World Bank Reprint Series: Number 481 R E P 4 8 I Michael M. Cernea Bridging the Research Divide Studying Refugees and Development Oustees Reprinted with permission from In Search of Cool Ground: War, Flight & Homecoming in Northeast Africa. Copyright © United Nations Research Institute for Social Development (UNRISD) 1996. Michael M. Cernea Bridging the Research Divide Studying Refugees and Development Oustees I MICHAEL M. CERNEA 21 Studying Refugees & Development Oustees Countries in a post-war situation, and the processes of war-to-peace transition, place the social problems of relocating war refugees and displaced population high on the world's agenda, with increasing visibility and urgency. At the same time, problems with a strikingly similar socio-economic content are brought up by the very numerous instances of 'planned' population displacement caused by development policies and programmes. Within the social science literature dealing with displaced populations, however, there exists an excessive separation between, on the one hand, the study of post-war and disaster-related refugees and, on the other hand, the study of populations uprooted by development projects. Separation entails weak communication. Indeed, these two bodies of social science research virtually do not 'speak to each other'. Following up on my earlier attempt to bring this issue into discussion (Cernea, 1990), the present chapter argues that the dichotomy must be overcome. Both bodies of literature stand to gain from breaking out of their relative isolation. Empirically, they could enrich each other by comparing their factual findings. Theoretically, they could broaden and refine their generalizations by integrating their sets of concepts. Methodologically, they could sharpen their enquiry by exchanging research techniques and approaches. And in the puiblic arena, they could more effectively help displaced popula- tions of all sorts by mutually reinforcing their policy and operational recommendations. The reasons for overcoming this persisting dichotomy are intrinsic to the situation of displaced populations. In this chapter I will discuss some commonalities and differences between situations where populations flee from wars, natural disasters or political events, and situations in which groups of people are forcibly displaced by planned develop- ments. Since both categories are displaced populations, for the purposes of this article I will generally use the term refigees to describe the first category and the term oustees' to describe the second. The chapter first considers the different causes that lead to the creation of either refugees or oustees, and distinguishes sub-categories within these two broad groups; second, it considers the magnitude of refugee and oustee flows, their human rights implications and the impoverishment they cause; third, it examines the different types of assistance provided through international aid to refugees and oustees; and fourth, it explores the similarities or differences that make the comparative analysis between refugees and oustees relevant. The chapter also discusses the policy principles that 'The tcrm oustees is borrowed from the Indian litcrature on involuntary population displacement, where it is commonly used to describe people forcibly 'ousted' from their habitat through govemment intcrvention, generally for the purpose of some development-required change in land or water use; it may not be the best conccpt, and it is rarely used in other countries, but it is employed here as a convenient abbreviation instead of 'development-displaced people'. 293 296 MICHAEL M. CERNEA development programmnes must follow to reduce forced displacements and provide for the social reinsertion of the people they displace. When discussing refugees, the focus of the present chapter is on internal rather than on international refugees. The sharp increase in the number of internally displaced people worldwide during the last two decades, especially in Africa and Asia, has brought these issues into the limelight. Particularly in post-war situations in Africa and recently in eastern Europe, the refugee problem has taken the dimension of a human and political disaster with vast international implications. A Questionable Research Dichotomy Social science research on displaced people has grown faster over the last decade than during any other comparable period. Yet its findings are far from being integrated theoretically or practically. Analytical specialization has also led to a rather rigid research dichotomy: namely, the literature on 'refugees' coexists side by side with a literature on 'oustees' or on 'development-caused involuntary resettlement.' There is little com- munication and mutual enrichment between them. Concepts and propositions are not interlinked, and empirical findings are rarely compared and integrated. For instance, most of the writings on refugees omnit oustee groups from the typology of displaced populations. In turn research on oustees foregoes the opportunity of doing comparative analysis by studying refugees. As a result, the chance for more in-depth treatrnent is being rnissed. The split among these related research fields is not the only one. We can speak, in fact, of a three-way split, rather than a dichotomy. Indeed, another distinct research field dealing more or less with the same processes is the one known as disaster-research. I do not propose to expand here on this 'third domain', which has certainly built an imnpressive body of research on its own. Yet, the victimns of disasters, the populations upon which the disaster research literature bears, are partly the same as in the case of the two other bodies of research literature. The disaster research literature typically centres on people affected by natural catastrophes, many of whom become refugees. It deals comparatively less with refugees from political events such as civil wars or ethnic persecution. Even more seldom does it study oustees from man-made disasters, such as failed resettlement from development programmes.2 This additional split is another ground for an integrative effort, particularly for integrating research findings. The reasons for distinguishing between various populations - refugees and oustees - are certainly valid, and will be emphasized further in this chapter. But do these reasons also justify the absence of comparative analysis, or the lack of overarching concepts and theories encompassing both (and perhaps other) categories of populations? In my view, certainly not. To take an example: as a discipline, anthropology owes a large part of its general knowledge about resettlement to African early experiences with displacements caused by high dars (Brokensha, 1963-4; Cemea, 1994). The Volta resettlement from Ghana's Akosombo and Kpong reservoirs, the resettlement of the Gwembe Tonga in Zambia, or the relocation of the Egyptian Nubians from the Aswan dams, are the best known cases and, alone, have yielded a sizeable body of social science research and publications. Several African scholars have dedicated multi-sided research to the resettlement of the people from the Kainji Lake Basin in Nigeria (Oyedipe, 1983). Social geographers and other researchers have focussed on the Niger River displacements. Recent studies on involuntary resettlement in Africa have brought up new issues, such as the impact of 2 The limniting effect ofthis orientation on the disastcr research itsclfare worth a separate analysis. Bridging the Research Divide 295 reservoirs not only through displacement but also through the dramatic modifications they cause to the traditional recessional agriculture downstream (Horowitz, et al. 1993; Horowitz and Salem-Murdock, 1990). Yet, research on Africa's refugees from wars, ethnic persecution or natural disasters seldom reaches back towards the early literature on development-related displacements on the same continent, for the comparative analysis of distinctions, similarities, or lessons about coping strategies. The fact that books and conferences on refugees usually overlook development- related displacement, and conversely that students of the latter leave refugees out of their research horizon, has a threefold weakening effect. First, it is weakening the political influence of such research on governments; second, it is weakening the design of operational programmes to assist these populations; and finally, it is weakening the effort to build social theory by generalizing larger bodies of empirical findings that result by aggregating what has been called 'contiguous problem analysis' (Dubin, 1978). What makes the comparison between refugees and oustees possible and useful? Both populations undergo a major disruption in their patterns of social organization and cul- ture. They are facing the same task of physically and culturally surviving this disruption by reorganizing their economy and their ways of life. This makes their situations eminently comparable and apt to reveal diverse, innovative, and largely yet unresearched social mechanisms for absorbing disruption shock, for coping, and for re-establishment. Moreover, comparing the response strategies and behaviours adopted spontaneously by refugees and oustees to regain normalcy can yield lessons for assistance approaches. Stepping stones towards this goal have been laid through some scholarly contribu- tions within sociology, particularly by Drabek (1987) and Kreps (1983, 1987). Drabek constructed an inventory of sociological findings about human responses to disasters. The inventory clarifies and codifies responses by type of social group and phase of the disaster's cycle. This scholarly tool is usefil to both researchers and planners: it can help either model building or developing operational approaches for prevention or mitigation. However, this inventory is limited to natural disasters and some aspects of technological hazards. By design Drabek's inventory does not incorporate findings about responses to such man-made, planned events as reservoir submergence. An inventory of the latter category of human responses, facilitating comparisons with responses to natural calamities, would be extraordinarily useful. Another factor favouring such a bridging of the research divide is the recent rapid expansion of the literature on development-caused displacement and relocation, particularly over the last eight to 10 years. This rapid expansion has changed, in my view, the state of the art.3 The expansion itself also embodies this new state of the art. Three features characterize this expansion and stand out: first, the growth of resettlement research by researchers from developing nations; second, the operational bent of most of this literature, its proactive stand and interest in the practical aspects of how resettlement should or should not be done; last, but not least, the policy debate, the transition from description and analysis prescribing solutions and to policy formulation, including the 'translation' of research findings from anthropology and sociology on resettlement into recommended approaches and institutional policies on resettlement (for details, see Cernea 1993a). Developing a systematized inventory of human and institutional responses to development-caused displacement is now an endeavour waiting for a challenge-tempted author. An intellectually stimulating step in this direction is the annotated bibliography on development-caused population displacement by Guggenheim (1994). This bibliography lists some 700 tides, a surprising number which mirrors research expansion in this area. Other stepping stones towards integrating the study of both 'refugee relocation' and 3 The last review of the state of the art was done about 20 years ago (Scudder 1973); taking it as a point of refcrence would allow for an intcresting stock-taking assessmcnt of the vast recent progress. MICHAEL M. CERNEA 'development relocation' have been laid by some scholarly attempts to develop a single analytical framework for both. For instance, Scudder and Colson (1982) delineated such an anthropological framework around the concept of stress and proposed to distinguish among successive temporal stages of the process of dislocation-relocation (see also Scudder, 1991; Scudder