University of Sussex
Falmer, Brighton BN1 9SJ
These working papers provide a means for UNHCR staff, consultants, internsand associates to publish the preliminary results of their research on refugeerelated
As the UN Refugee Convention passes its 50th anniversary, the nature and scope of
the ‘international refugee regime’ continues to be a matter of debate. The last decadehas seen a number of arguments to extend the regime, and/or the Convention. Mostrecent amongst these is the growing consensus that ‘internally-displaced persons’
(IDPs) should be brought under some form of international protection and/or
assistance (Holbrooke, 2000). Another strong group of candidates for inclusion hasbeen those displaced by development projects (Cernea and McDowell, 2000).
This paper, though, is concerned with a third group, who although their case for
consideration has been somewhat sidelined in recent years, nonetheless represent an
important group of interest to many policy-makers at international level:
Estimates of the number of ‘environmental refugees’ in the world vary widely, as do
definitions and typologies of such flows. The term was first popularised by Lester
Brown of the Worldwatch Institute in the 1970s, but perhaps the most quoted
contributions on the subject are those of El-Hinnawi (1985) and Jacobson (1988). Thelatter’s estimate of 10 million environmental refugees has been repeated by numerousauthors, albeit witho ut independent verification of its accuracy.
More recently, Myers and Kent (1995, 18) have described environmental refugees as‘persons who no longer gain a secure livelihood in their traditional homelandsbecause of what are primarily environmental factors of unusual scope’. Myers (1996) has suggested the total number of environmental refugees may be as high as 25
million, putting this group numerically well ahead of the ‘political’ refugees currently
of concern to UNHCR. Nonetheless, the term has been vigorously criticised by,
amongst others, McGregor (1993) and Kibreab (1994) for being poorly defined and
legally meaningless and confusing.
This paper seeks to go further in questioning the value of international policy-makers
focusing on ‘environmental refugees’ as a significant group of migrants, deserving of
the world’s attention. It is argued that although environmental degradation and
catastrophe may be important factors in the decision to migrate, and issues of concern
in their own right, their conceptualisation as a primary cause of forced displacement isunhelpful and unsound intellectually, and unnecessary in practical terms. Particular
reference is made to three categories of supposed ‘environmental refugees’: thosefleeing ‘desertification’; those displaced (or potentially displaced) by sea level rise;
and victims of ‘environmental conflict’. Following on from this, possible reasons for
focusing on ‘environmental refugees’ as a policy strategy are subjected to criticalscrutiny.
Environmental change and environmental refugees: the evidence
An initial difficulty in dealing with ‘environmental refugees’, or ‘environmentalmigrants’, is that there are perhaps as many typologies as there are papers on thesubject. El-Hinnawi (1985) and Jacobson (1988) started with three sub-categories of environmental refugee, namely temporary displacement due to temporary
environmental stress; permanent displacement due to permanent environmentalchange; and temporary or permanent displacement due to progressive degradation of
the resource base. In contrast, IOM/RPG (1992) drew distinctions between emergency
vs. slow-onset movements, temporary, extended and permanent movements, and
internal and international movements. Suhrke (1993) divided her discussion into
migration stimulated by deforestation, rising sea levels, desertification and drought,
land degradation, and water and air degradation, before proceeding to identify
environmental pressure points at which the combination of such factors establishes asusceptibility towards environmental migration.
Trolldalen et al. (1992) distinguished between refugees from natural disasters;
degradation of land resources; involuntary resettlement; industrial accidents; theaftermath of war; and climatic changes. More recently, the ball returned to IOM(1996), which used a six-fold division similar to that of Trolldalen et al., but also
drew an overall distinction between ‘natural’ and ‘man-made’ causes.
Whatever the precise definition or number of ‘environmental refugees’, a common
feature of the literature is to talk of ‘millions’ of displaced people, and their dramaticimpact on host regions, such that regional security is threatened. The image is one of
mis- or over-use of the environment leading to progressive decline in the resourcebase, and possibly contributing to further dramatic (and unintended) environmentalcollapse. Environmentalists and conflict specialists see common cause in discussion
of ‘environmental refugees’; even if the linkages between environmental change,
conflict and refugees remain to be proven. It is the purpose of the first part of thispaper to examine the evidence for such linkages.
At first glance, the data available on environmental refugees appears quite impressive.
A number of areas of the world are cited by a range of authors as being affected by
environmentally-induced migration, ranging across Asia, Africa and Latin America.
In Ultimate security, Myers (1993d, 189) starts a chapter on the potential for
displacement due to sea-level rise (see below) with concern about the plight of
Haitian boat people, ‘abandoning their homelands in part because their country hasbecome an environmental basket case’.
Homer-Dixon (1994, 22) draws, amongst other examples, on the evidence from South
Asia, where the piecing together of demographic information and experts’ estimatesleads him to conclude that Bangladeshi migrants ‘have expanded the population of
neighbouring areas of India by 12 to 17 million’ over the last forty years, whilst ‘thepopulation of the state of Assam has been boosted by at least seven million’. El
Hinnawi (1985) and Jacobson (1988) cite additional examples of environmentalrefugees from across the Sahel and the Horn of Africa, as well as the Soviet Union
and the United States.
However, despite the breadth of examples provided in the literature, the strength of
the academic case put forward is often depressingly weak. Taking first HomerDixon’sexample of migration from Bangladesh to India, caused by ‘environmentalscarcity’, it is something of a surprise to find that, even in his own article, a number of
other explanations for migration vie with that of environmental degradation. Thusmigration is also associated also with rules on land inheritance, the system of water
management in Bangl adesh, the standard of living in India, and the encouragement of
migration by some Indian politicians eager to gain new voters.
