies on “islands of peace” in inter-communal conflicts in Ambon, Indonesia (1999–2002), and Jos, Nigeria (2001–2016), Krause attributes decisive importance to individual leaders and leadership groups.157 Co­hesion and organisation are also identified as impor­tant success factors for collective civilian peace efforts.158 Nevertheless such efforts are not always (per­manently) successful.

Neutrality of the civilian population is rarely in the interest of the conflict parties, which employ violence to enforce loyalty. Under such circumstances neutral­ity can even become a form of resistance.159 In any case, the costs of neutrality for the civilian population increase when conflict parties use violence. This can lead civilians to take sides, either under coercion or in the expectation of protection.160 Violence can also con­tribute to the polarisation and radicalisation of pre­viously neutral individuals, potentially leading them to support the other side. Here the strategies of neu­trality and inconspicuousness encounter their limits.

Avoidance by changing patterns of mobility

When facing threat in violent conflicts, many attempt to avoid the gravest risks temporally or spatially. In Sri Lanka poorer households that could not afford to flee any great distance and whose livelihood was tied to their place of residence avoided the fighting by flee­ing temporarily into the forest or neighbouring villages and returning as soon as the situation per­mitted. Those who no longer lived in their home villages also returned temporarily to tend the crops on which they depended. Altered mobility in this case also involved avoiding areas declared off limits by the rebels, such as the forest, and paying small bribes to pass checkpoints.161

In northern Uganda the civilian population sought to minimise the risk of encountering armed actors through temporary avoidance: for example hiding out or spending the night in makeshift shelters in the forest. They were always ready to flee at short notice, and used changing routes to their hiding places to avoid leaving detectable trails. Another especially striking strategy from that conflict was so-called “night commuting”. At the height of the conflict in 2003–2005 tens of thousands of children and young people walked to the town centres every night to sleep in shelters, bus stations and hospitals where they were safe from LRA attacks and kidnappings.162

Altered patterns of mobility in Mogadishu included staying at home, only moving about the city when absolutely necessary, avoiding places with fighters or fighting, and even relocating within the city.163

As well as completely new patterns of mobility, necessitated for example by checkpoints, or avoiding mobility altogether, certain coping strategies in this category are closely related to flight and especially to internal displacement. Here Schewel’s definition of immobility as “spatial continuity in an individual’s center of gravity over a period of time” is helpful.164 What differentiates the strategies discussed here from forced and internal displacement is that the duration and/or distance is shorter, and the “center of gravity” remains unchanged. This initially temporary and geo­graphically limited mobility can change over time, for example if the conflict escalates. It may then also become part of the preparations for flight or internal displacement.

Dialogue, protest and resistance

The civilian population is not necessarily passive and helpless, but may also try to enter into dialogue with violent actors in order to reduce violence. Local lead­ers are frequently involved, despite the associated per­sonal risks.165 In the separatist conflict in Aceh, Indo­nesia (1976–2005), village heads emphasised their neutrality in order to negotiate with the conflict parties.166

Examples of peaceful resistance and protest are re­ported from many contexts. In Aceh, the more subtle versions included flying white or United Nations flags in areas controlled by the rebels (rather than the rebels’ flag). Considerably more risk was taken by villagers who protected two Javanese neighbours from the sepa­rat­ists and ensured that they were able to re­main. An­other example of civilian resistance is the human rights activists in Aceh who criticised both the insur­gent GAM and Indonesian government and army.167

Silent resistance should not be overlooked. When Mosul was controlled by the so-called Islamic State between 2014 and 2017 some parents kept their chil­dren out of schools run by the violent actors in order to avoid ideologisation and indoctrination – even if that meant that their children did not learn to read and write.168

Public resistance often leads to subsequent flight. The Geneva Convention of 1951 provides protections for persons fleeing persecution. Implicitly, that means there are people who, when faced with the choice to stay or leave, choose the former (at least initially) while making their opposition known.169 Strategies of resistance can exacerbate personal risk and end in flight – or detention and death. Given their obvious risks, strategies of resistance do not always serve to en­hance individual and collective safety, quite the contrary. The fact they are pursued at all by some civil­ians in most violent conflicts points to the pos­sibility that stayees may possess higher risk tolerance.


Civilians in conflict contexts employ various forms of self-defence against armed actors, for example found­ing their own militias or even joining fighting groups in order to reduce their risk.170 One striking example is Colombia’s unarmed “Guardia Indígena”, a com­mu­nity organisation that has negotiated with violent actors in the country’s long civil war.171

It is not always easy or even possible to distinguish between civilians and combatants. A case study from the first phase of the civil war in Liberia (1990–1996) shows how difficult it is to draw a line between self-defence and collaboration. The reasons for this in­clude close social and economic contact with fighters, the phenomenon of community members being sent to join different armed groups in order to strategically protect the whole village, and participation in trade in looted goods. Similar grey areas appear where indi­viduals switch between fighting and civilian roles (and may even find membership of an armed group to be empowering).172

In many conflicts civilians form armed self-defence forces. In certain cases these formations are promoted by the state and may be part of a counter-insurgency strategy.173 Sometimes such militias become involved in crimes, in particular drug trafficking, kidnapping and blackmail, or become a conflict party in their own right, seeking control over territory.174 Given that they often exercise massive violence against the civil­ian population,175 their protective function is re­stricted to their members.

Interim conclusion

Those who stay in regions affected by acute violent conflict employ a spectrum of strategies to protect themselves. This applies to stayees who are voluntarily or acquiescently immobile, and to a certain extent also to the involuntarily immobile. The latter may not see any realistic perspective of leaving, but must survive nonetheless.

Multiple survival strategies may be followed simultaneously or successively depending on the context and dynamics.

Between the poles of collaboration and resistance or self-defence – as per Hirschman’s classic categories –we find a multitude of coping strategies. Here, they have been sorted according to how passive or active (or even potentially violent) they are. Neutrality, avoidance and resistance are non-violent, while collaboration and self-defence may involve direct or indirect use of violence. Frequently multiple pro­tec­tion strategies are employed simultaneously or suc­cessively depending on the context and dynamics – which stayees know well and follow closely.

Stayees are not simply helpless victims of violent conflict. They possess and exploit options and agency. Their strategies are not always without danger to them­selves and others. Informers can trigger acts of violence, militias created for self-defence have been accused of human rights violations.

Despite cases of civilian complicity and profiteer­ing, it must not be forgotten that a large proportion of the affected civilian population still needs support. Supplying this is one of the central tasks of humani­tarian aid and – as the two become ever more closely coordinated – also increasingly of development co­operation.

