Scope of the question

The differences between camp and settlement approaches to refugee assistance are behind what Kibreab once called the "most sustained single controversy in African Refugee Studies" (Kibreab 1991) which surrounds the comparative advantages of self-settlement to organized settlement and refugee camps. It is a debate with very real implications. Although these numbers should be treated with caution, according to UNHCR (2002 est.) there are currently some 5.8 million refugees hosted in camps and centres around the world. This includes over 50 per cent of all UNHCR-assisted refugees in Africa (a total of 2,169,558 people), and 35 per cent of refugees in Asia. Clearly, camps and, albeit to a much lesser degree, planned rural settlements, constitute the main method of refugee assistance in the developing world, with the notable exception of Latin America.

Indeed, refugee camps easily qualify as the most conspicuous element of refugee assistance. They shape most Western images of the refugee phenomenon in developing countries – reflected for instance in the fact that awareness-raising campaigns by Médecins Sans Frontières ( MSF) involve a travelling exhibition reproducing a refugee camp. It is notable though that even though camps are often seen as a third-world phenomenon, increasing use of detention centres in the West seems to reintroduce "camp-based" answers to refugee issues here too.

On the other hand, large quantities of refugees still self-settle all over the world, despite the fact that increasingly restrictive policies by host governments have not only reduced the number of spontaneously settled refugees but also have meant that these situations can no longer be studied without attention to the potential risks such studies can entail for their subjects. At the Arusha Conference in 1979, figures of self-settled refugees in Africa were estimated to be 40 per cent of the total (Rogge 1987) , and Chambers (1979) claimed them to reach 77 per cent in the same year. These numbers are of course notoriously imprecise partially because self-settled refugees tend to live outside the assistance circuit of international agencies. Karadawi notes that only up to 40 per cent of those self-settled may receive material assistance (Karadawi 1983) whereas Cuenod estimated that no more than 25 per cent of African refugees lived in settlements where they could receive aid (Cuenod 1989) .

User's guide

This review falls necessarily short of providing an exhaustive overview of all cases and arguments that defend, define, or denigrate different forms of refugee settlement. Instead it tries to provide a useful overview of the main issues concerned and to guide further study. Geographically it is skewed towards Sub-Saharan Africa. This is for the simple reason that Africa is host to both more refugees and more refugee-camps than any other region. Section One offers some methodological caveats and deals with definitional issues. Section Two is organized by issue area and deals with directly comparative issues. The admittedly awkwardly named rubric "social aspects" covers socio-economic as well as socio-psychological issues. Readers more interested in some of these aspects are asked to also refer to FMO guides on psycho-social issues and gender. Given the disproportionate amount of research done on refugee camps, Section Three references literature that deals more or less exclusively with self-settled refugees and organized rural settlements. In each section, web-based sources are provided for further study. A bibliography of referenced and/or other important paper-based sources is provided at the end of this document.

[TOP]

Terminology and conceptual issues

The debate about the costs and benefits of different forms of refugee settlement was revived in the 1990s but still retains much terminological confusion. In the standard literature, the terms "camps" and "settlements" tend to be used interchangeably. The catalogue of the Refugee Studies Programme in Oxford, for instance, distinguishes between "organized settlements", which include closed camps; "camps", which include settlement literature; and "assisted self-settlements". Far from revealing inaccuracy on the part of the author, librarian, or practitioner, such definitions indicate how effectively blurred are the distinctions between these groups. UNHCR itself has differentiated between "permanent camps" and "camps". It calls the "Rhino-camp" (official name) in Uganda a "settlement" (official definition), but then lists Ugandan settlements as camps/centres in its statistical overview.

Moreover, different authors may situate the debate quite differently depending on the way the two categories (camps and settlements) are defined. There is often a tendency to define both according to the way they relate to an ultimate, durable solution: for some, camp and settlement approaches refer to two different stages in the refugee cycle, the former referring to temporary shelter, the latter to a durable solution, namely integration into the host country - which might or might not be preceded by a period of camp-based assistance. Others define camps as part and parcel of another durable solution, namely repatriation, while also holding settlements to be inevitably part of integrationist approaches.