At the same time, the source of Homer-Dixon’s demographic information (Hazarika,
1993) casts some doubt on the statistics too – in this case, between 12 and 24 million
such migrants. For example, Hazarika’s estimate of migration to Assam comes notfrom any direct measure, but from a comparison of the 1951 and 1991 Assam censusfigures, adjusted for the population growth rate in 1951, which shows a notionalexcess population. Yet this increase could be accounted for in a number of ways, with
likely candidates including: a rise in the population growth rate after 1951; undercounting
in 1951 (eminently plausible in a remote region); or over-counting in 1991
(also eminently plausible given the link between population size and allocation of
Other accounts of ‘environmental refugees’ are little better. A search through thereferences cited in another recent review paper by Ramlogan (1996) reveals littleconcrete evidence either. Thus for migration from ‘natural disasters’, we are pointed
again to work by Hazarika (1993), as well as to Mattson and Rapp (1991), Sanders(1990-91) and an unspecified report from the United States Embassy in Addis Ababa.
Yet Mattson and Rapp are respectively a climatologist and a geomorphologist, who
merely state that ‘refugee migration is linked to drought and famine’ rather than
demonstrating the linkage; Sanders indiscriminately describes as ‘environmentalrefugees’ some 4.1 million rural-urban migrants in Northeast Brazil in the 1960s and afurther 4.6 million in the 1970s, even though he admits that many areas not affected
by drought also lost population as a result of poverty; whilst for Ethiopia, Ramlogan
simply repeats an observation originally cited by Jacobson (1988) that one million
people ‘were about to move because of famine conditions’ -without actually saying
whether they did (Ramlogan, 1996, 83).
Similar difficulties emerge for the effects of ‘long-term environmental degradation’,
where Ramlogan curiously cites the ‘Black triangle’ of the Czech Republic, Poland
and south-east Germany as a region of out-migration due to pollution (a judgementthat might be questioned by German authorities who have spent the last few decadestrying to deal with in-migration!). And on the aftermath of war, Ramlogan makes theextraordinary assertion that the failure of Afghan refugees to return to their country
from Pakistan is due to poor land productivity and the number of land mines – when
surely the continuing conflict in their country of origin, and the largely favourableeconomic conditions they have experienced in exile might be considered more (or as)
Such problems str ike to the core of the literature on environmental refugees, and
nowhere more so than in the generation of statistics on its prevalence. In turn, thegeneration of statistics is critically dependent on the definition of ‘environmentalrefugees’, a process which might well be seen as impossible given the multiple and
overlapping causes of most migration streams. In so far as distinctions between causes
can be drawn, the following sections consider three different types of ‘environmentalmigration’ and the evidence that has been put forward for the existence of thesephenomena. It is evidence that is far from convincing.
A ‘myth’ extended: desertification-induced displacement?
Out of the range of environmental migration ‘types’ cited above, perhaps one of themost pervasive in terms of popular recognition at least is that of the poverty-stricken
(and usually African) farmer who is finally forced to leave the land because of
drought, progressive impoverishment of the soil, and ultimately famine. The phrase‘desertification’, conjured up in the 1970s to evoke the relentless onward march of thedesert, but with its origins in colonial concern about mismanagement of theenvironment (Swift, 1996), evokes too the flight of humans towards less hostile lands,
or more likely to ‘refugee’ camps. Particularly in the Sahel, but also in other
‘marginal’ semi-arid areas across Central America, Asia and even southern Europe,
desertification-induced migration epitomises the ‘threat’ posed to industrialised
societies by an army of the poor and starving on the move. As Jacobson (1988, 6) putit:
Desertification … has irreparably damaged millions of hectares
of once productive land and made refugees out of millions of
sub-Saharan African farmers. Migration is the signal that land
degradation has reached its sorry end.
However, the evidence for ‘desertification’ causing migration in any straightforward
way is somewhat limited. First, it is important to note that the concept of
‘desertification’ itself has come under fire in recent years, particularly as availability
of satellite images of the region has improved. Thus work by Dregne and Tucker
(1988) and Tucker et al. (1991) has shown a highly elastic response of vegetation
cover to growing season rainfall, with the ‘desert margin’ in the Sahel fluctuating
from year to year as a result.
Williams and Balling (1996, 50) question as ‘equivocal’ evidence that human
activities have changed climatic patterns through influencing surface albedo, surfaceroughness, plant cover and soil moisture. Mortimore (1989) has noted thatmanagement practices that were thought to contribute to land degradation need to beplaced in historical context. Overall, there is increasing talk of the ‘myth’ of
desertification (Helldén, 1991; Thomas and Middleton 1994; Swift, 1996).
If one accepts the argument that desertification itself is largely a myth, then it is not,
perhaps, too great a step to suggest that desertification-induced migration is a myth
too. Nonetheless, even if there is no secular trend of declining vegetation cover and
land productivity in the Sahel, and vegetation recovers as rainfall increases, it ispossible that stress migration might result from a temporary decline in theproductivity of agricultural and grazing land during drought periods. Yet, for such
migrants to be termed ‘environmental refugees’, it seems reasonable thatenvironmental decline should represent the main (if not the only) reason for their
In practice, such evidence is hardly forthcoming. For example, in one review of
desertification-induced migration world-wide, Schwartz and Notini (1995) citeexamples from Mexico, Haiti, and the Sahel, as well as the cases of north-east Brazil,
and north-west India discussed above. But each case is problematic. In the case of
Mexico, after a review of general environmental problems in the country, Schwartzand Notini provide only a brief discussion of an attempt to statistically correlate areasof emigration with areas of ‘aridity’. They go on to admit that not all arid areas are‘degraded’, that not all migration from these areas is necessarily the result of
desertification. Their rather lame conclusion is that ‘our discussion with experts,
research, and analysis of the relevant statistics data will likely confirm thatdesertification is a factor contributing to migration from this region’ (Schwartz and
Notini, 1995, 82: my italics).
Elsewhere, and reliant on studies funded by the Universities Field Staff Internationaland published by the National Heritage Institute, the evidence is little moreconvincing: for example, in Haiti, it is stated that deforestation and soil erosion aresevere problems facing the country, but no clear link is demonstrated to migration.
Indeed, for the case of emigration, it is stated that ‘it is evident that most of Haiti’semigrants in recent years have been political and economic refugees’ (ibid, p.88, my
italics), influenced only ‘to some extent’ by environmental deterioration.