Recommendations for Humanitarian Aid and Development Cooperation

As the foregoing analysis demonstrates, people still possess choices even in violent conflicts. It is also clear that the international community plays an ambivalent role when it comes to protection. On the one hand, the central international actors, such as humanitarian organisations, peace missions and development institutions, are fundamentally aware that those who stay also have needs. On the other, people living in conflict zones cannot rely on inter­national actors to ensure their physical safety and protect them from harm. Because of a lack of aware­ness, stayees are still too frequently overlooked, treated as simply involuntarily immobile or inadequately distinguished from IDPs.

A number of general recommendations for humanitarian aid and development cooperation follow from the above analysis.

Consider the entire mobility spectrum and understand the (im)mobility per­spec­tive as added value

The “mobility bias” in politics and research needs to be acknowledged and challenged. (Im)mobility needs to be considered in all its facets and the (im)mobility perspective understood as a bonus because it helps us to understand complex contexts and improve the work conducted there, without requiring any fun­damentally new activities. External actors involved in humanitarian aid, development cooperation and peace-building already possess comprehensive experi­ence in supporting civilian populations during and after violent conflicts. An additional (im)mobility perspective opens up a new understanding of people in conflict zones. It can contribute to ensuring that those who stay for whatever reason are more quickly identified, more thoroughly taken into account, and better supported by humanitarian and above all devel­opment actors. Even the fundamentally positive engagement of German development cooperation in forced displacement contexts since 2015 is not immune to overlooking those who stay.

It is important to be aware of the entire spectrum of (im)mobility, which extends well beyond refugees and IDPs: from voluntary immobile and acquiescent to involuntary immobile and immobilised. The objec­tive must be to ensure that basic services are available to all who need them, and that those affected are able to make their own (im)mobility decisions. Even if lack of resources, growing needs and/or difficult circumstances sometimes prevent this being properly realised, it does offer basic guidance for action.

Support for stayees must not hinder or prevent the possibility of later flight.

This certainly does not absolve humanitarian aid and development actors of the responsibility to iden­tify the immobilised. Depending on context and needs, external (humanitarian) actors can make a signifi­cant contribution here, by negotiating access to humanitarian aid and/or organising evacuations. People who wish to leave but are unable to do so on their own should be assisted to evacuate – ideally even in situations where armed groups immobilise civilians against their will. But those who wish to stay should also receive access to humanitarian aid where they are without needing to relocate to IDP camps, as should refugees and their host communities. That said, on-the-ground support for stayees must not hinder or prevent the possibility of later flight. Nor should acknowledgement of stayees lead to neglect of the forcibly displaced and returnees.

Actors following comprehensive approaches should provide aid regardless of migration status without asking whether a person is a stayee, refugee, IDP or returnee. Instead they should apply vulnerability criteria or use area-based approaches as an alternative to status-based support.

The HDP nexus, which seeks to integrate humani­tarian, development and peace-building approaches, also offers a conceptual starting point for situations of forced displacement and for post-conflict contexts. The objective here is to ensure that all those affected by a conflict – regardless of their mobility status –receive the attention they need through cooperation between all relevant actors.

Stayees are unevenly affected by conflict. Those who live far from the fighting can often feel comparatively safe and may not even think about leaving. Vul­nerability remains the central criterion for sup­port.

Acknowledge civilian agency

If international humanitarian aid and development actors are to protect and support civilian stayees, it is important to know and understand their survival strategies.176 Public discussions and international actors working on the ground still too frequently assume that civilians in such situations have no agency.177 Stayees must be taken seriously as actors and their specific circumstances need to be better understood, with all the facets of voluntary, acquies­cent and involuntary immobility. Only then can sup­port measures be developed to help all those affected by violent conflict, whether they decide to stay or flee and whenever they do so. This must include hearing what civil society voices are saying about dangers, con­veying their warnings to international and UN bodies, and ensuring that independent and impartial information on the state of the conflict is available to those affected (for example by radio or text message).178

The humanitarian aid “grand bargain”179 also points in the same direction, with its promises to pro­vide aid increasingly through and to local organisations and to involve recipients more closely in the deci­sion-making processes. It is crucial to avoid actions that undermine the survival strategies of the civilian population. Knowledge of these strategies is neces­sary, for example to recognise changing patterns of mobility or to avoid obstructing dialogue initiatives.

Knowing about the strategies also means being pre­pared to encounter civilian collaboration and per­pe­trators, understanding the circumstances that create them and what that means for a society during and after conflict. Confidence-building, reckoning and recon­ciliation are fundamental to sustainable peace-building. Mental health and psychosocial support (MHPSS) for the forcibly displaced has rightly received great attention in recent years. German humanitarian aid and development actors should ensure that the same is also offered to stayees.

Do No Harm: Protecting from the helpers

Where members of peace missions, humanitarian aid workers and development experts are deployed in frag­ile contexts and violent conflicts, reports of sexual­ised violence and exploitation frequently emerge. Truly comprehensive protection of the civil­ian population must therefore also take account of this potential violation of the do-no-harm principle.180 This problem affects forcibly displaced persons as well as stayees, making countermeasures all the more im­portant.

Ensuring that civilian populations and those work­ing with them know their rights and duties would be a first step, along with enforcing the rules and punishing violations – starting for example with the standard prohibition of romantic or sexual relation­ships with members of the civilian population. The recommendations on preventing sexual exploitation and abuse (SEA) published by UN organisations, task forces and other bodies are all very similar: estab­lish complaints offices, standardise procedures for in­vestigating allegations, employ qualified staff to pre­vent retraumatisation, protect and support victims materially and psychologically, and punish perpetrators. To date, however, implementation has been in­adequate, despite the additional risks and the nega­tive consequences for those affected by sexualised violence, who are overwhelmingly women and children.

The recent abuse and exploitation scandals (#Aidtoo) have naturally been discussed in German humanitarian aid and development circles. But there have been no real consequences, partly because the allegations principally concerned British organisations. The German government should consider which of the measures taken by the British government it could usefully adopt. Germany not only has a repu­tation to guard, as one of the biggest humanitarian and development donors. It also has an opportunity to set standards in a key sphere of feminist foreign and development policy.

