The question not asked : when does displacement end?

"A drop reflects the ocean", the old saying goes. This simple question - "When does displacement end?" - similarly reflects an ocean: the ocean of displacement issues. This question sounds simple but it isn't. In fact, it probes the depth and length of the entire involuntary displacement-resettlement continuum.

Population displacements differ by cause, content, the way they unfold and the way they end. Answers about 'end' must vary depending on the initial cause of involuntary displacement - which may be war, a natural calamity or a necessary development project. This article refers to the end of one type of displacements - those caused by public sector projects that install new infrastructure, build plants, construct highways or accomplish other needed developments.

Because such displacements are deliberate and planned in advance, their 'end' would logically have to be defined also in advance, and under the same plan. Paradoxically, however, these plans define only the start of displacement but not its time-bound end. Why this inconsistency?

Resettlement Action Plans (RAPs) are important instruments and contain many provisions indispensable for protecting those displaced. Instituting RAPs as mandatory safeguarding tools in development projects marked a major progress compared to the unregulated approaches of the pre-1980 period. Surprisingly, however, despite the great detail prescribed for designing such RAPs, they indicate only a time-bound beginning but not a time-bound end to displacement. In short, the question about the time-bound completion of the state of displacement is not asked. And an answer is not given. Yet this is a fundamental question. It must be asked. It must also be included in planning. And it must be answered.

However, before asking about the end, there is also another tricky question that needs clarification: when does displacement begin?

It is commonly assumed that displacement 'begins' when people are forced to physically leave their habitat. This is correct only in the case of displacement by natural disaster or by war, when the onset of population displacement is sudden and people must flee immediately. However, in displacements caused by development projects, the 'displacement' often begins long before the actual physical removal of people, as the onset of the economic effects of expropriation very often precedes physical relocation.

Indeed, in projects that expropriate large areas for 'right-of-way' - such as dams and reservoirs, highways, strip mining, etc - the decision about the forthcoming land take is made long before actual population removal. The legal principle of 'eminent domain' is applied and the new category of development-displaced people is created as a legal category, the result of implementing the expropriation decision. This decision leads to legal public notification of area boundaries and entails a 'cut-off' date and legal prohibition of new constructions and of new investments in the condemned areas, to avoid increases in compensation costs. This in turn induces depression - causing drops in land prices, halt in housing and enterprise construction, freezing of public investments for public services expansion, etc. The 'to-be-displaced' inhabitants begin to suffer adverse economic consequences long before being physically displaced. This period of pre-project 'condemnation' may last many years, until the project actually starts. These are years when the ugly manifestation of relative impoverishment begins.

This is why recognising the real 'beginning' of development-induced displacement is no less valid than asking about its 'end'. In light of empirical observations and social analysis, the conventional wisdom about the 'beginning' of displacement must be amended. The displacement clock starts earlier than is commonly assumed.

Criteria for defining the end of displacement: policy objectives

Let us return now to the initial question and examine our finding that Resettlement Action Plans generally tend to avoid this explicit question and its answer. RAPs fail to explicitly 'plan' a target end-date of displacement, although projects have to plan for all their other components. Information available for RAPs is plenty - on project conditions, area characteristics, inputs, options, projected outcomes, etc. RAPs are in a position to set a time-bound benchmark as a goal, within the project period or slightly after it, but this does not happen in the current format of RAPs. Let us therefore examine two elements: the criteria for defining the end of the state of displacement and the measurements to ascertain that criteria are met.

Consensus is growing among IDP researchers about the need for specific criteria on determining when displacement ends for various IDP groups. The criteria would clearly differ by the cause, content and type of displacement for different sub-categories of IDPs. For refugees, explicit "cessation criteria" determining when their state as refugee comes to an end have been already long defined, as early as in the 1951 Refugee Convention. In the case of some IDP sub-categories (resulting from conflicts, for instance) the problems are complex and under review, with different view points expressed, but for the category of development-displaced people the criteria may be easier to outline.