In other work too, the evidence that is presented for migration as a result of droughtand desertification is generally only the existence of migration from regions that areprone to such processes. A causal link to drought is seldom established, whilst in
some cases, not even the existence of ‘excess’ migration is demonstrated. ThusJacobson (1988) cites a number of Sahelian states in which rural-urban or north-south
migration occurred during the drought period of the mid-1980s, or in which
significant populations became dependent on food aid, and all of this is taken as primefacie evidence that these groups have been forced from desert margins because of
However, within the Sahel, and indeed in other semi-arid regions, there is a tradition
of migration that extends back over decades, and often centuries, and which rangesfrom nomadic pastoralism to long-distance trade, as well as the permanent relocation
of individuals and families. In turn, these migrations, though rooted certainly in thedifficult environmental conditions of the region, and the need to diversify incomeearning
opportunities, are not necessarily related to a decline in those conditions(Cordell et al., 1996; Rain, 1999). Indeed, there is now an increasing body of
literature on migration, both internationally from, and locally within the Sahel region,
which suggests that a simple link between poverty, environmental degradation and
migration is hard to sustain.
In the Senegal River Valley, one of many source areas for migration to Côte d’Ivoire,
Lericollais (1989) notes that migration has long reflected a household strategy to copewith environmental risks, which although severe, are not necessarily regarded asworsening. Studies have identified how migration plays a cultural role in thetransition to manhood, as well as being economically linked to the generation of
sufficient revenue to buy livestock (USAID, 1990; Velenchik, 1992).
Factors such as the decline of markets for traditional cash crops (which include gumarabic and cotton), the development of Senegal’s groundnut basin, and subsequentmechanisation of agriculture in the delta provide additional and more recentmotivations to move out of the middle and upper parts of the Senegal River Valley
(see Adams, 1977). Moreover, such conclusions are not limited to the western Sahel,
but can be extended across the continent. As David (1995, 18) notes from an
empirical study based in Senegal, Mali, Burkina Faso and Sudan:
Migration does not necessarily signify a rejection of a rural
livelihood. Rather, it demonstrates that the survival strategies of
rural Sahelians are not only rooted in their immediate vicinity,
but are also linked into economies in other rural and urban
locations. It is precisely this inter-linkage which supports ruralcommunities and helps them to survive in such climatically unstable environments.
The picture is one of migration as an essential part of the economic and socialstructure of the region, rather than a response to environmental decline -a picturereinforced by numerous other studies that have confirmed the critical role of migrantremittances in household and regional economies (Condé and Diagne, 1986; Horowitzet al., 1990).
The situation appears similar in other semi-arid regions of the world allegedly proneto desertification and related migration. For example, Glazovsky and Shestakov
(1994) argue that currently 40 per cent of the population of the former Soviet Union
are living in areas characterised by ‘acute ecological situations’ that are adequately
described as desertification. But they also admit that migration from such areas is notnew, including as ‘desertification-induced migration’ such movements as themigration of Mongolian tribes northwards in the second century B.C. due to drought,
or the removal of population from the Khoresm oasis in the first century A.D. after
the invasion of nomadic tribes which destroyed irrigation systems.
This notion of ‘environmental refugees’ hardly tallies with arguments about recentdestruction of the ecological balance by modern society; rather, migration is aga in
perhaps better seen as a customary coping strategy. In this sense, movement of peopleis a response to spatio-temporal variations in climatic and other conditions, rather than
a new phenomenon resulting from a physical limit having been reached.
For the ‘environmental refugees’ thesis to be plausible in the Sahel and other semi arid
regions, what is required is not simply evidence of migration from what havealways been harsh, marginal environments; rather evidence is needed of an increase in
migration at times, or in places, of more severe environmental degradation. Such aprocess is hinted at in discussions of ‘stress migration’ in the Sahel, one of five phasesof response to famine identified by Cutler (1984), the others being sale of stock, wagelabour, borrowing of cash or food, and the sale of valuables. Yet as Pottier (1993)
observes, there are a number of analytical question marks both over developing
typologies of responses to famine, and especially over assuming that these occur in asequence, the last, and most severe of which is migration.
For Pottier, migration is not an ‘end result’ which can be labelled simply as a‘problem’, but often forms part of the solution to famine for those concerned. In each
case, the dynamic causes and consequences of migration need to be investigated, notassumed. Nor are migrants from drought necessarily ‘refugees’ even in the broad
sense of the word. Indeed, Turton and Turton (1984, 179) reported how the Mursi of
Ethiopia responded to the 1970s drought through a strategy of ‘spontaneousresettlement’ in which they systematically avoided distributed relief at institutionalfeeding points -which might ‘have turned large numbers of Mursi into permanentrefugees in their own country’.
Some of the evidence that does exist specifically on migration responses to
environmental stress points at least in part in the opposite direction. Thus a study by
Findley (1994) of emigration from the Senegal River Valley in Mali shows thatduring the drought of the mid-1980s, migration actually declined rather than
increased. In turn, there was a clear reason for this, since to migrate requires an initialcash investment to pay for travel and associated expenses on arrival, and an economicdownturn reduces the ability of families to ma ke such an investment. However, there
was an interesting nuance to this finding, in that whereas mainly male migration
(defined as departure for a period of six or more months) declined, the process of
circulation (defined as departure for less than six months, and involving many morewomen and children) did increase during the most severe period of the drought.
In a similar vein, Davies (1996) talks of the difference between ‘coping’ strategies(such as temporary circulation) and ‘adaptation’ to drought, the latter involving morepermanent and irreversible changes in livelihood, and usually an increase in poverty
and vulnerability. It is less than clear that migration represents a prominent form of
‘adaptation’ in the Sahel.
Refugees from rising seas?
Where there is perhaps some more justification of the notion of environmentalmigrants (if not ‘refugees’) is in the case of more dramatic and permanent changes to
the environment associated with catastrophic events such as floods, volcanoes and
earthquakes. Sometimes such natural events involve temporary displacement, as in thecase of the Kobe earthquake of 1995, where, according to the Japan Times, an initialfigure of displaced of over 300,000 fell to below 50,000 within three months of thetragedy. Simi larly, the floods of March 2000 in central and southern Mozambique sawthe forced displacement of up to a million people, but a few months on from thistragedy, most had been able to return to their homes.