A comprehensive perspective on return and reintegration

A safe and dignified voluntary return is one of the three durable solutions for refugees. Supporting large return movements when conflicts end and refugee status expires is an important task of humanitarian aid and development cooperation in the Global South. In light of the observation that stayees are frequently overlooked (although a degree of progress has been made), several recommendations can be made. For peace-building and reconciliation it is imperative that aid and development programmes serve the needs of all relevant groups – including stayees and IDPs – rather than concentrating solely on returnees.181 To address the competition for scarce sources such as in­come that frequently occurs, aid programmes should exploit or create economic opportunities, for example through local procurement, and expanding basic ser­vices for all residents regardless of migration history and refugee status. Systematic investment in basic services can also contribute to reducing social ten­sions, if returnees are seen as bringing hope for devel­opment and prosperity rather than as a burden. Such interventions must be clearly and comprehensibly com­municated.182

Conflicts over land and compensation or return of property are almost inevitable when refugees return. Affordable, unbureaucratic and gender-equitable pro­cesses, easily accessible mediation mechanisms, and properly resourced courts are all helpful for conflict resolution.183

The purpose of such measures is not simply to pro­mote fairness. As examples from Bosnia, Burundi, Iraq and South Sudan demonstrate, competition over land, employment and social norms between stayees and returnees can cause new conflicts and hamper devel­opment.184 External actors providing support need to go further than the standard do-no-harm ap­proach and conflict sensitivity. They must also scru­tinise the psychosocial consequences of their actions and ensure that adequate attention is paid to the psychological impacts of violence and trauma among those who stay as well as those who leave. Ger­man development cooperation has a long record of sup­porting return and reintegration in the Global South, including in cooperation with UNHCR. It should en­sure that the comprehensive multi-sectoral approach is applied across the board. That means explicitly in­cluding stayees, as well as returnees and the places they return to.


In Germany and the EU the question of populations that remain in conflict zones has tended to be over­shadowed by a narrow focus on international forced displacement following the refugee movements of 2015/16. Protecting civilian populations in armed conflicts is by no means a new issue, but political attention has to date been sporadic. It is quite pos­sible that the question will come to the fore again in connection with the current growth in numbers of refugees and seemingly exhausted reception capacity. Support for stayees implies a possibility of reducing international migration and keeping people in their countries and regions of origin. That makes improving protection for people in situations of violent con­flict attractive to governments that want to prevent refugees arriving in their countries. From the human rights perspective, it is obvious that protecting civil­ian populations must not serve to prevent flight or to undermine the right to asylum. For those forced to flee asylum remains vital. The German government should both continue to support countries and regions hosting refugees and IDPs and ensure that its humanitarian aid and development cooperation are better tailored to the needs of stayees.

The existence of stayees appears to demonstrate that it is fundamentally possible for civilians to remain where they live even during armed conflicts. This in turn might call into question why refugees and asylum-seekers need protection at all. Analysis of the macro-, meso-, and micro-level factors and the complex decision-making processes and trade-offs involved in fleeing or staying exposes that assertion as a gross oversimplification. Certain individuals possess the agency and opportunity to decide whether to stay or leave, while others are involuntarily im­mobilised. A proper understanding of the different contexts can contribute to providing enough support and protection for people to stay voluntarily. The inter­action of individual markers and environment has different effects from person to person, and many will continue to need support. At the political level, peaceful resolution of armed conflicts remains the priority, along with efforts to uphold humanitarian international law, in particular to protect civilian populations and sanction violations.



Armed Conflict Location and Event Data Project


Ejército de Liberación Nacional/National Liberation Army (Colombia)


Gerakan Aceh Merdeka (Free Aceh Movement) (Indonesia)

HDP nexus

Humanitarian-development-peace nexus


Internally displaced person


Lord’s Resistance Army (Uganda)


Liberation Tigers of Tamil Eelam (Sri Lanka)


Mental health and psychosocial support


Sustainable Development Goals


United Nations High Commissioner for Refugees



 Prakash Adhikari, “Conflict-induced Displacement, Under­standing the Causes of Flight”, American Journal of Political Science 57, no. 1 (2013): 82–89 (82).


 United Nations High Commissioner for Refugees (UNHCR), Global Trends 2022 (Copenhagen, 14 June 2023), 3.


 Other terms are occasionally found in the literature, for example “stayers”, see Diana Mata-Codesal, “Is It Simpler to Leave or to Stay Put? Desired Immobility in a Mexican Village”, Population, Space and Place 24, no. 4 (2018): e2127 (2).


 Kerilyn Schewel, Understanding the Aspiration to Stay: A Case Study of Young Adults in Senegal, Working Papers, no. 107 (Oxford: University of Oxford, January 2015), 4; Schewel, “Under­standing Immobility: Moving beyond the Mobility Bias in Migration Studies”, International Migration Review 54, no. 2 (2020): 328–55 (331); Elisabeth Gruber, “Staying and Immobility: New Concepts in Population Geography? A Lit­erature Review”, Geographica Helvetica 76, no. (2021): 275–84 (280f). Bakewell uses the term “sedentary bias” to describe the tendency of development organisations in Africa to make staying put the objective of their interventions, see Oliver Bakewell, “‘Keeping Them in Their Place’: The Am­bivalent Relationship between Development and Migration in Africa”, Third World Quarterly 29, no. 7 (2008): 1341–58. Malkki, in a much earlier contribution, identifies sedentar­ism as a dominant perspective that tends to obscure other ways of life, see Liisa Malkki, “National Geographic: The Rooting of Peoples and the Territorialization of National Identity among Scholars and Refugees”, Cultural Anthropology 7, no. 1 (1992): 24–44 (31).


 The focus on mobility in migration research mirrors a broader “mobility turn” in the social sciences, see Mimi Sheller and John Urry, “The New Mobilities Paradigm”, in Environment and Planning A: Economy and Space 38, no. 2 (2006): 207–26; Schewel, Understanding the Aspiration to Stay (see note 4), 332.


 The same focus on mobility to the neglect of immobile groups and individuals is also noted in the field of environmental migration, see Caroline Zickgraf, “Immobility”, in Routledge Handbook of Environmental Displacement and Migration, ed. Robert McLeman and François Gemenne (Abingdon and New York: Routledge, 2018), 71–84 (74).


 Will H. Moore and Stephen M. Shellman, “Fear of Per­secution: Forced Migration, 1952–1995”, Journal of Conflict Resolution 48, no. 5 (2004): 723–45 (724); Adhikari, “Conflict-induced Displacement” (see note 1), 82; Abbey Steele, “Seek­ing Safety: Avoiding Displacement and Choosing Destina­tions in Civil Wars”, Journal of Peace Research 46, no. 3 (2009): 419–29 (420).


 United Nations General Assembly, New York Declaration for Refugees and Migrants, Resolution Adopted by the General Assembly (New York, 3 October 2016).


