2016-global-report-internal-displacement-IDMC

Honduras SPOT LIGHTUnearthing a hidden displacement crisisLong notorious for some of the highest homi- The aim was to collect information on peoplecide rates in the world,119 Honduras has recently who had changed their place of residence withinexperienced enough of an increase in displace- Honduras between 2004 and 2014 for “specificment caused by criminal violence to bring the reasons relating to violence and general crime”.issue into the political limelight. To begin tackling For the purposes of the survey, the specificits impacts, the government created the Inter- reasons were “forced recruitment, extortion,Agency Commission for the Protection of Persons murder, threats, injury, sexual violence, insecurityDisplaced by Violence (CIPPDV) in late 2013, and in the community [conflict, shootings], kidnap-tasked it with driving “the creation of policies ping, forced disappearance, torture, discrimi-and the adoption of measures to prevent forced nation, arbitrary detention and dispossessiondisplacement caused by violence, as well as to of land and dwellings”. People who reportedcare for, protect and find solutions for displaced having changed their place of residence becausepeople and their families”.120 of robbery or assault were not classified as displaced.122As a first step, the commission launched aresearch project in 2014 to reveal the country’s A total of 2,138 households were surveyed acrossinvisible displacement crisis and determine its 20 municipalities identified as having the highestscope and scale. An inter-agency team was concentrations of displaced people.123 Based oncreated to carry out a study, led by CIPPDV and an extrapolation of the findings it is estimatedsupported by the National Statistics Institute, the that they are home to around 174,000 IDPs,Jesuit Reflection, Investigation and Communica- including children born in displacement.124 Oftion Team, the UN Refugee Agency (UNHCR) and those surveyed, 67.9 per cent said their decisionthe Joint IDP Profiling Service (JIPS).121 to move was influenced only by violence and insecurity, and without consideration of other factors that usually determine migration, such as employment or living conditions.12525% 20% 9,000 8,000 Number of displaced households 7,000 Percentage of total number of displaced households 6,000 5,00020% 4,000 3,00015% 12% 13% 12% 2,000 10% 11% 1,00010% 5% 4% 4% 5% 4% 5%0% 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014Source: Survey and Enumeration of Households Affected by Internal Displacement in 20 Municipalities in Honduras (Nov-Dec 2014) OFF THE GRID: The world’s overlooked IDPs 47

Two girls play in the A deeper examination confirms a correlation ment, in accordance with the Guiding Principlesneighbourhood of between the intensity of criminal violence and on Internal Displacement, with the aim of iden-Comunidad Bordos population movement. As depicted in the graph, tifying the range of people who may receive theLlanos de Sula, San displacement levels remained relatively stable State’s attention”.130Pedro Sula, just a few between 2004 and 2008, but rose noticeablyblocks away from a between 2009 and 2013.126 This trend coincides The UN special rapporteur on the human rightsmurder scene where to some extent with the rise in homicide rates, of IDPs stressed the point further during hispolice found buried an indication of the degree of violence to which official visit to Honduras in November 2015.131bodies. Photo: Oscar people were exposed.127 He welcomed the government’s recognitionLeiva Marinero/ of internal displacement, and highlighted theCatholic Relief Services, The dynamic is contradicted somewhat by a spike need for concerted action to tackle its causesJuly 2014 in displacement in 2014, when homicide rates and protect IDPs’ rights. He urged the govern- fell.128 This anomaly may in part be explained by ment “to strengthen its efforts to stop an internal respondents’ tendency to report more recent displacement epidemic” caused by organised and events to a greater extent than those that gang-related crime and violence. happened long ago.129 Alternatively, the surge in the number of IDPs may reflect a broader reality of people fleeing a general deterioration of their security and daily lives. Be that as it may, the progressive increase in the number of people displaced accentuates the Honduran authorities’ need to create a clear and shared conceptual framework within which to understand why and at what point people flee areas plagued by criminal violence. The need is reiterated in one of the study’s main recom- mendations, “to establish a definition of who may be considered a victim of forced displace-48 G20R1ID6

Findings from interviews conducted across This lack of protection leaves people with littlethe Northern Triangle countries of El Salvador, choice other than to embark on dangerous migra-Guatemala and Honduras indicate that internal tions, risking trafficking and murder, to neigh-displacement has become so prevalent that it bouring countries or onwards to the US.136 Therecould be considered a “household phenom- was a major spike in unaccompanied minorsenon”.132 A 2012 survey asked respondents in entering the US through its southern borders inthe three countries whether they had ever felt 2014, the majority fleeing poor and violent townsthe need to change neighbourhood for fear of in El Salvador, Guatemala, and Honduras. Thereviolence, to which 13.5 per cent answered “yes”. was also a resurgence in the number of children and their families arriving in the US in search ofThe figure does not reflect the incidents of actual safety in the second half of 2015, reflecting thedisplacement, but it does give a general sense ongoing danger in these countries.137of the insecurity people perceive. It translates toaround four million people, roughly the entire The media reported regularly on their plightpopulation of Puerto Rico, living with a need during the first half of the year, and the USor desire to move their families because of the president, Barack Obama, spoke of an urgentthreat they feel from criminal violence. humanitarian situation.138 Some analysts went further, describing the surge as “only the tip ofFalling through the cracks: the iceberg of a deeper new humanitarian crisisdata gaps and their in the region”.139consequences This stresses the urgency of engaging in aThe conceptual and information gaps on displace- more holistic and evidence-based humanitarianment in the region described above reflect a approach to migration through Mexico andreality in which people fleeing the effects of Central America and displacement across thecriminal violence fall through the cracks. National region. This should be based on reliable dataresponses to date have focused almost entirely and clear concepts, using new and broad inter-on combatting criminal behaviour through the pretations of the Guiding Principles and otherjustice and security sectors, leaving families who legal frameworks of what constitutes internalflee the violence with little recourse.133 displacement associated with criminal violence.Scant and anecdotal information reveals signifi-cant vulnerability across the region. The afore-mentioned 2012 study in Mexico confirmedthat IDPs faced three major problems comparedwith the local resident population: less accessto the labour market, education and adequatehousing.134 A more recent profiling of IDPs inHonduras also confirmed that displaced house-holds were in more precarious positions thantheir counterparts in the general population interms of access to housing and social services.135 OFF THE GRID: The world’s overlooked IDPs 49

This woman used to Displaced byhave 200 goats andsheep and ten camels, drought-related disasterswhich were her meansof income. Most of The dynamics and impacts of displacement asso- Complex causality underthe livestock died as a ciated with slow-onset disasters, including those drought conditionsresult of pasture and where drought plays a part, are relatively poorlywater shortages. She understood and reported on. The estimation The scale of displacement associated withnow lives in the Hariso methodology used to generate the global figures drought, as for other types of hazards, is largelydisplacement centre for displacement related to disasters triggered by determined by the underlying vulnerability ofin the Siti region of rapid-onset hazards is not well suited to assess people to shocks and stresses that compel themEthiopia, with only one that associated with drought and its complex, to leave their homes and livelihoods. As such,camel and ten goats multi-causal and often delayed impacts. drought’s role has to be understood in combina-and sheep. She says: tion with other social, demographic, political and“We live by water, our Were estimates to include drought-related crises, economic drivers of displacement and disastercattle live by water. the number of people displaced by the impacts risk.140Without water we are of weather hazards would be even higher. Thisno more.” knowledge gap creates an important blind spot in The combined and complex impacts of thesePhoto: Abiy Getahun/ the recognition of displacement associated with factors take time to manifest themselves,Oxfam, December 2015 slow-onset disasters, and in the development of meaning people may be displaced months after policy and operational responses to address the the onset of drought.141 needs of some of the world’s most vulnerable populations and solutions to their plight. Hazard events such as floods and earthquakes create direct physical threats and immediate impacts that trigger displacement. Drought contributes more indirectly to displacement risk, largely through the erosion of food and liveli- hood security among vulnerable populations to the point where fleeing their homes becomes a survival strategy, often of last resort.50 G20R1ID6

Defining droughtThe nature of drought makes estimating its severity and impacts challenging. It is a relative ratherthan absolute condition that occurs in both high and low rainfall areas, and its characteristics varysignificantly from one region to another. Its point of onset and end are difficult to determine, and itseffects accumulate slowly over long periods of time. They also tend to be more diffuse and spreadover a wider geographical area than those of other hazard types.Common definitions of drought, as described below, put the emphasis on its climatic causes – mete-orological drought – which are directly related to precipitation levels. Like other “natural” hazards,however, it has both natural and social dimensions, and its impacts on individuals, households andcommunities can only be understood in relation to demographic, socio-economic and politicalfactors that increase the exposure and vulnerability of people.Other definitions, including agricultural and hydrological drought, highlight the interaction of naturalconditions with human and social factors such as the management of water supplies and changesin land use and land cover.According to the Intergovernmental Panel on Climate Change, some parts of the world have expe-rienced more intense and longer droughts since the 1950s.142​These are likely to intensify in the21st century in some seasons and areas, with the potential for adverse impacts on many sectors.143Meteorological drought Hydrological droughtA precipitation deficit over a Below-average water levels inpre-determined period of time lakes, reservoirs, rivers, streamsthat varies by location according and groundwater affect non-to user needs or applications. It agricultural activities such asis commonly measured according to a tourism, urban water consumption, hydro-threshold of lower than normal or expected electric power production and ecosystemlevels of rainfall. conservation. As with agricultural drought there is no direct relationship with precipita-Agricultural drought tion levels and there may be a considerable time lag before effects are observed in theInsufficient soil moisture to hydrologic system.144support crops, forage growthand pasture. The infiltration of Environmental droughtprecipitation into the soil is often not direct,and depends on slope, soil type and other A combination of the abovefactors. It can take several weeks or monthsbefore shortfalls begin to produce soil mois-ture deficiencies and lead to stress on crops,pastures and rangeland. OFF THE GRID: The world’s overlooked IDPs 51

The most significant factors that drive displace- Recurrent drought may also contribute to longer- ment are those that leave exposed and vulnerable term processes of environmental degradation communities unable to manage severe or recur- such as increased soil erosion, the deterioration of rent drought impacts. Drought is a particular rangeland, deforestation and biodiversity loss. As concern for communities whose food and liveli- seen among pastoralist communities in the Horn hoods depend on rain-fed agriculture, pasture of Africa, this in turn may ultimately force them and rangeland, and whose basic survival is put to seek alternative livelihoods and places to live.149 under increasing stress when conditions over- whelm their normal coping strategies. In many countries, drought and other natural hazards also become intertwined with the Around 84 per cent of the damage and losses impacts of conflict, driving insecurity that is both drought causes worldwide are to agriculture, a cause and a consequence of displacement. In particularly livestock and crop production. Other Somalia, prolonged drought between 2010 and sectors including health, nutrition and water 2012 on top of political instability, conflict and and sanitation are also affected.145 Agricul- widespread poverty precipitated a complex emer- tural drought reduces crop yields and livestock gency and famine that led to huge displacement headcount. It may lead to a fall in wages and both internally and across the country’s borders. employment among farmers and labourers, while inflating food prices as commodities become The UN system applies the term famine only to scarce in local markets. the worst cases when certain mortality, malnutri- tion and hunger thresholds are exceeded.150 May These pressures reduce households’ purchasing to October 2011 was such a period in Somalia, capacity and access to food, deplete their savings and more than 265,500 people were displaced and may force the sale of vital productive assets. during it. This “distress migration of whole fami- Over time, they reduce the quantity and quality lies” was mostly from the agro-pastoral and of their food consumption, and food insecurity pastoral livelihood areas of the country where and malnutrition increase, particularly among the drought was the predominant driver.151 most vulnerable households.146 Our research in parts of northern Kenya, southern The voluntary to forced Ethiopia and southern Somalia highlights a range migration continuum of human factors that combine with drought to contribute to the displacement of pastoral- The distinction between displacement and volun- ists – not from a sedentary home but from their tary migration is often unclear, particularly when traditional and primary source of livelihood.147 population movements are associated with slow- onset disasters or gradual environmental change. Driving factors include the amount of grazing In practice, displacement sits on a continuum land available, pastoralists’ ability to access it, ranging from predominantly forced to predomi- herd size and composition, livestock marketing nantly voluntary movements, where the former strategies, remittance flows, market prices and emphasises push factors to leave and the latter the scale and type of humanitarian interven- pull factors at the intended destination.152 tions.148 Other underlying factors include high fertility rates and the growth in pastoralist popu- Population movements associated with rain- lations, which increase exposure levels. fall variability and environmental change take different forms across the displacement-migra- Linkages to other hazards tion continuum, and the patterns reflect the diverse coping strategies that households and Drought also contributes to the likelihood of communities employ. In slow-onset crises, popu- other types of environmental hazards occurring. lation movements are more likely to be dispersed, In combination with high temperatures, it can with individuals and households leaving over increase the risk of wildfires and the displacement extended periods of time rather than in large of people whose homes are exposed to them. groups over short timeframes. This is another Drought periods that precede heavy rainfall may important factor that makes the displacement also increase the risk of flooding because desic- they cause more difficult to identify. cated land is less absorbent.52 G20R1ID6

Displacement might be identified as a tipping Making distinctions between IDPs and migrants Wayuu children inpoint where abnormal movement patterns indi- in slowly evolving crises may be both arbitrary Colombia’s desertcate the breakdown of normal coping strategies and impractical in operational terms. That said, region of La Guajiraunder severely stressed conditions. Following recognising people as internally displaced as spend most of theirlow rainfall and failed harvests in Niger in 2010, opposed to voluntary migrants helps to identify day looking formany poor households anticipated that normal them as people in need of particular attention water in dried up,migration strategies to meet seasonal food short- from governments, humanitarians and develop- saline or otherwiseages would not be sufficient and moved their ment organisations, and who should be priori- contaminated wells.households to search for work in urban areas.153 tised for protection and assistance. Most of what they draw up is brownishField research in Bangladesh, Ghana, Guatemala, Using the language of displacement can signal the sludge unfit for humanIndia, Peru, Tanzania, Thailand and Vietnam severity of people’s vulnerability and the urgency consumption.conducted for the Where the Rain Falls project of their needs. In the case of repeated displace- Photo: C. George/ECHO,found that household members highly dependent ment, it may highlight populations in need of solu- December 2015,on rain-fed agriculture and with few local options tions to reduce chronic disaster risk. Identifying https://flic.kr/p/FaA1hzand resources to diversify their livelihoods were people as displaced can also alert authorities andmost directly affected by rainfall variability and humanitarians to the potential existence of equallydrought, and were forced to migrate in search of or even more vulnerable people from the samefood or work to support their families.154 disaster-hit areas who have been unable to leave and are in need of protection.156Such people might be considered as being atthe displacement end of the continuum. Less In the context of slow-onset disasters and gradualclearly displaced, but still trapped in vulnerable environmental change, the evidence points to thesituations, are those with more labour migration usefulness of an integrated framework for analy-options to cope with seasonal hunger, though sis based on the voluntary-to-forced continuumstill without being able to escape cycles of depri- of population movements, within which the iden-vation.155 tification of people as displaced from situations of severe distress or crisis remains important. OFF THE GRID: The world’s overlooked IDPs 53

Improving data collection Our experience shows the importance of moni- toring displacement situations regularly over time, The collection of reliable data on displacement particularly IDPs’ protection risks and evolving and other population movements, and the needs vulnerabilities during prolonged, recurrent or of people in gradually deteriorating conditions, protracted displacement.160 As seen in some east is vital to timely and well targeted operational African countries and Yemen in 2015, unresolved and policy responses. A preliminary review of displacement also makes food insecurity worse reporting on drought-related disasters in 2015 because planting and harvesting are disrupted revealed, however, that little such information is while farmers are absent from their land.161 This being shared. One exception is discussed in the in turn increases the risk of further displacement. spotlight on Ethiopia. Improved monitoring and reporting of displace- ment associated with drought-related crises Data on population movements and IDPs’ needs would also enable better understanding and has many uses. The identification of unusual or preparation for its short and long-term impacts intensified migration patterns can serve as an on food insecurity. indicator within early warning systems of the need for action that may pre-empt or at least Good data is also important for understanding mitigate a crisis. Displacement data is also useful past and future displacement trends and their in formulating social impact indicators within many underlying drivers, and as a basis for early warning and information systems, and in investing effectively in measures to avert disas- guiding the development of plans and policies on ters, mitigate their effects and support sustain- drought prevention and preparedness. able recovery. Our research in the Horn of Africa found that the ability to understand such trends This is recognised in legal frameworks such as the was hampered by the paucity of historical and Kampala Convention, which refers in article 4.2 current data on drought impacts. Even where to the establishment and implementation of early data is collected, as in Somalia and Ethiopia, the warning systems, disaster risk reduction strate- multidimensional nature of people’s displace- gies and disaster preparedness and management ment is rarely captured, which also limits under- measures as ways of preventing and preparing standing of IDPs’ needs and potential solutions. for displacement.157 Governments also need such data to monitor and report on progress against disaster risk reduc- tion and development objectives at the national and international level, including targets under the 2015 Sendai framework and the Sustain- able Development Goals.158 Parallel processes to develop indicator frameworks for both policy agendas were underway at the time of writing, with specific indicators for measuring displace- ment associated with disasters on the table.15954 G20R1ID6

