Internal Displacement: Return, Property, Economy

Authors


Although there is an increasing recognition of the right to return in the international arena, the problem of conflict-induced displacement remains a debacle in the world. While almost fifty per cent of the peace agreements signed since 1991 include provisions on the right to return of refugees and internally displaced people (IDPs), seventy per cent of the countries involved still have ongoing internal displacement problems. The global number of IDPs has increased from 23.7 million in 2005 to an estimated 24.5 million in 2006 mainly due to the fact that more people were newly displaced in 2006 than were able to return during the same year. While 4 million people -- more than twice as many as in the previous year -- were displaced as a result of armed conflict in 2006, only 3.6 million were able to return to their homes. While quantitatively analyzing different conditions that the literature cites as affecting the rate of return of IDPs through a dataset that I had compiled, I will argue that although internal displacement is a security problem in its nature, its solution lies in establishing effective property repatriation mechanisms for the displaced as well as creating economic opportunities.

The argument here is based on a simple assumption: if IDPs are to return, then they need a place to which they can return. Therefore, they not only need to be granted the right to return; they also need their property rights to be restored, or in cases where property rights did not exist prior to conflict, granted. There are two relationships between property rights and the return and resettlement of IDPs. While it is easy to assume that having a place to return to will actually facilitate return, there is also a more complex connection. Property rights not only give people a place to live and/or work, but following Hernando de Soto’s (2000) innovative study of the poor, they are also the key to the creation of capital, and therefore provide the means to build new lives elsewhere. Thus, property rights are important as catalysts for either the return or the resettlement of IDPs as well as being crucial means of reintegration of returned or resettled IDPs.

The growing attention to the problem of internal displacement has not produced any systematic global studies with comparative insights. Although competing hypotheses about return of IDPs are developed, broad patterns are not identified in the literature. In order to fill in this gap in the literature, this paper subjects the competing hypotheses about return to an aggregate statistical test. The aim is twofold. The first one is to detect those variables that affect the patterns of return in the entire universe of countries with internal displacement problem, and the second is to test my own hypothesis on property rights and return against existing propositions. Accordingly, in order to test the competing hypotheses quantitatively, I have compiled a new dataset incorporating all the countries with problems of conflict-induced internal displacement.

The paper consists of four sections. While the first section presents a literature review on the concepts such as internal displacement and return the second gives details about the dataset employed to test the hypotheses, the sources of the data, possible problems with the data, and possible solutions to alleviate these problems. The dependent and independent variables are defined and the modes of measurement explained. The third section introduces the methods used to analyse the data and presents the statistical results, while the final section summarizes the findings and is followed by a brief conclusion.

Internal Displacement and Return

The return of refugees and IDPs has been an objective of almost 50 per cent of the peace agreements signed in the post-Cold War era and policy-makers and academics all believe that peace requires refugee repatriation, and that every peace agreement must insist on it (Zolberg et al., 1989). However, evidently, bestowing the right to return is not in itself a sufficient condition for successful return, but there are also certain conditions that must be met in order to achieve the result. Unfortunately, the literature on the conditions for IDP return is relatively small and largely idiographic. It consists of many single case studies and a few comparative analyses, all of which are generally lacking in global scope.

However, Deininger et al.’s (2004) case study, which analyses the conditions that have affected the desire to return of the displaced population in Colombia, delivers valuable results, and the literature on ending civil wars provides supporting arguments for each of their findings, which are as follows: First, the desire to return is affected by economic prospects in the reception and origin sites: the study demonstrates that economic calculations have a great impact on IDPs’ decision to return or resettle. The study shows that possession of assets in the place of origin, especially access to land, creates a strong desire to return regardless of the land size. Based on similar economic calculations, possible employment opportunities at the original site also seem to positively affect IDPs’ desire to return. If return is positively correlated with economic prospects in the origin sites, then the importance of economic recovery and sustainable development as priorities of successful peace implementation in conflict areas is multiplied (Barakat and Chard, 2002; Smoljan, 2003; Woodward, 2002).

Second, the desire to return is affected by the socio-demographic characteristics of the household: the study also reveals that more vulnerable households (defined as households that belong to an ethnic minority, with one parent, with female heads or large dependency ratios, or that have large numbers of members who are younger than 14 years of age) demonstrate less of a desire to return. It is important to note that, as Cohen and Deng point out, “(d)isplacement changes the structure and size of households and changes family patterns and gender roles,” and “the number of female-headed households increases significantly” (1998: 26). Thus, the phenomenon of displacement itself affects the motivation to return.

Third, the desire to return is affected by how traumatic the displacement process was: when displacement is caused by traumatic events, families are less willing to return. Research conducted among the Madurese IDPs in Indonesia provides information that yields a similar conclusion on the relation between traumatic events and return: those who have lost relatives are less likely to go back to their homes (Turnip, 2003). Furthermore, in a country like Bosnia and Herzegovina where ethnic cleansing and war led to the death or disappearance of approximately 100,000, return does not become the most plausible option (Barakat and Chard, 2002; Tyler, 2003). Thus, the focus of peace operations on justice is further emphasized in relation to IDP return. As Mendeloff explains, “A sense of justice is often necessary for the personal, psychological healing that allows for reconciliation” (2004: 360). Therefore, a feeling of justice is necessary to help overcome the trauma for the return of IDPs.

