World Ba!lk Reprint Series: Number 480 Michael M. Cemea The So~iological Action-Research 9£ Development-Induced Population Resettlement Reprinted with permission from Romanian Journal of Sociology 6(2) : 97-120. Copyright© Romanian Journal of Sociology, 1995. Michael M. Cernea The Sociological Action-Research of Development-Induced Population Resettlement THE SOCIOLOGICAL ACTION-RESEARCH OF DEVELOPMENT-INDUCED POPULATION RESETTLEMENT MICHAEL M. CERNEA Member Correspondent of the Romanian Academy Senior Adviser for Sociology, The World Bank, Washington, DC The sociological study of development-caused population displacement and resettlement has considerably expanded during the last decade. The expansion of this social research area is a direct response to the controversies mounting in many countries in connection with government programs for hydropower, infrastructure, and urban growth. When people are involuntarily displaced, the social, economic and cultural impacts they suffer are usually profoundly destructive. Increased recognition of such adverse consequences has resulted, inter alia, in a broader call for social science presence in the analysis, planning, mitigation and evaluation of displacement and resettlement processes. This paper discusses the findings and impacts of a recent vast study of development-caused population resettlement. This was action-research study, carried out in 1993-1994 by a multidisciplinary research task force led by the author of this article, under the auspices of the World Bank. The study encompassed 192 ·· development projects in 39 countries, entailing the development and resettlement of various population gropus estimated to total about 2.5 milion people. These projects were co-financed by the World Bank and were implemented, in part or in full, during the 1986-1993 period. Over 100 of these projects are still ongoing in 1995/1996. An additional number of resettlement operations that occurred outside Bank projects and for which data were available, were also considered by this study. The 192 projects entailed the displacement and resettlement of about 2.5 milion people over 8-10 years. Altogether, the action-research covering these projects addressed: policy issues in resettlement; the socioeconomic nature of displacement; impoverishment risk through displacement; institutional capacities and operational performance in implementing resettlement, including planning, financing, and monitoring; and options for impoving resettlement strategies and performance. The report on this study was completed in 1994.1 . Note:. The views and interpretations expressed in this article are those of the author and should not be necessarily attributed to the institutions with which he is affiliated. Correspondence regarding this article can be sent to the author at: The World Bank, 1818 H Street, NW, Washington, DC 20433, USA. 1 The full report on this broad action-resear~ is entitled Resettlement and Development. The . Bankwide Review of Projects Involving Involuntary Resettlement 1986-1993. The World Bank, · Environment Department, April 8, 1994. The Task Force that coordinated the Bankwide action- research and prepared the report was led by Michael M. Cernea. The present article draws on this report. The full report can be obtained by writing to the World Bank Public Information Center, 1818 H.St., NW, Washington, DC 20433, USA, or to the author of this article. Romanian Journal of Sociology, VI, 2, P. 97...;120, Bucharest, 1995 1-- 98 MlCHAEL M. CE&";'"EA 2 'WHY INVOLUNTARY RESETTLEMENT MAY BECOME NECESSARY Involuntary population displacement and resettlement has a companion of development throughout history. It has been indelibly written into the evolution of industrial as well as developing countries. Installing major hydropowei dams, irrigation and drinking water systems, urban construction and expansion, or extending highway networks has entailed displacements fraught with harships and deprivation. The beautiful master-plan of today's Paris, and particularly the urban marvel of its unique Place de L'Etoile, were made possible by painful population relocation in the days of Baron Haussmann. New York's impressive Cross Bronx Expressway slashed through many neighborhoods, of which some - as hindsight teaches us - could have ·been saved (Caro 1975). It is known that the drama of forced urban displacement in Boston's West End has generated an entire sociological literature (Gans 1959, 1968). Modem hydroelectric complexes in Romania, such as the Bicaz power plant or the dam on the Danube at Portile de Fier - have required the uprooting of many villages or the flooding of an island. This list could be extended with many other internationally known examples. Involuntary displacements continue to occur in all countries for reasons related to·· the betterment of living conditions, introduction of · needed environmental infrastructure, or expansion of public services. During the recent few decades, the scale of development-related population displacement has grown rapidly particularly in developing countries, due to the accelerated provision of infrastructure and growing population densities. Developing countries invest around $200 billion per year in new . infrastructure. Such developments often involve changes in land and water use patterns, and in some instances this requires that people be displaced. The numbers of people needing resettlement can be reduced and, with more creative engineering and better social design, progress in limiting displacement has been made. But the need for resettlement cannot be eliminated. Nor should it be assumed that the volume of resettlement will be much lower in the future. Around the world more than 2 bilion people still lack access to