Drought-related disasters in2015/2016Drought risk was amplified in 2015/2016 by the effects of the El Niño weather phenomenon on rain-fall patterns, which brought drier-than-normal conditions to many regions of the world (see El Niñospotlight). Its impacts were felt most strongly in eastern and southern Africa, south and south-eastAsia, areas of the Pacific, Central America and the Caribbean and highland areas of South America.Sub-Saharan Africa is particularly susceptible to drought-related disasters, which affect the foodsecurity, nutrition and health of vulnerable people. More than 60 per cent of the region’s populationlives in rural areas. The agriculture sector employs 60 per cent of the workforce and accounts for 25per cent of GDP, rising to 50 per cent when the agribusiness sector is included.162 It has experienceda high number of increasingly frequent droughts.From May 2015 to early 2016, delayed onset and lack of rainfall in eastern Africa led to drought. Therewas a significant increase in the number of people affected by food insecurity and high malnutritionlevels across Djibouti, Eritrea, Ethiopia, Somalia, South Sudan and other areas where people werealready suffering the cumulative impacts of recurrent poor growing seasons.163As of December 2015, around 18.5 million people were estimated to be food insecure across theregion, a 64 per cent increase on August estimates.164 Displacement in this context, along with watershortages, poor sanitation and hygiene conditions and high malnutrition levels added to the risk ofwater and vector-borne diseases. The resurgence of Rift Valley fever in Kenya, Somalia and Tanzaniawas of particular concern (see Ethiopia spotlight).165El Niño also made pre-existing drought and poor harvests worse in southern Africa, where manyregions recorded the lowest rainfall in at least 35 years between October and December 2015.166Around 28 million people were food insecure by early 2016, according to the Southern AfricanDevelopment Community (SADC).167 The effects of the severe 2015 drought will affect harvestssignificantly and have a devastating impact on food security over the year to come.In south and south-east Asia, El Niño affected agriculture, water resources and food security, causinga weak monsoon season and associated drought in countries such as Indonesia, the Philippines andSri Lanka.Lack of rainfall also had a severe impact on agriculture and food security in the Pacific, particularlyparts of Fiji, Papua New Guinea, the Solomon Islands and Vanuatu. The Marshall Islands became thefirst country to declare a state of emergency in early February 2016 as a result of a severe shortageof fresh water on many atolls, caused by persistent drought.168In parts of Central America, the Caribbean and highland areas of South America, insufficient anderratic rainfall from March 2015 led to drought conditions and deepening food insecurity. In CentralAmerica, El Niño contributed to another year of drought, one of the most severe in the region’shistory.Countries in the northern part of Central America have faced chronic drought and dry spells and cropfailures for three consecutive years, and communities in the region’s “dry corridor” of El Salvador,Guatemala, Honduras and Nicaragua are experiencing one of the worst droughts in decades, withan estimated 3.5 million people food insecure. OFF THE GRID: The world’s overlooked IDPs 55

SPOT EthiopiaLIGHT Extreme conditions, extreme measures Ethiopia suffered one of its worst meteorolog- Climate trends across decades and extreme ical droughts for 50 years in 2015, following the variability in rainfall from season to season play failure of two consecutive rainy seasons.169 More an important role in aggravating the drivers of than 80 per cent of the country’s agricultural disaster and displacement risk. Food insecurity is yield and the employment of 85 per cent of the verging on chronic as farming areas that receive workforce depend on adequate rainfall.170 The sufficient rain have shrunk over the past 20 drought contributed to the lowest soil moisture years.175 Most food is consumed by the families levels in at least 30 years, crop failure, below- who produce it.176 average vegetation cover and severe water short- ages in pastoral and arable farming areas. The livelihoods of around seven million pasto- ralists have been jeopardised by the cumulative Devastated livelihoods and high inflation have impacts of more frequent drought on livestock combined to increase food insecurity and malnu- losses, rising cereal prices and lower returns when trition rates, particularly in central and eastern they sell or trade their animals.177 Natural cycles areas, forcing many people to leave their homes such as El Niño will continue to contribute to in search of food, water or work. extreme precipitation patterns, and most global climate models project an increase in the occur- Recurrent and severe drought has also contributed rence of both drought and floods in Ethiopia over to competition and clashes between communi- the coming decades.178 ties under highly stressed conditions over access to scarce water and pasture.171 Communal land Displacement in 2015 and tenure systems grant pastoralists equal rights early 2016 to exploit resources, but in practice the use of grazing areas is regulated between and within Drought contributed to the internal displacement tribes. When drought pushes a tribe to migrate of more than 280,000 people between August into another’s area, tensions between pastoral- 2015 and February 2016, according to IOM. The ists or between pastoralists and settled farmers figure includes at least 147,996 people displaced can arise.172 by severe food insecurity in the drought-affected and predominantly pastoralist regions of Afar Drought and other climate-related hazards do and Somali. not act alone in driving disaster and displacement risk. Their impacts are determined in large part The government and its humanitarian part- by structural issues that drive vulnerability and ners have also noted displacement caused by exposure such as poverty, demographic trends, communal conflict in these areas, related to the weak institutions and environmental degrada- effects of drought on competition for pasture tion. Despite rapid economic growth, the reduc- and water.179 The overall figure also includes tion of extreme poverty, slowing population around 67,800 people displaced by communal growth and improved social safety nets over the conflict associated with the drought in Oromia past decade, Ethiopia remains one of the poorest and Somali over the same period.180 countries in the world.173 The figures do not, however, capture displace- Its population is still set to double in less than 30 ment associated with drought among all affected years, putting further pressure on livelihoods and populations, such as those in the East and West natural resources through deforestation, over- Hararge districts, because data collection is grazing, soil erosion, desertification and poor limited to specific areas by the resources avail- farm management practices. Development is able.181 Further tracking of household mobility unevenly distributed, leaving vulnerable people strategies, such as men migrating without their and emerging regions disadvantaged and at higher risk of displacement.17456 G20R1ID6

families in search of work, in some cases possibly The most acute effects of the 2015 droughtcrossing borders, would be of great benefit. It continue to be felt, with the potential forwould help to inform immediate and long-term hundreds of thousands more people to becomeprotection and assistance interventions to save displaced in 2016 if early and adequate humani-lives, reduce morbidity, protect and restore tarian assistance is not mobilised.190 Between 50pastoralist and arable livelihoods, and prepare and 90 per cent of crops and livestock have beenfor and reduce the impact of further shocks and lost in some areas.191displacement.182 The government and its humanitarian partnersAs of mid-December 2015, around 72,700 have also highlighted the vital importance ofpeople categorised as “drought displaced” were ensuring access to safe drinking water, withoutstaying in makeshift shelters at 24 sites in the Siti which the potential for large-scale displacementsarea of northern Somali.183 Most of the sites were of whole communities would be high.192 Otherspontaneous collective settlements or centres, likely impacts of displacement on pastoralistsgenerally organised along ethnic or family lines, and agro-pastoralists in remote areas include theand a third were scattered individual shelters. disruption of their children’s education, psycho-Most of the IDPs were from pastoralist commu- social and mental health issues, and less accessnities in Siti who had remained in the area while to health and nutrition services.193moving between districts and villages, often withtheir remaining livestock.184 The 2016 humanitarian funding appeal for $1.4 billion, including food aid for 10.2 millionThe main reason given at all sites for not being people, was only 37 per cent met as of the endable to return home was lack of food, and most of January.194 The government is prioritisingif not all IDPs also said they had lost livestock. vulnerable segments of the population includingReasons not captured are likely to include lack of people displaced by the effects of drought andaccess to water points, grazing land, veterinary woman and child-headed households.195 Betterservices, livestock markets, cash and credit.185 data collection and monitoring of displacementIDPs at all sites bar one had been displaced for and the needs of people affected by droughtthe first time, which further emphasises the would go a long way to ensuring that the govern-severity of the situation in 2015.186 ment and its operational partners and donors have the information to make this a reality.Eighty-five per cent of the IDPs had characteris-tics that added to their assistance and protection Both humanitarian and development organisa-needs. The disaggregated data reveals that 72 tions also need to make concerted efforts toper cent were under the age of 18, including facilitate longer-term recovery and developmentaround 16,000 infants under the age of four. It solutions. The coordinator of the UN’s responsealso showed nearly 200 people suffering from in Ethiopia has said the government’s leadershipchronic disease or serious medical conditions, and integration of the humanitarian responseand nearly 300 with physical or mental disabili- into its national development systems providesties. There were more than 3,200 people aged 60 a good basis for the long-term efforts needed toor over, nearly 2,500 pregnant or breast-feeding strengthen people’s resilience to future shocks.196women and nearly 1,000 households headed byone person, most often a woman.187 Without a sustained focus, however, on improving livelihood security for people alreadyIn the severely drought-affected regions of Afar, displaced and those who may become so, theSomali and Oromia, as of the end of the year risk of the current crisis becoming prolonged andthere were a variety of groups of IDPs displaced repeated is likely to increase.at different times and for various reasons. In theKilibati area of Afar, more than 14,500 peoplefleeing “drought” joined 5,700 people displacedby the effects of a volcanic eruption in Eritrea,of whom 1,800 had been living in displacementsince 2010.188 In Siti, around 71,200 people“displaced by drought” joined more than 5,600people displaced by communal conflict a monthearlier, and another 7,600 displaced by communalconflict up to two and a half years earlier.189 OFF THE GRID: The world’s overlooked IDPs 57

Displaced by development projects Displacement associated with development There is a widespread assumption that those projects is not currently covered in global displaced are immediately resettled. This, displacement data. That said, such projects however, rarely happens and many are left to have historically forced large numbers of people search for improvised options on their own. A off their land “in the public interest” across the 2011 study conducted across ten Indian states world, as states exercise their power to further found that only 17 per cent of people displaced development through compulsory acquisition by development projects had been resettled.203 based on the legal principle of eminent domain. When resettlement is provided, those affected The movement of people whose land is acquired are rarely included in the design, planning and for a development project is forced, because they management of their move. Displacement tends are not given the choice to remain in their home to weaken their tenure security, because inade- areas. Even if their rights are fully respected in the quate or non-existent compensation and income process of acquisition and resettlement, a person mean they are unable to buy or rent housing or removed to make way for a development project land. They face going into debt to make up the qualifies as an IDP.197, 198 difference or accepting sites unsuitable for relo- cation, which makes them vulnerable to further Dispossession and displacement associated with upheaval.204 Indigenous people, women, children development projects is often a slow process that and elderly people are more exposed and endure begins long before people physically move. Some the adverse effects of this type of displacement leave when the project is announced in an attempt disproportionately.205 to mitigate their losses, while living conditions for those who remain deteriorate as investments and the provision of services in the area diminish.199 In some cases, people receive notice of less than a day, making it a brutally sudden process. This in addition to a lack of, or conflicting information, inadequate compensation, asymmetric resettle- ment negotiations and the dismantling of their communities that put those affected under signif- icant psychological stress.200 People displaced by development projects suffer a range of human rights violations. The coerced and involuntary removal of people from their homes is a violation of the right to adequate housing. Those affected also lose access to land and natural resources, which leads to a breach of other human rights including access to food, live- lihoods, education, water and healthcare. Their physical security is at risk if they resist displace- ment, or when force is used during the evic- tion process. Other impacts include increased morbidity, restricted mobility and the loss of social support networks.201 Decades of study have shown that displacement associated with development projects leads to impoverishment and disempowerment.20258 G20R1ID6

development projects thatdisplace peopleThe table below shows the kind of development projects that displace people. Their aims tendto be economic gain, infrastructure renewal or conservation, and they may be led, financed andimplemented by governments, the private sector, development finance institutions or a combina-tion thereof.The developer is responsible for ensuring human rights are respected throughout the project period,and the state must ultimately protect against abuses by the public and private sector, including busi-nesses and their contractors.206 Displacement should be followed by resettlement to a new locationwhere those affected are helped to improve, or at least restore, their lives.Some features of displacement are common across all sectors, but many characteristics are moreparticular. For example, mining and dams usually both displace large numbers of people, but theeffects and economic options for IDPs’ solutions are significantly different. Solutions to displacementmust be tailored to the specifics of each sector.Project type ExamplesWater supply Irrigation dams, reservoirsTransport Roads, railways, canals, airports, portsEnergy Hydropower dams, thermal power plants, exploitation of oil and gasMining Metals, gemstones and non-renewable resourcesEnvironmental protection Nature parks, forest reserves, wildlife sanctuaries, embankment fortificationUrban renewal Public transport, housing projects, parks, markets, new town- ships, city beautification, sewage systemsInfrastructure for social services Hospitals, public health centres, schools, collegesCommercial infrastructure Special economic zones, info-technology parksMega-events Olympic Games, World Cup, Eurovision Song ContestIndustrial construction Steel, cement and aluminium factoriesAgriculture and forestry Biofuels, food production, logging, cattle raisingClimate mitigation Reforestation, carbon sequestration OFF THE GRID: The world’s overlooked IDPs 59

Overlooked and unprotected tion’s housing and land rights network’s eviction impact assessment tool.210 Information on the Over the past three decades there has been number and outcomes of such cases, however, growing acknowledgement that people displaced has not been collated. Remedies and assistance by development projects suffer adverse conse- received is often due to the efforts of local civil quences and require protection. In 1980 the society and the displaced themselves. World Bank adopted its first formal policy on projects that involve involuntary resettlement, Those who resist displacement or call for better and today every major multilateral develop- protection have been threatened, intimidated, ment bank has established minimum guidelines verbally and physically assaulted, sexually to oversee displacement associated with their harassed, arrested and even killed.211 Neither projects. Complaint mechanisms have also been the World Bank nor the International Finance instituted. Corporation have responded meaningfully to such abuses associated with projects they have More than 80 private banks and financial insti- financed.212 Some 758 complaints have been tutions have adopted the Equator principles, a filed against 11 development banks since 1994, framework to manage social and environmental and displacement was an issue in 35 per cent risks associated with development projects. of them.213 However, as finance institutions are Responsible business conduct is framed by the largely immune from prosecution in national Organisation for Economic Cooperation and courts,214 complaints procedures may still improve Development (OECD)’s guidelines for multi- claimants’ situations since they are otherwise national enterprises, the International Labour rarely recognised or provided adequate remedy Organization (ILO)’s tripartite declaration of for the harm done. principles concerning multinational enterprises and social policy, and the UN Global Compact, a International humanitarian organisations have worldwide initiative that aims to help companies considerable experience in responding to internal operate responsibly and support society. displacement, but they are not on the front lines assisting people forced to flee their homes by The Great Lakes Pact and the Kampala Conven- development projects. This may be due to lack tion also provide for the protection of people of awareness, limited resources, restricted access displaced by public and private development and a wish to avoid jeopardising their relations projects.207 Domestic laws and policies on internal with the authorities.215 International development displacement such as those in Peru, Kenya agencies and private companies are reluctant to and Nepal have been adopted with provisions monitor and assist IDPs after their projects are devoted to development projects, as have land completed. acquisition laws in China, India and Mozambique that improve protection for the dispossessed.208 Startling estimates despite Such legal frameworks, however, are often over- incomplete data ridden by superseding legislation or executive powers, undermining the protections they offer, Several types of data source exist on people or they are simply ignored. displaced by development projects. The devel- oper’s and financier’s documents contain reset- At the UN, the 1998 Guiding Principles, the 2007 tlement figures, and in rare cases official gazette Basic Principles on Development-Based Evictions notifications and local land records may reveal and the 2011 Guiding Principles on Business and the amount of private land acquired for a project. Human Rights all aim to protect people displaced Rough estimates can be achieved by multiplying by development projects. The UN Development the area of acquired land with its average popu- Programme has also advised the Indian govern- lation density. Local media articles and people ment on resettlement practices.209 with community knowledge may also provide figures based on their own surveys, monitoring Despite these laws, policies and guidelines, the and memory. Satellite imagery, when available, displaced have few options when it comes to can corroborate other data. holding developers to account for the negative impacts they experience. Some have taken their A global figure calculated by collating these cases to court with the help of human rights sources would still, however, be an under- advocates. One was brought in Kenya with data collected using the Habitat International Coali-60 G20R1ID6

estimate. Project documents of governments, tion reports include the number of people the Residents of thecorporations and multilateral finance institu- project displaced, but such figures are inconsist- Baprolla resettlementtions are not always readily made available, or ently quoted as households, families, people or site in Delhi with staffdo not report figures routinely or with consistent cases, which does not allow credible estimates members from theterminology. Those that are reported may be to be compiled. Housing and Landunderestimates to increase the chances of the Rights Network. Theyproject being approved and funded. The actual The World Bank itself reported in 2012 that most are among 500 familiesnumber of people displaced is rarely reported completion reports did not provide substantive living on the siteonce a project is completed.216 information about resettlement outcomes.219 following their eviction Such information for projects implemented by from slums in the IndianThose who use land with informal tenure or under other multi-lateral financial institutions, private capital in 2015. The sitecustomary law may be excluded, because reset- companies and governments is also not available. is on the edge of thetlement figures in some cases only take in those In the absence of data on the number of people city, far from schools,with individual land titles, and gazette notifica- a project displaced, where they went and their health care facilitiestions and local land records do not reveal the land new situations, it is not possible to evaluate how and job opportunities.acquired from those without such deeds. People project implementors have upheld their human Its isolation andliving beyond the development site but displaced rights obligations or the outstanding protection poor lighting makeby the indirect effects of a project, and those and assistance needs of the displaced.220 it a dangerous placeliving in displacement years after its completion for women and girlsare also unlikely to be fully captured.217 The most frequently cited global estimate for after dark. In theory people displaced by development projects is 15 the residents have aA review of public World Bank documents for 969 million people a year since the mid-2000s.221 This ten-year lease, but itprojects citing possible resettlement between number, provided by Michael Cernea, the lead is unclear what their2004 and 2013 is revealing. Only 43 per cent author of the World Bank’s study of displacement tenure status will beforecast a number of people to be affected.218 associated with projects it supports and a global beyond that.Terms such as “physically displaced”, “economi- expert on displacement and resettlement caused Photo: IDMC, Marchcally displaced”, “resettled” and “affected by development, was born out of a previous esti- 2016people” are used interchangeably and may not mate of 10 million people displaced annually byindicate actual displacement. Some comple- dams, urban and transport development projects OFF THE GRID: The world’s overlooked IDPs 61

published in 1996.222 The estimate was increased In addition to the fact that those displaced get to 15 million to account for mining and other left behind, there is an inflated sense of progress, sectors and the general proliferation of devel- because indicators that track development such opment projects worldwide. It is still, however, as the Sustainable Development Goals and the considered conservative. upcoming New Urban Agenda capture gains but not setbacks. Accumulated figures for people displaced by development projects appear only to be avail- The high cost of poorly handled displacement and able for China and India. In China, the total is resettlement extends well beyond those directly 80 million between 1950 and 2015,223 and in affected. In Bangladesh, Indonesia, Nigeria, the India 65 million between 1947 and 2010.224 Philippines and Sudan, resistance, tensions and These are considered under-estimates because, conflict have erupted as a result of mismanage- among other things the figure for India does not ment, corruption and the unequal distribution cover all states, and the figure for China omits of benefits.230 There is also the risk of communal the impact of extractive industries. Reports of violence in resettlement areas between local the number of people displaced differ widely, communities and those newly displaced, and of documents are not always public and those that human rights abuses as people attempt to claim are published are not always reliable.225 their rights. This may lead to further displace- ment. The two countries are also the only ones for which protracted displacement figures are avail- Large, carbon-intensive energy sector projects able, and only then for certain types of projects. such as oil extraction, coal mining and biofuel Two decades after their resettlement, at least plantations also generate greenhouse gas 46 per cent of the 10 million people resettled emissions that contribute to global warming, to make way for reservoirs in China were still increasing the risk of disasters and future living in “extreme poverty”,226 while in India 75 displacement. per cent of those displaced by dams were still impoverished.227 The planning that goes into development projects provides an opportunity to mitigate In the absence of data for all development dispossession and displacement and prepare sectors, dam and reservoir construction appears for durable solutions from the outset. Though to displace most people worldwide.228 A report few in number, examples of good practice do published in 2000 estimated that such projects exist. Those displaced in Indonesia developed had forced between 40 and 80 million people to aquaculture and fisheries in new reservoirs, in flee their homes since 1950. Anecdotal evidence Senegal they gained access to irrigated land and suggests fewer people are displaced by mining in Norway they received a percentage of revenue than by dam construction and urban renewal.229 from electricity sales. Displacement undermines Projects should be undertaken with political development gains commitment, adequate skills, sufficient finan- cial and institutional resources, a participatory Rather than being priority beneficiaries on approach and respect for human rights. They account of their losses, the displaced usually should have in-built resettlement and rehabili- pay the price for development projects and end tation programmes in line with international up worse off. Displacement deepens inequality, standards, as well as mechanisms for monitoring decimates communities and undermines devel- progress towards durable solutions. This would opment gains by making the very poverty that help to ensure that the displacement they cause such projects purportedly seek to alleviate worse. results in beneficial and sustainable development Impoverishing and disempowering people in the for all.231 name of development also allows human rights abuses to continue unchallenged and demon- strates the failure of states to ensure the rights of IDPs.62 G20R1ID6