Fourth, the desire to return is affected by the duration of displacement: return becomes a less desirable option when periods of displacement are long; households choose to resettle in the new place of residence rather than to confront an uncertain situation in their villages of origin. The findings of the civil war literature on the duration of conflict and return are prolific: conflicts are likely to last longer “when rebel groups are weaker relative to the government, have fewer options for substituting political means for violent ones, and control territory and operate in the periphery” (Gleditsch et al., 2005). There are also studies asserting that the most durable settlements are likely to be those that conclude civil conflicts of low intensity that have lasted for extended periods of time (Hartzell et al., 2001). Furthermore, the literature also emphasizes that wars in which land constitutes an important factor of conflict between a peripheral ethnic minority and state-supported migrants of a dominant ethnic group tend to last for long periods of time (Fearon, 2004). However, in certain cases, like Cyprus, the length of the period of displacement might not have any effect on the desire to return. Similarly, despite their decades-long partition due to the enduring Israeli-Palestinian conflict, permanent separation from home is unimaginable for many Palestinians (Sayigh, 2004).

Besides the factors asserted by the Deininger et al., study, I test another factor of effective property repatriation mechanisms. The restoration and effective enforcement of property rights of IDPs is an essential factor in the successful return, and resettlement, of these people because it: ensures that IDPs have a place to live in case of return and secures capital (that is, a means of rebuilding their lives), also in case of settling somewhere else. The idea that property rights are an important determinant of the settlement of the IDP problem and an element of sustainable peace was inspired by the case of Titina Loizidou, a Greek Cypriot IDP who had been internally displaced since the Turkish occupation of the north of Cyprus in 1974. On December 18, 1996, the European Court of Human Rights agreed that the right of Ms. Loizidou, “to the peaceful enjoyment of (her) possessions’” guaranteed under Article 1 of Protocol No. 1 of the European Convention for the Protection of Human Rights and Fundamental Freedoms was violated by Turkey, and in December 2003, Ms. Loizidou received damages of € 1.2 million from Turkey. The decision of the Court was widely publicized in the Turkish press. The case was an illustration that after more than three decades, there were still 210,000 IDPs in Cyprus, and getting their pre-1974 properties back was almost a precondition for peace on the island. Thus, restoration and effective implementation of property rights was an essential factor in the successful return and resettlement of IDPs of Cyprus, and in the creation of a sustainable peace on the island.

The importance of the property rights is also asserted by the Universal Declaration of Human Rights where “Everyone has the right to own property alone as well as in association with others. No one shall be arbitrarily deprived of his property”. Principle 21 of the Guiding Principles on Internal Displacement reiterates the same right, but still, the legal framework regarding the property rights of IDPs is not very concise. As Bailliet writes, “there is disagreement over whether the right to property is a civil and political right covering registered property, as opposed to a socio-economic right applicable to the customary claims of indigenous people/farmers who link the land to their rights to food, housing and the right to life itself. General human rights instruments do not set forth a right to restitution of property and the soft law is vague” (2003: 10). Furthermore, Paragraph 2 of Principle 29 of the Guiding Principles declares:

Competent authorities have the duty and responsibility to assist returned and resettled internally displaced persons with recovery, to the extent possible, of the property and possessions they left behind or were dispossessed of upon their displacement. When recovery of such property and possessions is not possible, competent authorities shall assist these persons in obtaining appropriate compensation or other forms of just reparation or shall themselves provide such recompense.

On the face of it, this bold statement tackles the problem of uncertainty that IDPs face in relation to their right to property at the time of return or resettlement. An examination of the case law of the general human rights instruments shows that the right to restitution or compensation was usually decided in the affirmative (Kälin, 2000). In reality, Principle 29 confers the choice of the possible forms of reparation upon the state without any guarantees that land will be returned (Bailliet, 2003). This can become a major problem for IDPs in cases such as Colombia where, as Pettersson argues, civilians are constantly faced with threats of violence meant to displace them from their homes and lands, while the government shows a lack of willingness to respond to IDP demands for resettlement or compensation (2000). For example, we now know that the civilians in Colombia faced a double threat (1) from left-wing rebels, that is, the Revolutionary Armed Forces of Colombia (FARC) funded by kidnapping and extortion of rich Colombians as well as drug trafficking, and (2) from the right-wing paramilitaries that were organized in the 1980s to help defend rich Colombians. Moreover, government officials are known to collude with paramilitaries against civilians. Members of President Alvaro Uribe’s congressional coalition have been jailed in 2007 for backing right-wing paramilitary groups that were involved in a massacre of civilians in the northern hamlet of La Rochela.

However, the picture does not have to be so bleak. Resolution 1998/26 of the UN Sub-Commission on Protection and Promotion of Human Rights, “housing and property restitution in the context of the return of refugees and internally displaced persons,” reaffirms “the right of all refugees, as defined in the relevant international legal instruments, and IDPs to return to their homes and places of habitual residence in their country and/or place of origin, should they so wish”. Similarly, the UN Economic and Social Council is making a continuous effort to outline principles on housing and property restitution for refugees and displaced persons (E/CN.4/Sub.2/2005/17). In the “Foreword” to the Handbook on Housing and Property Restitution for Refugees and Displaced Persons: Implementing the ‘Pinheiro Principles’, the heads of OCHA/IDD, UN HABITAT, UNHCR, Food and Agriculture Organization (FAO), OHCHR, Norwegian Refugee Council (NRC), and the Internal Displacement Monitoring Center (IDMC) reiterate their commitment to promoting durable solutions for IDPs and refugees, “including their right to return to the homes and properties from which they fled or were forced to leave due to armed conflict and human rights violations” (2007).