 Steffen Angenendt and Nadine Biehler, On the Way to a Global Compact on Refugees: The “Zero Draft”: A Positive, but Not Yet Sufficient Step, SWP Comment 18/2018 (Berlin: Stiftung Wissenschaft und Politik, April 2018).


 Bundesministerium für wirtschaftliche Zusammen­arbeit und Entwicklung, “Menschen auf der Flucht” (Berlin, 2023). In the meantime the funding instrument has been renamed the special initiative for displaced persons and host countries (Sonderinitiative “Geflüchtete und Aufnahme­länder”).


 Roger Zetter et al., “Violence, Conflict, and Mobility: A Micro-level Analysis”, in A Micro-level Perspective on the Dynamics of Conflict, Violence, and Development, ed. Patricia Justino et al. (Oxford: Oxford University Press, 2013), 206–27 (225); Ana M. Ibáñez and Andrés Moya, “Who Stays and Who Leaves during Mass Atrocities?” in Economic Aspects of Genocides, Other Mass Atrocities, and Their Preventions, ed. Charles H. Anderton and Jurgen Brauer (Oxford: Oxford University Press, 2016), 251–73 (260).


 A hierarchisation can also be identified among those who flee. International refugees receive greater attention than IDPs, with stronger institutional and legal protections (although efforts are under way to address this), see Anne Koch, On the Run in Their Own Country: Political and Institutional Challenges in the Context of Internal Displacement, SWP Research Paper 5/2020 (Berlin: Stiftung Wissenschaft und Politik, May 2020).


 Jørgen Carling, “Migration in the Age of Involuntary Immobility: Theoretical Reflections and Cape Verdean Experiences”, Journal of Ethnic and Migration Studies 28, no. 1 (2002): 5–42


 Hein de Haas, “A Theory of Migration: The Aspirations-Capabilities Framework”, Comparative Migration Studies 9, no. 1 (2021): 8.


 Schewel, Understanding the Aspiration to Stay (see note 4).


 Carling, “Migration in the Age of Involuntary Immobil­ity” (see note 13), 12f.


 Schewel, Understanding Immobility (see note 4), 338, 343.


 Ibid., 329 (italics in original). The definitions of mobility and immobility are not uncontested, see Gruber, “Staying and Immobility” (see note 4), 280.


 Schewel, Understanding the Aspiration to Stay (see note 4), 336f.; that is also the assumption of Carling’s original model: Carling, “Migration in the Age of Involuntary Immobility” (see note 13), 8.


 Schewel, Understanding the Aspiration to Stay (see note 4), 335f. The categorisation is still under discussion in migration research. De Haas points out that it could represent a retro­spective rationalisation of lack of mobility options, and argues that further research is needed, see de Haas, “A Theory of Migration” (see note 14), 23.


 Schewel, Understanding the Aspiration to Stay (see note 4), 339f. Media reports of evacuations of older and disabled people in Ukraine in summer 2022 mentioned individuals whose mobility was restricted by age or illness (no capability) or evacuees were reluctant to leave (aspiration to stay), see Maria Varenikova and Ivor Prickett, “Evacuating the Vulnerable amid the Terror of War”, New York Times International (online), 8 June 2022; Anastasia Magasowa, “Die letzte Ver­bindung”, taz.de, 7 June 2022.


 Schewel, Understanding the Aspiration to Stay (see note 4), 335f; Carling, “Migration in the Age of Involuntary Immobility” (see note 13), 37.


 Richard Black and Michael Collyer, “‘Trapped’ Populations: Limits on Mobility at Times of Crisis”, in Humanitarian Crises and Migration: Causes, Consequences and Responses, ed. Susan F. Martin et al. (London: Routledge, 2014), 287–305 (293, 298).


 The UNHCR describes resettlement, integration and voluntary return as durable solutions for refugees, see UNHCR, “Was wir tun”, 2023. Resettlement refers to relocation of refugees from initial host countries to third states under the auspices of UNHCR.


 Bram Ebus, “Five Years after ‘Peace’, the Colombian Communities Living in Forced Confinement”, The New Humanitarian, 25 November 2021.


 Stephen C. Lubkemann, “Involuntary Immobility: On a Theoretical Invisibility in Forced Migration Studies”, Journal of Refugee Studies 21, no. 4 (2008): 454–75 (455).


 Even if the Haitian gangs are entangled with political actors, and instrumentalised by them, the Global Initiative against Transnational Organized Crime argues that they are motivated more by profit and power than by political ideol­ogy, see Gangs of Haiti: Expansion, Power and an Escalating Crisis (Geneva: Global Initiative against Transnational Organized Crime, October 2022), 6.


 “Gang Violence Traps Thousands in Haitian Town”, Deutsche Welle (online), 13 July 2022.


 Charli Carpenter, “Civilian Men Are Trapped in Ukraine. Human Rights and Humanitarian NGOs Should Pay Attention to Kyiv’s Sex-selective Martial Law”, Foreign Policy, 15 July 2022.


 Black and Collyer, “‘Trapped’ Populations” (see note 23), 288.


 Ibid., 287, 293; Ibáñez and Moya, “Who Stays and Who Leaves during Mass Atrocities?” (see note 11), 258.


 The term “‘trapped’ populations” (see note 23), which Black and Collyer apply to the conflict context, originates in research on environmental and climate-related migration, where it refers to people who live in vulnerable regions but lack the resources to leave, see, for example, UK Government Office for Science, Migration and Global Environmental Change: Future Challenges and Opportunities: Final Project Report (London, 2011), 14; Zickgraf, “Immobility” (see note 6), 72.


 Gruber, “Staying and Immobility” (see note 4), 278.


 Schewel, Understanding the Aspiration to Stay (see note 4), 336f.; Gruber, “Staying and Immobility” (see note 4), 277; de Haas, “A Theory of Migration” (see note 14), 16; Zickgraf, “Immobility” (see note 6), 75.


 Pratikshya Bohra-Mishra and Douglas S. Massey, “Indi­vidual Decisions to Migrate during Civil Conflict”, Demog­raphy 48, no. 2 (2011): 401–24 (402).


 Duygu Ozaltin et al., “Why Do People Flee? Revisiting Forced Migration in Post-Saddam Baghdad”, Journal of Inter­national Migration and Integration 21, no. 2 (2020): 587–610 (589).


 Ibáñez and Moya, “Who Stays and Who Leaves during Mass Atrocities?” (see note 11), 252; Moore and Shellman, “Fear of Persecution” (see note 7), 725.


 Steele, “Seeking Safety” (see note 7), 420.