Brazil SPOT LIGHTOlympic Games preparations displace thousandsin Rio de JaneiroSports mega-events such as the Olympic Games and individualised procedures.239 Some Vila Autó-commonly displace people, both to make way for dromo residents have received supposed marketvenues, accommodation, tourism-related infra- rate compensation240 as a result of well-organ-structure and transport, and also to improve the ised resistance, 241 while others within and outsideinternational image of the host city by eliminating the favela struggled to secure their promisedunsightly slums from areas exposed to visitors payment.242 In almost all cases, the compensationand television audiences. 232 233 does not cover the cost of an adequate home and the accompanying new expenses, leaving thoseIn Rio de Janeiro, around 6,600 families were affected in debt.evicted or under threat of eviction in 2015 tomake way for the 2016 Olympic Games.234 The Many people under threat of eviction have foughtvast majority of those affected were living in to ensure their rights are respected. Resistancefavelas or informal settlements, and were relo- has led to confrontations with officials, humili-cated from their homes in central areas of the ation and mistreatment, physical injuries duringcity to distant suburbs. Given that 60 per cent of municipal guard assaults and death threats.243the Rio 2016 Olympic Park area will be condo- People who resisted eviction longest came underminium developments sold on the open market most pressure, and some settlements had aafter the Games,235 return is not an option for constant municipal guard presence that residentsthose displaced. deemed oppressive.244 The pressure to get Rio ready for the Olympics did not allow time forThe evictions process began in 2009 when the institutions and procedures to be reformed. Oncity won the bid for the Games and was inter- the contrary, it has enabled abuses to occur.twined with preparations for the 2014 WorldCup. Residents under threat have been unable to As a result, communities have been forced toaccess official information about the urbanisation relocate to low-income housing projects onprojects or the process of their removal. Options the poorer outskirts of the city, where there isoffered by the city have not been publicised and little or no urban infrastructure.245 The commuteresidents were neither consulted on nor partici- to the city centre from some relocation areaspated in discussions on possible alternatives to is more than two hours by public transport,evictions. Together with two Rio universities, demonstrating that rather than benefit fromsome residents of Vila Autódromo, one of the urban improvements, those displaced suffer theirlargest favelas to be demolished, presented an impacts. Despite legislation known as the Leialternative to their eviction to the city authorities, Orgânica, which prohibits moving urban dwellersbut their proposal was rejected.236 more than seven kilometres from their original homes, many housing complexes are around 50Nor have many families received adequate notice kilometres away.246of their eviction. There was a surge in “flash evic-tions” across various favelas in 2015, in which Surveys of the displaced in two relocation areasmunicipal guards arrived to demolish homes or and anecdotal evidence shows a deteriorationbusinesses with no warning to residents and their in their access to livelihoods. Distance is anbelongings still inside.237 Residents who remained obstacle to maintaining their current jobs, andfeared leaving their homes and also saw the value there are no means of subsistence, few employ-of their property and due compensation decrease ment opportunities and little access to marketsas the demolitions progressed. Some were also in the new areas.left without access to water and electricity.238 Given that communities were not resettled asThe amounts paid in compensation have varied a whole, social networks were also broken up.between communities, and between households Some women resettled alone, sometimes withwithin the same communities, as a result of weak children, because their partners did not want to OFF THE GRID: The world’s overlooked IDPs 63

A girl watches as do so. With little or no support, their opportuni- Evictions in Rio go beyond the Olympic Games.the Vila Autódromo ties to work and socialise outside the home are The city has a long history of removing low-neighborhood limited, leading to isolation and mental health income communities from desirable areas.252 Theassociation building issues.247 Schools and health centres have also city government has used its hosting of a series ofis demolished been difficult to access in some cases, either high profile events over the past decade to justifysoon after dawn. because they are remote or because provision is the relocation of the urban poor from prime loca-Photo: Megan tied to place of residence.248 tions for middle and upper class housing.253 TheHealy / CatComm Olympics and others have contributed to prop-/ RioOnWatch, Access to the resettled communities is difficult erty speculation and gentrification, a patternFebruary 2015 because some have been overtaken by organ- seen in many cities that host mega-events.254 ised criminal groups, which tax residents and put families at risk of violence.249 Some have Recurring patterns of human rights abuses linked been forced out of their new homes as a result to such events can be prevented. They should be of intimidation and threats.250 Removed from planned and staged with a more comprehensive communal ties, and given that many moved from and consistent approach to managing social risks areas where such groups were less active, the and adverse human rights impacts.255 Bidding displaced lack the networks and strategies to documentation should set better terms for devel- protect themselves. opment strategies to avoid evictions, and where that is not possible, to minimise them and ensure The urban poor have suffered the most direct they are carried out in line with international impacts of the evictions.251 The majority took standards and respect for human dignity. place in areas with great potential for increases in land value, and as such the process has made economic and social inequalities worse by rein- forcing discrimination. Already living in a precar- ious situation, the displaced have been pushed further into deprivation. With no monitoring of, or response to their needs resulting from their displacement, further impoverishment and marginalisation is likely to result.64 G20R1ID6

ConclusionThis report’s findings illustrate once again that ment, the factors that lead to it becomingthe global phenomenon of internal displacement protracted and obstacles to solutions are oftenshows few if any signs of abating. The responses political in nature.of national governments and the internationalcommunity to date have all but failed to limit its As highlighted above in the spotlight on Nepal,scope, let alone reverse the upward trend. there is an increased risk that displacement asso- ciated with the April and May 2015 earthquakesAnnual figures for new displacement associ- will become prolonged because political insta-ated with conflict and violence have been on an bility and weak governance mean that IDPs haveupward trend since 2003, and as of the end of not been properly protected and assisted.2015 there were more such IDPs than at any otherpoint since IDMC began monitoring in 1998. If In countries such as DRC, endemic violence,that were not enough, disasters displaced more insecurity and poverty caused repeated displace-than twice as many people as conflict during ments in 2015. People become more vulnerablethe year. each time they are displaced, setting the scene for further displacement in the future as the resil-For many people, 2015 will not have been the ience of individuals, households and communitiesfirst time they were forced to flee their homes. is eroded.Others will have been displaced several timesduring the year. Evidence shows that displace- Our analysis of the main drivers of displacementment often persists for years, and sometimes associated with disasters shows that economicdecades. The longer it lasts, the more likely and political factors play a key role here too.IDPs are to fall off the radar of data collectors, People’s exposure and vulnerability is drivenresponders and the media. by urban, demographic and economic growth, and developing countries bear the brunt of theThese latest estimates paint only part of the phenomenon.global picture of displacement. They do notinclude all of the people displaced by violence Inequality in these countries makes displacementperpetrated by gangs and criminal groups, nor do a greater concern for the less well-off and thosethey capture those forced to flee their homes by subject to socio-economic discrimination andprojects undertaken in the name of development, marginalisation. Displacement can be a symptomor by disasters associated with slow processes of of pre-existing patterns of social exclusion,environmental change such as drought, sea level affecting the poor just as much in low incomerise and desertification. countries as in their middle and high income counterparts.Progress has been made in conceptualising anddefining some of these phenomena, which is a Development projects undertaken by govern-vital first step towards global data collection. ments and private companies can be drivers ofOnce we begin to monitor them more systemati- displacement that impoverish and marginalisecally, the numbers will rise and the picture will people. As illustrated by the case of peoplelikely become more complex. It is also true to say forcibly evicted from their homes to make waythat the figures that are published are almost for facilities for the Rio Olympics, the protection,inevitably under-estimates. restoration and improvement of the lives and livelihoods of those obliged to resettle tend toFindings in 2015 corroborate IDMC’s previous be inadequate. Through displacing communities,analyses, which point to a correlation between projects can undermine development goals.displacement, political instability and incomeinequality. The drivers and triggers of displace- Global report on internal displacement 2016 65

Raising awareness of the nature and dynamics IDPs in urban settings as compared with the rest of internal displacement in all its forms is key of the urban poor is limited at best. to helping policy-makers and practitioners collect the right kind of data, and target limited The same holds true for our understanding of resources to where they are most needed. It is the impacts of displacement on other vulnerable particularly important to provide insights into groups across the world, because much of the displacement as a multi-dimensional and cross- data currently available is not disaggregated by cutting issue of direct relevance to other global location, age, sex, ethnicity or religion. challenges, from humanitarian action and peace- building to disaster risk reduction, climate change There is also still only limited understanding adaptation and sustainable development. of the causal relationships and feedback loops between displacement and its drivers, including The World Humanitarian Summit in Istanbul political instability, income inequality, urban provides an important opportunity to transform growth and climate change. Quantifying the policy and operational approaches to displace- economic cost of displacement across different ment and to recognise the phenomenon as the countries and contexts would make a compel- complex political and development challenge ling case to governments and policy-makers for that it is. The pledge to leave no one behind in incorporating responses into their longer-term implementing the 2030 agenda for sustainable development plans. development places new obligations on states to ensure that people affected and displaced Identifying the exact tipping points which compel by conflict and disasters can benefit from, and IDPs to cross international borders in search contribute to sustainable long-term development. of safety and refuge would also provide vital insights into what needs to be done to protect Displacement is more visible than ever as an issue and assist people at their points of departure, that requires more action, from the commit- transit and arrival. ments under the 2015–2030 Sendai framework for disaster risk reduction to the decisions by As the global monitor of internal displacement, parties to the UN framework convention on we intend to expand our provision of knowl- climate change and national adaptation plans. edge relevant to policy-making and operational The risk of displacement created by the exposure planning in an effort to advance current and and vulnerability of increasing numbers of people future global commitments to reduce the risk of in rapidly growing and poorly planned urban displacement, and find lasting solutions for the areas cuts across these agendas, and is one of millions of IDPs worldwide. Our ability to do so the central themes of a new framework that is will depend on the breadth and strength of our expected to emerge from the UN conference on partnerships, and on states’ continued commit- housing and sustainable urban development in ment to support these efforts. October 2016. These policy frameworks provide important entry points for addressing displacement in a more comprehensive and joined-up way. In order to be successful, however, they need to be informed by robust evidence and their implementation meas- ured against accurate and realistic targets and benchmarks. For this to happen, a solid global baseline and frequently updated quantitative and qualitative data are needed to inform the processes every step of the way. This includes building a better knowledge base on IDPs’ movements, the conditions in which they live and the vulnerabilities they may have as a result of their displacement. We know that a large proportion of IDPs live outside of camps, and that they increasingly seek safety in towns and cities, but our knowledge of the needs of66 G20R1ID6

Notes 21. OCHA, Humanitarian Needs Overview 2016 - DRC, November 2015, available https://goo.gl/grtN6WPart 1: On the GRID 22. IDMC, Afghanistan: New and long-term IDPs risk1. IISS, Armed Conflict Survey 2015, available at becoming neglected as conflict intensifies, 16 July https://goo.gl/PHj8Dn (purchase required) 2015, available at http://goo.gl/HRarpn2. OCHA, Humanitarian Needs Overview 2016, No- 23. Ibid vember 2015, p.8, available at http://goo.gl/4pLzx3 24. OCHA, Humanitarian Needs Overview 2016 -3. Ibid Colombia, October 2015, available at http://goo.gl/4. Protection cluster Yemen, Task Force on Population mLkMeK 25. Ibid Movement, 6th Report, 10 December 2015, avail- 26. OCHA, Humanitarian Needs Overview 2016 - CAR, able at http://goo.gl/ygSK0O October 2015, available at http://goo.gl/GbmRAe5. European Commission, ECHO Daily Flash, 9 Novem- (French) ber 2015, available at http://goo.gl/gBe8sG; OCHA, 27. Reuters, Five humanitarian crises largely overlooked Humanitarian Bulletin Yemen, 30 November 2015, in 2015, 16 December 2015, available at http://goo. available at http://goo.gl/EdsRDM; Protection gl/5wgnp0 cluster Yemen, Task Force on Population Movement 28. OCHA, Humanitarian Needs Overview 2016 - South 6th Report, 10 December 2015, available at http:// Sudan, November 2015, available at http://goo.gl/ goo.gl/ygSK0O RwfjpV6. BBC, Yemen crisis: How bad is the humanitarian 29. The displacement data available for the 2008-2015 situation? 15 December 2016, available at http:// period is not sufficient to be able to draw conclu- goo.gl/JDuYOg; World Bank data, 2014, available sions about general historical trends. To do so at http://goo.gl/fOf8sK would require historical data spanning decades. We7. Christopher Ward, The water crisis in Yemen: Man- are continuing to develop a preliminary historically aging extreme water scarcity in the Middle East, IB modelled dataset in order to better understand Tauris, 30 August 2015, p.xxviii both past and future trends and to reduce high8. WEF, Global Gender Gap Report, Yemen Country levels of uncertainty within the model, largely Profile 2015, available at http://goo.gl/0TZqH3 related to the quality and significant reporting bias9. OCHA, Humanitarian Dashboard, 6 November in data sources used to date. Further iterations of a 2014, available at http://goo.gl/97U63n strengthened model, working in collaboration with10. HRW, Yemen: Coalition Blocking Desperately the UN office for disaster risk reduction (UNISDR) Needed Fuel, 10 May 2015, available at https://goo. will be produced in the coming year. Meanwhile, gl/8aIfbt, the Guardian, Saudi-led naval blockade we continue to work with the assumption that dis- leaves 20m Yemenis facing humanitarian disaster, 5 placement risk will increase in future, along with the June 2015, available at http://goo.gl/m7NQLI widely expected rising trend in disaster risk.   How11. Protection cluster Yemen, Task Force on Population much it will increase, however, will be greatly Movement 6th Report, 10 December 2015, p.3, dependent on the extent to which policy and meas- available at http://goo.gl/ygSK0O ures to reduce disaster risk are successfully invested12. Ibid in by governments and international donors.13. HRW, UN: Rights Council Fails Yemeni Civilians, 2 30. Shelter cluster, Nepal factsheet, November 2015, October 2015, available at https://goo.gl/iCnQLb available at http://goo.gl/UgjJgP; Government of14. Protection cluster Yemen, National Protection Nepal, Post Disaster Needs Assessment 2015, p.4, Cluster - Yemen Protection Strategy September available at http://goo.gl/6Vy39f 2015, pp.7,8, available at https://goo.gl/MniOce; 31. IOM, DTM Round 1 Report, Earthquake Response Protection cluster Yemen, Task Force on Population Nepal, 2 to 7 May 2015, available at https://goo.gl/ Movement 6th Report, 10 December 2015, p.5, soCwZA available at http://goo.gl/ygSK0O 32. Ibid; IOM, DTM Round 2 Report, Earthquake Re-15. Protection cluster Yemen, Yemen Protection Strat- sponse Nepal, 15 June 2015, available at http://goo. egy, September 2015, p.7, available at https://goo. gl/vTwDZY gl/MniOce 33. Ibid16. Ibid, pp.7,8 34. Ibid17. ECHO, Yemen - Tropical Cyclone Chapala, 5 No- 35. CARE, Housing, land and property issues in Nepal vember 2015, available at http://goo.gl/u9qk9o and their consequences for the post-earthquake18. Protection cluster Yemen, Task Force on Population reconstruction process, February 2016, available at Movement 6th Report, 10 December 2015, avail- http://goo.gl/wp3SvK able at http://goo.gl/ygSK0O 36. Shelter cluster Nepal, Winterisation update, No-19. OCHA, Humanitarian Needs Overview 2016 - vember 2015, available at https://goo.gl/eKkncM Ukraine, November 2015, available https://goo.gl/ 37. Groupe URD and British Red Cross, Urban Prepar- wPtCXS edness: Lessons from the Kathmandu Valley, 2014,20. OCHA, Humanitarian Needs Overview 2016 - Ni- pp.14-15, available at http://goo.gl/LpWFQn geria, November 2015, p.7, available http://goo.gl/ 38. BBC, Where is Nepal aid money going? 21 May Wn6wKk 2015, available at http://goo.gl/A477KS 39. WEF, Building Resilience in Nepal through Public- Private Partnerships, October 2015, p.8, available at http://goo.gl/JsLUFG 40. OSOCC, Situation Analysis Nepal Earthquake, 15 May 2015, available at http://goo.gl/YIi6bT Global report on internal displacement 2016 67