reconsidered and amended this framework in his 1993 study, in which he addressed also the refugee-oustee question). In a less formal elaboration, Oliver-Smith and Hansen (1982) suggested that such population movements might be treated as degrees on a continuum of migration processes, ranging from voluntary to involuntary migration - although attributing the concept of 'migration' to planned dis- locations might be a bit of an overstretch. By and large, however, the positive appeal launched by these integrative approaches has still not triggered systematic comparative research. Such investigation still remains to be done. A recent excellent monograph on the resettlement programmnes in Ethiopia (Pankhurst, 1992) is one of the few exceptions in the way it attempts to capture the linkages and real-life overlaps between state- induced resettlement and refugee-causing natural circumstances. Causal Agents of Population Dislocation Any examination of comnnonalities and differences between refugees and oustees should start with the causes that account for such processes of social disintegration. Explaining causality is essential for understanding the origin, identity and composition of various displaced populations, their basic needs, and the ways in which they can be assisted. Among the circumstances leading to forced displacements, we generally distinguish four mnain types or clusters of causal agents: a natural environmental disasters: droughts, famines, floods, earthquakes, etc.; b wars/political turnoil; c persecution: ethnic/racial/religious; and d development policies and programmes causing major changes in land/water use. Events belonging to the first three types trigger massive refihgee flows. Conversely, the fourth type listed above - purposive change processes - often results in the creation of oustee populations. This typically occurs when major infrastructure construction programmes demand compulsory population displacement for acquiring 'right of way'. Such displacements are being carried out through state authority. For want of a more adequate concept, populations forced to move by development programmes are described under a variety of terms: as 'oustees' - the term used in this chapter - or 'evictees', 'displacees', 'resettlers', 'relocatees', 'project affected persons', etc. Within these large groups - refugees and oustees - further conceptual and practical distinctions can be made. Distinctions within refugee populations A distinction made between refugees displaced by one of the first three types of causal agents ('a', 'b' or 'c') is whether or not they have crossed an international border, or have remained in their country of origin. This is a circumstantial distinction rather than a theoretical one, but it is quite important in its practical consequences. Internal refugees are those who, having abandoned their houses and lands, still remain within the borders of their country. International refugees are those who have crossed an international border. Indeed, in the case of racial persecution or civil war, the 'push' factors are much the same for the refugees who cross a national border and seek shelter Bridging the Research Divide 297 elsewhere, as for those who leave their shelter but cannot or choose not to leave their countries and thus become 'internal' refugees.4 The trauma of being an 'internal refugee', a refugee in one's own land, is not necessarily less than being a refugee on foreign soil. The distinction between internal and trans-border refugees is consequential, however, for resouirce allocation and assistance strategies. While there are established inter- national structures mandated to assist international refugees, much less institutionalized support is available for internal refugees. The Georgetown Declaration (1988) correctly called attention to the fact that, despite the huge number of internal refugees, 'at present no international agency has responsibility for ensuring the adequacy of protection and assistance, including health care, for internally displaced persons'. Cross-border refugees are recognized by international agencies as refugees. However, the same agencies tend not to recognize as 'refugees' those displaced people who do not cross borders, namely the internally displaced people, even though both the triggers and the consequences of their displacement may be identical. In practical terms, this difference in formal recognition results in huge disparities in the levels of public and private assistance provided to such groups. These circumstances have prompted refugee experts to state that in the area of refugee assistance 'the challenge for the future is presented by internally displaced people, [where] roles and mandates are undear' (Keely, 1988). Distinctions within oustee populations Returning to ouistee populations, we can make a different kind of distinction among them. According to the degree to which they endure losses, populations displaced by infrastructure construction (or comparable programmes) can be subdivided into the following sub-categories: a people who lose their houses; b people who lose land (in full or in significant part), but not their houses; and c people who lose both houses and farm land or other productive assets. Compensation and other rights and economic entitlements are often denied to some of the people displaced by government-sponsored development plans through a 'simple' device: they are omitted from oustee statistics and, implicitly, from resettlement planning. This happens more often to people in the sub-category (b) above. Technical and planning agencies tend to ignore them on the grounds that their houses are not destroyed and therefore they are not technically displaced. In practice, however, the loss of even one significant part of a household's farmland (e.g., some 25 per cent of the total holding) may make that farm economically unviable and thus compel the family to move away; it may cause marginalization and the sliding of a family from barely above to below the poverty threshold (Cernea, 1990b). For instance, this happens to many people when vast networks of irrigation canals are constructed: the 'right of way' swallows part of a family's cultivated land, but leaves untouched its house, located farther away (Morse et al., 1992). Those who lose only their houses but not their lands (category 'a') may sometimes avoid becoming oustees if alternative house plots can be secured nearby, in the vicinity of their lands and kin. Labelling (designation of status) is a powerful tool of practical policy used by govern- ments for allocating or withholding resources from certain groups, as was correctly argued in the case of refugees (Zetter 1988, 1990). Mitatis mutandis, the denial of oustee status to those in sub-group (b) above, (e.g., as is happening in some programmes in Indonesia or India) is another case of labelling manipulation. This is not just an 4 In certain regions, particularly in Africa where boundaries often split areas inhabited by the same ethnic group, some refugees may not even be initially aware of whether or not they have crossed a border. 29)S MICHAEL M. CERNEA innocuous semantic matter, but a matter of recognizing rights. Mislabeling is a way to belittle the adverse effects of projects through the device of refusing to recognize the losses, rights, and legitimate entitlements of part of the population affected. Under the provisions of the World Bank's policy regarding involuntary resettlement (see World Bank 1990; Cemea 1988, 1995), all three subgroups [isted in (a), (b), and (c)] are defined as project-dislocated populations, and their socioeconomic re-establishment must be carried out by the same projects that displace them. There are differences in the degree to which one or another category is affected by specific losses and this is to be reflected in their overall compensation packages, but without entailing a denial of their status as development affected oustees. The Magnitude of Forced Displacements The magnitude of forced displacement is often not filly realized. Although in some post-war countries large return movements have occurred, the total number of affected people continues to increase. In the early 1980s there were about 11 million cross- border refugees. In 1993, the UNHCR estimated that the figure had risen to 18.2 million (UNHCR, 1993: iii). Subsequently there have been further huge cross-border movements, particularly in Africa. Figures are less easy to obtain for internally displaced populations, but these too seem to be rising. It is now increasingly being recognized that they are even higher than for those displaced internationally. In 1993, the UNHCR estimated that there were about 24 million internally displaced people worldwide, and this figure excludes those forcibly displaced by development programmnes. There are, in fact, no readily obtainable aggregate estimates for this latter group. Lack of information about development-induced displacement is partly a consequence of countries not publishing official statistics about those affected - something which partly explains the insufficient public awareness of the issue. Nevertheless, extrapolated information from recent research makes it clear that such displacement occurs on a very large scale. Some of the most massive and visible individual cases of development related displacement are those associated with dam and reservoir construction. Table 21.1 shows the size of displacements caused, or expected to be caused, by the construction of some major dams in various countries. But these projects are not the only significant causes of development displacement. A recent study of all World Bank-assisted projects entailing resettlement during 1986-1993 has found that the transportation, water supply and urban infiastructure sector now has the largest number of projects involving resettlement in the Bank's project portfolio, representing above 51 per cent (World Bank, 1994). This compares with agricultural projects (14 per cent) and energy and industry projects (about 29 per cent). While these proportions refer to the Bank's project portfolio, they reflect the overall world trend of rapidly increasing urban population displacements (Cemea, 1993b). In addition, massive forced dislocations are entailed by other types of development projects, such as thermal power plants or open-strip mnines. Quite often, seemingly innocuous projects such as establishing an agricultural research centre or a national park, building a hospital complex in a city or constructing a drinking water storage basin, involve land expropriation which deprives many families of their habitats or productive assets. In the absence of country-based statistics, the only way to assess the magnitude of worldwide development-caused displacement is to develop an estimate based on (i) available displacement information from the better known projects, and (ii) investmnent trends in infrastructure. With respect to the former, it needs to be born in mind that project-constructing agencies commonly underestimate the number of people adversely Bridging the Research Divide 299 Table 21.1 People affected by dam-caused displacement Dam Country Number of People Already Built Dongpinghu China 278,000 Sanmenxia China 319,000 Srisailam India 100,000 Assad Syria 60,000 Keban Turkey 30,000 Portile de Fier Romania/Yugoslavia 23,000 Victoria Sri Lanka 31,500 Akosombo Ghana 84,000 Kossou Cote d'lvoire 85,000 Kainji Nigeria 50,000 High Aswan Egypt 100,000 Nangbeto Togo/Benin 12,000 Saguling Indonesia 55,000 Danjiangkou China 383,000 Sobradinho Brazil 60,000 Itaparica Brazil 50,000 Mangla Pakistan 90,000 Currently under Construction Three Gorges China 1.1 million Xiaolangdi China 181,000 Shuikou China 70,000 Ertan China 30,000 Tehri India 105,000 Narmada Sardar Sarovar India 250,000* Almatti India 160,000 Narayanpur India 60,000 Yacyreta Argentina/Paraguay 50,000 Kayraktepe Turkey 20,000 Under Design Kalabagh Pakistan 80,000 Gandhi Sagar India 100,000 Soubre C6te d'lvoire 40,000 Komati Swaziland 20,000 ttf laJ IChisapani) Nepal 50,000 Note: This table is\ based on data from project documents and/or public sources. Some of these projects were co- financed by ti'e World Bank, while others were financed entirely from domestic or other international sources. For most irrigaizion dams already built, data refer to reservoir displacements, without including the number of people displace,d by the canal/road networks in the irrigation command areas. This number in,cludes 127,000 people displaced by the reservoir, reported in 1993 by the Narmada Control Authority, and people affected by significant loss of land to the canal and road networks downstream, as documented b/f the Sardar Sarovar independent Review (see Morse, et al., 1992). 