 Moore and Shellman, “Fear of Persecution” (see note 7), 725; Adhikari, “Conflict-Induced Displacement” (see note 1), 82; Ozaltin et al., “Why Do People Flee?” (see note 36), 589; Christian A. Davenport et al., “Sometimes You Just Have to Leave: Domestic Threats and Forced Migration, 1964–1989”, International Interactions 29, no. 1 (2003): 27–55 (31); Stefanie Engel and Ana María Ibáñez, “Displacement Due to Violence in Colombia: A Household‐level Analysis”, Economic Develop­ment and Cultural Change 55, no. 2 (2007): 335–65 (338).


 Lidia Ceriani and Paolo Verme, Risk Preferences and the Decision to Flee Conflict, Policy Research Working Paper, no. 8376 (Washington, D.C.: World Bank, March 2018), 1; Clionadh Raleigh, “The Search for Safety: The Effects of Conflict, Poverty and Ecological Influences on Migration in the Developing World”, Global Environmental Change 21 (19 October 2011) (online), 4, doi: 10.1016/j.gloenvcha.2011. 08.008.


 Adhikari, “Conflict-induced Displacement” (see note 1), 82; Ceriani and Verme, Risk Preferences and the Decision to Flee Conflict (see note 40), 1; Black and Collyer, “‘Trapped’ Populations” (see note 23), 287; Ibáñez and Moya, “Who Stays and Who Leaves during Mass Atrocities?” (see note 11), 261.


 Mara Redlich Revkin, “Competitive Governance and Displacement Decisions under Rebel Rule: Evidence from the Islam­ic State in Iraq”, Journal of Conflict Resolution 65, no. 1 (2021): 46–80 (50).


 De Haas, “A Theory of Migration” (see note 14), 4f.


 Davenport et al., “Sometimes You Just Have to Leave” (see note 39); Raleigh, “The Search for Safety” (see note 40); Erik Melander and Magnus Öberg, “Time to Go? Duration Dependence in Forced Migration”, International Interactions 32, no. 2 (2006): 129–52; Melander and Öberg, “The Threat of Violence and Forced Migration: Geographical Scope Trumps Intensity of Fighting”, Civil Wars 9, no. 2 (2007): 156–73.


 Adhikari, “Conflict-induced Displacement” (see note 1), 82; Davenport et al., “Sometimes You Just Have to Leave” (see note 39), 31; Shane Joshua Barter, “Zones of Control and Civilian Strategy in the Aceh Conflict”, Civil Wars 17, no. 3 (2015): 340–56 (353); Raleigh, “The Search for Safety” (see note 40), 1.


 Davenport et al., “Sometimes You Just Have to Leave” (see note 39), 31; Raleigh, “The Search for Safety” (see note 40), 1; Melander and Öberg, “The Threat of Violence and Forced Migration (see note 44), 158; Melander and Öberg, “Time to Go?” (see note 44).


 Hans-Georg Bohle, “Geographies of Violence and Vul­nerability. An Actor-oriented Analysis of the Civil War in Sri Lanka”, Erdkunde 61, no. 2 (2007): 129–46 (141).


 Adhikari, “Conflict-induced Displacement” (see note 1), 83; Moore and Shellman, “Fear of Persecution” (see note 7), 727f.; Ozaltin et al., “Why Do People Flee?” (see note 36), 591.


 Nicholas Van Hear, “Reconsidering Migration and Class”, International Migration Review 48, no. S1 (2014): 100–21 (106).


 Unlike in Kalyvas’s well-known and useful model of geographical zones of control in irregular civil wars (Stathis N. Kalyvas, The Logic of Violence in Civil War [Cambridge: Cam­bridge University Press, 2006], 88), the focus here is on the type of violence, in order to include different types of con­flict.


 Raleigh, “The Search for Safety” (see note 40), 8; Schewel, Understanding the Aspiration to Stay (see note 4), 8; Ozaltin et al., “Why Do People Flee?” (see note 36), 600.


 Raleigh, “The Search for Safety” (see note 40), 4.


 Carling’s “aspiration/ability model” also posits that factors at different levels influence the desire to migrate. He locates personal wishes at the micro level, and the social, economic and political context (the “emigration environ­ment”) at the macro level, see Carling, “Migration in the Age of Involuntary Immobility” (see note 13), 13. Schewel, on the other hand, assigns the reasons for preferring immobility to three categories (factors that retain, those that repel and “internal constraints’ on decision-making”), see Schewel, Under­standing Immobility (see note 4), 339. The classification employed here is independent of both those proposals.


 Nathalie E. Williams, Living with Conflict: The Effect of Community Organizations, Economic Assets, and Mass Media Con­sumption on Migration during Armed Conflict, dissertation (Ann Arbor: University of Michigan, 2009), 71.


 Etienne Piguet, “Theories of Voluntary and Forced Migra­tion”, in Routledge Handbook of Environmental Displace­ment and Migration, ed. McLeman and Gemenne (see note 6), 17–28 (18).


 Justin Schon, “Focus on the Forest, Not the Trees: A Changepoint Model of Forced Displacement”, Journal of Refugee Studies 28, no. 4 (2015): 437–67 (439).


 Ana Arjona, “Civilian Cooperation and Non-cooperation with Non-state Armed Groups: The Centrality of Obedience and Resistance”, Small Wars and Insurgencies 28, no. 4–5 (2017): 755–78 (765).


 See Zickgraf, “Immobility” (see note 6), 79, on “place attach­ment”.


 On the question of “agency” see also de Haas, “A Theory of Migration” (see note 14), 14–16, 30.


 Ozaltin et al., “Why Do People Flee?” (see note 36), 591.


 Jørgen Carling and Kerilyn Schewel, “Revisiting Aspira­tion and Ability in International Migration”, Journal of Ethnic and Migration Studies 44, no. 6 (2018): 945–63 (958).


 Carling, “Migration in the Age of Involuntary Immobil­ity” (see note 13), 9; Piguet, “Theories of Voluntary and Forced Migration” (see note 55), 19; Zickgraf, “Immobility” (see note 6), 75.


 Schewel, Understanding the Aspiration to Stay (see note 4), 7.


 Raleigh, “The Search for Safety” (see note 40), 4.


 This limitation applies to functionalistic models in gen­eral, see de Haas, “A Theory of Migration” (see note 14), 8. The problem appears insoluble without further research placing greater weight on mobility/immobility aspirations.


 Adhikari, “Conflict-induced Displacement” (see note 1), 83; Ozaltin et al., “Why Do People Flee?” (see note 36), 591; Sue Lautze, Saving Lives and Livelihoods: The Fundamentals of a Livelihoods Strategy (Boston: Feinstein International Famine Center, Tufts University, 1997).