electricity and are forced to use sticks and dung for their energy needs; 1.7 billion lack sewerage systems and 1 billion lack access to clean piped water, resulting in the unnecessary death of 2-3 milion infants and children each year. Food production will need to double again in the 40 years - an impossibility without additional investment in irrigation. And the world's population will grow by almost 1 billion each decade over same period. Substantial further investments in infrastructure, with their attendant population relocations, will thus be absolutely essential if poverty is to be reduced. The action-research which I summarize in this article generated the first worldwide estimate of. dev~lopment-caused displacements, collecting and correcting data from a broad variety of sources. According to our research, the displacement toll of the 300 large dams that, on average, enter into construction every year, is in excess of 4 million people. In turn, we estimated that the urban development and transportation programs being started each year in developing countries entail the involuntary displacement of some additional 6 million people. I 3 Induced Population Rciettlement 99 According to our research, over the past decade an estimated 80 to 90 million people have been resettled as a result of infrastructure programs for dam construction, and urban and transportation development. If we include other sectors of the economy, the aggregate number is certainly much higher, but data for other sectors are not yet acccurate enough. Within the worldwide resettlement processes, the programs financed by the World Bank ar.rount for 1-3% of the total (three percent in dam construction and one percent in urban development). Projects in he Bank's active portofolio in one year alone (1993/1994), the year of the action-research, involved the resettlement of 2 million people over an eight-year period. In these programs, the Bank has the responsibility to help borrowing countries restore, and if possible improve, the livelihoods of displaced and relocated people. Over the past fourteen years since the Bank estabilshed its path-breaking and social science-grounded resettlement policy, it has worked with governments to promote better domestic policies and legal framewoks for resettlement. Improving resettlement is difficult for developing country governments, particularly in low income countries with land scarcity, which face competing needs, resource limitations, and constraints on institutional capacity. Often implemen- tation performance in resettlement is lower than expected. The social costs of inadequate resettlement and its adverse effects can be very high, resulting in increased poverty for many people. This is especially serious since many of those affected are already very poor. This heightens the moral imperative of ensuring sound policies and effective implementation. However, if done right, resettlement programs can become an element of a nation's strategy to reduce poverty. This requires not only sound policies and adequate resources, but also a change in mind- set - towards recognizing resettlers' entitlements to share in the benefits of the projects which cause their displacement. Improving resettlementprocesses in thus an urgent priority, and also a task that .calls on the skills of sociologists and other applied researchers. CONCEPTUAL ISSUES Social scientists doing research on development processes are empirically informed about how development projects can make some people worse off. But social scientists all too often speak to themselves: historically they have been much better at recording development's tragedies than helping prevent them. I have often argued (Cernea 1985, 1993) that public policy responses to difficult development issues could be improved through increased action-research. But it is important to state that social scientists themselves have to do much more to equip governments · and private organizations with adequate policy, strategy and practical advice. During the 1970s and first half of the 1980s, development research has given relatively little analytic attention to involuntary displacement and resettlement, despite a vast literature on the sociology of voluntary settlement. This situation has considerably changed in the latter half of the 1980s and the first half of the 1990s, as demonstrated clearly by an annotated reference bibliography for develoopment research on resettlement which includes over 700 selected titles (Gu1rn:enheim · 100 MIC.HAEL M. CERNEA 4 1994). The debate within social sciences has been joined, and in fact has been exponentially broadened, by the much larger debate in the international media by public interest groups, and by NGO activists, a debate in which India's Narmada Sardar Sarovar project became a high-visibility peak point, but is not the only focus. The national and international deba.tes around involuntary resettlement have been much amplified by the growing political resistance and active opposition developed I in several countries by the key actors of this process - the populations to be forcibly i displaced. ' Processes of involuntary population displacement and resettlement are " frequent enough, big enough, complex and consequential enough, to merit the mobilization of the social science conceptual and operational tools available to analyze them. Social scientists can and should assist specifically in (a) identifying the disruptions caused by displacement; and (b) recommending policies and practical approaches to prevent avoidable displacements from happening and to mitigate the harmful effects of the unavoidable ones. My argument in ~his papaer is that social scientists, working either in academia or in governmental and development