300 MICHAEL M. CERNEA affected by their activities. As a rule, the real numbers of displaced people are likely to be more accurate for internationally assisted projects than for domestically financed ones, which tend to be less scrutinized (and more haphazard when it comes to resettlement operations). Nevertheless, it is essential to include domestic projects because they undoubtedly account for the bulk of forced displacements. Therefore, an estimate of the aggregate magnitude of development displacement worldwide needs to be based on the better verified final counts of people displaced by projects receiving international assistance, combined with an assessment of how that data relates to data on other infra- structural investments. The study of World Bank-financed projects causing displacement, mentioned above, has made possible such an estimate of the overall displacements taking place in the developing world because of dam construction and urban and transportation projects. The study found that in the early 1990s about 300 new high dams (above 15 m) are entering construction every year and that they entail the displacement of some 4 million people. If dams with a height of less than 15 m are included, the total would be considerably greater. In addition, the projects for urban and transportation infrastructure being started every year entail the displacement of some 6 million people, on average. Thus, in both sectors at least 10 million people are being displaced by every year's new cohort of development programmnes. The overall total for all sectors is certainly greater because it should include the oustees from other projects such as thermnal plants, open- strip mining, parks and forest reserves, etc., for which overall estimates are not yet available.5 These aggregate numbers are staggering. Over the last decade, at least 80 to 90 million people have been displaced by programmes in only two development sectors. Population displacement by development programmes is now a worlduide problem, of a magnittude previotisly uinsuispected. Moreover, ongoing industrialization, electrification and urbanization processes are likely to increase, rather than decrease, the number of programmes causing involuntary population displacement over the next 10 years. Particularly in countries engaged in large hydropower progranmes, there are annual 'waves' of newly displaced people. Recent statistics from China, for instance, show that over the last three decades, from the late 1950s to 1989, the number of people displaced by water conservation projects alone exceeded 10 million (Chao, 1990). For India, a 1989 study by Fernandes, Das and Rao estimated that during the last four decades some 15.5 million people were displaced by projects (dams, mines, industries, wildlife reserves, etc.); most of them still have to be 'rehabilitated'. In a follow-up analysis (1991), Fernandes revised upward that initial assessment to about 18.5 million people. Massive displacements also occur under another category of state-sponsored programme aimed at large-scale population redistribution. Such resettlements - many compulsory, and some voluntary but state-promoted - are allegedly initiated to improve the condition of those relocated. Whether or not they actually improve it is a matter to be assessed ex-post in each case. Lassailly-Jacob (1994) has pointed out that three !ri5or reasons are typically advanced for such massive population redistribution prograrames: (i) maximized utilization of a nation's resources; (ii) strategic considerations and, counter- 5 In the methodology for making this estimate wc took into account all available infor'mation on dam construction in the developing world provided by the Intemational Commission on Large Dams (ICOLD) and the World Register of Dams (see ICOLD 1988, 1993; also Mermel 1990). The ICOLD statistics indicate that from 1951 to 1982 an average of 340 dams entered the construction stage.every year. During the early 1990s worldwide construction of high dams has averaged about 300 new diams per year. China alonc has been starting construction of about 150 dams a year in the early 1990s. Becaxuse ICOLD statistics are silent on population displacements, affected populations were taken from World F3ank statistics for the projects it finances. Weighted averges by classes of dam height were used in extmapol-ating population data from the universe of Bank-financed dam projects to the overall number of dams2 reported by ICOLD statistics. Bridging the Research Divide 301 insurgency measures; and (iii) apartheid policies. There have been numerous examples of such programmes, some of them dating back to colonial times. Openly recognizing strategic and counter-insurgency reasons, the French army undertook a massive policy of population regrouipement (regrouping) during the 1954-61 Algerian War of Indepen- dence, to prevent the rural population from actively assisting the guerrillas or from being exposed to fighting. Apartheid policies in South Africa were accompanied by forced mass population relocation into areas knows as homelands (de Wet 1993). The Bulusu of East Kalimantan (Indonesia) and other interior Dayak groups have been forcibly relocated to government-established resettlement centres, with the stated purpose of providing them with services unavailable in their area of origin (Appell, 1985). Other population redistribution programmes have claimed to have agricultural develop- ment goals. Examples indude the relocation of the Azande population in Sudan (1946- 50), the mostly coerced villagization programmne in Tanzania that brought some 5 million scattered peasants into nucleated settlements, or the involuntary relocation of farmers in Ethiopia away from drought-prone and famine areas (Dieci and Viezolli, 1992). The evidence from such schemes is often mixed and controversial. Some of the social scientists who conducted empirical studies have concluded that, in fact, these relocations were politically motivated operations disguised by authorities under a 'development' rhetoric (Clay, 1986, 1988; Kesmanee, 1988; De Mars, 1992). Other researchers contend that more complex economic, social and ecological factors may be at work in such situations. In his subde analysis of the Ethiopian resetdement programmes, Pankhurst argues that the 'stereotypes of resettlement as either purely induced by famine or enforced by the government [are] equally misleading simplifications' (Pankhurst, 1992: 9) . In condusion, the displacement of oustees from development programmes fully invites comparative analysis with the displacement of refugees by sheer magnituide and not only by the natuire of the social problems engendered. The fact that planned programmes often produce long-term benefits (usually for other people) does not make the disruptions any less, or the hardship any lighter, for those who have the misfortune to be uprooted. Human Rights, Entitlements, and Impoverishment Situations engendering refugees, as well as planned displacements caused by develop- ment raise important issues of human rights. Both types of situations cause the impoverishment of populations which often are already marginal or poor. It is important to at least briefly define the main legal points involved, particularly as they regard the oustees - because the discussion of refugees' human rights and poverty conditions has already a large literature supporting it (Zolberg et al., 1989; Downing and Kushner, 1988; Zetter, 1988; Harrell-Bond, 1986). The magnitude and disastrous impacts of displacements worldwide, and the increased visibility of these impacts, have gradually contributed (not alone, obviously) to the significant trends observable in intemational policy and in aid-flows: one is the advent of the human rights argument more to the forefront of the policy dialogue; another is increased concern about the secondary impoverishment effects of programmes that are primarily supposed to help reduce poverty. Human rights and development Legal scholars, anthropologists, and sociologists increasingly explore the ethical, economic and juridical links between human rights and policies for inducing develop- ment (see Shihata 1988, 1993; Paul, 1988, 1989 and 1992; Downing and Kushner, 1988). Anticipating the direction in which the international situation is evolving, it can be predicted that the formulation of future development policies and programmes will 302 MICHAEL M. CERNEA 1 :< 21 . The building of the Kariba dam in the late 1950s has provided electricity and has afforded opporttunitiesfor commercialfishing: boats arriving at the dock of afishing company located on the Zimbabwean shore, after a night on the lake, 1992. (Tim Allen) 10 ' - , C ,,.nc l- - C -== ,~ ~ ~ ~ ~~~- X(j *'VJ' ' I'< J - C ' \ di .s, 21 .2 Many people have benefitedfrom the electric power produced by the Kaniba project, the vast majority of them living in uirban centres: downtown Harare, 1992. (Tim Allen) Bridging the Research Divide 303 increasingly have to take into account the protection of basic human rights, including the rights and entitlements of oustees. Three distinct sets of human rights issues are pertinent to the design and implementa- tion of development projects: a issues regarding general economic rights, particularly those pertaining to sharing equitably to the benefits of development; b issues related to protecting the basic human rights of, and providing emergency assistance for, people affected by civil war, ethnic persecution, or natural disasters; and c issues arising when development programmes adversely affect the rights and welfare of some population segments. The legal questions regarding the definition of rights and entitlements are central to how displacement and resettlement take place. The very decision by the state to expropriate and displace people raises issues regarding people's right to self-determina- tion and the right to control one's own life and fate. Such decisions are based on the state's right of eminent domain, which is to take away private property for the purpose of public use even without the consent of the property owner. The state is normally required by law to pay just compensation to the property owner. Practice shows, how- ever, in case after case, that only a small portion of the losses suffered by displaced persons are recognized in the formal economic and legal systems, or procedures, of most developing countries. Loss of usufruct rights, rights of way, customary rights to land, etc., are poorly or not at all recognized in many national legal systems. In addition, the perception of what should be defined as basic human rights is not just a legal matter. It is dependent on culture as well and on the development of a mature civil society. As Downing correcdy pointed out, the 'precise content of human rights logics varies between and within the same culture at different times. Yet, the logics also tend to share critical, perhaps universal dimensions' (Downing, 1988: 9). For instance, customary rights of indigenous and tribal people, including their access to common property resources (forests, grazing lands, etc.), are recognized in a few countries but not recognized in others. Moreover, even a recognized right to compensation for lost assets may not by itself constitute an adequate means to re-establish the standard of living of the affected population (Shihata, 1993). Under certain circumstances, the denial by the state of displaced peoples' rights and entidements in situations of forced displacement has led to active resistance against resettlement and to sharp political conflicts (Oliver-Smith, 1994). Legal experts and human rights activists are exploring the possibilities of a 'charter' to define and help secure human rights in development processes when it appears that people are put at risk (Paul, 1992 and personal communication). The state resorts to the exercise of eminent domain law when it constructs new infra- structure that is essential for national development, satisfies the needs of many beneficiaries, and generates gains unobtainable in any other way. Because such construc- tions entail the forced displacement of those 'in the way', the state voids the right of private individuals to hold on to their land and residences, compelling them to move elsewhere. Such conflicts, however, are not between private right and public wrong, or between public right and private wrong. Most often, this is a conflict between right and right, when the public sector acts on behalf of the rights of the larger numbers of people that expect to benefit from a certain development, while those adversely affected oppose it on the grounds of their own legitimate right to the place where they live and work. Thus, the fulfillment of the needs of the many6 clashes with the rightfil entidements of the few. This conflict between right and right is not insoluble. It can be negotiated and over- 6 Assuming that the given development will indeed benefit many and not only the sclfish interests of a powerful minority. 