41. WEF, Building Resilience in Nepal through Public- 54. Centre for the Study of Labour and Mobility, Migra- Private Partnerships, October 2015, p.10, available tion and Resilience: Experiences from Nepal’s 2015 at http://goo.gl/JsLUFG; EERI, Nepal Earthquake Earthquake, June 2015, available at http://goo.gl/ and Aftershocks Building Codes, 27 July 2015, FNJucG available at http://goo.gl/93QdOS 55. AFP, Nepal quake survivors fight freezing tempera- 42. WEF, How can we build peace in post-earthquake tures, 5 February 2016, available at http://goo. Nepal? 16 October 2015, available at https://goo. gl/951h22; WEF, How can we build peace in post- gl/MI7GJn earthquake Nepal? 16 October 2015, available at https://goo.gl/MI7GJn 43. WEF, Building Resilience in Nepal through Public- Private Partnerships, October 2015, available at 56. CARE, Housing, land and property issues in Nepal http://goo.gl/JsLUFG and their consequences for the post-earthquake reconstruction process, February 2016, available at 44. IRIN, Fuel shortage threatens Nepal aid as winter http://goo.gl/wp3SvK comes, 12 October 2015, available at http://goo.gl/ PkmHvM; Reuters, Nepal ethnic minorities call off 57. WEF, Building Resilience in Nepal through Public- blockade over charter, 8 February 2016, available at Private Partnerships, October 2015, p.12, available http://goo.gl/dV9I1l; TIME, Nepal Has Finally Passed at http://goo.gl/JsLUFG; UN News Centre, Nepal’s a New Constitution After Years of Political Turmoil, emergency preparedness saved lives in earthquake 17 September 2015, available at http://goo.gl/ aftermath – UN health agency, 13 May 2015, avail- mc2jtL; HRW, “Like We Are Not Nepali”: Protest able at http://goo.gl/1igT39 and Police Crackdown in the Terai Region of Nepal, October 2015, available at https://goo.gl/BghwLo 58. CARE, Housing, land and property issues in Nepal and their consequences for the post-earthquake 45. Amnesty International, Nepal Earthquake Recovery reconstruction process, February 2016, available at Must Safeguard Human Rights, June 2015, p.9, http://goo.gl/wp3SvK available at https://goo.gl/QUl6Kg; Transparency International, Corruption Perception Index 2015, 59. WMO, Exceptionally strong El Niño has passed its available at https://goo.gl/YmJJYX; WEF, How can peak, but impacts still continue, 18 February 2016, we build peace in post-earthquake Nepal? 16 Octo- available at https://goo.gl/wiAUFL ber 2015, available at https://goo.gl/MI7GJn 60. NOAA, How ENSO leads to a cascade of global 46. Brookings Institution, A Critical Disconnect: The impacts, 19 May 2014, available at https://goo. Role of SAARC in Building the DRM Capacities of gl/0cu1w1 South Asian Countries, May 2015, p.23, available at http://goo.gl/T782z2v 61. NOAA, November El Niño update: It’s a small world, 12 November 2015, available at https://goo. 47. HRW, “Like We Are Not Nepali”: Protest and Police gl/Fgb0MB Crackdown in the Terai Region of Nepal, October 2015, available at https://goo.gl/BghwLo 62. WMO, Exceptionally strong El Niño has passed its peak, but impacts still continue, 18 February 2016, 48. IRIN, Fuel shortage threatens Nepal aid as winter available at https://goo.gl/wiAUFL comes, 12 October 2015, available at http://goo.gl/ PkmHvM 63. Ibid 64. NOAA, Frequently Asked Questions About El Niño 49. The Guardian, Running on empty: Nepal’s fuel smugglers defy border blockades, 3 February 2015, and La Niña, April 2012, available at http://goo.gl/ available at http://goo.gl/E32LI4; TIME, No Respite dAx2eY for Shortage-Hit Nepal as Political Efforts Fall Short; 65. WMO, Exceptionally strong El Niño has passed its 25 January 2016, available at http://goo.gl/NR8lpz; peak, but impacts still continue, 18 February 2016, Reuters, Nepal ethnic minorities call off blockade available at https://goo.gl/wiAUFL over charter, 8 February 2016, available at http:// 66. Ibid goo.gl/dV9I1l 67. South Sudan was included as a part of Sudan until 50. UN, Ongoing border blockade on imports sends food and fuel prices ‘skyrocketing’ in Nepal, 11 it declared independence in 2011. December 2015, available at http://goo.gl/msNFgt 68. DMC, DRC: Multiple crises hamper prospects for 51. Amnesty International, Nepal: Reconstruction must durable solutions, January 2014, available http:// not leave behind those most affected by the earth- goo.gl/I8Bwm1 quake, 24 June 2015, available at https://goo.gl/ 69. IDMC, Leaving no one behind: Internal displace- cKv8ea; NRC and Handicap International, Principles ment and the 2030 agenda for sustainable and pragmatism in crisis: Perspectives from the development, briefing paper, 25 September 2015, field, 2016, forthcoming available at http://goo.gl/FGYOed 70. IDMC, Global Estimates 2015: People displaced by 52. Inter-Agency Common Feedback Project, Commu- disaster, July 2015, pp.92-99, available at http:// nity Perceptions of Aid Distribution Fairness, Sep- goo.gl/gRqfKn tember 2015, available at https://goo.gl/gs54wG; 71. Asahi Shimbun, Housing relocation plans slashed in Amnesty International, Nepal Earthquake Recovery Tohoku disaster areas as costs skyrocket, 16 Febru- Must Safeguard Human Rights, June 2015, p.10, ary 2016, available at http://goo.gl/2yef8N available at https://goo.gl/QUl6Kg; IDSN, Waiting 72. Japan Reconstruction Agency, Current Status of for Justice in Response: Report of Initial Findings Reconstruction and Challenges, March 2016, avail- from Immediate Needs Assessment and Monitoring able at http://goo.gl/g6rUI5 Responses towards Affected Dalit Communities in 73. IDMC, Five years on for Fukushima’s IDPs: Life with Nepal Earthquake 2015, p.5, available at http:// radiological risk and without a community safety goo.gl/Jcm8NR; NRC and Handicap International, net, 11 March 2016, available at http://goo.gl/ Principles and pragmatism in crisis: Perspectives AUX5hW from the field, 2016, forthcoming 53. Amnesty International, Nepal Earthquake Recovery Must Safeguard Human Rights, June 2015, pp.11, 12, available at https://goo.gl/QUl6Kg68 G20R1ID6

74. Fukushima prefecture, Overview of Fukushima O’Brien, Statement to the Security Council on Syria, evacuee survey results, 27 April 2015, available at October 2015, available at https://goo.gl/dNRfvd https://goo.gl/SHKC3l (Japanese) 96. UN, Report of the Secretary-General on the im- plementation of Security Council resolutions 213975. UN, Haiti Humanitarian Response Plan, January- (2015), 2165 (2014), 2191 (2014), and 2258 (2015), December 2016, April 2016. Citing IOM data as of 23 March 2016, p.9, available at http://goo.gl/ February 2016. https://goo.gl/jBLEaA F4ochy76. UN, Haiti Transitional Appeal 2015-2016, 11 March Part 3: off the GRID 2015, pp.15-17, available at http://goo.gl/rJoo8Q 97. IDMC, Philippines IDP Figures Analysis, May 2015,77. Ibid, p.19 available at http://goo.gl/atkBaA; IDMC, Central78. Ibid, p.19 African Republic IDPs face significant challenges as79. UN News Service, Haiti: six years after quake, UN instability and political transition continues, May 2015, available at http://goo.gl/iTg2CN cites progress but warns of ‘deteriorating’ situation for those still in camps, 16 January 2016, available 98. Refugee Survey Quarterly, The New Wave: Forced at http://goo.gl/bl6y4C displacement caused by organized crime in Central America and Mexico, Vol.33, No.3, 10 June 2014,Part 2: inside the GRID pp.34-68, available at http://goo.gl/4CgOgI80. UN General Assembly resolution A/C.3/70/L.51/ 99. Wall Street Journal, Latin America is world’s most Rev.1, November 2015, available at http://goo.gl/ violent region, April 2014, available at http://goo.gl/ f8JmHD GUQ5P981. NRC, Fuelling the Fire: How the UN Security Coun- 100. The Telegraph, El Salvador becomes world’s most cil’s permanent members are undermining their deadly country outside a war zone, January 2016, own commitments on Syria, March 2016, available available at http://goo.gl/1VppLs at http://goo.gl/WRFme7 101. Refugee Survey Quarterly, The New Wave: Forced82. OCHA, Syria: Humanitarian Need Overview 2016, displacement caused by organized crime in Central October 2015, p.4, available at http://goo.gl/iFCeyJ America and Mexico, Vol.33, No.3, 10 June 2014, pp.34-68, available at http://goo.gl/4CgOgI; IDMC,83. Ibid, p.6 New Humanitarian Frontiers, October 2015, p.9,84. Under-Secretary-General for Humanitarian Af- available at http://goo.gl/yxmpTA; ACAPS, Other Situations of Violence in the Northern Triangle of fairs and Emergency Relief Coordinator Stephen Central America, July 2014, p.15, available at http:// O’Brien, Statement to the Security Council on Syria, goo.gl/xGJUDI October 2015, available at https://goo.gl/dNRfvd85. IRIN, No way out: How Syrians are struggling to 102. IDMC, New Humanitarian Frontiers, October 2015, find an exit, 10 March 2015, available at http://goo. p.9, available at http://goo.gl/yxmpTA gl/sqH9eO86. NRC, Fuelling the Fire: How the UN Security 103. Forced Migration Review, Criminal violence, Council’s permanent members are undermining displacement and migration in Mexico and Central their own commitments on Syria, March 2016, p.9, America, February 2014, available at http://goo.gl/ available at http://goo.gl/WRFme7 jc9MK887. OCHA, Syria: Humanitarian Need Overview 2016, October 2015, p.7, available at http://goo.gl/iFCeyJ 104. ACAPS, Other Situations of Violence in the North-88. Under-Secretary-General for Humanitarian Af- ern Triangle of Central America, July 2014, p.15, fairs and Emergency Relief Coordinator Stephen available athttp://goo.gl/xGJUDI O’Brien, Statement to the Security Council on Syria, October 2015, available at https://goo.gl/dNRfvd 105. Forced Migration Review, Criminal violence,89. Ibid displacement and migration in Mexico and Central90. NRC, Fuelling the Fire: How the UN Security America, February 2014, p.2, available at http:// Council’s permanent members are undermining goo.gl/jc9MK8 their own commitments on Syria, March 2016, p.5, available at http://goo.gl/WRFme7 106. IDMC, Forced displacement linked to transnational91. Ibid, p.6 organised crime in Mexico, May 2012, p.6, available92. Foreign Affairs, Aiding Disaster: How the United at http://goo.gl/0TdreE Nation’s OCHA Helped Assad and Hurt Syrians in Need, 1 February 2016, available at https://goo.gl/ 107. NRC, Displacement in the wake of violence in Latin YWDgLn America, 2014, p.11, available at http://goo.gl/93. IDMC, Global Overview 2015: People internally dis- C8A4RV; IDMC, New Humanitarian Frontiers, Oc- placed by conflict an violence, May 2015, available tober 2015, p.6,available at http://goo.gl/yxmpTA at http://goo.gl/fotxpj94. WFP, WFP Forced To Make Deeper Cuts In Food 108. UNHCR, Protections and Solutions Strategy for the Assistance For Syrian Refugees Due To Lack Of Northern Triangle of Central America, December Funding, 1 July 2015, available at https://goo.gl/ 2015, p.6, available at http://goo.gl/k60KqK wSlsQA95. Under-Secretary-General for Humanitarian Af- 109. IDMC, New Humanitarian Frontiers, October 2015, fairs and Emergency Relief Coordinator Stephen p.18,available at http://goo.gl/yxmpTA 110. CIPPDV, Characterization of internal displacement in Honduras, November 2015, p. 22, available at http://goo.gl/ W T YCYJ 111. ACAPS, Other Situations of Violence in the North- ern Triangle of Central America, July 2014, p.15, available at http://goo.gl/xGJUDI Global report on internal displacement 2016 69

112. Ibid 139. NRC, Displacement in the wake of violence in Latin 113. IDMC, Forced displacement linked to transnational America, 2014, p.10, available at http://goo.gl/ C8A4RV organised crime in Mexico, May 2012, p.12, avail- able at http://goo.gl/0TdreE 140. UK Government Office for Science, Migration and 114. Forced Migration Review, Criminal Violence and Global Environmental Change (2011), final project displacement in Mexico, February 2014, p.2, avail- report, pp.44-45, available https://goo.gl/0HafwD able at http://goo.gl/XErLQt 115. ACAPS, Other Situations of Violence in the North- 141. IDMC, Assessing drought displacement risk for ern Triangle of Central America, July 2014, p.15, Kenyan, Ethiopian and Somali pastoralists, April available athttp://goo.gl/xGJUDI 2014, available at http://goo.gl/nmSzK0 116. Refugee Survey Quarterly, The New Wave: Forced displacement caused by organized crime in Central 142. IPCC, 5th Assessment Report (AR5), 2014, working America and Mexico, Vol.33, No.3, 10 June 2014, group I, section 2.6.2.3, available at http://goo.gl/ p.36, available at http://goo.gl/4CgOgI P8ctCP 117. Ibid, p.37 118. The Telegraph, El Salvador becomes world’s most 143. IPCC, Managing the Risks of Extreme Events and deadly country outside a war zone, January 2016, Disasters to Advance Climate Change Adaptation, available at http://goo.gl/KrXTnA 2012, available at https://goo.gl/5rT2uC 119. UNODC, Global study on homicide 2013, available at https://goo.gl/4ZPnJq 144. Definitions draw on IPCC, 4th Assessment Report 120. Official Journal of the Republic of Honduras, Execu- (AR4), 2007, working group II, section 3.4.3, avail- tive Decree No. PCM-053-2013, November 2013, able at http://goo.gl/xilwWA; UNISDR, Drought risk available at http://goo.gl/qo0dWA (Spanish) reduction framework and practices: Contributing 121. CIPPDV, Characterization of internal displacement to the implementation of the Hyogo Framework in Honduras, November 2015, p.10, available at for Action, 2009, available at http://goo.gl/t3f6gh; http://goo.gl/ T XqrBm Wilhite D A and Buchanan-Smith M, Drought and 122. Ibid, p.11 water crises: Science, technology, and manage- 123. Ibid ment issues, 2005, chapter 1, available at http:// 124. Ibid, p.12 goo.gl/f0Q86q 125. Ibid, p.14 126. Ibid, p.13 145. FAO, The impact of disasters on agriculture and 127. Ibid, p.44 food security, 2015, p.xviii, available at http://goo. 128. Ibid gl/Ht8iju 129. Ibid, p.13 130. Ibid, p.16 146. Ibid 131. OHCHR, Honduras: Concrete action crucial to stop 147. IDMC, Assessing drought displacement risk for epidemic of violence-induced internal displacement – UN expert, November 2015, available at http:// Kenyan, Ethiopian and Somali pastoralists, April goo.gl/ahCyBs 2014, available at http://goo.gl/oYDDoW 132. ACAPS, Other Situations of Violence in the North- 148. Ibid ern Triangle of Central America, July 2014, p.16, 149. Ibid available at http://goo.gl/xGJUDI 150. FEWS NET, Integrated Phase Classification Version 133. Forced Migration Review, Criminal Violence and 2.0, available at http://goo.gl/Sm1cbB displacement in Mexico, February 2014, p.2, avail- 151. FEWS NET, Internal and External Displacement able at http://goo.gl/XErLQt among Populations of Southern and Central Soma- 134. IDMC, Forced displacement linked to transnational lia Affected by Severe Food Insecurity and Famine organised crime in Mexico, May 2012, p.7, available during 2010-2012, February 2014, available at at http://goo.gl/0TdreE http://goo.gl/vs9wJ4 135. CIPPDV, Characterization of internal displacement 152. Nansen Initiative, Agenda for the Protection of in Honduras, November 2015, p.14, available at Cross-border Displaced Persons in the Context of http://goo.gl/ T XqrBm Disasters and Climate Change, October 2015, avail- 136. PSC, International Verification Mission on the able at https://goo.gl/RssTH3 Human Rights Situation of the Honduran Migrant 153. USAID, Niger security alert: Production deficits Population and their Right to International Protec- provoke population displacement, January 2010, tion, September 2015, available at http://goo.gl/ available at http://goo.gl/3asuad qq28bP 154. UNU/CARE, Where the rain falls: Climate change, 137. Migration Policy Institute, Trends in unaccompanied food and livelihood security, and migration, No- child and family migration from Central America, vember 2012, available at http://goo.gl/5pldnT January 2016, available at http://goo.gl/4CUYXc 155. Ibid 138. New York Times, Children at the border, October 156. UK Government Office for Science, Migration and 2014, available at http://goo.gl/czQ2aO Global Environmental Change (2011), final project report, pp.108-109, available at https://goo.gl/ bJSZ5z 157. AU, Kampala Convention, October 2009, available at http://goo.gl/4Fmnjq 158. UN, Sendai Framework for Disaster Risk Reduc- tion 2015-2030, available at http://goo.gl/22Zq0H; UN, Transforming our world: the 2030 Agenda for Sustainable Development, October 2015, available at http://goo.gl/CqEPS3 159. See work of the Open-ended Intergovernmental Expert Working Group (OEIWG) on Indicators and Terminology Relating to Disaster Risk Reduction,70 G20R1ID6