More significant for discussion here, however, is the interaction of such ‘natural’ and
irreversible events with processes of human-induced environmental degradation: in
other words, examples where a failure to observe principles of good environmentalmanagement and sustainable development can be seen to have contributed to theenvironmental decline that is at the root of displacement. In this context, perhaps themost significant argument for ‘environmental refugees’ -and a main plank of theargument of writers such as Norman Myers – is the predicted effect of human-induced
climate change, and the impact this may have on sea-level rise, and increased flooding
of low-lying coastal areas.
A relatively simple assessment is needed to estimate the populations ‘at risk’, with
Jacobson (1988) for example suggesting that a one metre rise in sea level could
produce up to 50 million environmental refugees. Myers again quotes a higher figure,
with a forecast of 150 million environmental refugees by 2050 (Myers, 1993d, 191),
and it is this figure that is used by the Intergovernmental Panel on Climate Change(IPCC), the UN scientific body responsible for reviewing the causes and impacts of
climate change, in its calculation the costs of not responding (Bruce et al., 1996).
Myers (1996) has subsequently put the potential number at 200 million environmentalrefugees from sea-level rise alone.
Nonetheless, the question of predicting how many people might be forced to leavetheir homes as a result of shoreline erosion, coastal flooding and agriculturaldisruption linked to climate change is far from being straightforward. In particular,
although Myers identifies a number of parts of the world, including Bangladesh,
Egypt, China, Vietnam, Thailand, Myanmar, Pakistan, Iraq, Mozambique, Nigeria,
Gambia, Senegal, Columbia, Venezuela, British Guyana, Brazil and Argentina, asbeing threatened by ‘even a moderate degree of sea-level rise’ (Myers, 1993d, 19495),
and is able to point to figures for flood-related deaths in these regions, he doesnot identify any specific populations that have been forced to relocate from floodproneareas in the recent past as a result of sea-level rises that have already occurred.
The point is that there are many potential responses to increased flooding, of which
migration is only one. Some of the rural-urban migration that has occurred in areasprone to flooding has been to cities that are hardly better placed to withstand the effects of sea-level rise.
In general, calculating the population ‘at risk’ from sea level rise is a long way frompredicting mass flight of a ‘refugee’ nature with its attendant need for internationalprotection and assistance. For example, in a study of response to floods in
Bangladesh, Haque and Zaman (1993) point out that there are a range of adaptiveresponses by local populations, which include forecasting, the use of warning
systems, flood insurance, relief and rehabilitation efforts. Interestingly, they note that‘in contrast to the English meaning of “flood” as a destructive phenomenon, its usagein Bengali refers to it as both a positive and a negative resource’ (Haque and Zaman,
Earlier work by Zaman (1989, 197) stressed how in Bangladesh, ‘whilst erosion
removes land, new land appears elsewhere’, which can be ‘used immediately after itre-emerges’. As a consequence, although 61 per cent of his study population in thedelta had been displaced, 90 per cent of these households had moved less than two
miles from their original location.
Environmental conflict and refugee movements
In addition to the possibility of a direct link between deteriorating environmentalcircumstances or dwindling natural resources and induced migration, a further
postulated cause of ‘environmental refugees’, and a link back to the literature on
‘political refugees’, is the notion that environmental degradation is increasingly at theroot of conflicts that feed back into refugee movements. This has become a major
theme of the literature on ‘conflict studies’ as East-West rivalry is no longer aconvenient explanation of war, and other factors behind conflict and forced migration
need to be found.
However, a review of major conflicts that have caused large-scale forced migration
during the 1990s, for example, provides little evidence of the generation of
environmental ‘hotspots’ that have developed into war. Thus of the eleven distinctconflicts identifiable as being behind ‘recent’ forced migrations (i.e. since 1990),
some, far from reflecting disputes over declining natural resources, could be better
described as conflicts in whi ch the protagonists are attempting to control already or
potentially-rich natural resources.
The Gulf War of 1991 occurred as a result of one oil-rich nation seeking to control itsoil-rich neighbour; the current war in Sudan is also at least partly about control of
oilfields in the south and the building of a canal to open up the southern region
(Collins, 1990), whilst Azerbaijan and Kazakhstan, currently undergoing oil-led
booms, are hardly the poorest of the former Soviet republics. Both of these latter
conflicts, and others, ranging from the former Yugoslavia and the Great Lakes to
Bhutan and Burma, might be seen to have more to do with the rise of ethnic (and/or
religious) nationalism than overtly environmental conflict.
Of course, in some cases, and particularly in the ‘complex political emergencies’ of
the Great Lakes, Sierra Leone/Liberia, and Somalia, environmental issues can be seen
to have some relevance in the development of hostilities, and a case can be made thatenvironmental degradation forms an important root cause of the conflict. In Rwanda,
an extreme position is put by Diessenbacher (1995, 58), who argues thatoverpopulation not only caused the genocide in Rwanda, but had ‘an exponentialeffect on other influencing variables’.
Although his thesis does not rely on environmental degradation per se, but rather thefailure of the productivity of the environment to keep up with population growth, a clear link with environmental degradation is identified. In general, the imageportrayed of muc h writing on Rwanda since the genocide is of a poverty-stricken
country in which the conflict was somehow linked to the inadequacy and deterioration
of the resource base, such that the war was partly a struggle over scarce natural
However, an alternative perspective quite reasonably locates the recent conflict in
Rwanda in a political struggle for power, in which ethnicity and access to naturalresources were both mobilised as issues by powerful élites (Lemarchand, 1994;
Prunier, 1995; Reed, 1996). Equally, the history of the region, and especially thehistory of colonial policy of ‘divide and rule’ of populations that had previously lived
together over centuries (albeit not always in perfect harmony) can also be seen ashighly relevant to the ge nesis of the conflict (Davidson, 1994; Mamdani, 1996).