Perhaps more appropriately, "camps and settlements" can be understood to cover three forms of assistance policies: (1) planned and (2) unplanned rural settlements which are based on various forms of officially recognized self-reliance, and (3) camps generally based on full assistance. This perspective postpones, (for purposes of definition only) the politically charged question of durable solutions, and instead concentrates on the different forms of assistance in situ. The UNHCR's Evaluation and Policy Unit has in some ways taken this last approach by introducing the umbrella terms of "protracted refugee situations". This approach bypasses many of the definitional issues involved. The terminology applies to both organized settlements and camps, as long as they exist for more than five years without clear prospects of finding a durable solution such as voluntary repatriation, local integration, or resettlement. The approach excludes spontaneous or self-settlement. It is in line with UNHCR statistical tables, which also generally combine camps and planned settlements in one category called "camps/centres"(even though here no time limit is specified).

Defining camps and settlements

For many observers, the Rwandan refugee camps in Zaire, and more specifically those three camps around Goma (Kibumba, Mugunga, and Katale) which in 1994 together hosted about 800,000 people (alongside a cholera epidemic), have certainly gained "paradigmatic status" – and fuel much of the scepticism of camps. Yet the notion of camps covers a much wider range of situations, and apart from the relatively clear-cut distinction between planned and self-settlement, definitions of refugee situations frequently lack objective criteria and clear demarcations. This is less important when one deals with immediate policy questions but inevitably skews any argument about policy-alternatives.

The table below indicates patterns of settlement in a continuum from integration/non-camps to segregation/closed camps (Van Damme 1998) :

Urban refugees & integrated rural refugees	Peaceful cohabitation	Spatial separation	Spatial segregation
Self-settled Rwandans in Rutshuru Zaire	Rwandans in small open camps, Uvira, Zaire	Rwandans in large open camps, Goma, Zaire & Benaco, Tanzania	Rwandans in closed camps, Ngozi, Burundi
Rwandans in West Tanzania 1959–94 (with qualifications)	Bangladeshis in India 1971–2	Sudanes in North Kenya	Salvadorians in Honduras
Urban refugees in Uganda & Kenya	Chadians in West Sudan	Somalis in East Kenya	Cambodians in Thailand
Mozambicans in South Africa	Ethiopians in East Sudan		Vietnamese boat people in Hong Kong
	Sudanes in North Uganda

In the wider literature, "camps" are rarely conceptually defined, even though a number of characteristics underlie the usage of the term. The terminology always implies a specific mode of assistance.

Stein, for instance, favours Murphy's (1955) emphasis on the effects of camps: "although the physical conditions of camps may vary widely, from hell to hotels, the effects tend to be uniform. The most important characteristics of the camps are: segregation from the host population, the need to share facilities, a lack of privacy, plus overcrowding and a limited, restricted area within which the whole compass of daily life is to be conducted. This gives the refugees a sense of dependency, and the clear signal that they have a special and limited status, and are being controlled."

What follows are five parameters which frequently underlie the usage of the terms "camps" or "settlements", and which serve to define refugee accommodation. Most of the below criteria are not dichotomous measures; many are quantifiable. Taken together they inform most of the typologies and choice of vocabulary recurrent in the literature. Hoerz (1995) uses them to describe a settlement continuum ranging from low to high spatial and economic integration of refugees with the surrounding population, similar to Van Damme's typology above. His categories move from "completely separate existence of refugees and locals (“closed camps”)"; "in camps but free to trade"; "in camps but free to move and trade"; "separate status but equal opportunities with locals" (e.g. agricultural settlements); and "integration of refugees and refugee settlements". A more directly descriptive typology, which uses some of the parameters below, can be found in Jacobsen (2001).