available at http://goo.gl/1OicoI; and the Inter- Document, January 2016, available at http://goo. agency and Expert Group on Sustainable Develop- gl/3yvz9E ment Indicators, available at http://goo.gl/AAzg38 181. IDMC correspondence with NRC Ethiopia country160. IDMC, Global Estimates 2015, People displaced office, March 2016 by disaster, available at http://goo.gl/fG8XGJ; and 182. Government of Ethiopia and UN, 2016 Ethiopia Hu- Global Overview 2015, People displacement by manitarian Requirements Document, January 2016, conflict and violence, available at http://goo.gl/ p.18, available at http://goo.gl/3yvz9E zbJVbh 183. IOM Ethiopia PEDTM dataset, 85 sites, as received161. FEWS NET, East Africa food security outlook up- from IOM on 22 February 2016 date: Conflict, insecurity, and displacement driving 184. Ibid, as reported for 18 out of 24 sites. food insecurity, June 2015, available at http://goo. 185. IDMC, Assessing drought displacement risk for gl/1YAIH8 Kenyan, Ethiopian and Somali pastoralists, April162. Ibid 2014, available at http://goo.gl/nmSzK0163. OCHA El Niño: Overview of Impact, Projected 186. IOM Ethiopia PEDTM Dataset, 85 sites, as received Humanitarian Needs and Response, January 2016, from IOM on 22 February 2016 available at http://goo.gl/CWfO7i 187. Ibid164. Ibid 188. IOM, Internal Displacement Monitoring Report,165. Ibid October-December 2015, annex I, available at166. OCHA, El Niño: Overview of Impact, Projected http://goo.gl/cpBIi0 Humanitarian Needs and Response, January 2016, 189. Ibid available at http://goo.gl/CWfO7i 190. Ibid167. Ibid 191. FAO, Escalation of food insecurity and malnutrition168. Radio New Zealand, Marshall Islands state of emer- in Ethiopia, video reference F0139, 2016, available gency called, February 2016, available at http:// at http://goo.gl/xQiK01 goo.gl/pNxinQ 192. Government of Ethiopia and UN, 2016 Ethiopia169. FEWS NET, Illustrating the extent and severity of Humanitarian Requirements Document, January the 2015 drought, 17 December 2015, available at 2016, available at http://goo.gl/3yvz9E http://goo.gl/dTE48Z 193. Government of Ethiopia and humanitarian part-170. Joint Government and humanitarian partners docu- ners, Humanitarian Requirements 2015, section ment, Ethiopia 2016 Humanitarian Requirements 3.3.4.3, p.13 and p.30, available at http://goo.gl/ Document, January 2016, available at http://goo. qhQV4e; Government of Ethiopia and UN, 2016 gl/3yvz9E Ethiopia Humanitarian Requirements Docu-171. Ibid, pp.35-36 ment, January 2016, p.29, available at http://goo.172. FAO, Land tenure and rural development, 2002, gl/3yvz9E p.7, available at http://goo.gl/O2uchd; Pastoral- 194. Government of Ethiopia and UN, 2016 Ethiopia Hu- ism: Research, Policy and Practice, Effects of cattle manitarian Requirements Document, January 2016, rustling and household characteristics on migra- p.29, available at http://goo.gl/3yvz9E; OCHA, tion decisions and herd size amongst pastoralists Weekly Humanitarian Bulletin: Ethiopia, 25 January in Baringo District, Kenya, 2011, available at http:// 2016, available at http://goo.gl/TlCIpw goo.gl/na1DJp 195. Government of Ethiopia and UN, 2016 Ethiopia173. World Bank, Ethiopia Overview, September 2015, Humanitarian Requirements Document, January available at http://goo.gl/sQRnkZ; UNDP National 2016, available at http://goo.gl/3yvz9E Human Development Report 2014, available at 196. Government of Ethiopia and OCHA, Ethiopia’s http://goo.gl/edSEz6; UNDESA data produced by drought: Early action in motion, undated video, Worldometers, available at http://goo.gl/jtWAUg available at https://goo.gl/YCSN8P174. UNDP National Human Development Report 2014, 197. Brookings-SAIS Project on Internal Displacement, available at http://goo.gl/edSEz6 Summary Report of the Conference on Develop-175. FEWS NET, A Climate Trend Analysis of Ethiopia, ment-Induced Displacement, December 2002, p.6, April 2012, available at http://goo.gl/ILAepV available at http://goo.gl/WWeNPC176. Ibid, p.2 198. IFRC, World Disaster Report 2012, Development177. IDMC, Assessing drought displacement risk for and displacement: hidden loser from a forgotten Kenyan, Ethiopian and Somali pastoralists, April agenda, October 2012, available at http://goo.gl/ 2014, available at http://goo.gl/nmSzK0 q7a0ne; World Commission on Dams, Dams and178. Bread for All, Climate Risks and Development Development: A New Framework for Decision- Projects: Assessment Report for a Community- Making, November 2000, p.132, available at http:// Level Project in Guduru, Oromiya, Ethiopia, 2009, goo.gl/j0hoc9 available at https://goo.gl/6mxJN9 199. Ibid, World Commission on Dams, p.99179. Government of Ethiopia and UN, 2016 Ethiopia 200. International Accountability Project, Back to Humanitarian Requirements Document, January Development: A Call for What Development Could 2016, available at http://goo.gl/3yvz9E Be, March 2015, p.40, available at https://goo.gl/180. IOM, Internal Displacement Monitoring Report, paOCMA; Patel S, Sliuzas R and Mathur N, The Risk October-December 2015, annex I, available at of Impoverishment in Urban Development-Induced http://goo.gl/cpBIi0; our use of the term “commu- Displacement and Resettlement in Ahmedabad, nal conflict” here is in keeping with the language Environment and Urbanization, Vol.27 No.1, April of the 2016 Ethiopia Humanitarian Requirements 2015, available at http://goo.gl/QIdqbv (registration required) Global report on internal displacement 2016 71

201. Kothari, M, The Global Crisis of Evictions and Dis- 212. HRW, At Your Own Risk: Reprisals against Critics placement: A Housing and Land Rights Response, of World Bank Group Projects, June 2015, p.3-5, December 2015, p.8, available at http://goo.gl/ available at https://goo.gl/K4Dd7u uNoGK0 213. Daniel C, Genovese K, VanHuijstee M and Singh S, 202. Cernea M, Reforming the Foundations of Involun- eds, Glass Half Full? The State of Accountability in tary Resettlement: Introduction, in Cernea M and Development Finance, January 2016, p.37, available Mohan Mathur H, eds, Can Compensation Prevent at http://goo.gl/PKqkhy Impoverishment: Reforming Resettlement through Investments and Benefit-Sharing, 2008, p.4; World 214. Ibid, p.14 Bank, Risks and Reconstruction: Experiences of 215. Ibid, IFRC; Smith A O, Applied Anthropology and Resettlers and Refugees, March 2000, available at http://goo.gl/Vww5eF; Forced Migration Review, Development-Induced Displacement and Resettle- Creating Poverty: the Flawed Economic Logic of ment, in Kedia S and VanWilligen J, eds, Applied the World Bank’s Revised Involuntary Resettlement Anthropology: Domains of Application, 2005, Policy, February 2002, available at http://goo.gl/ p.191; Dawson G and Farber S, Forcible Displace- sYMTNK; McDowell C, ed, Understanding Impov- ment Throughout the Ages: Towards an Interna- erishment: The Consequences of Development-in- tional Convention for the Prevention and Punish- duced Displacement, 1996; World Commission on ment of the Crime of Forcible Displacement, 2012, Dams, Dams and Development: A New Framework p.127, available at http://goo.gl/FuvNcj (registration for Decision-Making, November 2000, p.106, required); Terminski, B, Development-Induced Dis- available at http://goo.gl/j0hoc9; IIED, Avoiding placement and Resettlement, 2015, p.50, available New Poverty: Mining-Induced Displacement and at https://goo.gl/WlMk8H Resettlement, No.58, April 2002, p.3, available at 216. McDowell C and Morrell G, Non-Conflict Displace- http://goo.gl/qnZ1tV ment: A Thematic Literature and Organisational Review, August 2007, p.67-68 203. Fernandes, W and Bharali G, Uprooted for Whose 217. World Commission on Dams, Dams and Develop- Benefit? Development-Induced Displacement in ment: A New Framework for Decision-Making, Assam 1947-2000, 2011, p.19 November 2000, p.104-105, available at http://goo. gl/j0hoc9; Forced Migration Online, Types of devel- 204. International Accountability Project, Back to opment projects causing displacement, undated, Development: A Call for What Development Could available at http://goo.gl/49Vg32; Fernandes W and Be, March 2015, p.77, available at https://goo.gl/ Bharali G, Uprooted for Whose Benefit? Develop- paOCMA; ment-Induced Displacement in Assam 1947-2000, 2011, p.7 205. Kothari, M, The Global Crisis of Evictions and Dis- 218. ICIJ, Explore 10 Years of World Bank Resettlement placement: A Housing and Land Rights Response, Data, April 2015, available at http://goo.gl/kVvOv4 December 2015, p.7; Pandey B, Depriving the Un- 219. World Bank, Involuntary Resettlement Portfolio derprivileged for Development,1998, p.185; World Review, May 2012, p.ix, available at http://goo.gl/ Commission on Dams, Dams and Development: n6miT7 A New Framework for Decision-Making, Novem- 220. World Bank, World Bank Acknowledges Short- ber 2000, p.17, 115, 124, available at http://goo. comings in Resettlement Projects, Announces gl/j0hoc9; IIED, Avoiding New Poverty: Mining- Action Plan to Fix Problems, March 2015, available Induced Displacement and Resettlement, No.58, at http://goo.gl/F8H0oY; World Bank, Resettle- April 2002, p.11, available at http://goo.gl/qnZ1tV ment and Safeguards Action Plan Fact Sheet Update, December 2015, available at http://goo. 206. UN Guiding Principles on Business and Human gl/27sxr8;World Bank, Involuntary Resettlement Rights, the state duty to protect human rights, Portfolio Review, May 2012, available at http://goo. 2011, available at http://goo.gl/6YU49t; UN Basic gl/n6miT7; World Bank, Operational Manual: In- Principles and Guidelines on Development-based voluntary Resettlement (OP 4.12), December 2001, Evictions and Displacement, articles 11 and 71-73, available at http://goo.gl/ouQDFv available at http://goo.gl/aP1Sot 221. McDowell C and Morrell G, Non-Conflict Displace- ment: A Thematic Literature and Organisational 207. ICGL, Great Lakes Pact, 2006, article 1(4) and (5), Review, August 2007, p.67; Cernea M, IRR: An available at http://goo.gl/W3FxDY; AU, Kampala Operational Risks Reduction Model for Population Convention, December 2012, article 10, available at Resettlement,2007, p.2, available at http://goo.gl/ http://goo.gl/cf1LxY LZx4fB 222. World Bank, Resettlement and Development: The 208. Government of Mozambique, Regulation for Reset- Bankwide Review of Projects Involving Involuntary tlement Resulting from Economic Activities, August Resettlement 1986-1993, March 1996, p.x, avail- 2012; HRW, Human Rights Watch Recommenda- able at http://goo.gl/Bf81fg tions for Mozambique’s Resettlement Decree, Sep- 223. Cernea, M, Email communication with IDMC, 22 tember 2012, available at https://goo.gl/eHmygE March 2016; Tan, Y, Email communication with IDMC, 12 April 2016 209. UNDP, Volunteering to Help Resettle Families: A 224. Brookings-LSE, Progress in India: New Legislation Young Man Supports Displaced, undated, available to Protect Persons Internally Displaced by Develop- at http://goo.gl/nogyuy ment Projects, October 2013, available at http:// goo.gl/lt4miz 210. IDMC, Home Sweet Home: Housing practices and 225. Wilmsen, B and Webber M, Displacement and Re- tools that support durable solutions for urban IDPs, settlement as a Mode of Capitalist Transformation: March 2015, available at http://goo.gl/MIbmtt Evidence from China, in Price S and Singer J, eds, Global Implications of Development, Disasters and 211. International Accountability Project, Back to Climate Change, 2016, p.59 Development: A Call for What Development Could Be, March 2015, p.62,74, available at https://goo. gl/paOCMA;; World Commission on Dams, Dams and Development: A New Framework for Decision- Making, November 2000, p.103,106, available at http://goo.gl/j0hoc9; Forced Migration Review, Undermining Development: Forced Eviction in Bangladesh, December 2012, available at http:// goo.gl/ WdCbh472 G20R1ID6

226. World Commission on Dams, Contributing Paper to 244. Melissa Fernandez, Unsettling resettlements: com- Thematic Review I.3: Displacement, Resettlement, munity, belonging and livelihood in Rio de Janeiro’s Rehabilitation, Reparation and Development, 1999, Minha Casa Minha Vida in: Geographies of Forced p.5, available at http://goo.gl/6a18UO; Cernea, Eviction: Dispossession, Violence, Insecurity, eds. M, Can Compensation Prevent Impoverishment?, Katherine Brickell, Melissa Fernandez Arrigoitia and 2008, p.18-24 Alexander Vasudevan, Palgrave (forthcoming, 2016); Ibid, Amnesty International; Rio on Watch,227. World Commission on Dams, Dams and Develop- Vila Autódromo Target of the Ravenous Appetite of ment: A New Framework for Decision-Making, Olympic Construction, 25 February 2016, available November 2000, p.108, available at http://goo.gl/ at http://goo.gl/WFgrl2 j0hoc9 245. World Cup and Olympics Popular Committee of228. World Bank, Resettlement and Development: The Rio de Janeiro, Rio 2016 Olympics: The Exclusion Bankwide Review of Projects Involving Involun- Games, November 2015, p.20, available at http:// tary Resettlement 1986-1993, March 1996, p.91, goo.gl/1OJxfI available at http://goo.gl/Bf81fg; Forced Migration Online, Types of development projects causing 246. Melissa Fernandez, ‘Unsettling resettlements: com- displacement, undated, available at http://goo. munity, belonging and livelihood in Rio de Janeiro’s gl/49Vg32 Minha Casa Minha Vida’ in: Geographies of Forced Eviction: Dispossession, Violence, Insecurity, eds.229. Forced Migration Online, Types of development Katherine Brickell, Melissa Fernandez Arrigoitia and projects causing displacement, undated, available Alexander Vasudevan, Palgrave (forthcoming, 2016) at http://goo.gl/49Vg32 247. Ibid, p.13; UN Special Rapporteur on the right to230. World Commission on Dams, Dams and Develop- adequate housing, op. cit; COHRE, “Women and ment: A New Framework for Decision-Making, Human Rights: Forced Evictions Issue Brief,” 2013 November 2000, p.132, available at http://goo.gl/ j0hoc9 248. Ibid, Melissa Fernandez, p.13 249. Institute for Human Rights and Business, Lucy231. Ibid, p.109; World Bank, Resettlement and Devel- opment: The Bankwide Review of Projects Involving Amis, Conference Report: Human rights and mega Involuntary Resettlement 1986-1993, March 1996, sporting events, December 2015 p.75, available at http://goo.gl/Bf81fg 250. Amnesty International, The State of the World’s Human Rights 2015/2016, January 2016, available232. UNHRC, Report of the Special Rapporteur on at https://goo.gl/YShQG7 adequate housing as a component of the right to an 251. World Cup and Olympics Popular Committee of adequate standard of living, and on the right to non- Rio de Janeiro, Rio 2016 Olympics: The Exclusion discrimination in this context, 18 December 2009, Games, November 2015, p.9, available at http:// p.6, para.16, available at http://goo.gl/N4VLBX goo.gl/1OJxfI; UNHRC, Report of the Special Rap- porteur on adequate housing as a component of233. Ibid the right to an adequate standard of living, and on234. World Cup and Olympics Popular Committee of the right to non-discrimination in this context, 18 December 2009, available at http://goo.gl/N4V- Rio de Janeiro, Rio 2016 Olympics: The Exclusion LBX; Institute for Human Rights and Business, Lucy Games, November 2015, p.36, available at http:// Amis, Conference Report: Human rights and mega goo.gl/1OJxfI sporting events, December 2015235. Urban Geography, Gentrifications in pre-Olympic 252. Urban Geography, Gentrifications in pre-Olympic Rio de Janeiro, 1 December 2015, p.17 Rio de Janeiro, 1 December 2015, p.2236. Ibid, p.17 253. World Cup and Olympics Popular Committee of237. Rio on Watch, Tense Week Introduces New Policy Rio de Janeiro, Rio 2016 Olympics: The Exclusion of ‘Lightning Evictions’ Across Rio Favelas, 8 June Games, November 2015, p.19, available at http:// 2015, available at http://goo.gl/eTvUuQ goo.gl/1OJxfI238. Amnesty International, The State of the World’s 254. Urban Geography, Gentrifications in pre-Olympic Human Rights 2015/2016, January 2016, pp.27,93, Rio de Janeiro, 1 December 2015, p.17; Ibid, World available at https://goo.gl/YShQG7; World Cup and Cup and Olympics Popular Committee of Rio de Ja- Olympics Popular Committee of Rio de Janeiro, Rio neiro, p.8; Institute for Human Rights and Business, 2016 Olympics: The Exclusion Games, November Lucy Amis, Conference Report: Human rights and 2015, p.20, available at http://goo.gl/1OJxfI mega sporting events, December 2015, p.4239. Rio on Watch, Vila Autódromo Target of the Raven- 255. ILO/IOE, ITUC and OHCHR, Joint statement on the ous Appetite of Olympic Construction, 25 February occasion of Wilton Park meeting on Human Rights 2016, available at http://goo.gl/WFgrl2 and Mega-Sporting Events, 17 November 2015,240. Rio on Watch, Vila Autódromo Continues to Resist available at http://goo.gl/3mEI0f Olympic Eviction in the Face of Government Pres- sure, 14 January 2015, available at http://goo.gl/ AMI8WH241. Places Journal, The Displacement Decathlon: Olym- pian struggles for affordable housing from Atlanta to Rio de Janeiro, April 2013, available at https:// goo.gl/sL8zc5242. Rio on Watch, Candomblé Practitioner from Vila Autódromo Describes the Terror of the Eviction Process, 17 September 2015, available at http:// goo.gl/3o17VS243. Ibid; Amnesty International, The State of the World’s Human Rights 2015/2016, January 2016, pp.27, 93, available at https://goo.gl/YShQG7 Global report on internal displacement 2016 73

Methodological annex Introduction One of the innovations in our methodology relates to our assessment of confidence in the The figures included in this report are the result primary data and what it means for the estimates of IDMC’s most ambitious effort yet to present concerned. The confidence assessments signal our figures as transparently as possible. We our commitment to transparency while providing have also attempted to apply more methodo- a roadmap for future work to strengthen data logical consistency to our data collection and collection,  something we are committed to analysis and to document this process for our helping our partners achieve over the coming readers. These improvements have helped bring years. our reporting on displacement associated with disasters and that associated with conflict and This annex describes how we produce our violence together in one report. They have also displacement figures by explaining the source enabled us to make more rigorous comparisons data, calculations, definitions and decision rules between different displacement situations and we use in our analysis. Our aim is to provide get more out of our source data. maximum transparency so that readers under- stand the process, can replicate our work inde- The evidence presented here represents a base- pendently and make use of our data in innovative line, and indicates many areas in which we will ways. We will make our data publicly available need to improve our data gathering and analy- on our website for others to use freely. sis in order to paint a comprehensive picture of internal displacement. This section highlights We are also using innovative ways for policy- some of the main challenges we face and illus- makers, researchers, partners, the media and trates the most significant caveats to which we the public to interact with our data via an open call readers’ attention. portal, making it easier to produce customised reports and analysis. Our data on displacement associated with disasters for 2015 covers 601 sudden-onset Given the complexity of displacement, we are natural hazards in 113 countries. We are still in forced to rely on a variety of internal and external the process of developing and extending our sources in compiling our estimates. We have reas- approach to monitoring displacement associated sessed some of the criteria we use to maximise with drought and other slow-onset phenomena, the reliability and accuracy of source data, and which means we do not yet have global figures this report presents our figures in a way that for such disasters (see part 3). clearly indicates how recently it was updated. Our data on displacement associated with We currently use two similar but distinct meth- conflict and violence covers 52 countries and one odologies to produce displacement estimates disputed territory. We have data on several other related to conflict and violence, and disasters. countries, but we chose not to include it in our This annex describes both approaches. global figures for methodological consistency.74 G20R1ID6