Thus, there is a growing realization of and increased attention to the idea that resolving housing and property claims is an important element of post-conflict reconciliation and rehabilitation (Leckie, 2000, 2000a). Setting up institutions such as the Commission on Real Property Claims (CRPC) in Bosnia, the Housing and Property Directorate (HPD) in Kosovo, and the Land Claims Court in South Africa indicate a different, and a better, approach (Leckie, 2000ab; Lewis, 2001). Still, in many cases, thousands of IDPs cannot return to their original homes or get their property rights because of a lack of enforcement mechanisms (Chimni, 2002).

What is meant by return? Coming from a family of Palestinian refugees, Edward Said states in his memoir that “as any displaced and dispossessed person can testify, there is no such thing as a genuine, uncomplicated return to one’s home” (quoted in Oxfeld and Long, 2004). Post-conflict return is complex, and in many cases, extremely politically charged as well. Forced displacement by its nature is entirely politically manipulated (see Stedman and Tanner, 2003). For example in Bosnia and Herzegovina, during the war, dislocating people was a means of creating ethnically homogeneous territories; and in the post-conflict period some local officials resisted the return of displaced people who were from ethnically different backgrounds. Like the international promotion of minority returns in Bosnia and Herzegovina, in many cases return is used as a political tool in efforts to reverse the wrongs of the war, and in this case, to recreate ethnically heterogeneous communities, that is, to have ethnically mixed spaces (villages, towns, neighbourhoods, cities).

IDP return is an important concern in policy circles, but the academic literature on the concept is rather limited. It extensively borrows from the literature on diasporas that underlies the importance of homeland in relation to return, and the migration literature that depicts different kinds of spatial movements involved in returning. The general presumption about IDP return is that IDPs always want to return to their homes and that return is the best attainable solution. Chimni (2003) criticizes this current focus on return that is “driven by the objective of not promoting the goal of protection but of ensuring early return”. While many IDPs certainly do want to go home, the assumption that the needs and experiences of IDPs are homogeneous is mistaken (Sorensen, 2003). IDPs, like all other social groups, contemplate their alternatives based on their interests, and sentiments, and behave accordingly.

Thus, return is the best possible solution when it is voluntary. Recently, the concept of voluntary return began to entail more than the lack of physical oppression or open threats, but “the consultation/participation of displaced people in the process of making decisions about their return, resettlement and reintegration” (Santini, 2004: 53). There is a growing need to address the issue of agency as well as the problem of the lack of enforcement mechanisms. As stated earlier, an increasing number of peace agreements hold provisions on the return of the displaced, but face implementation problems in the post-conflict phase. At the same time, the basic assumption in almost all of these agreements is that the best solution for the displacement problem is return. Moreover, there has been an increasing emphasis on voluntary return and that IDPs should have the choice.

However, IDPs do not always want to go back home. As Walter Kälin, who took over the UN mandate as Representative of the United Nations Secretary-General on the Human Rights of Internally Displaced Persons in 2004 asserts, “Authorities are sometimes anxious to promote return as a symbol of normalization after the chaos brought on by a disaster. However, they should respect IDPs’ right to choose whether to return to their place of origin or to resettle elsewhere, and in either case are expected to assist them to reintegrate (Guiding Principle 28)” (2005: 23). Sorensen tells us, “While we tend to think of displacement as a temporary deviation from normal life, a disruptive event to be corrected, the possibility also exists that some people see displacement as an opportunity for change. People do not only look back; they also look to the future and try to plan for it” (2001: 8). Thus, IDPs do not always return home, but sometimes resettle based on strategic calculations of interest. Although conventional wisdom about IDPs considers them as “living in limbo” or feeling that they are “neither here nor there,” such stereotypes about IDPs are substantially challenged by the literature (Holtzam and Nezam, 2004).

Finally, there are five hypotheses to be tested in this article: 1) Economic prospects facilitates the rate to return; 2) More vulnerable households are more reluctant to return; 3) The rate of return is lower the higher is the intensity of conflict (i.e., atrocities/crime against civilians are in alarming numbers during war); 4) The rate of return is lower the longer is the wars’ duration (i.e., duration of displacement); 5) The rate of return is higher where there are effective property repatriation mechanisms. It is important to note here that the first four of these hypotheses are derivations from the Deininger et al., study that was analyzing the desire to return based on household surveys conducted in Columbia. In this article, the same independent variables will be utilized to test the rate of return in an aggregated dataset. The next section gives details about the dataset employed to test these hypotheses, the sources of the data, possible problems with the data, and possible solutions to alleviate these problems

Measuring the Variables

This section gives details on the sources of data and case selection, exposes potential problems with the data, and explains how the independent (economic prospects, vulnerability, intensity and duration of conflict, and the presence of effective property rights) and dependent (rate of return) variables were operationalized. The purpose of the section is to present the way by which these abstract concepts were converted into concrete tools for statistical analysis. The quantification of abstract concepts is at best an inexact and unsatisfactory method, and it is important to remember that while some of the indicators named here are close to being exact measures, others are necessarily estimates.