I advance for discussion two propositions about how we could define the end of displacement for the large sub-category of development-displaced populations:

First, the criterion for determining the end of displacement must derive from the policy that defines the objective of development-induced involuntary resettlement. If displacement results from deliberate policy, and is legally induced, displacement's end must be equally policy-induced and determined.

Second, in light of this criterion, displacement would end when the policy objective is reached - namely, when the displaced people achieve a livelihood level improved over their pre-displacement levels or (as a controversial minimal caveat still allows under some current policies) when they are restored at a level equal to their pre-project level (plus 'without-project' growth). Current resettlement policies of major development aid agencies, bilateral and multilateral, clearly define the basic objective of their involuntary resettlement policies as "improving, or at least restoring" livelihoods.

These two propositions, as answer to the legitimate question about 'when does displacement end?', imply also an intimate link between criteria and measurements. Criteria which are not measurable would be useless. Conversely, if measurements are not undertaken, theoretical criteria alone would be of little use.

Practice versus policy

What does current practice tell us? Time-wise, the displacement-resettlement continuum unfolds through three essential phases: 1) expropriation/displacement; 2) transfer process from old location to new site; and 3) resettlement and reconstruction of livelihood. The process initiated with the identification of land for expropriation and the physical displacement does not end at phase two. It truly ends only when the third phase - resettlement-reconstruction as per the policy requirement - is completed. This explains why confusing the end of displacement with the end of the physical transfer phase is a grievous mistake. Unfortunately, many government officials, planners and project managers are still making this confusion.

As a result, many people are displaced and relocated but remain not rehabilitated for years: for them, displacement has not ended. While in theory the first two phases of the continuum must be followed immediately by the third phase, in practice this very often does not happen, and those displaced do not advance through the third phase. Those affected are not sustainably reestablished, displacement is not ended, and they remain displaced and impoverished, left chronically behind those unaffected by the project.

As long as RAPs avoid setting a benchmark end-date (even a flexible one), there is no 'planned' milestone in a given project against which to measure performance and institute accountability. Leaving the end of displacement open-ended rather than planned in a time-bound fashion leaves those displaced in limbo and undercuts the safeguards principle.

One of the most telling statistics in this respect was provided by Indian researchers, who concluded that out of about 20 million people displaced in India over a 40 year period, 75%, that is 15 million, were relocated but not rehabilitated, emerging from the transfer phase worse off than they were before the projects and displacement started. (1)

Necessary remedies

What would be the practical and operational implications of defining displacement's end correctly, in light of basic policy objectives?

Empirical findings from research carried out by many anthropologists, geographers and sociologists led me to the formulation of the Impoverishment Risks and Reconstruction (IRR) model, which posits that overcoming the risks of decapitalising resettlers and impoverishing them is the core task in resettlement.(2) The basic poverty risks identified in the IRR model are: landlessness, joblessness, homelessness, marginalisation, food insecurity, increased morbidity, education loss, loss of access to common property resources and social disarticulation. If displacement is wrongly defined as ending with physical transfer, before reconstruction begins, those displaced are abandoned to fight on their own the risks imposed on them. Compensation for expropriated assets is not the full solution by far and has long been proven to be an insufficient remedy against impoverishment. To avoid such situations, I suggest three points.

First, a correct definition of what is the 'end' of development-induced displacement would need to be included in every project's RAP as an objective consistent with policy.

Second, the very content of the RAP should be tailored towards achieving this end, in a measurable way and with time-bound accountability for each project. This, again, is usually not happening now, perhaps with the exception of some internationally assisted projects subject to monitored safeguard policies. Pressure to reach the policy-defined end will force improvements in the entire preceding resettlement work. Projects guided by safeguard policies have indeed made great progress towards transparently accounting for the number of people to be displaced, mitigating many adverse impacts. But many such projects do not transparently account for how many of these displaced people end their displacement with their livelihoods improved by the project's end, and how many do not. To the credit of some agencies, a demand for some assessments was incorporated in the OECD and World Bank guidelines on involuntary resettlement. Yet, these assessments are not explicitly linked to reaching 'closure' and an end of displacement.