Indeed, the conflict itself appears now to be taking on a regional character, not limited
to the zone of high population density in the Great Lakes itself (Pottier, 1999).
Similarly, in the case of the war in Liberia and Sierra Leone, Richards (1996, 115)
reviews the evidence for an environment-conflict link, but concludes that ‘no directconnection between deforestation and the war is found’; in essence, although Liberiaand Sierra Leone have environmental problems, they do not have environmentalcrises. Instead, Richards argues, the causes of the war need to be sought elsewhere.
Regardless of the particular root of the war, an analysis such as that of Kaplan (1994,
46) which links together ‘disease, overpopulation, unprovoked crime, scarcity of
resources, refugee migrations, the increasing erosion of nation-states and internationalborders, and the empowerment of private armies, security firms, and internationaldrug cartels’, provides little causal explanation but much passion in an ‘analysis’ thatis symptomatic of much of the field.
Elsewhere in the world, and in earlier conflicts, once again, the evidence for
environmental pressure or degradation (or indeed population pressure itself) actually
causing conflict and forced migration itself is limited. Diessenbacher (1995) suggeststhat of the 181 wars and civil wars worldwide since 1945, 170 have occurred in placessuffering from population explosions. But such an association is not a substitute for
causal analysis, and in detail, it is a thesis that all too often breaks down.
For example, Lazarus (1991) quotes a report by USAID on El Salvador, which arguesthat conflict between the government and rebels in the late 1980s has resulted in
‘fundamental environmental as well as political problems’, but this is hardly evidencethat these problems fuelled a war so much rooted in the international politics of theCold War. In Mozambique, which saw at least three million people displaced abroad
and internally, conflict was again clearly rooted in the Cold War; here, it is interesting
that overwhelming perception of Mozambican refugees on return after the conflictwas that they were going back to a country with unlimited resources and few if any
environmental problems (Black et al., 1998a).
However, it is also quite ironic (and telling) that in one of Africa’s least populated
countries, pressure of population on resources has probably occurred, stimulated notby high population densities per se, but by granting of land concessions to privatecompanies (cf. McGregor, 1997). In Somalia, the history of western (and Soviet)
intervention is so long that it is practically impossible to disentangle from the troubled
history of this war-torn country (De Waal, 1997). The point is that in conflict, asmuch as in migration, it is difficult or impossible to isolate particular causes, outsidethe broader context within which these processes develop; indeed, conflict and
migration themselves are part of a dialectical relationship with this broader ‘context’,
such that a simple causal link from environmental degradation to conflict to migration
is hardly likely to be found.
Environmental explanations of migration: whose agenda?
The examination of statistics on ‘environmental refugees’, and of the detailed casestudies in which this category of forced migrant is supposed to be prominent, are notencouraging in terms of staking out a new area of academic study or public policy.
Yet, the list of international organizations that have stressed concern about‘environmental refugees’ remains impressive. Organizations from the InternationalOrganization for Migration (IOM) to the United Nations Environment Programme(UNEP), and the Intergovernmental Panel on Climate Change (IPCC) have shown an
interest in the concept, sponsoring a wide range of reports and initiatives.
Meanwhile, amongst others, Norman Myers in particular has been prominent in
popularising the term amongst dignitaries ranging from President Clinton to the then
United Nations Secretary General, Boutros Boutros-Ghali (Kibreab, 1997). AsKibreab points out, ‘prominent international personalities are irrelevant in
determining the explanatory or predictive value of a term’ (ibid, p. 21) – but they areimportant in allowing it to gain currency.
It is in this context that the final section of this paper turns to the question of why theterm ‘environmental refugee’ has been so seductive. For Kibreab (1997, 21), theanswer lies in the agenda of policy-makers in the North, who wish to further restrictasylum laws and procedures: thus the term was ‘invented at least in part to
depoliticise the causes of displacement, so enabling states to derogate their obligation
to provide asylum’.
Since current international law does not require states to provide asylum to thosedisplaced by environmental degradation, argues Kibreab, the notion that many or even
most migrants leaving Africa for Europe, or Central America for the US are forced to
move by environmental factors allows governments to exclude a significant number
from asylum. Academics have in turn been complicit in this process by endorsing theterm. This is a plausible explanation, given some force by Westing’s (1992, 205)
observation that the ranks of both recognised and unrecognised refugees ‘are being
swelled by environmental refugees rather than by political or social refugees’.
However, the notion that ‘environmental refugees’ have been talked up by northern
governments seeking to restrict asylum sits somewhat uneasily with the fact that much
of the literature on ‘environmental refugees’ has in practice argued for an extension of
asylum law and/or humanitarian assistance to cover those forcibly displaced by
environmental degradation, rather than endorsing a differentiation between ‘political’and ‘environmental’ causes as a matter of policy.
Thus a report by the World Foundation on Environment and Development and theNorwegian Refugee Council (an arm of the Norwegian government) argued for
establishing a system of protection for environmental refugees (Trolldalen et al.,
1992, 23), whilst the International Organisation for Migration (IOM) and the USbased
Refugee Policy Group (RPG) also concluded that new international instrumentswere needed to provide assistance and/or protection to a group currently ignored by
international policy (IOM/RPG, 1992, 30). Even if the practical impact of literatureon ‘environmental refugees’ has been to endorse northern states moves to restrict thedefinition of asylum still further, this does not appear to have been the consciousintention of many of those writing on the subject.
In fact, one of the ironies of writing on environmental refugees has been that whilstpurporting to highlight a ‘forgotten’ category of forced migrant, which is ignored by
international policy makers, this literature in practice serves only to differentiate a
single cause of migration, which often forms part of a set of reasons why an
individual or family may be forced to relocate. As McGregor (1993, 158) argues,
‘(t)he use of the term “environmental” can imply a false separation between
overlapping and interrelated categories’. But this separation is frequently not made in
practice by organisations such as UNHCR who already use their ‘good offices’ to
provide assistance to a range of groups in ‘refugee-like circumstances’.