Case Selection

Responding to the growing focus on the problem of internal displacement, in 1998 the UN Inter-Agency Standing Committee endorsed the outsourcing of the development of an IDP database to the Norwegian Refugee Council (NRC) with the objective of creating an authoritative source of information on internal displacement. The online Global IDP Database was launched in 1999. Managed by the NRC’s Geneva-based Internal Displacement Monitoring Center (IDMC), by the end of 2002, the database began to provide comprehensive and regularly updated information and analysis on the entirety of the global, conflict-induced, internally displaced population (Danevad and Zeender, 2003). IDMC is a non-governmental organization whose major donors are the Department for International Development (DFID) of the United Kingdom, the Ministry of Foreign Affairs of Norway, the United Nations Refugee Agency (UNHCR), Australia’s AusAID, and the Swedish International Development Cooperation Agency (SIDA).

The list of cases used in this study was compiled based on the information provided by the IDMC’s global IDP database in 2005. There are 49 countries in the list, including cases in which the problem still exists (as of February 2008). Since the concern of this article is to explain how the problem of internal displacement can be solved, their inclusion requires some explanation. There are three reasons for including these cases. The first is the lack of data in relation to the number of already resolved cases. Recognition of internal displacement as an international problem is a recent phenomenon and the data on the subject are rather new. The second is to avoid systematic bias in favor of cases in which the problem had been relatively easy to resolve. And the third reason is the author’s belief that internal displacement is better understood as a process rather than a static occurrence, and continuing cases of internal displacement offer a variety of necessary information.

Data Collection

Comprehensive data on most of the variables included in this study have not previously been collected. No existing datasets include information on rates of IDP return, conflict duration, property rights mechanisms, GDP per capita, and battle related deaths. This information had to be collected by researching every country on the list using primary and secondary sources. The result is the IDP Dataset. Variables were coded using multiple sources, the most useful of which were the United Nations Development Programme’s (UNDP) Human Development Report 2006, the World Bank, the Uppsala Conflict Database, the United States Institute of Peace’s Peace Agreements Digital Collection, and the IDMC’s Global IDP Database.

Potential Problems

There are three potential problems with the data. The first two are general criticisms of research on forced migration to which research on internal displacement is also susceptible, while the third is specific to internal displacement.

The first (and fundamental) potential problem is related to the unit of research. There are so many categories of forced migration that even a single household contains several or all categories. For example in Sri Lanka, a particular district may include a blend of households of which each might contain a number of categories of forced migration like asylum seekers, returning refugees, IDPs and/or others affected by war (Van Hear, 2000). Although the global IDP database does offer valuable data on particular countries, it is not disaggregated enough to provide information on specific districts or households.

The second potential problem is that of methodology. Some have argued that the existing research on forced migration cannot be trusted. For example, Castles argues that earlier sociological approaches, which were based on the principle of relatively autonomous national societies, were missing the fact that forced migration is a social process in which human agency and transnational social networks play a major part, and therefore generated misinformed policies (2003). Similarly, Jacobson and Landau claim that “much of the current research on forced migration is based on unsound methodology, and that the data and subsequent policy conclusions are often flawed or ethically suspect” (2003: 185). They underline “non-representativeness and bias, issues arising from working in unfamiliar contexts including translation and the use of local researchers, and ethical dilemmas including security and confidentiality issues and whether researchers are doing enough to “do no harm” as the main problems of the current research on forced migration (ibid). These authors criticize modest and small-scale qualitative approaches while advocating a more transnational and interdisciplinary undertaking. However, the data presented by the IDMC are trying to do exactly this by offering a global and interdisciplinary approach.

The third potential problem is specific to internal displacement. There are two main reasons for the general difficulty in tracking the number of internally displaced. To begin with, as Holtzman and Nezam point out, IDP “statistics are often inherently political, with a range of incentives on the part of various actors to inflate or deflate numbers” (2004: 8; Stoddard, 2004; Rasmusson, 2006). Such criticisms are answered by the IDMC database team as follows: “Having gained substantial experience in the politics behind IDP figures in various countries, the database team (of IDMC) is able to critically assess the reliability of the various sources and chose in most cases the most conservative figure among available estimates” (Danevad and Zeender, 2003: 9).

In fact, it is not only IDP statistics that are inherently political or biased. As Ward states, “Many population questions are bound up in national cultures, local laws and customs, religion, and even languages; each of these factors affect the technical measurement of populations… Demographic studies, however neutral they may be in any auditing or accounting sense, are invariably political in the wider meaning of the term” (2004: 192).

Another complication in tracking the number of IDPs is the problem of when a person ceases to be internally displaced. Whether people who were displaced but were capable of carrying on their activities in another place can continue to be called “displaced” is a recurring and problematic question of resource allocation (Deng, 1995; Bailliet, 2003). Accurate numbers are necessary in devising programmes and policies that deal with the needs and vulnerabilities of the IDPs; and it is important to know when internal displacement ends in order to know when to reallocate resources and change focus from displacement to other issues (Mooney, 2003ab). Unfortunately, there are no agreed-upon standards by which to judge, but only ad hoc and/or inconsistent determinations of cessation of IDP status which sometimes cause arbitrary decreases in the number of IDPs, and which poses the risk of premature or wrongful withdrawal of protection (Fernandez and Vidal, 2003; Bonoan, 2003).