Parameters for camps and settlements
1. Freedom of movement: the more this is restricted, the more a refugee settlement is generally seen to take on the character of a camp. Even though the cases where refugee movement outside designated areas is strictly impossible are rare, legal restriction and even lax and arbitrary enforcement have large implications for refugee livelihoods. This characteristic of camp situations is amply documented and moreover echoes refugee perceptions: the Rwandese refugees in Tanzania, who were studied by Malkki, protested against the misnomer "settlement" for their location, arguing that "it is a camp because we cannot leave when we want to" (Malkki 1995) .
2. Mode of assistance/economics: one may distinguish between camps based on relief handouts and food distribution with little possibility for refugees to engage in subsistence farming or other economic activities, and, on the other hand, situations in which refuges can engage in a wider range of economic activities. Measurable indicators may be plot size in camps and the range of de facto restrictions on work. In camps, generally only limited income-generating programmes are permitted, while self-settled refugees will tend to be more integrated into the local economy, be it with or without governmental permission.
3. Mode of governance: this indicates the mechanisms of decision-making within or over the refugee community. Chambers (1979) uses this parameter when he states that settlements are like camps when hierarchies are external and abusive, defining camps and settlements in terms of different mechanisms of power. Camps are thus notably distinguished from self-settlements by the parameters of control: the restrictions on socio-economic, political, and cultural freedoms that are placed on their inhabitants over and above those existing for local populations. At the extreme end, Hyndman defines camps as "sites of neo-colonial power-relations where refugees are counted, their movements monitored and mapped, their daily routines disciplined and routinized by the institutional machinery of refugee relief agencies" (Hyndman 1997) .
4. Designation as temporary locations/shelter (irrespective of their actual longevity): an early UNREF document notes "there is no standard definition of “refugee camp” ... It would seem, however, that the term “refugee camp” designates a group of dwellings of various descriptions …which, mainly because of the poor conditions of the dwellings but also for other reasons, are meant to provide temporary shelter" ( UNREF 1958) . Being considered temporary is both a characteristic of a camp and itself shapes policy responses regarding economic and social freedoms of refugees.
5. Population size and/or density: this indicator, connected to questions of freedom of movement, planning, and economics, is also a useful definitional guide. Clark and Stein, in their detailed evaluation of UNHCR settlement history, refer to the key obstacle related to "settlements" as overcrowding arising from the temptation to "fill up and over", which turns them into camps or transit centres (see also Black 1998) . Following the large-scale influx of Burundi and Rwandese refugees into Tanzania in the mid 1990s, a number of sites that were initially conceived on the model of the older rural refugee settlements of the 1970s were reduced in plot size and turned into camps in order to accommodate the increasing numbers of new arrivals.

Key introductory texts and some methodological caveats

As far as there is a real debate about the alternatives of camps and organized and self-settlement, two different sets of debates are often mixed. One concentrates on the causal effect of different settlement patterns measured by a variety of social and economic indicators. The second is concerned with the factors that cause different settlement patterns. Most texts presented here shift back and forth between these two sets of causal analysis and for analytical purposes it is useful to be aware of the distinction.

Not many texts systematically compare the effects of camp and settlement situations on refugee welfare, host economies, and political structures, or general levels of security and conflict. This is partially due to both a lack of available research and its relatively slow consolidation. Another reason is the general tendency within refugee studies to eschew potentially problematic comparisons in favour of in-depth case studies. While this has much to do with refugee studies' disciplinary origins in anthropology, there are other methodological issues that make structured comparison difficult. These include, among others:

1. Differences in population: it is repeatedly the case that the most vulnerable and weakest stay within the camps and the more able refugees avoid them.
2. Third variables: success or failure of planned or self-settlement may be contingent on a variety of variables, such as familiarity with the host country and its population, the degree of hospitality encountered, and the economic resources and land generally available. Increasingly, studies that focus on refugee impact on local communities emphasize the importance of local context for success and failure of the pursuit of an ever-wider range of (refugee) policy aims.
3. Interdependence of cases: in many cases, refugees may live in different settlement patterns co-existing in the same host country, and linkages may exist between them. In such instances, refugees might be doubly based, using both the camp and the outside to ensure their personal or family livelihoods and/or survival (Hyndman 1997) . There is indeed consistent evidence of this phenomenon even though its significance is understandably difficult to gauge.