To monitor and report on displacement associ- Relating others’ data toated with conflict and violence, we collect data IDMC’s data modelon the countries affected and present nationallyaggregated figures for: In order to obtain a comprehensive and accu- rate picture about the state of displacement at ||New incidents of displacement from 1 January any given point in time, we have generated a to 31 December 2015 unique data model (see figure A.1). One of the challenges we face in producing displacement ||IDPs who returned, integrated locally or figures is how to relate our partners’ primary and settled elsewhere between the same dates, secondary data to it. and when available, for those who crossed an international border and those who were In order to account comprehensively for the born or died in displacement number of people displaced in a given situation, we would have to populate each component of ||The total number of IDPs as of 31 December the model, updating the information as quickly as 2015 the situation evolved. We are currently working with partners such as IOM, OCHA and UNHCRWe use an event-based methodology to esti- to do just that, in an effort to better reflect themate the number of people displaced by disasters dynamics of displacement.during the course of the year, and derive aggre-gated figures for new displacement for each of The purpose of our data model is to betterthe countries affected. capture all incidents of new displacement, or “flows”, during the year as information becomesWe have monitored displacement associated with available, the number of IDPs reported to haveconflict and violence since 1998 and that associ- found durable solutions or to have crossed anated with disasters since 2008. Over time, we international border, the number of children bornhave continuously sought to improve the ways in displacement and the number of IDPs whowe collect and analyse our data. Over the past have died.eight years, we have successfully obtained dataon ever larger numbers of new displacement Figure A.1: IDMC’s displacement data modelevents associated with disasters, accountingfor more small to medium-sized events than in IDPs settledprevious years (see table A.1). Reporting on these elsewhereevents helps paint a more comprehensive picturein terms of the number of people displaced glob- Failed settlement Settlementally. It also provides the empirical evidence base elsewhere elsewhereto understand them and how they differ frommega-events.Table A.1 Categories of events by magnitude Returns ReturneesEvent size Number of people Internal displacement Failed returns / Locally displaced returnee integratedSmall to medium Fewer than 100,000 Children born IDPs (includes secondaryLarge 100,000 to 999,999 to IDPs and tertiary displacement) displacement IDPsVery large One to three millionMega More than three million Local integration Deaths Cross-border Failed local flight integration Cross-border return to displacementAs a result of this year’s methodological improve- People displacedments, including the standardised application of across bordersthe rules and criteria used to analyse displacement (e.g., refugees)associated with conflict, comparisons betweencountries are now more valid than before. METHODOLOGICAL ANNEX 75

The model is an ideal vehicle for compiling quarterly basis, in which case we publish the displacement estimates, but in reality we have sum of the estimates reported. For Afghanistan found it difficult to populate systematically. We we received data from UNHCR and the govern- seldom receive comprehensive data from our ment on newly profiled IDPs by the month of partners for all of its components. This is often their displacement during 2015, which we aggre- because the type of data specified is simply not gated to arrive at an annual estimate (see table collected or, when it is collected, it is not disag- A.2). The number of newly displaced people in gregated. A primary data source may report the December is an under-estimate because of the extent to which the number of IDPs has declined time lag between the displacement event and during the course of the year, but may not specify the IDPs’ being profiled. the reason for the decrease. The remainder of this annex explains how we Table A.2. Monthly data on new displacement in Afghanis- account for the main flows we report, and how tan (Source: UNHCR and the Government of Afghanistan) they influence our estimates. It also explains how we have selected countries and events to include Month New displacement and why we have excluded some countries we reported have reported on in the past. It also outlines how we assess and express our confidence in Jan 2015 30,697 the source data. Feb 2015 12,923 Accounting for displacement associated with conflict and Mar 2015 8,335 violence Apr 2015 54,686 We produce our figures for displacement associ- ated with conflict and violence via country-level, May 2015 11,504 or situational monitoring. That is, we learn of a displacement situation and begin collecting data Jun 2015 25,895 on it over time. Jul 2015 57,014 We have historically published three main figures – the total number of people displaced as of the Aug 2015 30,374 end of the year, the number of people newly displaced during the year and the number of Sep 2015 30,564 people who returned during the year. Where possible, we have also reported on the number Oct 2015 49,902 of IDPs who have settled elsewhere or integrated locally, those who have sought safety by contin- Nov 2015 19,693 uing their flight across an international border and the number of births and deaths in displace- Dec 2015 3,822 ment. We calculate our figures as follows: TOTAL 335,409 New displacement It should be noted that “new displacement” We may calculate the new displacement inflow is something of a misnomer in that data may for a given year, represented by the orange capture the same people being displaced more “internal displacement” arrow in figure A.1, in a than once during the year. Given that we are number of ways. unable to track individual IDPs, it is often not possible to determine the extent to which this is If our partners provide us with data on new the case for the numbers reported. displacement once a year, we simply report the annually aggregated figure. More often, however, The current lack of disaggregated data on IDPs they provide us with such data on a monthly or who fail to achieve durable solutions, and on cross-border returns to displacement, also means that such inflows are taken as incidents of new displacement. Capturing the end of displacement We calculate annual return flow estimates in a similar way to those for new displacement. For Afghanistan, the aggregated return flow for 2015 represents the sum of the reported monthly figures (see table A.3).76 G20R1ID6

Table A.3. Monthly data on returns in Afghanistan (Source: Births and deaths in displacementUNHCR and the Government of Afghanistan) We only account for births and deaths in displace-Month Reported returns ment when our partners provide data, and weJan 2015 None reported managed to obtain it disaggregated by sex andFeb 2015 None reported age for 20 out of 53 countries in 2015. GivenMar 2015 None reported the shortage of disaggregated data and the factApr 2015 None reported that IDPs’ fertility and mortality rates may notMay 2015 None reported correspond with national figures, we do not tryJun 2015 300 to extrapolate births and deaths in displacementJul 2015 30,329 from national demographic data.Aug 2015 2,914Sep 2015 None reported Depending on the scale and duration of displace-Oct 2015 66,323 ment, the lack of primary data on these flows canNov 2015 19,386 represent a potentially significant blind spot. InDec 2015 10,136 protracted crises such as Macedonia’s, reportedTOTAL 99,059 changes in the size of the displaced population may depend more on demographic trends thanThe same procedure applies to reporting data on returns, local integration and settlement else-on local integration and settlement elsewhere, where, given the lack of progress in these areas.when it is available. It is important to note thataccounting for returns, local integration and Total number of IDPsresettlement reduces the number of IDPs wereport, but it does not necessarily mean that The inflows and outflows described above allthey have achieved durable solutions to their influence the total number or “stock” of IDPs atdisplacement. Data to assess the sustainability a given moment in time – 31 December 2015 inof these processes is not available at the global the case of this report. We estimate the numberlevel, nor are there universally accepted indicators of IDPs at the end of the year by triangulatingfor measuring their progress. data reported from one or more sources with a mathematically derived estimate based on the “flow” data available on new displacement, returns, local integration, settlement elsewhere, cross-border flight and births and deaths in displacement.Cross-border flight of IDPs We arrive at the total number of IDPs as of 31 December 2015 by taking the total at the endWhen possible, we deduct the number of IDPs of 2014 and adding or subtracting flow data aswho flee across an international border. In order follows:for us to be able to do this, those collectinginformation about refugees and asylum seekers Total number of IDPsDec 2015 =need to register whether people had already Total number of IDPsDec 2014been displaced prior to fleeing across the + [Births2015 + new displacement2015]border. Failure to do so risks double-counting. – [Returns2015 + settlement elsewhere2015 +The number of refugees and asylum seekers is l+ocdaelaitnhtse2g01r5a]tion2015 + cross-border flight2015currently subtracted from their country of origin’sgeneral population but not its displaced popula- The equation is technically incomplete becausetion. it does not take into account the “counterflows” represented by failed returns, local integrationThe spotlight on Syria in part two of this report and settlement elsewhere, or cross-borderexplains the widespread concern that this issue returns into displacement. Given, however, thathas led to inflated internal displacement figures data is not collected and these phenomenathat combine numbers on IDPs and refugees, are accounted for as new rather than repeatedparticularly in highly dynamic and politically sensi- displacement, the equation serves its purpose.tive crises. METHODOLOGICAL ANNEX 77

In reality, the lack of coverage of the components ||New displacement includes repeated displace- of our data model and the way outflow data is ment. This is the case every year in DRC and aggregated mean the actual equation for most in many other contexts. countries is often simply: ||Double-counting. In Myanmar, a small TotaTlontaulmnbuemrboefrIDofPIsDDePcs2D01e5c = number of IDPs may have been counted more + RNeetwurndsis2p01l5acement2015 than once by two or more sources. – 2014 ||Partners change their data-collection meth- The mathematical formula for estimating the odology, as in DRC, or the scope of their stock of IDPs is at best a modelled approximation. geographical coverage, as in Nigeria. We compare this with the data we obtain from our sources, and when we do they do not always ||We change our primary source because of correspond. There are number of reasons for this: the lack of available data or doubts about their credibility. ||The initial value – the estimate for the end of the previous year – is incorrect and needs ||There is a lack of data on a flow that signifi- to be revised. In Afghanistan, delays in the cantly affects the number of IDPs in a country. profiling of IDPs meant that people displaced Data on the number of refugees and asylum in 2014 were captured well into 2015, which seekers from Syria does not indicate whether meant we had to retroactively revise our they had previously been displaced internally. December 2014 estimates. Reflecting the date of sources When situations remain unchanged from one year to the next, or when flow data is not available, we base our end-of-year estimates on the data provided by our partners. In many countries, however, it has not been updated for several years. In countries with complex or multiple displacement crises, such as Chad, Iraq and Myanmar, data for one crisis may be regu- larly reported, while for others it may be outdated or missing. If there is no credible evidence that IDPs in such situations have returned, integrated locally or settled elsewhere, we have in the past included them in our global figures. In the interests of transparency, this year’s report stratifies the stock of IDPs based on when the primary data was collected (see figure A.2). The length of the bar as a whole represents the total number of IDPs for whom we were able to obtain data. The right-hand section represents data which is increasingly out of date. Figure A.2. Different strata for stocks of IDPs, ordered by the date of the source data New Displacement Displacement displacement prior to 2015 prior to 2015 in 2015 IDPs displaced in 2015 who remained IDPs displaced prior to 2015 about whom IDPs displaced prior to 2015 about whom in displacement at year’s end there was updated data in 2015 the most recent data is from before 2015 Reported return, settle- Reported return, settle- Reported return, settle- ment elsewhere, local ment elsewhere, local ment elsewhere, local integration or cross- integration or cross- integration or cross- border flight of IDPs border flight of IDPs border flight of IDPs who were displaced in who were displaced who were displaced prior to 2015 (based on 2015 prior to 2015 data from 2015) (based on data from 2014 or older)78 G20R1ID6

Accounting for displacement ||Under-reporting of small-scale events: Small-associated with disasters scale displacements are far more common, but less reported on. Disasters that occur inOur estimates for displacement associated with isolated, insecure or marginalised areas alsodisasters are generated by event rather than by tend to be under-reported because accesscountry. We monitor and collect displacement and communications are limited.information from our partners and internationalmedia outlets for all reported disasters. We apply ||“Invisible” IDPs: There tends to be significantlyno threshold when doing so, either in terms of more information available on IDPs who havethe number of people involved or the distance taken refuge at official or collective sites thanthey have travelled. Our database includes on those living with host communities and inrecords of one up to 15 million IDPs. other dispersed settings. Given that the vast majority usually fall into the second category,We generate a single estimate for each event for figures based on data from collective sites arethe total number of people displaced. It is impor- likely to be substantial underestimates.tant to note that our figures do not necessarilycapture the peak number of IDPs, but instead ||Real-time reporting is less reliable, but lateraim to provide the most comprehensive figure for assessments may underestimate: Reportingthose displaced with minimal double-counting. tends to be more frequent but less reliable in the most acute and highly dynamic phases ofIn order to generate our estimates, we collect a disaster, when peak levels of displacementdata from a number of reports on the same are likely to be reached. It becomes moredisaster, each specifying whether its figures refer accurate once there has been time to maketo individuals or households, the reporting terms more considered assessments.and sources used, the publisher, the title of thesource document and the date of publication. ||Estimates based on later evaluations of severely damaged or destroyed housing willThis dataset allows us to better interpret the be more reliable, but they are also likely tocontext of the figure in each report. In deter- understate the peak level of displacement,mining our estimates, it is vital that the data given that they will not include people whoseselected represents the most comprehensive homes did not suffer severe damage but whofigure from the most reliable source available fled for other reasons.for that event. Our estimates for many disasters are calculatedWhen possible we triangulate the figures using by extrapolating from the number of severelycompeting reports. In most cases, however, our damaged or destroyed homes or the numberestimates are derived from a single report. In of families in evacuation centres. In both casesothers, they are the aggregation of a number of we multiply the housing and family data by thereports that together cover the wide geograph- average number of people per household.ical area affected by a disaster.Reporting biasWe are aware that our methodology and datamay be subject to different types of reportingbias, some of which are detailed below: ||Unequal dissemination of data: Global reporting tends to emphasise large events in a small number of countries where interna- tional agencies, funding partners and media have a substantial presence, or where there is a strong national commitment and capacity to manage disaster risk and collect information. METHODOLOGICAL ANNEX 79

Estimating average household size Primary sources often report the number of homes rendered uninhabitable or the number of families displaced, which we convert into a figure for IDPs by multiplying the numbers by the average household size (AHHS). There is, however, no universal dataset with updated and stand- ardised AHHS data for all countries. In its absence, some global disaster datasets have opted to apply an average across all countries or groups of countries. The Emergency Events Database (EM-DAT) uses an average family size of five for developing countries and three for industrialised countries.1 Our 2014 and 2015 Global Estimates reports relied mainly on two international datasets containing household size information, the United Nations Statistics Division (UNSD)’s population data for 2013, with source data from 2000 to 2011; and its population data for 1995, with source data from 1989 to 1993. We also used data retrieved directly from the websites of a few national statistics offices, but for the 121 countries where this was not possible we had to estimate the AHHS by adding a constant to the fertility rate. Given the potentially significant effect of AHHS on our estimates, we have improved our meth- odology for the 2016 GRID in several ways. We searched for more datasets on household size, and found a number compiled by international organisations based on census data: seven UNSD population datasets published between 2009 and 2015; Eurostat data published in 2016; the OECD Family Database published in 2015; and the World Bank’s 2012 world development indicators. We also used two other datasets that rely mostly on national census data, Euromonitor’s World Economic Factbook 2014 and an academic dataset based mostly on official data compiled by Official Statistics of Finland.2 To these we added data from USAID’s Demographic and Health Surveys (DHS), which are available for 75 developing countries and, while not being full national censuses, are based on nationally representative samples – usually between 5,000 and 30,000 households. They are also designed to be comparable internationally.3 Merging these 13 datasets into a relational database allowed us to identify gaps and discrepan- cies, which we addressed by searching the websites of national statistics offices to glean official figures not yet disseminated, and by receiving data from partner organisations such IOM and JIPS operating in the field. Where several sources were available, we analysed the differences between them, which turned out to be minor for most countries. We investigated larger discrepancies in a few developing countries with large AHHSs further. Table A.4 illustrates how we prioritised the various datasets. This approach allowed us to increase the scope of our AHHS dataset from 215 to 251 countries and territories, without having to rely on estimates based on fertility rates. To compensate for the differences in data collection dates, which were more than ten years ago for a few countries, we built a statistical model of the change in household size over time. We calibrated it using two datasets with multiple data points and good international and intertem- poral comparability: the DHS dataset, keeping values that were measured at least 10 years apart, and another OECD dataset.4 The time elapsed between two measurements was found to have a significant influence on AHHS, which generally decreases over time. Demographic and economic indicators such as changes in the fertility rate, population or GDP per capita growth and regional indicators were found to be less significant.80 G20R1ID6

Table A.4. Prioritisation of the various AHHS data sources (the most recent data was selected for each row)Data source PriorityNational statistical office websites and recent field data from partner organisations OneCensus data from intergovernmental organisations and DHS data less than five years old TwoWorld Economic Factbook 2014 ThreeCensus or survey data more than six years old and academic datasets such as that of FourOfficial Statistics of FinlandDepending on the size of the displacement event, even a small change in the household size figurecan make a huge difference to the final estimate. In 2015 we calculated estimates for Pakistan’sKhyber Pakhtunkhwa (KP) province and Federally Administered Tribal Areas (FATA) based on thenumber of families registered as displaced. We then received new information that led us to usean average household size of 6.2 rather than 5.2 people, which in turn produced an estimate forthe number of IDPs nearly 200,000 higher than that we would have published last year basedon the same source data (see table A.5).Table A.5: The impact of household size on displacement estimates for KP province and FATA in PakistanNumber Former average Former IDMC New average Updated Varianceof families household size estimate household IDMC estimate betweenregistered as (people per (rounded to size (people (rounded to the twodisplaced family) nearest 1,000) per family) nearest 1,000) estimates191,018 5.2 993,000 6.2 1,184,000 191,000We also use average household size data widely in compiling our estimates for displacement asso-ciated with disasters. The most striking example of its impact on our estimates in 2015 involvedthe earthquakes in Nepal. Had we used the old 2014 average household size of 5.4, we wouldhave arrived at an estimate of 3,294,000 IDPs. Based on new, more accurate data, our actualestimate is 2,623,000, a difference of 671,000.We also revised some of the estimates we published for disasters in 2014 based on updatedaverage household size information, in order to ensure that our trend analyses are as accurateas possible (see table A.6).Table A.6: Revised estimates for displacement associated with disasters in 2014, based on new average household sizedataEvent Former Former IDMC New Updated Variance average estimate average IDMC estimate betweenEthiopia - Awash household (rounded to household (rounded to the tworiver flood size (people nearest 1,000) size (people nearest 1,000) estimatesNiger - rainy season per family) 63,000 per family) 36,000 – 27,000displacement 7.831 63,000 4.5 47,000 – 16,000South Sudan - se- 46,000 30,000 – 16,000vere floods in War- 7.7512 5.8rup state 7.62 5.0 METHODOLOGICAL ANNEX 81