In 2007, the Brookings Institution – University of Bern Project on Internal Displacement published a booklet called When Displacement Ends: A Framework for Durable Solutions, which outlines the conditions under which IDPs attain a durable solution and no longer need to be the focus of specific attention (Brookings, 2007). Mainly, the Framework determines the point “when IDPs no longer have needs that differ from the population around them” (ibid: 5). Although the intention of developing a framework to identify these conditions is praiseworthy, there is a problem with comparing the needs of the IDPs with the population at large. For example, applying such a criterion in a poor country like Guatemala (one of the poorest countries in the world) would lead to IDPs being treated as just another group of poor people instead of a special category. At least in terms of restitution of property rights, IDPs should be treated differently from someone arguing with his neighbor over a fence between their farms. Subjecting IDPs, who had to leave their homes because of the threat of violence, to the same property rights mechanisms as non-IDPs is simply unjust. Similarly, applying such an approach in Cyprus would simply cause 210,000 IDPs to lose their IDP status, ignoring the conditions that caused them to become IDPs in the first place. Even after three generations, the property problem in Cyprus is still a potential source of conflict. In cases like Cyprus, regardless of how long the IDP problem is outstanding, no end to the conflict can occur until IDPs’ special property rights are addressed. Thus, cases like Cyprus bring about the need to engage the time issue within a framework of justice.

Measuring the Dependent Variable

Although there has been abundant confirmation of IDP return globally, measuring return has always been difficult (Oxfeld and Long, 2004). As early as 1885, Ravenstein observed in Laws of Migration that the data for estimating counter-stream migrants were scarce (ibid). In order to see whether the existing theories predict the conditions under which the internal displacement problem can be settled either through return or resettlement, I have created a single dependent variable called “Rate of Return” (RATE hereafter). RATE is calculated based on the following formula:

image

For each case in the dataset, the number of IDPs in 2005 is subtracted from the number of IDPs in 2000, giving us the number of IDPs that are no longer in IDP status. In some cases where the number of IDPs increased, the result is negative. The difference is divided by the number of IDPs in 2000 in order to obtain a ratio for each case. While the main source of data for the numbers of IDPs in 2000 and 2005 was the IDMC, in some cases, UNHCR, the United Nations Office for the Coordination of Humanitarian Affairs (UN OCHA), the US Committee for Refugees and Immigrants (USCRI) and government sources were also used in order to ensure their accuracy. In cases where there were large discrepancies between the sources, their average was used.

Measuring the Independent Variables

There are five hypotheses to be tested in this section requiring us to operationalize at least five independent variables of intensity of conflict, duration of conflict, economic prospects, vulnerability, and property rights mechanisms:

  • H1: Economic recovery facilitates the desire to return.
  • H2: More vulnerable households are more reluctant to return.
  • H3: The rate of return is lower the higher is the intensity of conflict, i.e., atrocities/crime against civilians are in alarming numbers during war.
  • H4: The rate of return is lower the longer is the wars’ duration, i.e., duration of displacement.
  • H5: The rate of return is higher where there are effective property repatriation mechanisms.

Inspired by the model developed by the Hartzell et al., study, the intensity-of-conflict variable is measured by the total number of war-related deaths during the course of each conflict (2001). For each case, the number of war related deaths is divided by the duration of the conflict in years and then logged to reduce variance. In the majority of the cases, the Uppsala Conflict Database (UCD), which includes 122 armed conflicts in the period 1989–2006, and which is updated on a yearly basis, is the source of information on both the number of war related deaths and the duration of the conflict. Despite the fact that the UCD is updated every year, in cases that were still too recent to have been included in the database, the IDMC database and case study material were used to identify the number of war related deaths and the dates of conflict.

Gross domestic product (GDP), in purchasing power parity (PPP) terms in US dollars divided by midyear population, is the main economic performance measure in the IDP database. In the majority of the cases, the GDP per capita is obtained from the Human Development Report 2006 whose data are available online at the official website of the UNDP. In other cases, the CIA World Factbook is utilized as the main source of data. In order to reduce variance in the GDP per capita (PPP US$) measures, a GDP index was calculated following the UNDP model of formulation. The UNDP rationale in formulation of a GDP index is that in the UNDP Human Development Index “income serves as a surrogate for all the dimensions of human development not reflected in a long and healthy life and in knowledge. Income is adjusted because achieving a respectable level of human development does not require unlimited income” (UNDP, 2006). Accordingly, a logarithm of income is used where US$ 40,000 is the highest goalpost (maximum income identified by the UNDP needed to achieve a respectable level of human development) corresponding to 1 in the index, and US$ 100 is lowest goalpost (minimum income) corresponding to 0 in the index (ibid). For example, Algeria, with a GDP per capita of US$ 6,603 in 2005, had a GDP index of .70.

image

Due to criticisms received from colleagues in the field who argued that GDP is not a good measure to represent the actual situation of the people but only of the countries, other indicators such as the UNDP human development index, HDI, and the Gini coefficient index were also utilized to measure economic prospects. The UNDP was the main source used to code these variables.