Despite these limitations, the debate has continued, and recent key introductory texts and bibliographic references can be found in the websites below (also see Kibreab 1991) . For an anthropological account comparing the evolution of refugees (in Tanzania) who lived in planned settlements with those who lived unofficially in a more urban context, see also Malkki's by now well-known study Purity and Exile (1995) .

Generally speaking, criticism of camp-based solutions is based either on arguments that emphasize questions of economic or social development, or that are rooted in a rights-based critique which takes as a starting point the many restrictions on socio-economic and political freedoms that accompany camp-based refugee assistance. Also, many studies deal with a combination of the two.

Where it focuses on questions of development or resource management, proponents of various forms of planned or self-settlement emphasize participatory approaches and call for a capacity-based developmental model to replace the traditional "relief model" (seen to underlie camps) which is said to encourage passivity and hopelessness. As Crowley (1991) notes, "Although the welfare model has long been discredited as paternalistic and self-serving in the context of development, it's still dominant in the ethos and practice of emergency relief." Especially in the past, concern with integrationist approaches to refugee assistance had this clear developmental focus (Betts 1984) . Major attempts to introduce alternative approaches are put under umbrella terms such as Refugees and Development ( RAD), Refugee Affected Area-Approach ( RAA), and Integrated Rural Development ( IRD), and emphasize zonal development and, increasingly, a range of participatory, community-based methods of assistance. Most recently, High Commissioner Ruud Lubbers revived some of these ideas when talking about the "four Rs" (repatriation, reintegration, rehabilitation, and reconstruction) in his vision for " UNHCR 2004".

Rights-based critiques tend to focus on the breaches of refugee rights, both political and socio-economic, that accompany various assistance methods and generally conclude that camp-based solutions undermine the rights refugees are supposed to enjoy as both refugees and as human beings.

In sum, camp critiques point to the way camp settings prevent integration of refugees and host populations, increase dependency on relief aid, and ignore the resources and capacities of refugees themselves, as well as neglecting the repercussions of a refugee influx on the host populations.

On the other hand, "defenders" of camps emphasize their advantages in facilitating organized repatriation of refugees, attracting international assistance due to the higher visibility of impact, and their superior ability to monitor and target recipients and distribute aid faster and more effectively, especially in the short-run and in immediate emergency situations. They point out that in many refugee-hosting countries, international standards of assistance are most easily upheld in a controlled setting. This is in particular the case for curative health care and (primary) education facilities.

However, "in principle" some basic agreement exists among both policy-makers and academics about the frequent undesirability of refugee camps (see the UNHCR Emergency Handbook). The crux of the debate is therefore about two questions:

1. How to evaluate the trade-offs between the recognized negative effects of camps and their advantages under a range of financial, political, and time constraints that prevent the pursuit of an ideal assistance programme.
2. The degree to which alternatives to camps are politically and financially feasible. Here the debate about camp or settlement solutions frequently ends in a common agreement on the undesirability of camp approaches, only to usher in a debate about their necessity for political and logistical reasons. This second aspect deals no longer with the effects of settlement patterns, but concentrates on the factors that initially cause and later sustain them. Most texts in this review deal at least implicitly with the second set of questions. There is in addition a small but increasing amount of studies that deal with the determinants of refugee policy and the way it is defined by host states and UNHCR as well as the role local, regional, and global factors play in shaping it.

[TOP]