To achieve clear awareness of results, RAPs would have to schedule and carry out a sample survey study as part of the project completion report to determine a) whether, and how many, of those displaced have ended their displacement in substantive terms, recovering and improving their livelihoods and b) how many have not ended it during the project's limited duration and therefore must be still seen as needing assistance. Without such a study the safeguarding discourse cannot really conclude whether projects have succeeded in achieving the basic objective of the resettlement policy or whether these projects leave the process unfinished.

The recently revised World Bank policy on resettlement states that "upon completion of the project, the borrower undertakes an assessment to determine whether the objectives of the resettlement instrument have been achieved".(3) This positive proposition is a step in the right direction, yet it is more descriptive than prescriptive and is rarely carried out. The OECD guidelines prescribe that "all resettlement plans… should include a target date when the anticipated benefits to resettlers and hosts are expected to be achieved"(4). But unfortunately it is widely known that key bilateral aid agencies of OECD countries only weakly, if at all, implement their resettlement guidelines, and the correct concept of a 'target-date' is forgotten and absent.

My third and last point responds to the possibility that such a survey would find that many of those displaced have not 'ended' their displacement. Resettlers must be provided with resources and opportunities to share in the project's benefits and truly recover. Project agencies should design follow up measures to pursue assistance until the policy goal is accomplished. There are numerous operational ways in which this can be done measurably. Otherwise, exposing people to added risks and discontinuing reconstruction assistance before their economic displacement ends would only mean swelling the ranks of the poor with newly impoverished people.

I started this article by saying that this simple question - when does displacement end? - reflects the vast ocean of displacement problems. Indeed, it led us to examine in turn the beginning of displacement, its risks, the phases of the process, its impoverishment effects, the policy that would counteract the risks, criteria and measurements - all this in order to understand what its end is and ought to be. Thus, it is a difficult but very worthwhile question. The question about 'end' precipitates clarity. This key question must be asked in every single instance of displacement, in every RAP. And - most importantly - the question is not a riddle: it does have a clear answer. Current practice must be improved to accomplish the policy-required and defined answer.


Michael M. Cernea is Research Professor of Anthropology and International Affairs, George Washington University, in Washington DC, and Senior Adviser-Consultant of the World Bank. Email:


  1. Fernandes Walter, 'Development-Induced Displacement in Eastern India' in Dube S C (ed) Antiquity to Modernity in Tribal India, vol 1, 1998, Inter-India Publications: New Delhi; Singh, Shenkar and Banerji P (eds) Large Dams in India, 2002, New Delhi: IIPA. (Fernandes quotes the economist Vijay Paranjpye who estimates the dam-displaced people in India at over 21.6 million.)
  2. Cernea Michael M The Risks and Reconstruction Model for Resettling Displaced Populations, 1997, World Development 25:10:1569-87; Cernea Michael M 'Risks, Safeguards and Reconstruction: A Model for Population Displacement and Resettlement' in M Cernea and C McDowell (eds) Risks and Reconstruction: Resettlers' and Refugees' Experiences, 2000, Washington: World Bank.
  3. World Bank, Operational Policy 4.12 : Involuntary Resettlement, 2001,Washington DC
  4. OECD-DAC, Guidelines for Aid Agencies on Involuntary Displacement in Development Projects, 1992, Paris.

Opinions in FMR do not necessarily reflect the views of the Editors, the Refugee Studies Centre or the University of Oxford.
FMR is an Open Access publication. Users are free to read, download, copy, distribute, print or link to the full texts of articles published in FMR and on the FMR website, as long as the use is for non-commercial purposes and the author and FMR are attributed. Unless otherwise indicated, all articles published in FMR in print and online, and FMR itself, are licensed under a Creative Commons Attribution-NonCommercial-NoDerivs (CC BY-NC-ND) licence. Details at



facebook logo Twitter logo RSS logo

Forced Migration Review
Refugee Studies Centre
Oxford Department of International Development
University of Oxford
3 Mansfield Road, Oxford OX1 3TB, UK  +44 (0)1865 281700
skype: fmreview