GDP, HDI, and the Gini coefficient index are not specific data of economic prospects available for the IDPs, but measures of economic recovery in general. The shift from economic prospects to economic recovery in general was mandatory. It was a result of lack of specific data on the economic situation of IDPs, and was based on the assumption that the average achievements in a country would definitely reflect also on its IDPs. Measuring economic recovery for IDPs by national figures can be unstable, e.g., in cases where there are ethnic minorities, and the government discriminates against (or favors) them with respect to the living standards of the majority, countrywide figures would not reflect the situation of the IDPs. Thus, the results of the quantitative analysis should be treated accordingly.

As stated earlier, specific data on internal displacement are hard to find. Thus, data on vulnerable internally displaced households defined by Deininger et al. as households that belong to an ethnic minority, with one parent, or with female heads or large dependency ratios that have large numbers of members who are younger than 14 years of age is almost impossible to measure as such specific information on IDPs is not available for each country. The IDMC report of Global Overview of Trends and Developments in 2004 identifies 17 countries (35% of the cases) with a high proportion of households headed by women. In the IDP dataset this information is turned into a nominal measure where “1” represents cases with a large proportion of households headed by women and “0” represents the cases without such information. Realizing that this is not a perfect measure, and inspired by the Deininger et al., study, in this article vulnerability is also operationalized by the measure of the percentage of the population below the age of 15. The data were taken from the UNDP and checked against other sources such as the database of the United Nations Population Fund and the CIA World Factbook. Although this is not a specific measure of the IDP population but the whole country, the assumption here is that the general figure of the population below the age of 15 that represents the dependency ratios in each country could also be demonstrative of the IDP population.

The most difficult determinant to measure was property. The crucial problem here was to assess whether there were effective mechanisms to deal with property issues in any given country. Thus, the question revolves around the problem of implementation. Implementation is the carrying out of an authoritative decision, the “missing link” between policy input and output (Berman, 1978: 158, 160). As Berman defines it, “(i)mplementation analysis is not about whether a policy’s goals are fit and proper, which is a matter of values; nor does it concern itself with how they were chosen, which is a study of policy-making. … Implementation analysis is, in short, the study of why authoritative decisions (policies, plans, laws, and the like) do not lead to expected results” (1978: 160). In a more recent article, underlying the increasing attention paid to policy effectiveness, Barrett identifies four factors that contribute to “implementation failure:”

  • • Lack of clear policy objectives; leaving room for differential interpretation and discretion in action;
  • • Multiplicity of actors and agencies involved in implementation; problems of communication and co-ordination between the “links in the chain;”
  • • Inter- and intra-organizational values and interest differences between actors and agencies; problems of differing perspectives and priorities affecting policy interpretations and motivation for implementation;
  • • Relative autonomies among implementing agencies; limits of administrative control (Barrett, 2004: 252).

For each country in the IDP dataset, the first two factors of Barrett’s list were utilized to assess the presence of effective property rights mechanisms. Accordingly, a two-tier evaluation was conducted: first, documents reflecting policy objectives (in this case mostly peace agreements) were reviewed; and second, a survey of responsible actors and agencies was taken. Consequently, a nominal measure of the presence of effective property rights mechanisms was created. Countries that officially refer to property issues and designate accountable actors and agencies are represented with a “1”, while others are represented with a “0”. Only 29 per cent of the cases in the dataset met the two criteria for effectiveness.

In addition, two more measures of property rights, the Wall Street Journal and the Heritage Foundation’s Property Rights Index (PRI hereafter) and Gwartney and Lawson’s measures of property in the Economic Freedom of the World: 2005 Annual Report (GL hereafter) were incorporated into the IDP dataset as a check on the accuracy of the measurement of the property determinant discussed above. Due to the fact that the PRI and GL data were positively correlated and the correlation was very significant (with a Pearson correlation coefficient of .721 at significance level of .000), the GL data were removed from the dataset in order to avoid a problem of multicollinearity. The main reason why the GL data, and not the PRI, were removed is that the PRI covered more of the country cases in the IDP dataset (covering 37 cases in comparison to the GL’s 34). In three cases (Côte d’Ivoire, DR of Congo, and Myanmar) where GL had data, and PRI did not, the GL measures are integrated into the data.

The PRI ranks each country based on (1) the degree to which private property rights are protected, (2) the degree to which the government enforces laws that protect private property, (3) the possibility that private property will be expropriated, (4) the independence of the judiciary, (5) the existence of corruption within the judiciary, and (6) the ability of individuals and businesses to enforce contracts. The PRI variable is incorporated into the IDP dataset as a ranking of 1 to 10 where the more the legal protection of property, the higher the score; and similarly, the smaller the chances of government expropriation of property, the higher the score.

The PRI is treated as a nominal variable in the IDP dataset for two reasons. First, it is not possible to know if the distance between the ranks is equal or not. In other words, a country with a rank of 3 is not necessarily two ranks higher than a country with a rank of 1. Second, the IDP dataset includes data for only 37 of the 161 countries in the PRI, causing gaps in the rankings of the countries in the IDP dataset, that is, there are no countries in the dataset with rankings of 6 or 7. Therefore, it is more appropriate to treat the PRI variable in the IDP dataset as nominal. Accordingly, as with any nominal variable in regression analysis, the PRI variable is recoded into a set of dummy variables, each of which has two levels: 0 and 1.