 Melander and Öberg, “Time to Go?” (see note 44), 132; Ibáñez and Moya, “Who Stays and Who Leaves during Mass Atrocities?” (see note 11), 252; Moore and Shellman, “Fear of Persecution” (see note 7), 742.


 Raleigh, “The Search for Safety” (see note 40), 3; Stathis N. Kalyvas and Matthew A. Kocher, “How ‘Free’ Is Free Rid­ing in Civil Wars? Violence, Insurgency and the Collective Action Problem”, World Politics 59, no. 2 (2007): 177–216 (186).


 Raleigh, “The Search for Safety” (see note 40), 4.


 Melander and Öberg, “The Threat of Violence and Forced Migration” (see note 44), 159.


 Burcu Karakas, “Russians Fleeing War and Repression Seek Solace in Istanbul”, Deutsche Welle (online), 2 July 2022.


 “Factbox: Where Have Russians Been Fleeing to since Mobilisation Began?Reuters, 6 October 2022.


 Melander and Öberg, “The Threat of Violence and Forced Migration” (see note 44), 166.


 Davenport et al., “Sometimes You Just Have to Leave” (see note 39).


 Kalyvas and Kocher, “How ‘Free’ Is Free Riding in Civil Wars?” (see note 68), 186.


 Laia Balcells and Abbey Steele, “Warfare, Political Iden­tities, and Displacement in Spain and Colombia”, Political Geography 51 (2016): 15–29 (16).


 Betcy Jose and Peace A. Medie, “Understanding Why and How Civilians Resort to Self-protection in Armed Conflict”, Inter­national Studies Review 17, no. 4 (2015): 515–35 (517); Mary Kaldor, New and Old Wars: Organized Violence in a Global Era (Stanford, Cal.: Stanford University Press, 1999), 2.


 Susanne Schmeidl, “Exploring the Causes of Forced Migra­tion: A Pooled Time-series Analysis, 1971–1990”, Social Science Quarterly 78, no. 2 (1997): 284–308 (302f.).


 Anna Lindley, “Leaving Mogadishu: Towards a Sociology of Conflict-related Mobility”, Journal of Refugee Studies 23, no. 1 (2010): 2–22 (11).


 Raleigh, “The Search for Safety” (see note 40), 6f.


 Hannah Arendt, On Violence (Orlando, Fl.: Harcourt Brace Jovanovich, 1970), 51.


 Reed M. Wood, “Rebel Capability and Strategic Violence against Civilians”, Journal of Peace Research 47, no. 5 (2010): 601–14 (602); Ana Maria Ibáñez and Andrés Moya, “Do Con­flicts Create Poverty Traps? Asset Losses and Recovery for Displaced Households in Colombia”, in The Economics of Crime. Lessons for and from Latin America, ed. Rafael Di Tella et al. (Chicago: University of Chicago Press, July 2010), 137–72 (149); Jen Ziemke, From Battles to Massacres, PhD dissertation (Madison: University of Wisconsin-Madison, 2008).


 Ibáñez and Moya, “Who Stays and Who Leaves during Mass Atrocities?” (see note 11), 251.


 Clionadh Raleigh et al., “Introducing ACLED: An Armed Conflict Location and Event Dataset”, Journal of Peace Research 47, no. 5 (2010): 651–60 (651).


 Raleigh, “The Search for Safety” (see note 40), 4–6; Stathis N. Kalyvas, “Micro-level Studies of Violence in Civil War: Refining and Extending the Control-Collaboration Model”, Terrorism and Political Violence 24, no. 4 (2012): 658–68 (660f.).


 Melander and Öberg, “The Threat of Violence and Forced Migration” (see note 44), 166.


 Andrew R. Morrison, “Violence or Economics: What Drives Internal Migration in Guatemala?” Economic Develop­ment and Cultural Change 41, no. 4 (1993): 817–31 (828).


 Bohra-Mishra and Massey, “Individual Decisions to Migrate during Civil Conflict” (see note 35), 403, 419–422.


 Lindley, “Leaving Mogadishu” (see note 79), 7ff.


Shell-Shocked: Civilians under Siege in Mogadishu, New York: Human Rights Watch, vol. 19, no. 12(A) (August 2007), 3; “So Much to Fear”: War Crimes and the Devastation of Somalia (New York: Human Rights Watch, December 2008), 4.


 Lindley, “Leaving Mogadishu” (see note 79), 10ff.


 Ibid., 10f.; Schon, “Focus on the Forest, Not the Trees” (see note 56), 443.


 Ibáñez and Moya, “Who Stays and Who Leaves during Mass Atrocities?” (see note 11), 256.


 Kalyvas and Kocher, “How ‘Free’ Is Free Riding in Civil Wars?” (see note 68), 188.


 Steele, “Seeking Safety” (see note 7), 422.


 Wood, “Rebel Capability and Strategic Violence against Civilians” (see note 82), 601; Ibáñez and Moya, “Who Stays and Who Leaves during Mass Atrocities?” (see note 11), 253; Steele, “Seeking Safety” (see note 7), 422; Kalyvas, The Logic of Violence in Civil War (see note 50).


 Steele, “Seeking Safety” (see note 7), 422; Kalyvas, The Logic of Violence in Civil War (see note 50).


 Arjona, “Civilian Cooperation and Non-cooperation with Non-state Armed Groups” (see note 57), 765.


 Mara Revkin and Delair Jebari, West Mosul: Perceptions on Return and Reintegration among Stayees, IDPs and Returnees (Bagh­dad: International Organization for Migration, June 2019), 39.


 Arjona, “Civilian Cooperation and Non-cooperation with Non-state Armed Groups” (see note 57), 766f.


 Kalyvas and Kocher, “How ‘Free’ Is Free Riding in Civil Wars?” (see note 68), 186.


 Balcells and Steele, “Warfare, Political Identities, and Displacement in Spain and Colombia” (see note 76), 16; Davenport et al., “Sometimes You Just Have to Leave” (see note 39), 32.


 Balcells and Steele, “Warfare, Political Identities, and Displacement in Spain and Colombia” (see note 76), 16.


 Mathias Czaika and Krisztina Kis-Katos, “Civil Conflict and Displacement: Village-level Determinants of Forced Migra­tion in Aceh”, Journal of Peace Research 46, no. 3 (2009): 399–418 (402).


 Erin Baines and Emily Paddon, “‘This Is How We Sur­vived’: Civilian Agency and Humanitarian Protection”, Secu­rity Dialogue 43, no. 3 (2012): 231–47 (232, 241).