IDMC’s data collection, As such, we interpret IDPs to include not only citi- analytical process, definitions zens of the country in which displacement takes and decision rules place, but also non-nationals such as migrants and asylums seekers in Libya, and Palestinian Country identification refugees in Syria and Lebanon; refugees who have returned to their home country but have IDMC collects and presents data on IDPs for been unable to go back to their habitual place each country it monitors based on internation- of residence, such as Afghan refugees returning ally recognised borders or, in the case of foreign from Pakistan; and stateless people such as the occupation such as Palestine, on demarcation Rohingya who have been displaced by conflict lines. The 2016 GRID dataset for displacement or violence. associated with conflict also includes the Abyei area, which is disputed between Sudan and Forced displacement should not only be associ- South Sudan and whose final borders are to ated with the notion of a fixed place of residence, be determined in a referendum. We report on but also flight from traditional “living spaces” that displacement in new states created by secession, support people’s livelihoods, such as pastoralists’ such as Kosovo and Timor Leste, when they have grazing areas. Given that the concept of habitual broad international recognition. residence is intimately linked to the issue of liveli- hoods, people who have lost them as a result of People displaced within areas of an internation- their displacement – such as pastoralists in Somalia ally recognised state under foreign occupation and elsewhere in eastern Africa – are considered are considered IDPs, irrespective of their location IDPs. We consider a person to be displaced regard- with respect to the de facto borders or the terri- less of how far or for how long they flee. torial claims of the occupying power, providing the original international borders still have broad In accordance with the Inter-Agency Standing international recognition. Examples are eastern Committee Framework on Durable Solutions, Ukraine, Crimea, South Ossetia and the Turkish displacement is deemed to end when IDPs have Republic of Northern Cyprus. returned home, integrated locally in their place of refuge or settled elsewhere in the country in a For the purpose of this report, countries are sustainable way, and no longer have vulnerabili- defined as independent nation states, including ties linked to their displacement. We acknowl- their overseas territories. Our dataset for displace- edge this concept, but for the purpose of our ment associated with disasters includes some monitoring and reporting, we do not count countries, such as Taiwan, that do not have broad returnees as IDPs, and subtract the figure from international recognition. The inclusion of such our total estimates, whether they are known to countries and other contested territories does not have achieved a durable solution or not. This is imply any political endorsement or otherwise on because it is not possible in the vast majority IDMC’s part. of cases for us to properly gauge the extent to which IDPs have achieved a lasting end to their To make analysis of the dataset easier and more displacement or not. effective, we use UN country terminology and the three-digit ISO country code. For areas such On the other hand, we consider children born in as Abyei, which have no standard ISO code, we displacement to be IDPs, and they are included created one. in our estimates. This is particularly pertinent in countries such as Azerbaijan, where displacement Definition of an IDP has lasted for decades. As such, the number of IDPs in these countries may increase over the We use the definition of an IDP contained in years as a result of demographic trends, despite the 1998 Guiding Principles. The criteria related the fact that the original trigger has long ceased to the “forced” nature of displacement “within to cause any new displacement. internationally recognized borders” is clearly fundamental in determining whether the person For countries that have been divided into two is an IDP, but the Guiding Principles do not set internationally recognised states, such as Sudan other criteria by which to identity a person fleeing and South Sudan, we do not consider people their “home or place of habitual residence”. whose former place of habitual residence is in one of the new entities and refuge in the other as IDPs (see box below). For instance, we do82 G20R1ID6

not consider a person who fled from what was Disaggregated dataformally southern Sudan to northern Sudan anIDP following the creation of South Sudan, but We systematically seek to obtain not only disag-people displaced within either Sudan or South gregated quantitative data from our sources onSudan are considered IDPs. a possible increases and decreases in figures, but also other kinds of information, such as data disag-Data sources gregated by sex and age (SADD). Such information is vital in guiding an appropriate and effectiveOur ability to report on displacement and provide response to IDPs’ protection and assistance needs.reliable estimates is contingent on the availabilityof sources, and their willingness to gather and Little SADD is available for displacement associatedshare data. We draw on information produced with either conflict or disasters. The main reasonor compiled from a wide range of source types. is that specific information on IDPs’ sex, age andGovernments might be expected to have the disabilities is more easily captured in organisedprimary responsibility for counting IDPs, but many settings such as relief camps, while in many casesothers are involved in data gathering, including a significant majority of IDPs live in dispersedinternational organisations, community-based settings among host families and communities.organisations, specialised websites, thematicdatabases, local authorities, national Red Cross We also aim to gather and report disaggregatedand Red Crescent societies and private sector information by geographical area and time periodinstitutions. Such sources play a significant role, in order to paint the most comprehensive andparticularly when governments lack the capacity dynamic picture of displacement and provide aor will to collect the data or when their estimates sound basis for more complex research and analysis.are unreliable. Even when disaggregated data is available,Different sources gather different data for however, it tends not to represent a statisticallydifferent purposes, with different methodologies significant portion of the overall data collected.and for different objectives. These include opera- More is vital if we are to accurately inform thetional planning, which is influenced by consid- identification of, and response to the specificerations of timely funding. Divergent objectives needs of different groups of IDPs.often affect the way in which data gatherersestimate target populations or beneficiaries. Normalising displacement data by country population sizeWe are aware that various data sources may alsohave an interest in manipulating or tweaking To illustrate the magnitude of internal displace-the number of IDPs. They may choose to do ment at the country level, we normalise the data toso in order to call international attention to a account for population size using the UN Populationcrisis, maximise the amount of external assis- Division’s population estimates for each country. Intance received, downplay the scale of a conflict doing so, a clear distinction has to be made betweenor disaster if the government is held account- the notion of population and inhabitants. Whenable, or because of political sensitivities such as displacement is acute, including refugees fleeingto deflect international attention. across international borders, the population in a country at a given time may be significantly lowerIn order to mitigate this potential bias, whenever than the official figure. Syria is the most graphic casepossible we triangulate the data by using several in point, but the issue also affects other countriessources and prioritising those we have historically such as Libya and Somalia, for which there are nodeemed to have been most objective. up-to-date and reliable national population figures. As such, the ratios of IDPs to population and inhab-Language bias also affects our ability to source itants will differ, but both provide useful informationdisplacement data comprehensively. We can only for research and analysis.obtain and analyse information in the languagesin which we speak and read. Our staff andnetwork of partners speak most languages, butwe inevitably fail to capture some information,particularly for parts of Asia. METHODOLOGICAL ANNEX 83

Table A.7. Comparison of main monitoring attributes for displacement associated with conflict and disasters Displacement monitoring Conflict and violence Disasters attribute Event-based No Yes Geography or situation-based Yes No Global coverage Yes Yes Quantitative threshold No No Enables reporting of number, or Yes No, lack of data stock of IDPs Covers incidents of new displace- Yes Yes ment Includes other inflows and outflows Yes, subject to availability No, lack of data that determine the number of IDPs of data Includes SADD Yes, subject to availability Yes, subject to availability of of data data Figures disaggregated based on age Yes No, not applicable of source data Application of average household Yes Yes size data Methodological challenges In some countries only one data source is avail- particular to displacement able, while in others there may be several. For associated with conflict each country listed in the 2016 GRID dataset, we systematically looked for several sources. We gather data from primary and secondary We always strive to identify new data sources, sources on the number of people displaced even for countries and situations where others by international and non-international armed already exist. This enables us to crosscheck, but it conflict and other situations of violence. We aim may also create confusion because sources rarely to include all people forcibly displaced in such explain their methodologies. contexts. When different sources are available, or when Our monitoring is based on the sourcing and a new source provides information, we may analysis of other’s primary and secondary data. still decide to base our estimate on only one Data sources tend to be numerous during human- source. That decision may vary from year to year itarian crises and visible emergencies, when they depending on objective criteria, such as their compile information to target assistance, as in geographical and temporal coverage, or their Syria. During protracted and neglected crises, perceived reliability (see confidence assessment displacement data tends to be unavailable or out- section below). Or we may aggregate different of-date, as in Armenia, Cyprus, Georgia, Togo data from separate sources, which may help us and Turkey. extend the geographical coverage of our esti- mates. As such, our figures are more likely to Sources do not often use the same definition of take into account and reflect both qualitative an IDP as the Guiding Principles. Nor do they use and quantitative uncertainties. the same methodologies, which creates a serious challenge when compiling our estimates. In In many countries affected by conflict and several countries, including Afghanistan, Bosnia violence, no agencies or mechanisms collect and Herzegovina, DRC, Georgia, Pakistan and data on the number and kind of people who Ukraine, only IDPs who have been officially regis- have sought refuge in urban areas, those who tered with the authorities are counted. are hosted by relatives or other families or those who have fled to remote areas. This leads to the number of IDPs being under-estimated.84 G20R1ID6

On the other hand, some sources identify IDMC’s previous estimate for Sri Lanka includedreturnees as still being IDPs and include them in IDPs who had returned, but who had nottheir figures, which in terms of our methodology achieved a durable solution. This year, in keepingconstitutes an over-estimate and a particular with the rule we apply to other countries, wecomputing challenge, given that we subtract subtracted these returnees, which reduced ourreturnees for reporting purposes. For example, estimate by nearly half.Selection of countries in the GRID dataseton displacement associated with conflictand violenceThe 2016 GRID dataset contains information Interpretation ofon 52 countries and one disputed region, the information received fromAbyei area, where we have received or been sourcesable to obtain information on displacement.The inclusion of a country in the dataset is An in-depth reassessment of the sources avail-not contingent on a quantitative threshold able for all of the countries we included in ourfor the number of IDPs. It depends only on 2015 Global Overview and a close examina-the availability of credible data. The fact that tion of the data led to the following countriesa country is not included should not be taken being excluded from this year’s report:as implying that no displacement has takenplace, but rather that no information has been || Eritreaforthcoming, or that the displacement is not || Laoscaused by conflict or violence. || Liberia ||Timor LesteOur 2016 GRID estimates include a numberof changes from our 2015 Global Overview Analysis of primary causesthat result from the systematic and consistent of displacementapplication of decision rules to all situationsof displacement. Unlike instances in which A thorough review of our data and contextualwe have simply updated or revised a previous analysis revealed that in some cases, the mainfigure based on new evidence, they represent causes of displacement were not linked tosubstantive departures from previous prac- conflict but to other triggers such as forcedtice and concern whether to account for, and eviction. We found that such triggers were thereport on certain situations and caseloads at only cause of displacement in the followingall. countries, so we removed them from our 2015 dataset for displacement associated withSuch decisions were based on issues related to conflict and violence:a more consistent interpretation of informa-tion received from our sources, our analysis || Turkmenistanof the primary causes of displacement and || Uzbekistangeopolitical considerations that affect the || Zimbabwedefinition of international borders that areessential to determine whether someone isan IDP, a refugee or stateless. These borderissues cover foreign occupation, the creationof new states and unilateral secession.As a result, in some cases we have madequantitative changes to previous estimatesfor the same stock of IDPs, while in otherswe chose not to include certain countries inthe 2016 GRID. METHODOLOGICAL ANNEX 85

For other countries, where we found that as Sudan and South Sudan, we consider all the causes of displacement varied between people displaced within each of the new enti- different caseloads of IDPs, we subtracted ties as IDPs, and we produce separate esti- those IDPs whom we ascertained had not mates for each one. People who fled within been displaced by conflict. This led to a the previously undivided state and who reduction in the total number of IDPs for the crossed the border that delineates the new following countries: entities are no longer counted as IDPs. || Indonesia As such, we no longer count people who fled ||Papua New Guinea from Timor Leste to West Timor when the || Liberia former was established in 1999. Their number has been subtracted from our 2015 estimate Geopolitical parameters for Indonesia. a. Foreign occupation c. Unilateral secession We consider people displaced within areas For regional entities such as Abkhazia and of an internationally recognised state under South Ossetia, which have unilaterally seceded foreign occupation as IDPs, irrespective of outside an internationally supported process, their location with respect to the de facto we do not count IDPs within them sepa- borders or the territorial claims of the occu- rately from those in the state the entity has pying power, providing the original interna- seceded from. In cases where a majority of UN tional borders still have broad international member states have established diplomatic recognition. relations with a seceding entity, however, we do produce estimates for IDPs who have fled As such, our 2015 estimate of the number of within it. IDPs in Cyprus does not only include Greek Cypriots who moved to the southern part of For the purpose of the GRID only, we no the island at the time of Turkey’s invasion in longer count people as IDPs if they have 1974, as was the case in the past. It also incor- crossed what has become a de facto interna- porates estimates for Turkish Cypriots who tional border and find themselves in different moved from southern to northern Cyprus at entity from the one in which they were origi- the time. This interpretation and accounting nally displaced. As such, our estimate for is consistent with the methodology we have Kosovo refers only to people who have fled used for other occupied areas, such as Crimea within the territory itself. Given the Serbian and other parts of eastern Ukraine. government reported all IDPs in the country as having come from Kosovo, Serbia is not included in the 2016 GRID. b. Creation of new states These decisions not to continue counting people we previously considered IDPs in no For countries that have been divided into way implies that they no longer have vulner- two internationally recognised states, such abilities related to their displacement.86 G20R1ID6

Geographical scope and coverage To do so, we will soon begin piloting a hybrid monitoring methodology that combines event-Our methodology aims to capture the full based and country-based monitoring of displace-geographical scope of displacement and strives ment situations as they evolve over time. The ideato monitor and report on all situations across the is to identify displacement events in near-realentirety of each country we cover. In many, such time, manually verify those we deem to haveas DRC, Syria and Yemen, however, data sources led to people fleeing and then to engage part-do not cover all of the regions where displace- ners in the field to collect time-series data. Forment took place. As a result, displacement figures the purpose of initiating a humanitarian alert, inonly reflect geographical areas where humani- some cases our partners in the field will also helptarian agencies have been operating, and the us to identify events that have the potential toobjectives of their response. trigger displacement.Humanitarian agencies often lack access to Methodological challengesconflict zones because of insecurity, which can particular to displacementlead to significant information gaps. Our sources associated with disasterstend to monitor and report on displacementmore easily in areas where IDPs are most visible, The 2016 GRID presents our latest findings onsuch as in camps. In most cases, however, agen- new displacement associated with disasters incies fail to record the geographical dynamics of 2015, and compares it with our historical datasetIDPs’ movements when registering them. In other for 2008 to 2015.cases, such as Myanmar and Syria, they collectdata in regions that overlap, often using differentmethodologies.Data gatherers are very likely to overlook IDPs Typological considerationsliving in more dispersed settings. These includepeople who move to urban areas where they The 2016 GRID estimates are based on newblend in with local inhabitants; those who flee displacement known to have taken place as ato remote areas, such as the bush in CAR or result of disasters for which natural hazards havethe forests of Côte d’Ivoire; and those who are been identified as the primary trigger. Whenhosted by other families or relatives, as in the available, we use the internationally acknowl-Philippines. They end up unreported, and the edged name of the hazard and categorise themscope and nature of such displacement cannot initially into four main types: geophysical, mete-be quantified and assessed. Their number and orological, hydrological and climatological. Thesefate remain unknown. are then refined into types, sub-types and sub- sub-types (see table A.8).Temporal scope and frequency of To better understand the complexities of thereporting phenomena, we plan to break down the different stages of a disaster by identifying its primary fromThe 2016 GRID dataset reports separately on its secondary, tertiary and subsequent triggers.the total number of IDPs as of 31 December2015, and the number of people newly displaced The 2015 dataset presents figures for displace-during the year. The former reflects the number ment associated with sudden-onset hazards, butof people still displaced at the end of the year, in future reports we intend to include that associ-but does not capture repeated displacement or ated with slow-onset hazards such as drought.other movements of people who fled or returned In 2014, we developed a model-based method-home during it. ology, which we used to monitor the displace- ment of pastoralists in the Horn of Africa duringThe figures reported are static, but IDPs’ move- the 2010 to 2011 drought, and we started toments are not. For this reason, we aim to improve collect data on slow-onset hazards in 2015.our methodology and increase not only itsgeographical, but also its temporal coverage.We plan to produce displacement figures morefrequently in order to capture the fluidity andcomplexity of IDPs’ movements. METHODOLOGICAL ANNEX 87

Table A.8. Typology of natural hazards* Hazard category Type Sub-type Sub-sub-type Geophysical Earthquakes, mass Ground shaking, tsunamis, Derechos, hailstorms, movements, vol- sudden subsidence, sinkholes, thunderstorms, canic activity landslides, rockfalls, ashfalls, rainstorms, tornados, lahars, pyroclastic flows, lava winter storms, dust Meteorological Storms, extreme flows, toxic gases, glacial lake storms, storm surges, temperatures outburst flows (GLOF), volca- haze, gales nic eruptions Hydrological Flooding, land- Fire whirls slides, wave action Extra-tropical storms, tropical storms including hurricanes Climatological Drought, wildfires and cyclones, convective storms, cold waves, heat- waves, severe winter condi- tions Coastal floods, riverine floods, flash floods, ice jam floods, avalanches – snow, debris, mudflows, rockfalls – rogue waves, seiches Forest fires, land fires –bush, brush and pasture * This typology is adapted from the classification system developed by the international disaster database (EM-DAT) maintained by the Centre for Research on the Epidemiology of Disasters (CRED) in Louvain, Belgium. Spatial and geographical coverage Protracted displacement in the aftermath of disas- ters is a highly challenging area. We produced a Our estimates aim to provide a global picture first scoping exercise in 2015, which aimed to of displacement associated with disasters, but shed light on the phenomenon by challenging the we face a number of challenges in compiling notion that people who flee a disaster are not likely them. Thanks to long-standing partnerships with to remain displaced for long. This false assump- organisations such as IOM, we have been able to tion is fostered by only occasional reporting of overcome some, and reach out at the national ongoing cases, often to mark the anniversary of and local level for information. Language barriers, a particular disaster. Our scoping exercise allowed however, are a major challenge, particularly for us to re-examine the issue, and conclude that events that occur in south and south-east Asia. there are likely to be many more people living in To address this, we constantly seek to improve protracted displacement than previously thought. our access to data by expanding our network of We plan to monitor and analyse the phenomenon reliable collaborators, with particular focus on in-depth using our data model. our partners in the field. Temporal coverage Terminology Our dataset records incidents of displacement We use the term “displaced”, but it is rarely if that began in 2015 and are supported by a reli- ever adopted consistently and unequivocally by able and comprehensive source. The main chal- different countries or sources (see table A.9). lenge we faced in collecting data for the year People displaced by floods in 2015 were referred were overlapping events, such as cyclone Komen to as “homeless” in Madagascar and as “moved” and Myanmar’s monsoon floods, which made in Iraq. Often, sources refer to people displaced it difficult to identify people displaced by each by disasters as “directly affected”. It is true that disaster because our sources provided a final IDPs are part of a wider population affected by aggregate figure for all events. a disaster, but not all those affected are IDPs. As such, additional analysis is required to make sense of the terms sources use, and to under- stand when and how they signal displacement.88 G20R1ID6