agencies, must expand their horizons to focus on public policy issues, not only on discreet project interventions. Policy development is the introduction of constitutive rules about how to direct a major social process by formulating basic goals and mobilizing compatible means. Policies provide guidelines for allocating development resources in general, as well ' as for structuring individual projects. Turning social science knowledge about , , resettlement into prescribed "dos" and "dont't's" - the substance of policy-projects :I financed by governments and international agencies is as important as explaining I resettlement's causes and outcomes. I I The way in which development oriented social science treats involuntary i population dispalcement is in fact a good test for its ambitious claim: that sociological/anthropological knowledge can improve the formulation of development policies and operational approaches, so as to substantially enhance the benefits of induced development. Internationally, applied and development- I oriented social researchers recognize the need - but not always the potential - for transforming· social science knowledge into the building blocks for public policy making. This matter is of great concern to social scientist in many countries. The social science definition of resettlement. What is usually described as "involuntary resettlement" consists of two distinct, yet closely related social processes: (a) displacement of people and (b) reconstruction of their livelihood; this reconstruction is somentimes called rehabilitation. Each has its own demands, risks, costs, logistics, and sociocultural and economic effects. Displacement concerns how land and other major assets are expropriated and people are removed, to allow a project intendent for the overall social good to proceed. In real life, this is not just an "expropriation", a simple transfer of property in exchange for compensation. In sociological terms, it is a process of unraveling the I I existing patterns of social organization and functioning of ongoing production systems and settlement units. Forced population displacement always creates a social crisis, and sometimes ~ political one as well. The disruptions it triggers are rarely equaled in the "normal" processes of development; I will return to this further below. 5 Induced Population Resettlement 101 Rehabilitation or reestablishment, in turn, refers to the fate of the displaced people after relocation and to the reconstruction of their socioeconomic organization. In theory, the two processes - expropriation and reestablisment - are segments of a single continuum; in practice, the first does not automatically bring . about second. When people are displaced by projects for "right of way", they lose . either their land - in full or in part - or their dwelling, or both. As a consequence, the outcomes of resettlement may vary considerably from people's previous standards of living. Indeed, whether on not involuntary resettlement results in reestablishing people's incomes and livelihood depends largely on how displacement is planned and carried out. It also depends on whether resettlers are assisted to rebuild their livelihood. When people are displaced, production systems are dismantled, kinship groups are scattered, and long-established residential settlements are disorganized. People's lives are affected in very painful ways. Many jobs and assets are lost. Health levels tend to deteriorate. Linke between producers and their customers often are severed, and local labor markets are disrupted. Informal social networks that are part of daily sustenance systems - providing mutual help in child care, food ·security, revenue transfers, short-term credit, labor exchanges, and other basic sources of socioeconomic support - are dissolved. Local organizations and formal .and informal associations disappear because of the dispersion of their membres. ·Traditional community and authority systems can lose their leaders. Symbolic markers, such as ancestral shrines and graves, are abandoned, breaking links with the past and with peoples' cultural identity. The cumulative effect can tear apart the social fabric and local economy, and is profoundly disruptive to large numbers of people. The ma~n risk is impoverishment - through landlessness, joblessness, food insecurity, deteriorating health, or the loss of access to community assets. That is why carrying out resettle- .ment adequately is an impoverishment prevention and poverty reduction task. Displacing people involuntarily confronts government organizations with complex legal issues, The potential for violating people's individual and group rights makes compulsory relocation unlike any other development activity. The fact that the execution of some development project is delayed by courts, and that compensation levels are often raised significantly on appeal, reflects the recognition in legal systems that people cannot be arbitrarily displaced without just compensation, regardless of national need. When resettlement processes are carried out in a lawful manner that fully respect people's rights, opposition to projects by adversely affected people is reduced (althoug not eliminated) and • overall project implementation is likely to unfold more effectively. Resettlement · that reflects the needs and rights of affected persons is not just compliance with the law, but also constitutes sound development practice. The international debate on resettlement. Because of its complexity and I adverse effects, involuntary resettlement has become the focus of a wide I international debate, engaging and polarizing governments and nongovernmental .1 organizations, public option groups, parliamentarians, development agencies, and 