306 MICHAEL M. CERNEA come if the public sector, the state, is prepared to fully recognize the losses and pains inflicted on those called to mnake the immediate sacrifice and to accept uprooting. In practical terms the problem is not that the conflict is irreconcilable under any circum- stances, but that the premisses for resolution are often not equitably provided by the public sector. These prenisses consist not only of financial compensation. They also indude democratic consultation; provision of options to those affected through information, participation, and the building up of consent; recognition of local customs and traditions; and fair legal protection. Of course, these premisses also consist of considerable material resources necessary to reconstruct what is being lost - resources that must be contributed by the public sector and by the project's beneficiaries. Each real development that requires displacement usually does generate gains, but the problem is that the gains are only rarely shared by those who bear the pains. The core demand is to ensure that those who are the temporary victims of such projects are also enabled to have their basic rights to socio-economnic progress equitably satisfied. Impoverishment through displacement Like becoming a refiugee, being forcibly ousted from one's land and habitat by a darn, reservoir or highway is not only immediately disruptive and painfil, it is also firaught with serious long-term risks of becoming poorer than before displacement, more vulnerable economically, and distintegrated socially. Most of the research to date - both refugee research and oustee research - has vividly documented the adverse effects of forced dislocation. The most pervasive consequence is that many displaced people remain unable to reconstruct their income generation capacity. While it is easy to see why this happens in the case of refugees, the process is more complex and even more unacceptable in the case of oustees. They are supposed to receive compensation for their lost assets, and effective assistance to re-establish themselves productively; yet, this does not happen for a large portion of oustees. The empirical research evidence reveals a set of recurrent characteristics when the resettlement of oustees does not amount to income restoration and people are left worse off. While each of these characteristics is irreducible to the others, they have a common denominator: they converge into a process of impoverishment. Data from many empirical studies has been synthesized into a model of impoverishment risks resulting from forced displacements (see Cemea, 1990b; 1991). This model shows that development-caused displacements, if not mitigated by policies and approaches that restore oustees to their previous living standards, tend to lead to impoverishment by the following eight typical processes: a landlessness; b homelessness; c joblessness; d marginalization; e loss of access to common property resources; f food insecurity; g increased mortality and morbidity; and h social disintegration. Such adverse effects of development should not be denied or belitded, but must be acknowledged from the outset and addressed with appropriate mitigatory means. The evidence that substantiates these trends is massive, and the above features are derived from the factual descriptions of many cases, notwithstanding the variability of individual displacement situations (Scudder 1981, 1991, 1993; Appell 1985, 1991; Salem-Murdock, 1989; Horowitz, 1991; Ganguly Thukral, 1992; Partridge et ad., 1982; Christensen, 1982; Areeparompil, 1989; Fahim, 1988; Rew and Driver, 1986; de Wet, 1991; and many others). Bridging the Research Divide 305 A. more detailed review of this evidence (see Cernea, 1990b) showed that one unavoidable process in displacement, whose far-reaching pernicious effects are not yet sufficiently recognized, is the one listed last - social disintegration. Established integrated communities as well as social support networks territorial or kinship based, are dismantled through displace- -. ment. Anthropological and econormic . research in developing societies have demonstrated how important the informal networks among households are as a social support system, partly substituting for a. I ii insufficient cash income and for other resources. Informal social networks among households and families help cope with 21.3 As the waters rosefollowing the btilding poverty through informal loans, exchanges of the Kariba dam, thouisands of Tonga were of labour, food, clothing, durable goods, forced to leave their land, and were displaced gifts, etc.; they help in house-building, in into lessfertile locations. In many places they caring for children, in emergencies, etc. have been denied access to the lake shore, which The dismantling of these networks has become a tourist playgrotnd, and their crops through the dispersion of their members are vuelnerable to destnrction by protected animals, causes direct economic losses, in addition to notably elephants and buffalos. The time before social and cultuiral ones. This is a loss of social displacement is recalled wvith longing: a Tonga capital - of patterns of social organization elder, Zimbabwe, 1992. (Tim Allen) able to mobilize individuals in collective actions for common interests, which, if dismantled, are hardest to rebuild. Such loss of already existing social integration - and social capital is compounded in deve-- lopment projects which relocate people in a dispersed manner, rather than in groups and social units. Taken together, the eight key dimen- sions of impoverishment through displace- ment provide a warning model that can inform policies and plans for dislocation, and thus help prevent and rnitigate such consequences. It is crucial to emphasize that impoverishment through involuntary displacement is not inevitable. But avoid- ing it, as will be discussed in the last section , of this chapter, requires a convergence of w -- policy, planning, financial resources, and .X social-technical expertise. - X r 21.4 Since the displacement, many Tonga Emergency Assistance to have had tofarm in locations tnsuitablefor Refugees agriculture: among the stalks of her dehydrated millet, a Tonga woman picks the leaves of wild Growing public recognition of the human shrubs to cookfor suipper, Zimbabwe, 1992. rights issues, ofthe impoverishment processes, (Tim Allen) 306 MICHAEL M. CERNEA and of the political and ethical implications of the drama of both refugee and oustee populations, have led over the last two decades to increased financial and other assistance to displaced populations. Despite the essential differences in the causes of refugees' and oustees' uprooting, there are significant similarities in the consequences. Refugees and oustees resemble each other in that they are both victimized by events for which, as individuals, they cannot be held responsible. Both lose their houses and households; they temporarily or perma- nently lose their lands, water wells, workshops, vending stalls, or other assets.7 In both cases, their production systems are dismantled, their ways of making a living are disrupted and their very livelihood critically jeopardized. The supporting social networks in which their existence was embedded are unravelled. Both relocate in previously unknown places, among host populations often suspicious or directly hostile. Two new trends became manifest in the assistance to displaced populations, particularly during the 1980s. These trends were little analysed by the research com- munity. The first trend is a considerable increase in the ntumber and diversity of international aid projects targeted to disaster victims - refugees and non-refugees; the other is a significant orientation towards distinct project components geared to assist development oistees in their re-establishment. As 'vehicles' for such assistance, governments and aid agencies largely use two categories of interventions: (a) emergency assistance projects, for inmmediate and temporary help to refugees; and (b) resettlement projects (or project components) aimed at permanently relocating the people displaced by planned change. These trends open up new possibilities: on the one hand, refugee relief aid can be better informed with the experience of approaches that pursue not just short-term relief, but redevelopment, as do some projects for oustee resettlement; on the other hand, assistance given to oustees can be improved by learning from the coping mechanisms used by refugees to survive under harsh conditions (Christensen, 1982) and to regain self-sufficiency. The new trends are quite visible in the lending patterns of the World Bank, which are indicative of the presence of refugee and oustee issues on the international aid agenda. After its establishment in 1945, all the Bank's loans from 1947 to 1953 were given for post-war reconstnrction projects. Europe was then home to about 6 million refugees, and those emergency operations were addressed to a wide spectrum of needs and popula- tions, induding refugees. As these problems started to subside, however, in 1953 the World Bank ceased this type of operation. It replaced them with investment lending. Ornly after almost two decades, and because of a serious aggravation of the worldwide problem of refugees, did the World Bank resume emergency assistance operations in 1970. By this time, the refugee problems had shifted dramatically to developing countries, where the bulk of the refugees existed then and now. This assistance is primarily financial and technical. It provides an immediate response to disasters and has benefited large numbers of refugees fleeing natural or socio-political catastrophes (as well as people living in adjacent areas). Especially in the 1980s, the frequency of emergency loans and the amount of resources they channelled have risen sharply. For instance, 76 emergency projects were approved between 1974 and 1991, out of which 19 between 1989 and 1991. The resources channelled through such Bank-assisted projects between 1984 and 1988 were 3.5 times larger than in the prior five years (see Table 21.2). The diversity of these activities is significant: during 1989 alone, 18 loans were approved for post-disaster 7 Yet cven in this respect therc are some differences. Compared to other refugee-causing circumstances, development-related displacements inflict upon the uprooted population material losses that often are greater, or definitive. For instance, refugees fleeing their homelands because of civil war, persecution, or flooding may in some cases at least retum to their farming land; those displaced by dam reservoirs, however, will never get the same lands back, and simply cannot return to their prior homes. The compensation most receive for expropriation is supposed to help them regain substitute asseu. Bridging the Research Divide 307 Table 21.2 Emergency projects by type of disaster: 1969-91 FY69-73 FY74-78 FY79-83 FY84-88 