In this sense, then, Kibreab is correct to state that to focus on ‘environmental’ causescould lead to the withdrawal of asylum from those who currently receive it – exceptthat the focus here would be much more on large-scale forced migrations inside thedeveloping world (where UNHCR, for example, has much more room to manoeuvrein influencing which populations should receive protection and assistance, and wherestates have not traditionally screened individual asylum applicants), rather than on
asylum in the North, where rules are already very restrictive.
If academic and policy interest in the notion of environmental refugees is not overtly
motivated by a desire to restrict asylum, the question rema ins as to why so much
effort should have been spent in trying to separate environmental causes of migration
from other political, economic or social causes, even to the point of trying to rewritethe definition of a refugee in international law. Arguably, the answer lies not in
asylum literature or policy at all, but in environmentalist literature, as well as in thefield of ‘conflict studies’.
One of the major proponents of the notion of ‘environmental refugees’, Norman
Myers, comes not from a background in migration or refugees or asylum, but from thescience of ecology: in turn, the principle concern of his writing is not migration, butthe imminent threat of environmental catastrophe surrounding climate change (Myers,
1993c, 1993d), deforestation, and desertification (Myers, 1993a). In an article for themagazine ‘People and the Planet’, he points out that he ‘does not assert that theimmigrant problem should be perceived as some sort of threat’ (Myers, 1993b, 28).
Nonetheless, he goes on to suggest that ‘without measures of exceptional scope and
urgency, Europe may have to accommodate growing numbers of newcomers’, and
poses an ominous choice: ‘either to be more expansive in our attitudes towardsneighbouring countries that are also developing countries, or accept that Europe’sliving space will have to become more expansive to accommodate extra people’. In
other words, to do something about the rising tide of environmental refugees also
requires governments to do something about the causes of environme ntal degradation.
This in turn was a point that had not escaped the organisers of a 1994 UN symposiumon ‘Desertification and migration’ at Almeria, in which sponsors of the Convention to
Combat Desertification sought to generate northern (i.e. donor) interest by
highlighting the threat of mass migration to northern countries if nothing were done.
Thus the ‘Almeria Declaration’ produced by the symposium states:
The number of migrants in the world, already at very high levels,
nonetheless continues to increase by about 3 million each year.
Approximately half of these originate in Africa. These increasesare largely of rural origin and related to land degradation. It isestimated that over 135 million people may be at risk of being
displaced as a consequence of severe desertification. (INCCCD,
The driving forces behind this declaration were the representatives of southern rather
than northern governments; indeed the northern academics who attended wereprincipally responsible for ‘talking down’ the figure of populations ‘at risk of being
displaced’ from an initial one billion. It is also interesting to note Kibreab’sobservation that the term ‘environmental refugee’ originated with the United NationsEnvironment Programme – the first, and one of the few UN organisations not to belocated in the North, and seen by many as being more firmly aligned to African rather
than ‘northern’ interests within the UN.
Perhaps more important still in pushing the notion of ‘environmental refugees’ to
centre stage have been writers in the field of conflict studies, as attention has shifted
away from super-power rivalry as the major cause of conflict and forced migration
after the fall of the Berlin Wall and the end of the Cold War.
For example, reporting on a major project sponsored by the American Academy of
Arts and Sciences and the Peace and Conflict Studies Program of the University of
Toronto, Thomas Homer-Dixon (1994) presented three hypotheses on the relationship
between environment and conflict: (a) that environmental scarcity leads to simplescarcity conflicts between states; (b) that environmental scarcity causes largepopulation movement, which in turn causes group-identity conflict; and (c) thatenvironmental scarcity causes economic deprivation and disrupts social institutions,
leading to ‘deprivation’ conflicts. Although Homer-Dixon rejected the firsthypothesis, the latter two were upheld, focusing for example on the Bangladesh and
Northeast India (Assam) case, in which millions of environmentally-displaced peopleare said have contributed to communal conflict.
This theme was taken up by Suhrke (1993), who drew a distinction between
‘environmental migrants’, who respond to a combination of ‘push-pull’ factors prominentamongst them environmental factors -and ‘environmental refugees’,
suggesting that ‘(i)f it is to have a meaning at all, the concept of environmentalrefugee must refer to especially vulnerable people who are displaced due to extremeenvironmental degradation’ (ibid, p. 9).
This distinction is seen in part as having a temporal aspect, as the slow build-up of
environmental degradation is associated with ‘environmental migration’, to befollowed by the reaching of a threshold point at which sudden, absolute and
irreversible degradation induces a flow of refugees. However, such a distinction begsa number of questions, not least how a ‘refugee’ is defined; as McGregor (1993)
notes, the legal definition of a refugee – and ultimately the one that guides governmentand international policy – centres not on the speed of the onset of migration, nor
primarily on whether it is ‘forced’, but on the crossing of an international boundary
and consequent need for protection that cannot be, or is not, provided by the country
of origin. Thus in circumstances where an individual satisfies the criteria for being
labelled a ‘refugee’, the term ‘environmental’ becomes redundant.
In turn, it is unclear that the complex set of factors that lead to ‘environmentalmigration’ as defined by Suhrke would suddenly evaporate or crystallise into a single‘environmental’ cause at the time people become refugees. Although a distinction
could be sustained at the level of proximate causes of flight, this is unhelpful from an
academic point of view if it is accepted that the response to forced migration needs to
be guided by underlying, rather than simply proximate causes.
Taken as a whole, the impression gained from this brief review of existing literatureon ‘environmental refugees’ is one in which lists of factors have overcome theoreticalrigour. There are abundant typologies of ‘environmental refugees’ and ‘environmentalmigrants’, but little agreement on, or understanding of what these categories mightreally mean. Practical concern with the plight of poor people leaving fragile
environments has not translated into hard evidence of the extent or fundamental
causes of their problems. Moreover, there remains a danger that academic and policy
writing on ‘environmental refugees’ has more to do with bureaucratic agendas of
international organizations and academics than with any real theoretical or empiricalinsight.