 Adam G. Lichtenheld, “Explaining Population Displace­ment Strategies in Civil Wars: A Cross-national Analysis”, Inter­national Organization 74, no. 2 (2020): 253–94 (257).


 Davenport et al., “Sometimes You Just Have to Leave” (see note 39), 33.


 Lichtenheld, “Explaining Population Displacement Strat­egies in Civil Wars” (see note 106), 256f.


 Balcells and Steele, “Warfare, Political Identities, and Displacement in Spain and Colombia” (see note 76), 28ff.


 Melander and Öberg, “The Threat of Violence and Forced Migration” (see note 44), 158; Melander and Öberg, “Time to Go?” (see note 44), 146f.


 Czaika and Kis-Katos, “Civil Conflict and Displacement” (see note 104), 404; Balcells and Steele, “Warfare, Politi­cal Identities, and Displacement in Spain and Colombia” (see note 76), 16.


 Balcells and Steele, “Warfare, Political Identities, and Displacement in Spain and Colombia” (see note 76), 6.


 Steele, “Seeking Safety” (see note 7), 422, 424; Adhikari, “Conflict-induced Displacement” (see note 1), 88.


 Jose and Medie, “Understanding Why and How Civil­ians Resort to Self-protection in Armed Conflict” (see note 77), 516.


 Kalyvas and Kocher, “How ‘Free’ Is Free Riding in Civil Wars?” (see note 68), 203; R. Charli Carpenter, “‘Women and Children First’: Gender, Norms, and Humanitarian Evacua­tion in the Balkans 1991–95”, International Organization 57, no. 4 (2003): 661–94 (661f).


 Lindley, “Leaving Mogadishu” (see note 79), 15f.


 Czaika and Kis-Katos, “Civil Conflict and Displacement” (see note 104), 404f.; Mats Utas, “Victimcy, Girlfriending, Soldiering: Tactic Agency in a Young Woman’s Social Navigation of the Liberian War Zone”, Anthropological Quar­terly 78, no. 2 (2005): 403–30 (404f.).


 Bohra-Mishra and Massey, “Individual Decisions to Migrate during Civil Conflict” (see note 35), 420.


 Schewel, Understanding the Aspiration to Stay (see note 4), 339f.


 Benedikt Korf, Conflict – Threat or Opportunity? War, Liveli­hoods, and Vulnerability in Sri Lanka (Berlin: Humboldt-Uni­ver­sität, 2003), 14; Bohle, “Geographies of Violence and Vulnerability” (see note 47), 140; Heidelberg Institute for Inter­­na­tion­al Conflict Research, Conflict Barometer 2009 (Hei­delberg, 2009), 68f.


 Raleigh, “The Search for Safety” (see note 40), 8f.


 Korf, Conflict – Threat or Opportunity? (see note 120), 12; Czaika and Kis-Katos, “Civil Conflict and Displacement” (see note 104), 402; Kalyvas and Kocher, “How ‘Free’ Is Free Riding in Civil Wars?” (see note 68), 186.


 Korf, Conflict – Threat or Opportunity? (see note 120), 11.


 Zetter et al., “Violence, Conflict, and Mobility” (see note 11), 212; Williams, Living with Conflict (see note 54), 71.


 Revkin, “Competitive Governance and Displacement Decisions under Rebel Rule” (see note 42), 55.


 Lindley, “Leaving Mogadishu” (see note 79), 17.


 Ibid., 7, 13f.


 Adhikari, “Conflict-induced Displacement” (see note 1), 86f.; Bohra-Mishra and Massey, “Individual Decisions to Migrate during Civil Conflict” (see note 35), 420; Williams, Living with Conflict (see note 54), 92ff.


 Raleigh, “The Search for Safety” (see note 40), 9.


 Lindley, “Leaving Mogadishu” (see note 79), 15.


 Czaika and Kis-Katos, “Civil Conflict and Displace­ment” (see note 104), 400, 402, 415.


 Korf, Conflict – Threat or Opportunity? (see note 120), 11.


 Piguet, “Theories of Voluntary and Forced Migration” (see note 55), 19f.


 Lindley, “Leaving Mogadishu” (see note 79), 18.


 Adhikari, “Conflict-induced Displacement” (see note 1), 83.


 Anders H. Stefansson, “Sarajevo Suffering: Homecoming and the Hierarchy of Homeland Hardship”, in Home­comings: Unsettling Paths of Return, ed. Fran Markowitz and Anders H. Stefansson (Lanham, Md.: Lexington Books, 2010), 54–75 (58f.).


 de Haas, “A Theory of Migration” (see note 14), 23.


 Revkin, “Competitive Governance and Displacement Decisions under Rebel Rule” (see note 42), 47f.


 Ceriani and Verme, Risk Preferences and the Decision to Flee Conflict (see note 40), 15f.; David A. Jaeger et al., “Direct Evi­dence on Risk Attitudes and Migration”, Review of Economics and Statistics 92, no. 3 (2010): 684–89; Schewel, Understanding the Aspiration to Stay (see note 4), 342.


 Albert O. Hirschman, Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States (Cambridge, Mass.: Harvard University Press, 2004).


 Barter, “Zones of Control and Civilian Strategy in the Aceh Conflict” (see note 45), 348ff.


 Jose and Medie, “Understanding Why and How Civil­ians Resort to Self-protection in Armed Conflict” (see note 77), 528; Baines and Paddon, “‘This Is How We Survived’” (see note 105), 236.


 Baines and Paddon, “‘This Is How We Survived’” (see note 105), 239ff.


 Korf, Conflict – Threat or Opportunity? (see note 120), 2; Utas, “Victimcy, Girlfriending, Soldiering” (see note 117), 426; Oliver Kaplan, Resisting War: How Communities Protect Themselves (Cambridge and New York: Cambridge University Press, 2017), 4.


 Arjona, “Civilian Cooperation and Non-cooperation with Non-state Armed Groups” (see note 57), 766.


 Utas, “Victimcy, Girlfriending, Soldiering” (see note 117), 412f.; Arjona, “Civilian Cooperation and Non-coopera­­tion with Non-state Armed Groups” (see note 57), 756f.; Revkin, “Competitive Governance and Displacement Deci­sions under Rebel Rule” (see note 42), 50.


 Tanya Bandula-Irwin et al., “Beyond Greed: Why Armed Groups Tax”, Studies in Conflict and Terrorism (2022), 1–24.