FY89-91 No. US$m No. US$m No. US$m No. US$m No. USSm Civil war 4 165.6 2 60.7 7 257.5 2 123.0 5 245.0 Cyclone/hurricane 1 25.0 - - 5 76.2 3 25.6 3 254.0 Drought - - 8** 35.0 - - 3 400.0 1 16.0 Flood - - 3 135.0 1 7.0 5 345.5 5 301.1 Earthquake 2 50.0 2 86.5 3 85.0 4 580.0 5 469.3 Forest fire - - - - - - 1 56.9 - - Volcanic eruption - - 1 7.0 - - - - - - Total 7 240.6 16 324.2 16 425.7 18 1,531.0 19 1,285.4 Notes: I No emergency assistance loans were made by the Bank between Financial Year (FY) 1954 and (FY) 1969. For the period subsequent to FY69, this table includes only the Bank loans and projects labelled 'emergency operations', but, in addition, many other Bank projects also provided assistance to refugee populations. For instance, two important refugee assistance projects were targeted to Afghani refugees in Pakistan, but these are not included in the table. The emergency interventions reflected in this table have benefited, of course, also those disaster victims who did not become refugees. This table includes only new projects and new full scale loans/credits, which were additional to the emergency assistance provided through the speedy re- allocation of unused proceeds from other on-going Bank operations. * Includes six loans to the Sahelian countries in FY74 for a series of inter-related drought relief projects. reconstruction projects; of these, one was a post-civil war reconstruction project in Mozambique (urban rehabilitation); two were drought-assistance projects (India); five were post-hurricane projects (one in Costa Rica, one in Mexico, and three in Jamaica); two were post-earthquake projects (both in Nepal) and eight were post-flood assistance. Post-war reconstruction issues have received increasing attention in the Bank's work during the 1990s, particularly during 1993-1995. This has included experiments with new approaches in assistance projects, as well as in-depth field analysis of the demobiliza- tion and reintegration experiences in post-war countries such as Ethiopia, Namibia, Uganda (World Bank, 1995). A strategy discussion paper on 'post-conflict reconstruction' is currently preparing the grounds for broadening Bank lending and technical assistance to countries undergoing war-to-peace transition processes (Holtzman, 1995, draft). Strategy Alternatives in Assisting Refugees A key strategy issue in refugee assistance is always the trade-off between short-term relief needed immediately and investments for long-term redevelopment. The World Bank emphasizes the latter type of measures. Particularly after war or civil war periods - during which production activities and trade lines had been disrupted - Bank projects primarily aim to restore production and productivity and to rebuild the transportation systems required for recovery. This approach also creates employment. The Bank's 1984 policy guidelines on emergency lending, revised and improved in 1989 (World Bank, 1989), formally recommended to use, during post-disaster assistance,-what is being called the 'window of opportunity' for preventing/mitigating futu¸ disasters: namely, financing certain preventive measures which tend not to be seen as a priority in normal times, yet are likely to be adopted in addition to immediate relief8 In such situations, resources are rovided both for the reconstruction of damaged physical infrastructure and for building additional preventive physical and organizational protection systems. Africa is the continent where most refugee-assisted projects have been carried out. An S Weighing and deciding on trade-offs in such situations is vulnerable to imperfect initial information. Built- in flexibility for midstream adjustments is needed. This is why World Bank assistance usually needs to be complemented by short-tenn relief programmes by other agencies, including non-govemmental organizations. 308 MICHAEL M. CERNEA ;~~~~~~~~~~~~~~~~~~~~ 0¾/ '*S '-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2.5 In Mozambique, the World Bank has provided emergency assistance to suipport the government's immediate rehabilitation efforts throuigh quiickly disbuirsed hard cuirrency to meet priority needs: retuirnees to Mozambique, 1990. (S. Ernngton/UNHCR) example of a recovery programme after war and civil disturbance is the Bank's Recon- struction Credit to Uganda (FY80) given immediately after the overthrow of the military government in 1979. At that time Uganda's domestic economy was devastated, following large-scale physical destruction. Reconstruction was also complicated by social disruption, administrative disarray and continued political strife. In response to the urgent situation, the Bank approved a quick disbursing project to rapidly supply inputs for the productive sectors and to meet short-term foreign exchange needs. Unfortunately, however, policies carried out subsequently by the new government caused, in turn, other waves of refugees from political and ethnic oppression and appalling atrocities. Other comparable cases include projects in Mozambique (FY85), Ghana (FY83) and Zirnbabwe (FY81), where the Bank has financed recovery pro- granmmes and imports critical to begin a long-term process of rehabilitation and reform. When Mozambique joined the World Bank in 1984, its economy was suffering from the effects of the civil war with its refugees, inappropriate policies, natural calamities and structural imbalances of long standing. Against this backdrop, the government developed a policy and a two-year Economic Action Programme designed to stem the economnic decline. The Bank's emergency assistance supported the government's immediate rehabilitation efforts through quickly disbursed hard currency to meet priority needs as the government began the process of rehabilitation. The initial project made possible imports of equipment, spare parts and raw materials for the key sectors of industry,< agriculture and transport, and directly helped the reintegration of war-displaced people into productive activities. This approach has been replicated and validated in other, smaller scale projects assisted in Mozambique by bilateral donors, like the Italian- sponsored project near Gondola (Manica province) and the Swedish-sponsored seed production unit in Chimoio. Results from these two projects confirm (see Morna, 1990) Bridging the Research Divide 309 that production-oriented aid (inputs, seeds, etc.) enables former war-refugees to shake- off their dependency upon relief-aid, re-establish themselves on a productive basis and become again self-sustaining. More recently, in 1992 and 1993, the World Bank started work on a large-scale, long-term programme to assist Mozambique in addressing its gigantic current problem: reintegrating productively the 4 or 5 million of post-war refugees and intemally displaced people now returning (or already returned) to their sites of origin. This pro- gramme may well be the largest ever undertaken refugee resettlement assistance strategy. In many cases, internationally assisted proamrnmes for refugees have made the difference between life and death for uncounted numbers of people. Emergency programmes have, however, a limited scope, their implementation is often less than fully satisfactory, and what is needed is long-term resettlement-cum-development assistance. The least addressed dimensions in such programmes are usually the cultural and psycho- logical ones, including the patient social engineering work necessary to re-establish the refugee population in new, viable settlements, with access to productive activities, some employment, and services. Indeed, as Harrell-Bond observed (1986: 2), 'establishing a rural settlement involves social engineering: the reading of the literature produced by agencies responsible for refugee work suggested that the full implications of this reality has escaped attention'; these guidelines re-focus the Bank-provided assistance on restoring the self-sufficiency and autonomy of the affected population through redressing its productive capacity. Similarities and Differences: Oustees and Refugees The previous sections discussed several substantive matters about refugees and oustees such as causality, magnitude, impoverishment processes, rights and entitlements, assistance strategies which, each and all, offer the 'territory' for comparative research on refugees and oustees, identifying similarities and differences. These far from exhaust the possibilities for such comparative research, but point to fertile directions, and also allow for a few additional comments. Overstatements of similarities between refugees and oustees can sometimes be found in the literature, oversimplifying the differences. For instance, some social scientists describe communities displaced by development programmes as 'development refugees'. Without appropriate analysis, this description risks becoming more a metaphorical label than a rigorous concept, and is not helpful for practical purposes. Applying it wholesale to define all those displaced by development programmes would blur certain essential differences between oustees and refugees, and would be misleading. There are sharp dissimilarities between refugees and oustees in legal status and entitlements to assistance. One key difference between the situation of oustees compared with refugees is the responsibility of the state vis-a-vis development-displaced people. This responsibility is of a different nature to the state's responsibility towards refugees: the first is assumed deliberately through the state's decision to enforce the legal principle of eminent domain. Asserting this difference is a comerstone in the argument about the material entide- mentp of those uprooted by state-pursued change. When people are displaced by a public development project, it is incumbent upon the government to ensure that the oustees are fully compensated and enabled to reconstruct their productive livelihood. However, if such assistance is not provided, development-displaced people are likely to become destitute, landless or homeless. Their situation becomes, in essence, similar to that of refugees. This is not an abstract possibility: in many cases, people displaced through govemment- 31(0 MICHAEL M. CERNEA sponsored development progranmmes, who have not been adequately assisted to re° establish themselves economically, have actually fallen into a refugee-like situation. An exanple of people transformed into refugees by government-sponsored intervention comes recently from Uganda, from a 'forest management and conservation project' funded in part by European donors with an environmental justification. Under this project, about 35,000 people inhabiting the Kibale Game Corridor area were brutally expelled; in 1992, to accelerate that expulsion from certain areas, people's houses were bumed down and food stores were destroyed or looted. They became refugees; some were 'relocated' into camps in the Bugangaizi area, while the whereabouts of many thousands, forced to flee, are not known. A 'sliding process' whereby established inhabitants become oustees and soon refugees, happened, for instance, to many people displaced by the Srisailam darn in India, to people displaced by the Danjiangkou dam in China, or to many of the oustees from the Hirakud multi-purpose dam project in Orissa, India (Baboo, 1992). A paper prepared by Oxfam (1993) about the Hirakud dam oustees shows how 30 years after the initial displacement and inadequate resettlement of people from the Hirakud reservoir, some of the same families have not fully recovered, and are facing eviction once again, this time because of the construction of some thermal plants in the same area. It is, therefore, conceptually and practically useful to point out explicitly, rather than blur, both similarities and differences between the status of refugees (from war, ethnic persecution, drought, etc.) and the status of oustees from planned programmes. This pernits highlighting the circumstances in which the use of violence, or bad planning, or other forms of abdication of government responsibility, convert oustees into refugees. The purpose of resettlement policies must be to prevent absolutely the sliding of development- displaced people (oustees) into a refugee-like condition. The definitions currendy accepted internationally for the term refitgees tend to be fairly restrictive. They are (on purpose) not open to encompass either voluntary migrants or people dislocated by planned projects (see, for instance, the definitions discussed in Zolberg et al., 1989; or the definition used by the United Nations). This is under- standable for both practical and conceptual reasons. Since many countries and inter- national organizations grant special entitlements to refugees, every attempt to define the term is underlined by the concern to avoid confusing, for instance, a voluntary economnic migrant with a genuine refugee.' From Research to Practice: Reducing Displacements A consequential difference between refugees and oustees consists in how their condition sets in: sudden onset, rarely anticipated, versus slow, planned, well-in-advance-known onset. This provides options for assistance strategies to development oustees that range on a broad spectrum: from effective ways to sometimes avoid development-induced dis- placement, to other ways to reduice its size, and in most cases, to mitigate its negative impacts. 