This is not to say that environmental change – or indeed the existence of high risk
environments with highly variable climatic or other conditions – are not factorsbehind large-scale (and sometimes involuntary) migration. People have historically
left places with harsh or deteriorating conditions, whether this is in terms of poor
rainfall, high unemployment, or political upheaval, or some combination of these or
other adverse factors. Yet, without a firm definition of who is an ‘environmentalrefugee’, it is not easy to say that this category of people is increasing; whilst in amulti-dimensional world, in which people’s decisions to migrate (or stay) areinfluenced by a huge range of factors, an adequate definition does not seem very
If international protection and assistance were to be offered in the future, through theGeneva Convention or some other international instrument, to the supposedly
growing ranks of ‘environmental refugees’, the basis for such intervention would need
to be much clearer than it is at present. To what extent do those uprooted by
environmental disaster, whether temporarily or permanently, have particular
protection or assistance needs? Can it be said with any confidence that addressing the‘root causes’ of their flight (as UNHCR has sought to do for political refugees) would
be any more successful or relevant in reducing ‘environmental’ displacement?
Finally, if protection and assistance were extended by the international refugee regimeto ‘environmental refugees’, would this help or hinder the battle to focus the world’sattention on pressing environmental problems?
Adams, Adrian (1977) Le Long Voyage des Gens du Fleuve. L’Harmattan, Paris
Black, Richard. (1988). Refugees, Environment and Development. London, Longman
Black, Richard, Harrison, Elizabeth and Watson, Elizabeth (1998a). “NaturalResource Management Institutions in Post-Conflict Countries: A Framework
for Research”, Marena Project Working Paper no. 1, Brighton: University of
Bruce, J.P., Lee, H. and Haites, E.F. (1996). Climate Change 1995: Economic and
Social Dimensions of Climate Change. Contribution of Working Group III to
the Second Assessment Report of the Intergovernmental Panel on ClimateChange, Cambridge: Cambridge University Press.
Cernea, Michael, and McDowell, Christopher (2000). Risks and Reconstruction:
Experiences of Resettlers and Refugees. Oxford: Berghahn Books.
Collins, Robert. O. (1990). The Waters of the Nile: Hydropolitics and the JongleiCanal, 1900-1988. Oxford: Clarendon.
Condé, J. and Diagne, P. (1986). South-North International Migrations: a Case Study.
Malian, Mauritanian and Senegalese Migrants from the Senegal River Valley
to France. Development Centre Papers, Paris: Organisation for EconomicCooperation and Development.
Cordell, Denis .D., Gregory, Joel .W. and Piché, Victor (1996). Hoe and Wage: aSocial History of a Circular Migration System in West Africa. Boulder, Co:
Cutler, Peter (1984). “Famine forecasting: prices and peasant behaviour in northern
Ethiopia”. Disasters 8: 48-56.
David, Ros (1995). Changing Places? Women, Resource Management and Migration
in the Sahel. Case Studies from Senegal, Burkina Faso, Mali and Sudan.
London: SOS Sahel.
Davidson, Basil (1994). “On Rwanda”, London Review of Books 18 August.
De Waal, Alex (1997). Famine Crimes: Politics and the Disaster Relief Industry in
Africa. London: International African Institute and James Currey.
Diessenbacher, H. (1995). “Explaining the genocide in Rwanda: how populatoin
growth and a shortage of land helped to bring about the massacres and thecivil war”, Law and State 52: 58-88.
Dregne, Harold E. and Tucker, C.J. (1988). “Desert encroachment”, Desertification
Control Bulletin 16: 16-19.
El-Hinnawi, E. (1985). Environmental Refugees. Nairobi: United NationsEnvironment Programme.
Findley, Sally E. (1994). “Does drought increase migr ation? A study of migration
from rural Mali during the 1983-1985 drought”, International Migration
Review 28(3): 539-53.
Glazovsky, N.F. and Shestakov, A.S. (1994). “Environmental Migrations caused by
Desertification in Central Asia and Russia”, Paper presented to AlmeriaSymposium on Migration and Desertification, United Nations Negotiating
Committee for a Convention to Combat Desertification, February 1994,
Haque, C. Emhad and Zaman, M.Q. (1993). “Human response to riverine hazards in
Bangladesh: a proposal for sustainable floodplain development”, World
Development, 21(1): 93-108.
Hazarika, S. (1993). “Bangladesh and Assam: Land Pressures, Migration and EthnicConflict”, Occasional Paper of Project on Environmental Change and AcuteConflict, Washington, DC: American Academy of Arts and Sciences.
Helldén, Ulf (1991). “Desertification: time for an assessment?,” Ambio 20(8): 372-83
Holbrooke, Richard (2000). “A Borderline Difference: We Ignore Millions Who AreRefugees in Their Own Countries”, The Washington Post, 8 May.
Homer-Dixon, Thomas (1994). “Environmental scarcities and violent conflict:
evidence from cases”, International Security 19(1): 5-40
Horowitz, Michael, Salem-Murdock, M., Grimm, C., Kane, O., Lericollais, A.,
Magistr o, J., Niasse, M., Nuttall, C., Scudder, T. and Sella, M. (1990). Suivi
des activités du bassin du fleuve Sénégal, phase 1, rapport definitif,
Binghampton, NY: Institute for Development Anthropology.
INCCCD (1994). The Almeria Statement on Desertification and Migration, 11
February 1994. Châtelaine, Switzerland: Intergovernmental Negotiating
Committee for a Convention to Combat Desertification.
IOM (1996). Environmentally-Induced Population Displacements and Environmental
Impacts Resulting from Mass Migration. International Symposium, Geneva,
21-24 April 1996, International Organisation for Migration with United
Nations High Commissioner for Refugees and Refugee Policy Group.
IOM/RPG (1992). Migration and the Environment, Geneva and Washington, DC:
International Organisation for Migration and Refugee Policy Group.
Jacobson, Jodi (1988). Environmental Refugees: a Yardstick of Habitability. World
Watch Paper, no. 86, Washington, DC: World Watch Institute.