102 MICHAEL M. CERNEA 6 the media. Recent criticism of involuntary resettlement has often evolved into rejection of the goals and legitimacy of the program causing the resettlement, and sometimes of development itself. Two arguments are advanced. One denies, in principle, the acceptability of any involuntary resettlement. The second criticizes the low quality of specific resettlement operations. The social controversy around population displacement is understandable, yet the undifferentiated rejection of all resettlement is not a realistic position. In many development situations involuntary resettlement is unavoidable, and simply denying its necessity is myopically short sighted. Instead, it is more effective to ask crucial practicap questions: first, bow to minimize displacement's magnitude; and second, bow to respond effectively to the needs of the people being resettled. Our research tyam has explicitly stated that it shares the views of those critics, including World Bank critics, who deplore in an objectively documented manner the bad quality of resettlement operations. In fact, through its very decision to adopt a formal policy on resettlement issues, a policy that is based on sociological/anthropological research findings and that promotes equitable principles and approaches, the Bank has delivered the sharpest criticism of bad displacement practices that cause th·e impoverishment of those displaced. The World Bank's Resettlement Policy. Indeed, the Bank was the first international development agency to respond to the complexity and difficulty of development-cased displacements by adopting an explicit policy and institutional procedures to address displacement processes and resettlers' needs (World Bank 1980, Cernea 1988). Significantly, the very formulation of this policy is one of the instances in which research and policy proposals made by sociologiste inside the World Bank have resulted in the formal adoption by the institution of policy measures with wide international impact (I described these social science contributions in detail in other studies - see Cernea 1995b and 1995c). Through such sociological work, the theoretical conclusions and empirical findings of prior social science research on displacement/resettlement were translated into prescriptions that were adopted as policy. In summary, the basic elements of the World Bank's resettlement policy are: • Avoid or Minimize. Involuntary displacement should be avoided or minimized whenever feasible, because of its disruptive and impoverishing effects. • Improve or Restore Livelihoods. Where displacement is unavoidable, the objective of Bank policy is to assist displaced persons in their efforts to improve, or at least restore, former living standards and earning capacity. The means to achieve this objective consist of the preparation and execution by the borrowing country of resettlement plans as development programs. These resettlement plans should be integral parts of project designs. • Allocate Resources and Share Benefits. Displaced persons should be: (i) compensated for their losses at replacement cost, (ii) given opportunities to share in the benefits from the project that displaces them, and (iii) assisted in the physical transfer and during the adjustment period at the relocation site. 7 Induced Population Resettlement 103 9 Move People in Groups. Minimizing the distance between departure and relocation sites and moving people in groups can cushion disruptions and facilitate the resettlers' adaptation to the new sociocultural and natural environments. The tradeoffs between distance and economic opportunities must be balanced carefully. • Promote Participation. Resettlers' and hosts' participation in planning resettlement should be promoted. The existing social and cultural institutions of resettlers and their hosts should be relied upon in conducting the transfer and reestablishment process. • Rebuild Communities. New communities of resettlers should be designed as viable settlement systems equipped with infrastructure and services, able to integrate in the ,regional socioeconomic context. • Consider Hosts' Needs. Host communities that receive resettlers should be assisted to overcome possible adverse social and environmental effects from increased population density. • Protect Indigenous People. Tribal and ethnic minorities, pastoralists, and other groups that may have informal customary rights to the land or other resources · taken for .the project, must be provided \vith adequate land, infrastructure, and other compensation. The absence of legal title to land should not be grounds for denying such groups compensation and rehabilitation. The fundan1ental goal of the Bank's policy is to restore the living s;:andards and earning capacities of displaced persons - and when possible to improve them. Its provisions protect and enlarge the entitlements of displaced people and promote a safety net approach for restoring their livelihoods. By adopting this policy, the Bank rejected the argument that impoverishing resettlers was an unavoidable, if lamentable, facet of development. The Bank's policy calls for transforming people's in\;'oluntary resettlement into an opportunity for development and for enhancing _their prior living standards by enabling resettlers to share in the benefits of the development project that causes their displacement. Restoring previous standards of living is a formidable task in practice. The nature. ~nd the dimensions of the tasks at hand in implementing this policy must be well understood. As social scientists have argued, in many developing countries the absence of effectively functioning land and labor markets, the substantive and procedural inadequacies of compensation system for property appropriated by the state, and the lack of adequate social safety nets represents major factors limiting the opportunities for adequate socioeconomic resettlement. These are the three main reasons why the simpi'e compensation in cash of property losses (as provided generally by eminent domain laws) cannot realistically be expected at this time to result in satisfactory outcomes for project-affected people in developing countries. Preventing the impoverishment that displacement is likely to cause requires changes in the policies, legal frameworks, institutional capacities, and current practices of many developing countries, as well as of the countries in transition from planned to market based economies. This challenge is faced anew in every single development project that entails displacement. I . 