Like the measures of economic recovery (GDP, HDI, and the Gini coefficient), and the measure of vulnerability (population below the age of 15), PRI is also not a specific measure of the IDP population but the whole country. The assumption here in including the PRI in the analysis is based on the fact that although they comprise a special category, IDPs are also subject to the same property rights mechanisms with the rest of the population. However, like the national economic figures, the measure is unstable.

In the end, there are nine final measures that can be used to operationalize the five independent variables to be tested in the next section. In order to identify potential multicollinearity problems in the future, the nine variables are correlated (see Table 1). Analyzing Table 1 in detail shows that the Human Development Index and GDP Index are significantly positively correlated, with a Pearson Correlation of .91 and a significance level of .00 (denoted by double stars in the table). At the same time, the Property Rights Index (PRI) is also significantly correlated with the Human Development Index (.46 with significance level of .00) and the GDP Index (.56 with significance level of .00). The percentage of population under 15 is significantly negatively correlated with both the Human Development Index (−.81 with significance level of .00) and the GDP Index (−.79 with significance level of .00). Conflict duration in years is significantly correlated with the Human Development Index (.31 with significance level of .040) and with the Property Rights Index (.36 with significance level of .02). Finally, the variable of female-headed IDP households is significantly correlated with conflict intensity (.30 with significance level of .04).

Table 1. Correlations
  Conflict Duration in YearsConflict IntensityFemale Headed IDP HouseholdsGDP IndexGini IndexHuman Development Index 2004Population Under 15 (%)Property Rights Index 2007
  1. * Correlation is significant at the 0.05 level (2-tailed).

  2. ** Correlation is significant at the 0.01 level (2-tailed).

Conflict IntensityPearson Correlation
Sig. (2-tailed)
−.05
.72
       
Female Headed IDP
 Households
Pearson Correlation
Sig. (2-tailed)
−.10
.51
.30*
.04
      
GDP IndexPearson Correlation
Sig. (2-tailed)
.23
.12
−.10
.52
−.25
.09
     
Gini IndexPearson Correlation
Sig. (2-tailed)
.27
.14
−.09
.62
−.23
.21
−.02
.92
    
Human Development Index
2004
Pearson Correlation
Sig. (2-tailed)
.31*
.04
−.09
.55
−.22
.16
.91**
.00
−.12
.51
   
Population Under 15 (%)Pearson Correlation
Sig. (2-tailed)
.01
.93
.10
.50
.11
.47
−.79**
.00
.33
.07
−.81**
.00
  
Property Rights Index 2007Pearson Correlation
Sig. (2-tailed)
.36*
.02
−.26
.11
−.27
.09
.56**
.00
.00
1.00
.46**
.00
−.30
.06
 
Property Rights MechanismsPearson Correlation
Sig. (2-tailed)
−.02
.91
.21
.16
.10
.50
.03
.86
.10
.58
−.02
.92
−.09
.55
−.07
.68

Quantitative Tests

Multiple regression is the method used to identify the observed variability in the rate of return in different countries and to study the effects that individual independent variables (economic prospects, vulnerability, intensity, time and property rights) have on this phenomenon. Using the Statistical Package for the Social Sciences, SPSS, I tried to establish a model that is both explanatory and statistically significant.

The overall model consisting of the predictors of the PRI, property rights mechanisms, the Human Development Index, the Gini Index, female-headed IDP households, and conflict duration in years is statistically significant (p < .01), and explains 64 percent of rate of return (dependent variable) variance (R square of .64). The regression equation for the model is:

image

where pri is a matrix of indicator variables for each level of PRI; α is the intercept, and γ is a vector of coefficients for pri; βi,i = 1,2,…5, is the coefficients for the predictors of property, hdi, gini, femalhed, and confdura, respectively.

The variables of GDP Index and population under 15 are not included in the final formula due to the problem of multicollinearity that was predicted in the previous section. Since the GDP Index, population under 15, and Human Development Index are all highly correlated, only the Human Development Index is incorporated into the final model. To overcome the criticism from colleagues that was mentioned earlier, I preferred the Human Development Index to the GDP Index because it gave a more articulate picture of the actual situation of the people, and not only of the countries. Similarly, because conflict intensity was correlated with female-headed households, and because it corrupted the overall model, only the latter variable is included in the final model.

When the coefficients of the model are analysed more closely (see Table 2), it can be seen that the sign of both conflict duration in years and the Gini Index variables are negative, implying negative (and negligible) relationships with the dependent variable (with respective t-values of −.87 and −.70). The variables of female-headed IDP households, property rights mechanisms, the Human Development Index, and the PRI all have positive, significant effects on the dependent variable. It is important to be careful about the interpretation of the categorical PRI variable where the coefficient of each level is the difference between the effect of that level and the baseline level, which is pri = = 1. An initial look at the coefficient of the dummy variables shows that those countries that are ranked higher in the PRI have higher rates of IDP return.

Table 2. The Linear Regression Model
VariablesEstimateStd. Errort-valueSig.
  1. ** significant at 1%.