Kaplan, Robert D. (1994). “The coming ana rchy”, The Atlantic Monthly February
Kibreab, Gaim (1994). “Migration, environment and refugeehood”. In: Zaba, B. and
Clarke. J. (eds.) Environment and Population Change, Liège, Belgium:
International Union for the Scientific Study of Population, Derouaux Ordina
Kibreab, Gaim (1997). “Environmental causes and impact of refugee movements: a
critique of the current debate”, Disasters 21(1): 20-38.
Lazarus, D. (1991). “Climatic change creates environmental refugees”, IDOC
International 2/91: 1-2.
Lemarchand, René (1994). “Managing transition: Rwanda, Burundi, and South Africa
in comparative perspective”, Journal of Modern African Studies 32(4): 581
Lericollais, A. (1989). “Risques anciens, risques nouveaux en agr iculture paysanne
dans la vallée du Sénégal”. In Eldin, M. and Milleville, P. (eds.) Le Risque en
Agriculture. Paris: Orstom.
Mamdani, Mahmood (1996). “From conflict to consent as the basis of state formation:
reflections on Rwanda”, New Left Review 216: 3-36.
Mattson, J.O. and Rapp, A. (1991). “The recent droughts in western Ethiopia and
Sudan in a climatic context”, Ambio 20: 172-5.
McGregor, JoAnn (1993). “Refugees and the environment”. In Black, R. and
Robinson V. (eds.) Geography and Refugees: Patterns and Processes of
Change London: Belhaven, 157-70.
McGregor, JoAnn (1997). Staking their Claims: Land Disputes in Southern
Mozambique. LTC Paper no. 158, Land Tenure Centre, Madison, WI:
University of Wisconsin.
Mortimore, Michael (1989). Adapting to Drought: Farmers, Famines and
Desertification in West Africa. Cambridge: Cambridge University Press.
Myers, Norman (1993a). “Tropical forests: the main deforestation fronts”,
Environmental Conservation 20(1): 9-16.
Myers, Norman (1993b). “How many migrants for Europe?”, People and the Planet,
Myers, Norman (1993c). “Environmental refugees in a globally warmed world”,
Bioscience, 43: 752-61.
Myers, Norman (1993d). Ultimate Security: the Environmental Basis of Political
Stability. New York and London: W.V. Norton.
Myers, Norman (1996). “Environmentally-induced displacements: the state of the
art”. In: Environmentally-Induced Population Displacements and
Environmental Impacts Resulting from Mass Migration, International
Symposium, 21-24 April 1996, Geneva: International Organisation for
Migration with United Nations High Commissioner for Refugees and Refugee
Policy Group, 72-73.
Myers, Norman and Kent, J. (1995). Environmental Exodus: an Emergent Crisis in
the Global Arena, Washington, DC: The Climate Institute.
Pottier, Johann (1993). “Migration as a hunger-coping strategy: paying attention to
gender and historical change”. In Marcussen, H.S. (ed.) Institutional Issues in
Natural Resources Management. International Development Studies
Occasional Paper no.9, Denmark: Roskilde University, 201-33.
______ (1999). “The ‘self ’ in self-repatriation: closing down Mugunga camp, Eastern
Zaire”. In Black, R. and Koser, K. (eds.) The End of the Refugee Cycle:
Refugee Repatriation and Reconstruction. Oxford: Berghahn, 142-70.
Prunier, Gerard (1995). The Rwandan Crisis 1959-1994: the History of a Genocide.
London: Hurst & Co.
Rain, David (1999). Eaters of the Dry Season: Circular Labor Migration in the West
African Sahel, Boulder, CO: Westview.
Ramlogan, Rajendra (1996). “Environmental refugees: an overview”, Environmental
Conservation 23(1): 81-88.
Reed, W.C. (1996). “Exile, reform and the rise of the Rwandan Patriotic Front”,
Journal of Modern African Studies 34(3): 439-501.
Richards, Paul (1996). Fighting for the Rain Forest: War, Youth and Resources in
Sierra Leone. London: International African Institute and James Currey.
Sanders, T.G. (1990-91). Northeast Brazilian Environmental Refugees: Where They
Go. Parts I and II. Field Staff Report, no. 21, Washington DC: Universities
Field Staff International.
Schwartz, Michelle L. and Notini, J. (1995). “Preliminary report on desertification
and migration: case studies and evaluation”. In: Puigdefábrigas J, Mendizábal
(eds.) Desertification and Migrations. Logroño, Spain: Geoforma Ediciones,
Suhrke, Astri (1993). Pressure Points: Environmental Degradation, Migration and
Conflict. Occasional Paper of Project on Environmental Change and AcuteConflict, Washington, DC: American Academy of Arts and Sciences.
Swift, Jeremy (1996). “Desertification: narratives, winners and losers”. In Leach, M.
and Mearns, R. (eds.), The Lie of the Land: Challenging Received Wisdom on
the African Environment. London: International African Institute and James
Thomas, David S.G. and Middleton, Nick J. (1994). Desertification: Exploding theMyth. Chichester: Wiley.
Trolldalen, Jon Martin, Birkeland, Nina, Borgen, J. and Scott, P.T. (1992).
Environmental Refugees: a Discussion Paper. Oslo: World Foundation for
Environment and Development and Norwegian Refugee Council.
Tucker, C.J., Dregne, H.E. and Newcomb, W.W. (1991). “Expansion and contraction
of the Sahara desert from 1980 to 1990”, Science, 253: 299-301 .
Turton, David and Turton, P. (1984). “Spontaneous resettlement after drought: aMursi case study”, Disasters, 8(3): 178-89.
USAID (1990). Senegal Agricultural Sector Analysis. Dakar: United States Agency
for International Development.
Velenchik, A.D. (1993). “Cash-seeking behaviour and migration: a place-to-placemigration function for Côte d’Ivoire”, Journal of African Economies 2(3):
Westing, Arthur (1992). “Environmental refugees: a growing category of displaced
persons”, Environmental Conservation 19(3): 201-7.
Williams, M.A.J. and Balling Jr., R.C. (1996). Interactions of Desertification and
Climate. London: Arnold.
Zaman, M.Q. (1989). “The social and political context of adjacent to riverbank
erosion hazard and population resettlement in Bangl adesh”, Human
Organisation, 48(3): 196-205.