9 Zolberg's discussion (in Zolberg et al. 1989) of the refugee definition highlights the element of violence. He proposes a sociological distinction between three 'prototypes' under which refugees can be conceptually clustered, depending on the circumstances that led to their becoming refugees. These are the ativist, the talet-, and the victim. All three prototypes represent people uprooted from their habitat. The circumstances of their uprooting are different for each category, but one common characteristic of their situation is fiear of violence'. 'All three types of refugees have in common the fear of immediate violence - violence resulting from conflict between state and civil society, between opposing arnies, or conflict among ethnic groups or class formations that the state is unable or unwilling to control' (Zolberg, et al. 1989: 269). If the consequences of their situation are considered these three types all share similar refugee characteristics: abrupt destitution; residence loss; loss of economic self-sufficiency; cultural separation; identity deprivation; socio-psychological stress; and others. If adopted for practical purposes, this tryology has potential implications regarding the forms and amounts of assistance to be given to refugees who can be identified with one or another 'type'. Bridging the Research Divide 311 Drawing on the experience of projects causing such displacements, I will briefly refer to four important issues concerning means and solutions in addressing development- induced displacement: (a) avoidability of displacement through technical optimization and/or correct financial costing; (b) the need for national policies and legal frameworks governing forced displacement operations; (c) development-oriented strategies vis-a-vis relief approaches; and (d) resource allocation for reconstruction and the responsibility of national governments. Avoidability of displacement While development projects are and will be necessary, there are approaches and tech- niques apt to avoid, or often reduce, forced displacement. Development is not in itself a blind legitimizer of any displacement, regardless of size and consequence. On the contrary, induced development programmes must seek every opportunity to minimize disruption. For instance, experience has demonstrated that if planners carefully weigh trade-offi between dam height, power generating capacity, and the size of population dislocation, they may be able to reduce the latter. Small decreases in reservoir elevation may significandy cut down the number of people affected by submergence, with relatively litde loss of power or irrigation benefits. This was demonstrated by the designers of the Saguling dam in Indonesia, who ultimately selected a reservoir elevation of 645 metres instead of the initially envisaged elevation of 650 thereby reducing population displacement of some 90,000 people to 55,000 (Soepartomo and Tjiptohandojo, 1988). In the same way, optimization studies for road trajectories in Shanghai (China) under the Bank-assisted urban improvement project have resulted in alternative routing which avoided the displacement of several thousand families. Correct costing of resetding (both people and infrastructure) and the inclusion of the estimated costs in the new project, may also show from the outset that one or another project causing displacement is too cosdy and therefore should not be started. It is necessary to avoid engineering biases in the justification of such technical projects by opposing the fatalistic treatment of displacement as always unavoidable. Filling policy and legal vacuums The government's political and moral responsibility vis-d-vis development-dislocated people is higher, in all respects, than it is in the case of refugees from natural calamities, because the former displacement is initiated by the government and enforced through the state's power of eminent domain (Shihata, 1988). However, this responsibility is often not assumed through explicit policies. Indeed, many countries have laws permitting expropriation by the government, but do not have formal regulations or policies guaranteeing full resettlement and rehabilitation of oustees. The absence of such domestic policy frameworks allows many abuses and infringements of peoples' socio- economic rights to happen and remain uncorrected. The eminent domain laws and their prescriptions for compensation, which usually were crafted for situations of a limited expropriations, are utterly unfit in cases of massive expropriations. Until recendy, even some international donor agencies, such as the regional develop- ment banks for Africa, Asia and Latin America, and many bilateral donor agencies, did not have formal policy guidelines for projects that cause displacement. As a result, when these agencies financed projects entailing displacement, they funded the new physical infrastructure and hardware but only seldom provided equal guidance and financial support for resettling the oustees. Resettlement was being left to the local/territorial agency alone, which definitely has less resources and less overall responsibility. The practical consequence of this policy vacuum was that relocation tended to remain largely unplanned, underfunded, and poorly implemented. The first time when an explicit policy for addressing the socio-economic and cultural 312 MICHAEL M. CERNEA issues of involuntary resettlement was enunciated by an international development agency was in 1980, when the World Bank issued its formal policy guidelines on social issues in resettlement, a policy grounded in social science knowledge about resettlement (see World Bank, 1980, 1994; Cernea, 1988, 1990a, 1993a, 1995; Qureshi 1989). This policy is protective of the interests of those displaced and is regarded even by severe critics of the World Bank as promoting the highest standards in resettlement. u This institutionalized policy requires Bank staff and borrowing agencies to pursue resettlement standards of higher quality than would be pursued in the absence of such a policy. The two most important effects of this policy to date have been more systematic planning for relocation and more adequate resoturce allocation to the resettlers in a broad spectrum of projects in Africa, Asia and Latin America. Adequate implementation of this policy in projects, however, does not result automatically from just the adoption of sound policy. Departures from policy norms occur for various reasons, sometimes related to inadequate performance of Bank staff, yet more often because of disregard of project provisions or legal agreements by implementation agencies and borrowing governments. The Bank-assisted Narmada Sardar Sarovar dam project is only the most known case of departure from policy guidelines and poor implementation, but definitely not the only one in India (Morse et al., 1992). For instance, resettlement implementation has fallen below the standards established for resettlement in the Narayanpur and Almatti reservoirs in Karnataka state under the Upper Krishna II project; the provisions of the resettlement action plan, legally agreed upon between the Bank and the country, have been disregarded during implementation, so that twice - in November 1992 and again in September 1995 - the Bank had to suspend the disbursement of its loan for this project. Also in India the Bank had to decline extension of the credits for other projects (e.g., for Upper Indravati and for Subernarekha dams) also because of the borrowing agencies' failures in carrying out resettlement. Similar examples exist in other countries as well - Indonesia, Turkey, Brazil, Madagascar. This demonstrates that a good policy alone - and even a good plan alone - are not enough, if the political will, organizational capacity and financial resources are not adequate. Consistency of resetdement implementation with policy must be constandy enforced, and also monitored with the direct participation in such monitoring by the people affected and their organizations. There is a considerable gap between the resetdement policy norms adopted by the Bank, on the one hand, and the rather weak domestic norms (or sometimes complete policy vacuums) regarding displacement in many developing countries, on the other. For instance, it appears surprising that India and Indonesia two countries with many major on-going development programmes that cause large scale displacements - do not have a national policy regulating involuntary resettlement caused by public sector investments. The sarne can be said for many countries in Africa, such as Kenya, Cote d'lvoire, Togo and others, in which the state is promoting major projects causing displacement yet is not adopting adequate policies to regulate it and address its consequences. The weak (or absent) political commitment of governments to income restoration of those displaced is one of the root causes of the formidable problems created and perpetuated '° For instance, the by-now famous Morse review commission, which severely criticized the design, appraisal, and implementation of the Sardar Sarovar projects in India, concluded at the same time that '...the Bank had, in its 1990 and 1991 directives, set the highest standards of any aid or lending organization in the worldfor mitigating adverse consequences to human well-being caused by "involuntary resettlement"' (Morse et al., 1992: 37). Over the last few years (early 1990s) other intemational donors have followed suit. Significant, in this respect is the adoption by all the Organization for Economic Coopenrtion and Development (OECD) member countries, in December 1991, of unified policy guidelines for their aid agencies on involuntary displacement and resettlement in projects assisted by these countries (OECD, 1991). The OECD guidelines are virtually the same as the Bank's rcsettlement policy, taken as their model. The adoption by all major donors of similar resettlement guidelines reflects the broader trend towards recognizing that intemational aid flows should contribute to the protection of basic human rights. Bridging the Research Divide 313 for decades by forced displacements. In most African countries, compulsory resettle- ment, including the growing urban developments for infrastructural improvements, are carried out by government agencies in a policy vacuum. Certainly, laws and guidelines to empower the state to take away land 'needed for the public good' do exist. But sorely missing in the majority of African countries are explicit policies and legal frameworks to compel relevant state agencies to effectively address the vital issues of livelihood restora- tion and productive re-establishment of those displaced (Okidi, 1993). The expropria- tion laws generally lay down rules for the type of compensation that must be paid for the expropriated land. However, the notion of 'compensation' - payment for land taken for public use - is a narrow concept that differs in substance from the more exacting principle that the state has the obligation to restore people's economic well-being and capacities as productive agents. This distinction between mere compensation, on the one hand, and