104 MICHAEL M. CERNEA 8 For nations as for the World Bank, policy goals must be translated into systematic implementation. While implementation in Bank-financed projects has not consistently been at the level required by the policy, it has been improving. It also is a significant fact that treatment of affected populations under projects that apply Bank policy is, in general, considerably better than under programs not financed by the Bank - and the latter account for up to 97-98 percent of total involuntary resettlement in the developing world. Hence in countries where the Bank is financing projects which involve resettlement, the Bank also assists governments is establishing or improving domestic national policies that would I apply to all relevant domestic programs, the majority of which take place without Bank-financed assistance. KEY FINDINGS OF THE ACTION-RESEARCH The remainder of this article is devoted to summarizing the findings of our vast study on displacement causing development programs. Our action-research was carried out primarily in the field, jointly with many local sociologists and country officials, and included also secondary analysis of pre-existing data. About 150 small field teams that comprised at least one social specialist (usually a sociologist or anthropologist) were sent by the World Bank to analyze resettlement in the fielJ in over 100 projects in all countries under analysis. Given the applied character of our research, the types of questions we asked were primarly oriented towars assessing performance in the application of the social policy on displacement and resettlement. The study focused also on understanding the anatomy of some key socioeconomic processes occurring during forced displacement, such as impoverishment. But it also addressed the work of governments in developing countries, and the World Bank itself and its own staff, in preparing, appraising, implementing and supervising the resettlement programs. In this respect, this action-research generated a number of critical findings, which were reported in detail and candidly, and which triggered the adoption of important policy measures either by various governments or by the Bank, as well as improvements in our institution's working procedures. In this respect, the study demonstrated how useful applied sociological research can be to governmental agencies and policies. Quantified assessments. To begin with, the study carried out a "search and count" of projects that entail involuntary population relocation from among the "universe" of about 2000 ongoing Bank-financed projects. It was found that in the 1986-1993 period involuntary resettlement was a part of 192 projects, displacing a total of 2.5 million people over the life of those projects. Forty-six projects started in or after 1986 had closed by 1993, having resettled over half a million people. Under the 146 projects still active in 1993, nearly two million people are in various stages of resettlement. These projects represent eight percent of all the Bank's currently ongoing projects, and account for US$23 billion or 15 percent of overall Bank lending. The large majority of the Bank-funded projects (60 percent) are located in East Asia and South Asia, due to the rapid development of these 9 Induced Population Resettlement 105 countries and their high population densities. According to preliminary estimates, new projects proposed by member countries for Bank financing during 1994-1997 were expected to displace and resettle about 600,000 people. The scale of resettlement on a project basis varies from less than 500 to more than 200,000 people. However, the bulk of resettlement is concentrated in a handful of projects in India, China, Indonesia, and Brazil: eleven large projects situated in these countries account for over half-1.1 million- of all the people being resettled. I Projects in the South Asia and East Asia regions account for 80 percent of the population to be resettled. Latin America's share in the resettlement project I 1 portfolio is 9 percent and Africa's is 23 percent. However, the number of countries with multiple resettlement operations is expanding: the growing economies in 1 Pakistan, Bangladesh, Vietnam, and Indonesia expect significant i::J.creases in the I number of Bank-supported projects with resettlement. ! · The most significant sectoral shift among the various groups of projects entailing resettlement is the rapid rise of resettlement in transport, water and urban infrastructure projects. This group now represents the largest sector in terms of number of projects, accounting for 75 of the 146 ongoing projects involving resettlement. However, these resettlement operations are smaller than the average in hydropower or agriculture projects. Dams for irrigation, hydropower and drinking water are the single largestcause of displacement (63 percent of displaced people), and transportation corridors are the second largest cause (23 