 Utas, “Victimcy, Girlfriending, Soldiering” (see note 117), 409; Baines and Paddon, “‘This Is How We Survived’” (see note 105), 242f.


 Korf, Conflict – Threat or Opportunity? (see note 120), 20f.


 Ibid., 16, 19.


 Utas, “Victimcy, Girlfriending, Soldiering” (see note 117); Andrew Bonwick, “Who Really Protects Civilians?” Development in Practice 16, no. 3–4 (2006): 270–77 (275).


 Kalyvas and Kocher, “How ‘Free’ Is Free Riding in Civil Wars?” (see note 68), 179.


 Kaplan, Resisting War (see note 144), 3; Baines and Paddon, “‘This Is How We Survived’” (see note 105), 236.


 Jose and Medie, “Understanding Why and How Civil­ians Resort to Self-protection in Armed Conflict” (see note 77), 528; Baines and Paddon, “‘This Is How We Survived’” (see note 105), 236f.; Bohle, “Geographies of Violence and Vulnerability” (see note 47), 144f.


 Wood, “Rebel Capability and Strategic Violence against Civilians” (see note 82), 603.


 Kalyvas and Kocher, “How ‘Free’ Is Free Riding in Civil Wars?” (see note 68), 179.


 Jana Krause, Resilient Communities: Non-violence and Civil­ian Agency in Communal War (Cambridge and New York: Cam­bridge University Press, 2018).


 Kaplan, Resisting War (see note 144).


 Arjona, “Civilian Cooperation and Non-cooperation with Non-state Armed Groups” (see note 57), 761.


 Wood, “Rebel Capability and Strategic Violence against Civilians” (see note 82), 603.


 Korf, Conflict – Threat or Opportunity? (see note 120), 12, 14–16.


 Baines and Paddon, “‘This Is How We Survived’” (see note 105), 236ff.


 Lindley, “Leaving Mogadishu” (see note 79), 18.


 Schewel, Understanding Immobility (italics in original) (see note 4), 329.


 Conciliation Resources, ed., Local Engagement with Armed Groups: In the Midst of Violence, Accord Insight, no. 2 (London, 2015), 8.


 Barter, “Zones of Control and Civilian Strategy in the Aceh Conflict” (see note 45), 352f.


 Ibid., 351f.; Ibáñez and Moya, “Who Stays and Who Leaves during Mass Atrocities?” (see note 11), 255; Schon, “Focus on the Forest, Not the Trees” (see note 56), 12; Korf, Conflict – Threat or Opportunity? (see note 120), 11; Arjona, “Civilian Cooperation and Non-cooperation with Non-state Armed Groups” (see note 57), 760f.


 Revkin and Jebari, West Mosul (see note 99), 38.


 David Bartram, “Forced Migration and ‘Rejected Alter­natives’: A Conceptual Refinement”, Journal of Immigrant and Refugee Studies 13, no. 4 (2015): 439–56 (446); Carling and Schewel, “Revisiting Aspiration and Ability in International Migration” (see note 61), 957.


 Jose and Medie, “Understanding Why and How Civil­ians Resort to Self-protection in Armed Conflict” (see note 77), 528; Kalyvas and Kocher, “How ‘Free’ Is Free Riding in Civil Wars?” (see note 68), 183.


 Ebus, “Five Years after ‘Peace’” (see note 25); Heidel­berg Institute for International Conflict Research, Conflict Barometer 2021 (Heidelberg, 2022), 108.


 Utas, “Victimcy, Girlfriending, Soldiering” (see note 117), 409, 412f., 421–423.


 Mohammed Ibrahim Shire, “Protection or Predation? Understanding the Behavior of Community-created Self-defense Militias during Civil Wars”, Small Wars and Insurgen­cies 33, no. 3 (2022): 467–98 (468).


 Chelsea Estancona et al., “Civilian Self-defense Militias in Civil War”, International Interactions 45, no. 2 (2019): 215–66 (215f.).


 Shire, “Protection or Predation?” (see note 173), 468.


 Baines and Paddon, “‘This Is How We Survived’” (see note 105), 242.


 Barter, “Zones of Control and Civilian Strategy in the Aceh Conflict” (see note 45), 341.


 Baines and Paddon, “‘This Is How We Survived’” (see note 105), 242; Bonwick, “Who Really Protects Civil­ians?” (see note 151), 275.


 At the 2016 World Humanitarian Summit donor coun­tries and humanitarian aid organisations agreed to improve the efficiency and effectiveness of humanitarian aid. The agree­ment is referred to as the “grand bargain”, see Inter-Agency Standing Committee, “About the Grand Bargain”, https://interagencystandingcommittee.org/node/40190.


 The do no harm principle was first laid out in the epony­mous work by Mary B. Anderson, Do No Harm: How Aid Can Sup­port Peace, or War (Boulder, Col.: Lynne Rienner, 1999).


 Georgia Cole, “Working with ‘Stayee’ Communities: Learning from Eritrea”, Forced Migration Review 62 (2019).


 Gaim Kibreab, “When Refugees Come Home: The Rela­tion­ship Between Stayees and Returnees in Post-conflict Eritrea”, Journal of Contemporary African Studies 20, no. 1 (2002): 53–80 (69, 77); Revkin and Jebari, West Mosul (see note 99), 46f.


 Kibreab, “When Refugees Come Home” (see note 182), 59; Sonja Fransen, “The Socio-Economic Sustainability of Refugee Return: Insights from Burundi”, Population, Space and Place 23 (2017) 1, 16, doi: 10.1002/psp.1976; Revkin and Jebari, West Mosul (see note 99), 29.


 For further detail see Nadine Biehler, Anne Koch and Amrei Meier, Risks and Side-Effects of German and European Return Policy: Foreign Policy, Security and Development Trade-offs, SWP-Research Paper 12/2021 (Berlin: Stiftung Wissenschaft und Politik, December 2021), 37ff.

All rights reserved.

© Stiftung Wissenschaft und Politik, 2023

SWP Research Papers are peer reviewed by senior researchers and the execu­tive board of the Institute. They are also subject to copy-editing. For further information on our quality control procedures, please visit the SWP website: https://www.swp-berlin.org/ en/about-swp/quality-management-for-swp-publications/.

SWP Research Papers reflect the views of the author(s).


Stiftung Wissenschaft und Politik

German Institute for International and Security Affairs

Ludwigkirchplatz 3–4
10719 Berlin
Phone +49 30 880 07-0
Fax +49 30 880 07-200

ISSN (Print) 2747-5123

ISSN (Online) 1863-1053

Translation by Meredith Dale

(English version of SWP‑Studie 8/2023)