resettlement on a productive basis, on the other hand, is a critical one, yet it is conspicuously absent from the African policy literature. In some countries, the Bank-Borrower policy dialogue and the lessons derived from Bank-assisted projects have already led to the elaboration of improved or new domestic legislation. Yet at the same time other states and governments are clearly not willing to commit themselves to issuing binding policies. They offer only ad-hoc and limited improvements confined to one or another project. Furthermore, certain states deliberately go to great lengths to circumscribe the validity of such improvements only to internationally assisted projects, exempting the domestic projects from the improved rules at the expense of the people affected by these domestic projects. The Bank holds that the projects it finances should not engage in practices which violate universally recognized rights. In order to strengthen the legal guarantees in that respect, it actively encourages the enactment of domestic legal and policy frameworks in borrowing countries that would apply to resettlement tnder all projects, and not only under the few projects that receive international financial assistance. Researchers, legal experts, NGO activists and many others involved in resettlement have concluded that policy and institutional changes are widely needed in many developing countries (see Ganguly Thukral 1992; Shihata, 1991; Paul 1988, 1991; Dhagamwar, 1989). Inde- pendent legal scholars have noted that it cannot be said to be improper 'interference' for the Bank or International Development Association to insist, as a formal condition of a project loan, that project-consistent state law be adopted in borrowing countries to secure recognized rights for the people affected by the project; on the contrary, it would be inconsistent and 'unlawful to fail to insist on such protection' (Paul, 1988: 118). The independent adoption of national resettlement policies by governments of developing countries will go a long way towards safeguarding people's human and economic rights and speeding up the productive re-establishment of oustees. Some principles and elements of the Bank's resettlement policy might be used also as a guide in assisting refugees who return to their places of origin. Development versus relief Relief assistance to internally displaced populations is necessary, but not sufficient. On the 'development versus relief' issue, the debate has been going on for a long time. But the problem has not been resolved partly because humanitarian agencies are better at relief than development, partly because funders want operations to appear to be short- term interventions, and partly also because some specific situ'ations are such that immediate relief is imperative and the only practical possibility. The issue of development versus relief is not only a matter of resources, however, although the resources needed for the former are higher: it is also a matter of approach and strategy, a matter of how resources are deployed. Scarce assistance resources must be deployed as a leverage to elicit the capacities, energy and initiatives of those assisted, to 3'6 MICHAEL M. CERNEA strengthen their propensity to self-organization, and meet their basic interests in socio- economic re-establishment. The key policy objective in resettlement is restoring the income-generating capacity of resettlers. Therefore, because involintary displacement dismantles a previotus production system and way of life, all resettlement programmes muist be development programmes as well. Indeed, when resettlement is unavoidable, what is most needed is to ensure that the oustees' productive base and income-earning ability are reconstructed, that is, restored to the level they would have achieved without relocation, and improved whenever possible." Reaching this basic objective is largely dependent on the allocation of adequate resources for the re-establishment of people's productive capacity and on the participation of those affected in creating the new socioeconomic arrngements. Reconstruction and Resource Allocation The crafting of reconstruction strategies may improve through experience transfers between projects for resettling development oustees and for post-war refugee reinsertion. Specialists have signalled that 'too often the kind of agencies involved in the relief activity [for refugees] simply do not know much about longer term, developmental approaches' (Clark 1989: 10). Relief workers are asking pertinent questions about 'how can assistance be provided in ways that are as developmental as possible?'; or what ways can be found 'to build some kind of economic activity into the lives of refugees?' (Clark 1989: 10-11). In resettlement projects, the core of the resettlement plan is a development package. It must consist of provisions centred either on land-based strategies or on employment-based strategies. The most effective, and relatively less costly, are the land-based strategies. They often involve land reclamation, small-scale irrigation, tree-crops development, fisheries, social forestry and similar kinds of income-generating activities. Sometimes, land scarcity is a major limiting factor, given high population densities. Vocational training for jobs outside agriculture is used in employment-based strategies, but this avenue also requires investments in job creation. Vocational training alone, without actual employment in the newly acquired skills, does not restore the livelihood of those displaced. In all such activities, it is essential to stimulate the initiative of those to be resettled and assist their own efforts for self-resettlement among the host population at the new location. In fact, research on refugees has documented the effectiveness of self- resettlement, under certain circumstances. Hansen's research among Angolan refugees in Zambia found that self-resettled people achieved greater autonomy and integration in the long term than refugees that have spent an equal period of time in government- controlled refugee schemes (Hansen, 1990). Sirnilarly, an IDRC research in Tanzania, led by A. Chol and M. Mbago, among Burundi refugees from ethnic persecution suggests that self-settled refugee villages achieve better and more durable integration within the surrounding local host population than government-established refugee centres - this assessment being based on productive status, intermarriage with citizens of the host country, interpersonal relationships, education and stability (cf Harris, 1991). Such findings suggest that more encouragement given to the initiative, energy, and self- organizing capacity of oustees may unlock a potential insufficiently used in resettlement programmes. "This policy objective was formulated clearly in the Bank's 1990 operational directives: 'All involuntary resettlement should be conceived and executed as development programs, with resettlers provided sufficient investment resources and opportunities to share in project benefits. The displaced should be (i) compensated for their losses at full replacement cost prior to the actual move; (ii) assisted with the move and supported during the transition period in the resetdement site; and (iii) assisted in their efforts to improve their former living standards, income eaming capaciry, and production levels, or at least to restore them. Particular attention needs to be paid to the needs of the poorest groups' (World Bank, 1990: 3(b)). Bridging the Research Divide 315 Philanthropy from the private sector and the public, which is crucial for refugee assistance, cannot be counted upon as a resource in programme-caused dislocation. The government bears the primary financial and institutional responsibility for relocation in a manner that will restore the economic autonomy of those uprooted. The participation of NGOs in the planning and implementation of relocation is essential because NGOs are apt to express the needs and defend the rights of those displaced. Some NGOs are better placed than government agencies to mobilize and organize the energies of those displaced for reconstructive activities. In tum, bilateral donor agencies should never finance only the construction of new physical structures without also assisting the reconstruction of sustainable livelihoods for those displaced. In sum, there is much to explore about the similarities and differences between the condition of refugees and development oustees, and about how their needs can be addressed. The very nature of these similarities and differences offers the basis for more comparative analysis and for a better dialogue between the bodies of social science literatures devoted to the study of these populations. Their findings can complement and reinforce each other. Most important is the benefit for social action: comparative research will not only serve for in-depth conceptual developments, but will also inform policy and operational assistance more comprehensively and effectively. Acknowledgement This study was started as part of the author's research while a Visiting Scholar at Harvard University, Department of Anthropology, on leave from the World Bank. An abbreviated version, submitted to the refugee research seminar organized by UNRISD in Addis Ababa, Ethiopia, (1992), was expanded for this volume. The author expresses grateful thanks to I3. Maybury-Lewis. 1. Shihata, James C.N. Paul, Ruth Cernea, S. Guggenheim, B. Harrell-Bond, J. Clay, N. Yalman, R. Zetter, A. Hansen and Tim Allen, for com- ments or discussions that have benefited the study. Special thanks are due to Camille Weithers, who helped in the final editing of this paper. The views, findings and interpretations contained in this study are those of the author and should not necessarily be attributed to the institutions with which he is associated. References Appell, G.N. (1985) 'The Bulusu' of East Kalimantan: the consequences of resettlcment', in G.N. Appell (ed.) Modernization and the emergency oj a landless peasantry, studies in third world societies, Vol. 33, Williamsburg: College of William and Mary. AppeDl, G. N. (1994) 'Our vision of human rights is too small: anthropological perspective on fundamental human rights', in Hitchcock, R.K. and Morris, (eds.) International Human Righu and Indigenous People, forthcoming. Areeparampil, Matthew (1989) 'Industries, mines and dispossession of indigenous people: the case of Chotariapur'. in W. Femandes and E.G. Thukral (eds.) Development, Displacement and Rehabilitation, New Delhi: Indian Social Institute. Baboo, Balgobind (1992) Techinology and soial transformation. Thie case of the Hirakud multi-purpose dam project in Orissa, Concept Publishing Company, New Delhi, lndia. Brokensha, David W (1963-4). 