percent). In sectoral terms, agri;;ultura! projects c.'Jntinue to account for the largest segment of resettlers (52 percent). QualitatiYe findings. Beyond quantifying the extent and magnitude of resettlement operations, the study's most important findings resulted from the analysis of resettlement impacts and of the relationship between policy prescriptions and actual operations. · The single most important finding of the study is that good resettlement can 'prevent impoverishment and even reduce poverty by rebuilding sustainable livelihoods. Conversely, if resettlement is not done adequately, resettlers end up worse off than before. SociaJly responsible resettlement is also economically beneficial because the heavy costs of poorly handled displacement extend well beyond the immediately affected population - to the regional economy, and to the host population in relocation areas. Inadequate· resettlement induces local I resistance, increases political tensions, entails extensive project delays, and postpones project benefits for all concerned; the benefits lost because of such ~ avoidable project delays sometimes far exceed the marginal cost of a good ~ l resettlement package. Therefore, ensuring that involuntary resettlement is mininiized - and when unavoidable, is carried out without impoverishing the people I displaced - is fully justified on both economic and ethical grounds. i The largest resettlement operations move tens of thousands of people - often very poor people - long distances in a very short time, and reestablishing and improving their standard of living is a hard task. The inherent difficulty in reestablishing livelihoods and community services is compounded by the limited technical and institutional capacity of most developing countries, and by weak political commitment from some governments or executing agencies. 106 MICHAEL M. CERNEA 10 In essence, the field research and the various secondary analyses carried out during the 15 month of our study found that: a) The programs that follow Bank policy have demonstrably better resettlement results than projects that bypass these guidelines, and are superior to resettlement carried out outside Bank-financed programs. b) Resettlement performance is directly associated with the presence or absence of a domestic policy and organizational frameworks for resettlement. Countries, states, or sectors with an adequate resettlement policy generally achieve better outcomes in preventing impoverishment and restoring livelihoods than do countries or sector which lack such policies. c) The Bank's portfolio of resettlement operations is in much better health than in 1986, the year the previous resettlement review was completed - and radically different from the state of the 1980 portfolio - when the resettlement policy was issued. However, in a number of projects actual resettlement operations and outcomes are not consistent with the standards defined and demanded by the Bank's policy. d) The planning processes and criteria established through the Bank's policy have significantly improved the practices of some developing countries, other international donors, technical agencies implementing large projects, and the Bank itself. However, progress in this respect has been insufficient and uneven. Much more needs to be done to ensure consistency of planning, outcomes, and impact monitoring with policy goals. The in-depth study of resettlement experiences has generated important knowledge for both the World Bank and its member countries about ways to carry out, effective resettlement, and has identified many ways of correcting weak performance that are described in the full report. Despite the vast differences among countries and populations involved, much more has been understood about the major common factors that explain - by their presence or absence - why resettlement worked in some cases and failed elsewhere. The social dimensions of civil works construction projects require much greater attention by member country governments, as well as by the Bank (see also World Bank Operations Evaluation Department, 1993). Hence, beyond project-by-project activities, institutional development for addressing resettlement issues needs to be pursued as a valuable development objective in its own right. PERFORMANCE IN RESETI'LEMENT Evaluating the performance of the World Bank with respect to resettlement, the study concluded that the Bank made significant progress during 1986-1993 in three major areas: • Influencing the resettlement policies of a number of developing countries. • Assisting agencies from developing countries in avoiding unnecessary resettlement operati::ins or in reducing the scale of unavoidable displacements. • Providing assistance for improving the circumstances of resettling groups and their ability to restore their income. 11 Induced Population Resettlement 107 , At the same time, it was found that in programs implemented in a manner not consistent with policy and procedural guidelines, many people compelled to resettle ended up worse off. Influencing policies. Having been the first international development agency to adopt a resettlement policy, the Bank has promoted this policy with the developing countries whose projects include population displacement. One main I result of the Bank's catalytic impact is that several countries enacted or improved domestic policies and legal frameworks for resettlement. I I Resettlement works when governments want it to work. The responsability of J ' governments is to create adequate institutional capacity, that we defined as the synergy between policy, organizations, and resources. Where borrowing agencies do not genuinely concur with the Bank's resettlement policy objectives, resettlement i~ generally not