Table A.9. Explanation of reporting termsTerm ExplanationDisplaced Involuntary or forced movements, evacuation or relocation – when not specified – of individuals or groups of people from their habitual housingEvacuatedRelocated Voluntary and forced evacuations, both preventive and in response to onset of hazard Voluntary and forced relocations, both preventive and in response to onset of hazardSheltered People accommodated in shelters provided by national authorities or organisationsHomeless such as NGOs, the UN and IFRC People rendered homeless and without adequate shelterUninhabitable Limited to habitual place of residence, and includes houses, retirement homes, prisons,housing mental healthcare centres and dormitoriesOther Any term not mentioned aboveEven within the UN and coordinated interna- We also include shelters in refugee and displace-tional humanitarian reporting mechanisms there ment camps, for instance “collapsed tents” inis inconsistency in how different populations are in Jordan’s Zaatari refugee camp are counteddescribed and counted, with some estimates as uninhabitable housing. Such cases constitutebased on “people affected” and others on multiple displacement, in which people may have“people in need” or “people targeted”. fled conflict only to become displaced again when their camp is flooded.Many terms and expressions are specific tointernal displacement, and our database captures Evacuation datathe most common ones, as shown in table A.9.They may refer to individuals, groups of people We often use data on mandatory evacuationssuch as families or households, or housing. We and people staying in official evacuation centresuse the number of houses destroyed as a proxy to estimate event-based displacement. On thebecause it shows that at least one household has one hand, the number of people counted inbeen left homeless. We calculate the number of evacuation centres may underestimate the totalindividuals by applying the average household number of evacuees, as others may take refugesize available for each country (see box). elsewhere. On the other, the number of people ordered to evacuate may overstate the trueHousing information number, given that some are likely not to heed the order. The potential for such discrepanciesHousing information is vital in estimating is much greater when authorities advise ratherdisplacement associated with disasters. In 2015, than order evacuation, and as a result we do35.5 per cent of the sources we used for our esti- not incorporate such figures into our estimates.mates reported figures for uninhabitable housingwhen describing displacement. In order to use Quality assurance andhousing data as a valid proxy, we only consider independent peer reviewfigures for homes that have been damaged tothe extent they are no longer habitable.Terms that indicate the extent of damage include As in previous years, and in order to improve“houses at risk (of collapse)”, “houses severely our methodology, we submitted this year’s esti-affected/damaged” and “houses destroyed”. mates to a quality assurance process to verify theWe consider housing to be any place where data. The verification stage is as important as thepeople have established a habitual residence, data collection itself, because it allows possibleincluding retirement homes, prisons, religious discrepancies to be identified, and the data to beresidences and schools when dormitories are refined before it is finalised. This year’s processpresent. We include hospitals if the information was led in-house, and all of our entries have beenprovided suggests that long-term patients have double-checked.been displaced. METHODOLOGICAL ANNEX 89

For disaster events in 2015, all records with To limit evaluators’ bias and improve objectivity estimates of 500 IDPs or more have been fact- and consistency, clear decision rules are needed checked. In future we aim to extend the verifica- that limit the number of dimensions taken into tion process to the entire set of annual entries. account. To improve the Syria MSNA’s descrip- We have also submitted this methodological tive confidence scale and overcome its lack of annex to external peer reviewers, and elements proper aggregation, a technical note suggested of our methodology were reviewed in previous the application of a points-based index with years by a different set of independent experts. three criteria, effectively discarding four of the seven included in the original confidence scale.9 We will embed the external peer review and internal quality assurance processes into our There are ways of overcoming the limitations future work to ensure that the methods we use of point-based scores, but their complexity may to produce our figures are robust and that we render them opaque, adding another layer of have presented them accurately. potential confusion. Using only four indicators with two to five possible values for each, IOM Qualitative assessment of Iraq’s assessment framework yields up to 126 confidence in estimates for unique possible combinations.10 people displaced by conflict The challenge of applying nationally Building upon lessons from existing specific tools at the global level assessments It is difficult to extrapolate to the global level There have been several attempts recently to from confidence ratings designed for national design confidence assessment schemes to eval- circumstances. The three examples discussed uate data on internal displacement, part of a above all refer to situations in which a single broader movement in the field of humanitarian organisation or cluster designs the entire national needs assessments.5 The Task Force on Population data collection process. Movement in Yemen (TFPM), for example, has developed a confidence rating based on disag- At the global level, aggregation and cross-country gregation by sex and age, and the availability comparison is made more difficult by the number of data on districts of origin and displacement.6 of data sources and the fact that their motiva- tions for collecting information ranges from IOM Iraq calculates a confidence rating in rapid needs assessments to victim compensation order to produce an estimate for each location without any a priori global coordination. Sources’ in its displacement tracking matrix, based on methodologies also vary widely, from satellite the number of informants used, discrepancies imagery, registration, sampling, key informant between information from different sources, interviews and censuses, to name but a few. the accessibility of the location and the ability to independently validate the data received.7 This diversity stands in stark contrast to the The Syria multi-sector needs assessment (MSNA) standardisation of data in the three national gives a confidence rating for the population esti- examples mentioned above. As such, the same mates it provides, including the number of IDPs, set of criteria cannot easily be used to judge reli- using a six-point scale with up to seven criteria ability, and the diversity in which the results are for each point.8 reported makes it more difficult to make compar- isons between countries. Such assessments may seem reassuring, but IDMC’s confidence assessment if poorly conceived or implemented they may provide a false sense of certainty or confidence. We have made an initial attempt to design a They may hide the arbitrariness of the under- comprehensive framework to assess the confi- lying criteria and the way they are weighted and dence we have in the estimates we publish. The aggregated. They may also reflect the biases and methodology and results presented in this report challenges inherent in the various steps involved are the first steps of a process we will continue to in constructing an index and collecting the data. develop through several more iterations.1190 G20R1ID6

Given that we are as yet unable to apply many of ||In no country or displacement situation didthe criteria to our data on displacement associ- the data cover our model comprehensively.ated with disasters, we have only assessed our This means that information about somedata and 2015 estimates for that associated with flows is missing, resulting in a distorted orconflict in 11 representative countries (see table incomplete picture.A.10). In assessing our confidence in the data,we applied a common set of criteria based on: ||We were unable to receive data frequently enough to keep up with events as they ||The methodologies used to collect it unfolded on the ground, particularly for highly dynamic situations. Again, the likely ||Whether it could be independently validated result is a skewed picture of displacement that does not capture events which evolved ||The degree to which it is geographically or were resolved quickly. comprehensive in terms of the extent of the conflict and associated displacement The encouraging news is that in several of the 11 countries, the data we obtained was disag- ||Whether it is disaggregated by sex and age gregated both geographically and by IDPs’ sex and age. We have more confidence in these ||The frequency with which it was collected datasets and our estimates based on them than on those we derived by multiplying the number ||How extensively it covers the components of of destroyed houses or families evacuated by our data model average household size.For this initial assessment, we have not attemptedto weight or rank these factors, nor have weassigned quantitative point values for them orgenerated an overall score for each source andestimate. In order to do so rigorously, we will firstneed to empirically test the relative significanceof each of the factors.Some of the data gaps reported can be attrib-uted to the way governments and organisationscollect and disseminate data, but this is notalways the case. We try to be as comprehensiveas possible in our own data collection, but wemay overlook some sources that may address thegaps we report. As such, our assessment reflectsthe level of detail of the data we were able tocollect and process from various sources – notthe level of detail of all the data that exists orwas published by each provider.The assessment is shown in the table below, andreveals several features of our source data andthe estimates based on it:||In many cases we were unable to obtain thor- ough documentation of our providers’ data collection methodologies or protocols.||We often rely on only one source that we are unable to verify independently. METHODOLOGICAL ANNEX 91

Table A.10: Initial IDMC confidence assessment Bosnia and Colombia Indonesia Iraq Herzegovina Data on Methodology Registration Registration Unknown IOM DTM* displace- of the source(s) ment used Data could be No No No No triangulated nationally All relevant ar- Yes Yes Unknown No eas well covered Disaggregation Yes Yes No Yes of data in subna- tional adminis- trative entities Frequency of Yearly Yearly No update Bi-monthly reporting Yes Disaggregation Yes No Yes by sex Yes No No Yes Disaggregation Yes No Partial Partial by age No No No No No No Data on Returns No No No No end of Deaths No No No No displace- Births No No No ment and Cross-border other pro- movements No cesses Local integration No Settlement else- No where * International Organization for Migration’s displacement tracking matrix For each country or territory, data on new Bosnia and Herzegovina displacements and the number of IDPs as of the end of 2015 have been assessed together. In Source: Government – direct email contact with the many cases the same report is the source for Ministry for Human Rights and Refugees both numbers. The government gives information on “ceased dis- placement” without providing further details. The following notes expand upon and refine some of assessments in the table above.92 G20R1ID6

Nigeria Papua New Sudan Syria Thailand Ukraine Yemen Guinea MultipleIOM DTM Multiple (part- IOM DTM Lacks trans- Unknown Registration ly unknown) parencyNo No No No No No NoNo No No No No Uncertain NoYes No Yes Yes No Yes YesQuarterly No update Biannually Yearly Every No update Almost weekly two monthsPartial No Yes Yes No No YesPartial No Yes Yes No Yes YesPartial No Partial Partial No No PartialNo No No No No No NoNo No No No No No NoNo No No Partial No No NoNo No No No No No NoNo No No No No No NoColombia IndonesiaSource: Unit for Attention and Reparation of Victims Sources: Media reports; one source citing the Ministry(UARIV), government agency of National Development Planning (BAPPENAS), butMethodology: UARIV’s registration system counts the we were unable to trace the original document.number of people who have a claim as a victim of the Methodology: BAPPENAS’s methodology is unknown.country’s conflict, not the current number of IDPs. Geographic coverage: BAPPENAS’s reach is uncertain.Many people have been displaced more than once, Media reports mention new displacements in only aleading to multiple registration and double counting. few regions such as Aceh, Yahukimo and Karubaga.Our estimate aggregates data since 1985, with a peak Geographic disaggregation: The BAPPENAS data weperiod of displacement between 2000 and 2005. obtained was just one aggregated estimate.There is no information available on IDPs who mighthave found durable solutions to their displacementsince 1985. METHODOLOGICAL ANNEX 93

Iraq Syria Source: IOM Source: OCHA, which collates sources from various Geographic coverage: Access is limited in Anbar and entities Ninewa governorates, which Islamic State (also known Geographic coverage: A number of areas are hard to as ISIL or ISIS) controls. reach, particularly in the north-east of the country, End of displacement: Data on returns covered only and estimates are unreliable. part of the year. End of displacement: Data on returns and cross-border Other: There is no data on people displaced before movements is scarce. 2014. Other: Data collection takes place in a complex secu- rity environment, in which some stakeholders includ- Nigeria ing armed groups have motives for providing biased information. Source: IOM in collaboration with the authorities Geographic coverage: Access was not possible to 17 Thailand of the 27 local government areas (LGAs) in Borno state. In other states, it was only partial in some LGAs. Source: An International Crisis Group (ICG) report that Data disaggregated by sex and age: SADD published mentioned a few displacement cases by IOM is available for each site, but only for aggre- Methodology: We compiled data from sources cited gated figures for displacement associated with both in a single ICG report dated 2007, and which does not conflict and disasters, which does not permit a dif- focus on IDPs. The report in turn uses various sources ferentiated analysis. whose methodologies are unknown. End of displacement: In Adamawa, data on returns Geographic coverage: The report covers only parts of is only available for the northern part of the state. southern Thailand. Geographic disaggregation: The data covers only a Papua New Guinea limited number of displacements in a few locations. Sources: ICRC, media report Ukraine Methodology: Partly unknown. We compile data from sources using various methodologies. Source: Ministry of Social Policy Geographic coverage: Only places where ICRC and the Geographic coverage: The data has national coverage, media are present are covered. Ethnic clashes often but that for areas near the Russian border not under take place in remote areas where access is hindered government control is possibly less reliable. by insecurity and difficult terrain. End of displacement: Data disaggregated by region is Geographic disaggregation: Data is gathered in a only updated roughly once a week. The figures sometimes limited number of locations. show a decrease, which implies that the displacement of IDPs between regions and/or durable solutions are Sudan somehow taken into account, but no further details are available. Source: IOM Only Darfur and Kordofan are covered, which excludes Yemen areas such as Khartoum and the east of the country where displacement associated with conflict is likely Sources: UNHCR in the north of the country and IOM to have taken place. in the south coordinate a population movement task End of displacement: Detailed information is available, force, to which 22 organisations contribute data. including SADD, vulnerability and occupation, but Methodologies: UNHCR uses population movement again only for Darfur and Kordofan. tracking, and IOM its displacement tracking matrix. Geographic coverage: For around half of the country’s 21 governorates, data could not be collected in some districts. The largest gaps were in the Al Hudaydah, Hadramaut, Lahj and Shabwah governorates. End of displacement: IOM’s December 2015 report only covers some returnees who had fled disasters. Its February 2016 report does not disaggregate data temporally between 2015 and 2016, so it could not be used.94 G20R1ID6

Notes on IDMC’s confidence Notesassessment criteria 1. E M-DAT guidelines, available at http://goo.gl/SADD: The availability of SADD does not directly izFndCfactor into the calculation of the number of IDPs,but it can be considered a proxy for detailed data 2. Journal of the Royal Statistical Society, Averagecollection practices. The Task Force on Population Household Size and the Eradication of Malaria,Movement in Yemen used SADD as a criteria in 2012, available at https://goo.gl/gnWCnNits fifth, sixth and seventh reports in a decisionrule aimed at prioritising data.12 3. U SAID, The DHS Program: survey types, available at http://goo.gl/1mzLLO; USAID, The DHS Program:Geographically disaggregated data: Such methodology, available at http://goo.gl/pA2yyq; In-data is not, per se, an absolute requirement for ternational Journal of Epidemiology, Demographicaccurate national estimates of displacement. and Health Surveys: a profile, October 2012, p.1-12,In many countries, however, some of the enti- available at http://goo.gl/6kkmIBties that collect data only have access to someregions. Geographical disaggregation allows for 4. OECD, Doing Better for Families, 2011, available attriangulation and gaps to be identified, while its http://goo.gl/UlrLcKabsence can lead to possible double-counting.The Task Force on Population Movement in 5. ACAPS, How sure are you? Judging quality andYemen uses a similar rationale in its confidence usability of data collected during rapid needs as-rating to justify discarding data when location sessments, August 2013, available at http://goo.gl/information is incomplete. JcYxMkMultiple data sources: The availability of data 6. T ask Force on Population Movement in Yemen: 5thfrom a number of independent sources does not Report, October 2015, p.4, available at https://goo.guarantee higher quality or more accurate overall gl/6GtTbNresults. It can, however, prompt discussion ofthe various estimates available and the method- 7. IOM, Response to the Anbar Crisis in Iraq, Displace-ologies used to derive them. It also sometimes ment Tracking Matrix, Round II Report, April 2014,permits triangulation, which is useful in situa- p.3, available at http://goo.gl/Da8Z6etions for which displacement estimates are highlysensitive or more susceptible to data collectors’ 8. H umanitarian Liaison Group, Syria Multi-Sectoralbiases. Needs Assessment, October 2014, available at http://goo.gl/OuplUPTemporal dimensions: The frequency ofupdates is a relative criteria. Unfolding crises and 9. Benini A, Shikh Aiyob M, Chataigner P et al,rapidly changing situations such as those in Syria, Confidence in needs assessment data: the use ofIraq and Yemen require more frequent updates confidence ratings in the Syria multi-sectoral needsthan stable and often protracted situations such assessment (MSNA), a note for ACAPS and MapAc-as in Armenia and Cyprus. Yearly updates may tion, April 2015, available at http://goo.gl/Vo7W01suffice for some situations, but for others, it canexclude some of the shorter-term displacements. 10. IOM Iraq, displacement tracking matrix downloads, available at http://goo.gl/AHE54s 11. Benini A, Shikh Aiyob M, Chataigner P et al, Confidence in needs assessment data: the use of confidence ratings in the Syria multi-sectoral needs assessment (MSNA), a note for ACAPS and MapAc- tion, April 2015, p.14, available at http://goo.gl/ Vo7W01 12. Y emen Task Force on Population Movement, 5th Report, 14 October 2015, p.3, available at http:// goo.gl/TyCjPm; 6th report, 10 December 2015, pp.6-7, available at http://goo.gl/EDcbTE; 7th report, February 2016, p.17, available at http://goo. gl/oUffdiNext stepsOur confidence assessment is a work in progress,and we welcome input from partners interestedin contributing to its development. We plan toapply our criteria to all of the data we receiveand analyse so that our estimates are as accurateas possible. In doing so, our data users will bemade aware of the magnitude of uncertainty thedata contains, and the underlying reasons for it. METHODOLOGICAL ANNEX 95

TABLES Table 1: New displacement by country for disasters and conflict and total number of IDPs for conflict and violence Country or region Total number New New of IDPs as of 31 displacements in displacements in Abyei Area December 2015 2015 (disasters) Afghanistan 2015 (conflict) Albania (conflict) 335,000 71,000 Algeria 82,000 4,200 Angola 23,000 19,000 Argentina 1,174,000 71,000 5,600 Armenia 210,000 36,000 Australia 8,400 36,000 Azerbaijan 564,000 224,000 5,700 Bahamas 426,000 621,000 Bangladesh 2,800 Belize 98,000 531,000 Bhutan Bolivia 99,000 300 Bosnia and Herzegovina 2,900 Botswana 124,000 11,000 Brazil 452,000 Bulgaria 107,000 300 Burkina Faso 300 Burundi 6,270,000 59,000 Cabo Verde 7,800 800 Cambodia 3,700 Cameroon 1,500,000 3,100 Canada 200 Central African Republic 8,900 Chad 11,000 Chile 13,000 China 1,100 Colombia Congo (Republic of) 1,047,000 Congo, Democratic 3,602,000 Republic of the 4,600 106,00096 G20R1ID6