'Volta resettlement and anthropological research', in Human Organization, Vol. 22, No. 4, 286-90. Cemea, Michael M. (1988) Involuntary resettlement in development projects. policy guidelines in Worid Bank Financed Projects, World Bank, Washington, DC. - (1990a) From unused social knowledge to policy creation, Development Discussion Paper No. 342, May, HIID, Harvard University, Cambridge, MA. - (1990b) Poverty risksfrom population displacement in water rcsourrce projects, Development Discussion Paper No. 355, HIID, Harvard University, Cambridge, MA. - (1990c) 'Intemal refugee flows and development-induced population displacement', Journal of Refugee Studies, Vol. 3, No. 4: 320-39. - (1991) 'Involuntary resettlement: social research, policy and planning', in Cemea, M.M. (ed.), Putting peopl first. Sociological variables in ruMl development, 2nd ed., revised and enlarged, New York-Oxford: Oxford University Press. - (1993a) 'Anthropological and sociological research for policy fonnulation on involuntary resetdement', in M. M. Cemea and S. Guggenheim (eds), Antihropological approacihes to resettlement: policy, practice and theory, Westview Press, Boulder, CO. - (1993b) Urban environment and population resettkment, World Bank, Washington, DC. - (1994) 'African population resetdtment in a global contcxt', in C. Cook (ed.) Involuntary resettlemcnt in 316 MICHAEL M. CERNEA Afica, World Bank, Washington, DC. - (1995) 'Social integration and population displacement: the contribution of social sciences' in International Social ScienceJournal, Mouton, No. 143. Chao, Liang (1990) 'State plans new rules for dam resettlement, Chiina Daily, No. 2632, 4January. Christensen, Hanne (1982) Survival strategiesfor and by camp refugees: report on an exploratory sociologicalfield study in Somalia, UNRISD Report No. 82: 3, Geneva: United Nations. Clay, Jason W. (1986) Research Report: 'Refugees flee Ethiopia's collectivization', Cultural Survival Quarte71y, Vol. 10 No. 2: 80-5. Clay, Jason W. (1988). In J.W. Clay, S. Steingraber and P. Niggli (eds) The spoils offamine. Ethiopian policy and peasant agriculture, Cambridge, MA: Cultural Survival. Clark, Lance (1989) Internally displaced persons:framing tihe issues, Center for Policy Analysis and Research on Refugee Issues, Washington, DC. de Mars, William (1992) Decade of displacement in Etihiopia: internally displaed persons and humanitarian response, Refugee Policy Group, Washington, DC. de Wet, Chris (1993) 'Involuntary resettlement as imposed transformation' in M. Cemea and S. Guggenheim (eds.) Antihropological Approaches to Resettlement: Policy, Practice, and Theory, Westview Press, Boulder, CO. Dhagamwar, Vasuda (1989) 'Rehabilitation: policy and institutional changes required', in W. Fernandes and E.G. Thukral (eds) Development, Displacement and Rehabilitation, Indian Social Institute, New Delhi. Dieci, Paolo and Claudio Viezolli (eds) (1992) Resettlement and rural development in Ethtiopia, Milan: Franco Angcli. Downing, Theodore (1988) Human rights researcih: The chlallenge for the anthropologist, in T. Downing and G. Kushner (eds), Human righits and tribal people, Cultural Survival, Cambridge, MA. Downing, Theodore and G. Kushner (eds.) (1988) Human rights and tribal people, Cultural Survival, Cambridge MA. Drabek, Thomas E. (1987) Human systems responses to disasters. An inventory of sociologicalfindings, Springer Vcrlag, New York, Berlin and Heidelberg. Dubin, R. (1978) Theory building, Free Press, New York. Fahim, Hussein (1988). Egyptian Nubians: resettlement and years of coping, Salt Lake City: University of Utah Press. Fernandes, W. (1991) 'Power and powerlessness: development projects and displacement of tribals', in Social Action, Vol. 41, No. 3,July. Femandes, W., J.C. Das, Sam Rao (1989) 'Displacement and rehabilitation: an estimate of extent and prospects', in Fermandes, W. and E. Ganguly Thukral (eds) Development, displacement and rehabilitation, Indian Social Institutc, New Delhi. Ganguly-Thukral, E., (ed.) (1992) Big dams, displaced people, rivers of sorrow, rivers of change, Sage Publications, New Delhi and London. The Georgetoun declaration on health care for displaced persons and refugees (1988) Georgetown University, Washington, DC Guggcnhcim, Scott (1994) Involuntary resettlement: an annotated refrene bibliography for development researrh, processed. Hansen, Art (1990) 'Long-term consequences of two African refugee settlement strategies', paper presented at the meetings of the Society for Applied Anthropology, York, UK. Hanrell-Bond, Barbara E. (1986) Imposing aid: emergenmy assistance to refuges, New York: Oxford University Press. Harris, Craig (1991) Integmting refugees into lhost countries, IDRC Reports, Vol. 19, No. 3, OttaWa, Canada, October. Holtzman, Sreven (1995) 'Post-conflict reconstruction', World Bank, Social Policy and Resettlement Division, draft. Horowitz, Michael M., D: Loenig, C. Grimm and Y. Konate (1993) 'Resettlement at Manantali, Mali: short term success, long-term problems', in M. M. Cernea and S. Guggenheim (eds), Antiropological Approacles to Resettlement: Policy, nTeory, and Practice, Boulder, Colorado: Wesrview Press. Horowitz, Michael M. (1991) 'Victims upstream and down', inJournal of Refugee Studies, Vol. 4, No. 2, Oxford University Press. Horowitz, Michael M. and Muneera Salem-Murdock (1990). Senegal River Basin monitoring activity: synthesis report, Institute for Development Anthropology, Binghamton, New York. Keely, Charles B. (1988) Foreword to The Georgetoum Declaration of healthcare for displaced persons and refugees, Georgetown University, Washington, DC. Kesmanee, Chupinit (1988) 'Hill tribe relocation policy in Thailand', Cultural Survival Quarterly, 12(4). Kreps, Gary A. (1983) 'The organization of disaster response: core concepts and processes', InternationalJournal of Mass Emergencies and Disasters, Vol. 1: 439-67. -(1987) 'Classical themes, structural sociology, and disaster research', in Russell R. Dynes, De Marchi, and C. Pelanda (eds) Sociology of Disasters: Contribution of Sociology to Disaster Researrie, Franco Angeli, Milan. Lassailly-Jacob, Veronique (1994) 'Govemment-sponsored agricultural schemes for involuntary migrants in Africa: some key obstacles to their economic viability', in Adelman, H. and J. Sorenson (eds) Refugees: Development Aid and Repatriation, Westview Press, Boulder and York Lanes Press, New York. Mahapatra, L.K. (1983) 'Development for whom?' paper presented at the Intemational Congress for Anthropological and Ethnographic Sciences, Vancouver. McColm, R. Bruce (1989) 'The world's unwanted', T7he Wall StreetJournal, 28 September. Mermel, T.W. (1990) 'World Bank's share of worldwide dam construction', in G. Le Moigne, S. Barghouti and H. Plusquellec (eds) Dam Safety and tihe Environment, World Bank Technical Paper No. 115, Washington, DC. Bridging the Research Divide 317 Moma, Colleen Lowe (1990) 'Refugees return to the farm', Afiican Farmer,, 4, July. Morse, Bradford, J. Berger, D. Gamble, H. Brody (1992) Sardar Sarovar. The report of the independent review, Resources Futures International (RFI) Inc., Ottawa, Canada. OECD (1991) 'Guidelines for aid agencies on involuntary displacement and resetdement in development projects', OECD/GD(91)201. Paris, December. Okidi, C.O. (1993) Policy and legalframework on development driven involuntary resettlement in African countries (draft), Moi University, School of Environmental Studies, Eldoret, Kenya. Oliver-Smith, Anthony (1994) 'Resistance to resettlement: the formation and evolution of movements', in Researric in Social Movements, Conflicts and Chiange, Greenwich, CON: JAI Press. Olivcr-Smith, A. and Art Hansen (1982) 'Involuntary migntion and resettlement; causes and contexts', in A. Hansen and A. Oliver-Smith (eds) Involuntary migration and resettlement, Boulder, Colorado, Westview Press. Oxfam (1993) 'The Hirakud dam project', Public Policy Department, processed. Oyedipe, F.P.A. (1983) Adjustment to resettlment: a study of the resettled peopled in tihe Kainji Lake basin, University Press Limited, Ibadan, Nigeria. Pankhurst, Alula (1992) 'Resettlement and famine in Ethiopia. The villagers' experience', Manchester University Press, Manchester. Partridge, W.L., A.B. Brown and J.B. Nugent (1982) 'The Papaloapan dam and resettlement projcct: human ecology and health impacts'. In A. Oliver-Smith and A. Hansen (eds.) Involuntary Migration and Resettlment, Westview Press, Bouldcr, CO. Paul, James C.N. (1988) 'Intcrnational development agencies, human rights and humane development projects' in DenverJournal of International Law and Politics, Vol. 17, No. 1. - (1989) 'Rural development, human rights and constitutional orders in sub-Saharaii Africa', in Tnlin world legal studies - 1989: pluralism, participation and decentralization in sub-Saharan Africa, Valparaiso University, Indiana. - (1992) 'The human right to development: its meaning and importance', in theJohn Marshall Law Review, Vol. 25: 235. Qureshi, Mocen (1989) 'Resetdement, development and the World Bank' in 77Te Bank's World, No. 5, May. Refugee Policy Group (1992) Internally displaced persons in Africa: assistance challenges and opportunities, Center for Policy, Analysis and Research on Refugee Issues, Washington, DC. Rew, Alan and P.A. Driver (1986) Evaluation of the social and environmental impact of the Victoria Dam Project, Overseas Development Administration (ODA EV392), London. Salem-Murdock, Muneera (1989) Arabs and Nubians in New Ha!fa, Salt Lake City: University of Utah Press. Scudder, Thayer (1981) 'What it means to be dammed', Engineering and Science, Vol. 54, No. 4. Pasadena. - (1991) 'A sociological framework for the analysis of new land settlements, in Cemea M.M. (ed.), Putting peopkfirst. sociological variables in rural development, New York, Oxford University Press. - (1993) 'Development-induced relocation and refugee studies: 37 years of change and continuity among Zambia's Gwembe Tonga',Journal of Refugee Studies, Vol. 6, (forthcoming). Scudder, Thayer and Elisabeth Colson (1982) 'From welfare to development: a conceptual framework for the analysis of dislocated people', in A. Hansen and A. Oliver-Smith (eds.), Involuntary Migration and Resettlement, Westview Press, Boulder, Colorado. Scudder. Thayer (1973) 'The human ecology of big projects: river basin development and resettlement', Annual Review of Anthropology, 2. Shihata, Ibrahim F.l. (1988) 'The World Bank and human rights: an analysis of the legal issues and the record of achievements', DenverJournal of International Law and Politics, Vol. 17, No. 1, 39-66. Shihata, Ibrhim Fl. (1991) 'Some legal aspects of involuntary population displacement', in Michael M. Cemca and Scott E. Guggenheim (eds), 'Anthropological approaches to resettlement: policy, practice and theory', Westvicw Press, Boulder, CO. Soepartomo and P. Tjiptohandojo (1988) 'Comparison betweeni predicted and observed environmental impacts in Saguling Reserve', in CIGB-ICOLD, Sixteenth, International Congress on Large Dams, Vol. 1, San Francisco, June. UNHCR (1993) The state of the world's refugees. Penguin, New York. World Bank (1980) Operational Manual Statement 2.33, Social issues associated with involuntary resettlement in bank-financed projects, Washington, DC, February. - (1989) Operational Directive 8.50: Emergency recovery assistance, Washington, DC, November. - (1990) Operational Directives 4.30: Involuntary resettlment, Washington, DC,June. - (1994) Resettlment and development: the bankwide review of proiets involving involuntary resettlement 1986-1993, World Bank, Environment Department, Washington, DC. - (1995) 'From emergency to development: the demobilization and reintegration of ex-combatants in Ethiopia'; 'Beyond repatriationi: the demobilization and reintegration of ex-combatants in Namibia'; 'From swords to ploughshares: the demobilization and reintegration of ex-combatants in Uganda'; 'War-to-peace transition in Sub-Saharan Africa: lessons from the Horn, the Heart and the Cape'; Poverty and Human Resources Division, Technical Departnent, Africa Region, The World Bank. (These first three case studies, and the last study, a synthesis paper, were prepared by a core team led by Nat J. Colletta, with support from several Bank staff and consultants). Zetter, Roger W. (1988) 'Refugees and refugee studies: a label and an agenda',Journal of Refugee Studies, Vol. 1, No. 1. - (1990) 'Labeling refugees: forming and transforming an identity', paper given to the Program in Intemational Cooperation in Africa, Illiniois: Northwestern University. Zolberg, Aristide R., A. Suhrke, and S. Aguayo (1989) Escapefrom violence. Conflict and the refugee crisis in the developing world, New York: Oxford University Press. THE WORLD BANK Headquarters 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. Telephone: (202) 477-1234 Facsimile: (202) 477-6391 Telex: wUi 64145 WORLDBANK RCA 248423 WORLDBK Cable Address: INTBAFRAD WASHINGTONDC World Wide Web: http:// www.worldbank.org E-mail: books@worldbank.org European Office 66 avenue d'Iena 75116 Paris, France Telephone: (1) 40.69.30.00 Facsimile: (1) 40.69.30.66 Telex: 640651 Tokyo Office Kokusai Building 1-1 Marunouchi 3-chome Chiyoda-ku, Tokyo 100, Japan Telephone: (3) 3214-5001 Facsimile: (3) 3214-3657 Telex: 26838 The full range of World Bank publications, both free and for sale, is described in the annual Index of Publications, and of the continuing research program of the World Bank, in World Bank Research Program: Abstracts of Current Studies. The most recent edition of each is available without charge from: DISTRIBUTION UNIT, PUBLICATIONS OFFICE OF THE PUBLISHER THE WORLD BANK DEPARTMENT F 66 AVENUE d'IENA THE WORLD BANK 75116, PARIS 1818 H STREET, N.W. FRANCE WASHINGTON, D.C 20433 U.S.A. ISSN 0253-213