  2. N 32 R Square .64.

(Constant)−43.54**11.72−3.71.00
Conflict Duration in Years−.14.16−.87.39
Female Headed IDP
Households
7.01**3.282.14.04
Gini Index−.13.18−.70.49
Human Development
Index 2004
21.69**9.872.20.04
Property Rights
Mechanisms
7.91**3.282.41.02
pri==244.04**11.024.00.00
pri==329.78**6.114.87.00
pri==429.55**8.163.62.00
pri==532.67**7.304.47.00
pri==734.09**11.812.89.01

Results

The results of the above statistical analyses are interesting and controversial. To begin with, although the measure of the GDP Index was highly correlated with other determinants, and was thus taken out of the model, the economic measure of the Human Development Index did show a statistically significant relationship between economic prospects (measured with national figures of economic recovery due to lack of specific data on economic situation of the IDPs) and the rate to return. Accordingly, the higher the Human Development Index of a country, the higher the rate of return. The other economic measure included in the model (the Gini Index) has a negative effect on the dependent variable, which means that in countries where there is greater income inequality (where the Gini Index is higher) the rate of return is lower. However, the relationship is insignificant at a significance level of .49.

The measure for vulnerability (the percentage of population below age 15) was a problematic indicator as it was highly correlated with the other explanatory variables, but the second measure for vulnerability (female-headed IDP households) has a positive and significant relationship with the RATE variable, which indicates that the hypothesis that the more vulnerable IDP households (that is, those that are female-headed) are more reluctant to return is statistically showed to be false. However, this hypothesis could be better evaluated if there were more specific demographic data on internally displaced households and on the separate figures of resettled and returned IDPs.

Although not included in the final model (see Table 3), the hypothesis that the higher the intensity of conflict, the lower the rate of return could not be quantitatively supported. Statistical analysis shows that there is a positive (a higher intensity of conflict explains a higher rate of return), but insignificant correlation between conflict intensity and rate of return.

Table 3. Results of the Quantitative Analysis
 PropositionDeterminantMeasurementResults
H1Economic recovery facilitates the desire to return.Economic ProspectsHDI IndexProved (GDP Index omitted due to multicollinearity; and Gini Index has an insignificant effect.)
H2More vulnerable households are more reluctant to return.VulnerabilityFemale-Headed IDP HouseholdsDisproved (Population below age 15 omitted due to multicollinearity.)
H3The rate of return is lower the higher is the intensity of conflict, i.e., atrocities/crime against civilians are in alarming numbers during war.IntensityConflict Intensity (based on the number of battle related deaths and conflict duration)Positive (Disproved), but insignificant relationship
H4The rate of return is lower the longer is the wars’ duration, i.e., duration of displacement.TimeConflict duration in yearsNegative (Prove), but insignificant relationship
H5The rate of return is higher where there are effective property rights mechanisms.PropertyProperty Rights Mechanisms (Existence of a policy objective and a designated institution for implementation), Property Rights IndexProved by both measures

The correlation between conflict duration and rate of return is negative, proving the hypothesis that rate of return decreases the longer the war continues, though the correlation is insignificant at a significance level of .39.

The final hypothesis that rate of return is higher where there are effective property rights mechanisms is statistically significant by both measures of property rights mechanisms and the PRI. The relationship between the property rights mechanisms variable and the RATE variable is positive and significant, showing that the existence of a policy objective and a designated institution for implementation of property rights does increase the rate of return.

Along the same lines, the PRI variable shows that the higher a country is ranked in the PRI (that is, the higher the degrees to which private property rights are protected, the government enforces laws that protect private property, and the judiciaries are independent, while the lower the possibilities that private property will be expropriated, that corruption exists within the judiciary, and that individuals and businesses can enforce contracts) the higher the rate of return.

Conclusion

Although the results of the quantitative analysis support the arguments that the rate of return is higher where there are effective property rights mechanisms, and economic opportunities, the results should be treated with caution for many reasons. First of all, the IDP dataset suffers a “many variables/small-N problem”. As in the case of this analysis, the more complex the hypotheses the more variables are involved, and a comparison with a small number of cases (small-N) seems to provide too few degrees of freedom to mount a reliable test. Although the IDP dataset includes the entire universe of the cases of conflict-induced internal displacement, the N is still small. A potential way to overcome the problem and increase the N could be to make a yearly formulation of the RATE variable, meaning instead of just conducting a study between the years 2000 and 2005 over the entire period of 5 years, for each country case, a study of the periods 2000 to 2001, 2001 to 2002, 2002 to 2003, 2003 to 2004, and 2004 to 2005 could be conducted. However, the lack of accurately updated yearly data remains a problem.

Furthermore, the problem of missing data in the independent variables also limits the number of cases involved in the analysis. To illustrate, while there are 48 countries in the IDP dataset, the PRI involved only 40 of them. The same problem is also true for the Gini coefficient variable as the data were available only for 32 countries within the IDP dataset. Even the Human Development Index of the UNDP, a yearly-calculated index of human development by an organization with an extensive web of resources, did not have data on 5 of the countries in the IDP dataset.

The hypothesis that rate of return is higher where there are effective property rights mechanisms quantitatively tests better against existing propositions on IDP return. Although the statistical results presented in this chapter are encouraging for the argument being made, the explanation and confirmation of the theory still require further analysis with qualitative data, paving the way for new research. The actually existing cases of internal displacement are diverse and the inadequacy of the regression analysis can be solved through extensive qualitative evidence, which should be analysed within a comparative approach.

Ancillary