carried out well, regardless of the number of covenants, fallow-up supervision missions, or the frequency of Bank threats to siispend disbursements. Similarly, when the Bank itself does not consistently. adhere to its policy principles or procedures, project performance is weakened. The 1986 to 1993 period has been marked by some notable policy successes. For instance, with support from a Bank-financed sector project in 1990, Brazil's Electrobras Power Company developed satisfactory policy guidelines for resettlement in the country's power sector. Colombia adopted similar sectoral guidelines in 1992. China, with Bank involvement, improved its policy guidelines in the water resources sector. In 1993, India's National Thermal Power Corporation adopted a sectoral rehabilitation policy for all its operation, following detailed negotiations with the Bank as part of preparations for a sector loan. We concluded that the Bank has been far more effective overall - and immediate operations have benefited more - when it succeeded in reaching agreement with borrowing on the· broad domestic or sectoral policy framework relevant to Bank-assisted operations, than when its efforts were only confined to legal agreements for individual projects. In tum, the obligation laid down in individual loan legal agreements and the agreed upon "project policy" have sometimes formed the basis for discussing and improving more general domestic policy and legal frameworks. Major multilateral and bilateral donors also have recently issued resettlement guidelines similar to the World Bank's - for instance, the Inter-American Development Bank in 1990 - and the Asian Development Bank is now considering formally adopting similar guidelines. At the OECD's request, the Bank provided supp!Jrtin preparing resettlement guidelines, congruent with the Bank's guidelines, · and in 1991 the development ministers of all OECD countries sanctioned similar unified guidelines for their countries' aid agencies (OECD 1992). Avoiding or reducing displacement. The potential for reducing or eliminating. resettlement exists in many project proposals submitted to the Bank for financing. The engineering redesign of the Saguling high dam in Indonesia, for instance, lowered dam height by five me:ters, thus avoiding the displacement of 35,000 people. In the Cote d'Ivoir forestry project, the Bank proposed revision that reduced the Borrower's proposal to displace up to 200,000 people by 80 percent, to 40,000, and set substantially higher standards for those to be relocated. In Thailand, 108 MICHAEL M. CERNEA 12 resiting the Pak Mun dam and lowering its height has reduced displacement from 20,000 to 5,000 people. In Ecuador's Guayas flood-control project, the redesign of canal layouts completely eliminated the need to displace anyone. Restoring incomes. The ultimate test of consistency between resettlement operations and policy is income restoration and improvement. Resettling people productively on land and in jobs tends to restore income more effectively that only handing out compensations in cash without giving people institutional assistance. Incomes were successfully restored particulary when resettlers were enabled to share in the immediate benefits created by the very project that displaced them. Examples include helping resettlers develop aquaculture and fisheries in the new reservoirs (Indonesia), moving them into the newly irrigated command areas (Chin~, and some projects in India) or favoring them to cash in on the commercial opportunities created around the newly constructed infrastucture (Argentina, China, Turkey). In many projects, living standards for urban resettlers improved through more durable housing, more floor space per capita, and better access to sanitary services - even though this came at the cost of a longer commute. The performance of resettlement operations in restoring pre-move income levels is technically measurable, but few projects have included sufficient measures or methods to assess whether income restoration is being achieved. Though fragmentary, the weight of available evidence points to unsatisfactory income restoration more frequently than to satisfactory outcomes. Performance in income restoration is of particular concern in some countries (e.g., India and Indonesia), which have numerous projects entailing displacement but do not have a country- wide policy or legal framework prescribing the restoration of resettlers' incomes. We found that the key explanatory variable for success in restoring livelihoods on a productive basis is the government's com1.1itment to help resettlers, reflected in sectoral or national policy. THE RISK OF IMPOVERISHMENT Because the main social and economic risks arising from forced displacement is the impoverishment of the affected people, I will discuss this central issue in more detail below. Much of the research reported prior to our study in the sociological and social anthropological literature (Fahim 1983; Suarez and Cohen 1985; Fernandes and Thukral 1989) as well as studies by World Bank researches (Shibata 1991; Cernea 1988, 1991; Guggenheim 1989; Butcher 1990; Partridge 1990; etc.), have documented many adverse impacts of displacement on people's welfare, community life, and social identity. In a study devoted to impoverishment effects (Cernea 1990) I have proposed a "risk model" consisting of eight convergent subprocesses which result in lasting impoverishment, if the risks are not addressed from the outset through the very planning of the development program that causes displacement. The 1993-94 study examined the empirical evidence available from many resettlement operations in light of this impoverishment model. Indeed, to augment I I 'I 13 Induced Population Resettlement lo