CONFLICT TRENDS 06 2015
Visiting Address: Hausmanns gate 3
PO Box 9229 Grønland, NO-0134 Oslo, Norway
Peace Research Institute Oslo (PRIO)
from conlict, surely they will not be able to that have functioning state institutions, and
Understanding Attacks on
conlict, such as economic reconstruction and protect aid workers either. But in fact, this does therefore have a lower likelihood of experienc
institutional transformation (i.e. reform of not hold. We ind a tendency towards more ing government instability, are also associated
police, army, judicial system, elections) – are aid worker attacks in the most authoritarian with lower levels of risk to aid workers.
Humanitarian Aid Workers
not associated with a greater number of attacks countries, yet surprisingly, we do not ind any
on aid workers. This result is in line with recent evidence to support our hypothesis that in Limitations and forward research
research that shows that only PKOs with trans consistent states see more aid worker attacks.
formational mandates are efective at reducing This is especially interesting considering the Our study is among the irst to analyze the
the intensity of conlict in a country. In contrast, abundance of evidence showing these countries determinants of attacks against aid workers. As
traditional peacekeeping forces are not equipped in general to be more conlict-prone and instable such, there is scope for reinement.
to use lethal force to protect either themselves, or experience greater rates of criminal or social
civilians, or aid workers. violence. First, data improvements could improve the
quality and extent of reporting of attacks,
The type of conlict – and degree of While some have argued that particularly agreeing on what constitutes an incident, and
civilian targeting – matters less than ‘brutal’ or ‘humane’ types of insurgencies may more accurately deining and determining the
expected inluence attacks on humanitarian actors, we number of workers in the ield. With these im
www.prio.org/ConlictTrends
Conlict Trends Project
found no evidence indicating that the aims of provements, a better understanding of the risks
We found no evidence indicating that countries the rebels signiicantly inluence the degree involved at the sub-national level and of the There is an increasing demand for Brief Points
that experience one-sided violence (i.e. countries of aid worker attacks. Conlicts where rebels speciic types of humanitarian initiatives and
deploying humanitarian workers
where either the government or insurgents are
actively targeting civilians) have higher rates of
attacks on aid workers. The efect of one-sided
seek secession or regional autonomy, which in
turn are often conlicts in which rebels have
a strong regional presence and compete with
deployments could be made. As aid agencies
continue to expand across the globe – and to
new areas of conlict and insecurity, such as in to conlict-afected areas. However, • Violent conlicts see more aid worker
attacks
violence on aid worker attacks is essentially zero.
This is encouraging, as periods of one-sided
violence are situations where civilian popula
the government for service provision, do not
have more attacks on aid workers than conlicts
where the rebels are primarily seeking to
urban settings – the evolving and contested role
of humanitarian space is of high priority for
further study.
this need has expanded the risk
of violent attacks against staf in
• Countries with greater criminal
violence do not have more attacks on
tions are especially vulnerable and needing aid overthrow the government.
the most. That these situations do not appear to The next step in the study of aid worker attacks insecure ield settings. In this aid workers
be more dangerous to aid workers may further
encourage the international community to
provide more extensive support for vulnerable
More developed states have fewer
attacks, and country risk is important to
consider
is to dig deeper into the micro dynamics of
attacks. We have the data on conlict dynamics
to do such an analysis, but the crucial missing
brief, we identify six country-level • International military forces do not add
to aid worker risk, but International
populations. ingredient is ine-grained information from aid factors that can inluence attacks Peacekeeping Operations do
•
In general, we ind that the more developed the organizations themselves about who, where,
on aid workers. These six factors
ISBN: 978-82-7288-646-1 (print)
Design: medicineheads.com
Regime type and rebel organizational country, measured in terms of GDP per capita, and what aid workers are doing. Regardless, Groups who actively target civilians do
structure matter less than expected the fewer aid worker attacks. The same pattern by further reining the conditions for violence
help us to better understand the not attack aid workers more
978-82-7288-647-8 (online)
holds for other measures of state capacity and against humanitarians, aid organizations of
A long-standing literature has found that au
tocracies and democracies are roughly as good
at containing violence within their territories
state consolidation, namely ‘time since regime
change’ and ‘time in peace’. We ind that higher
capacity states see dramatically fewer attacks on
all types can better prepare for and prevent hu
manitarian attacks.
causes of aid worker attacks, and
may ultimately guide towards their
• More democratic and economically
developed countries pose fewer risks
– they have roughly equal levels of armed aid workers, presumably as they are better able Further Reading for aid workers
conlict. Semi-democracies, often called hybrid to protect aid workers present in the country. prevention in the future. This
or inconsistent states, in contrast have much
higher rates of conlict. We expected this to The same holds where countries are assessed policy brief is the irst from an
hold for attacks on aid workers as well. If semi-
democracies are unable to protect themselves
by international organisations as being of lower
political risk. This indicates that countries
emerging PRIO research program
on Humanitarian Security.
The Conlict Trends project aims to answer
non-proit peace research institute (estab
quences of and trends in conlict. The project
will contribute to new conlict analyses within Kristian Hoelscher Peace Research Institute Oslo (PRIO)
affect peace and conlict processes.
Jason Miklian Peace Research Institute Oslo (PRIO)
sues related to all facets of peace and conlict. Håvard Mokleiv Nygård Peace Research Institute Oslo (PRIO)
A Golden Age of Humanitarian Aid? Figure 2 shows the total number of aid worker has important implications for security and risk reduced the number of attacks to below what
attacks globally from 1997 to 2014, and the protocols in aid organizations; for donors and re- they would otherwise have been without such
The humanitarian aid sector has grown expo- number of battle-related deaths globally over searchers seeking to understand humanitarian improvements. Moreover, it is also possible
nentially over the past two decades. Global aid the same period. Here, the dashed vertical line insecurity; and in understanding how humani- that aid worker attacks are less of a ‘special’
spending by governments and private actors marks the start of the current conlict in Syria tarian agencies could and should engage with phenomenon that may be assumed, and merely
has increased 400% since 2000 to $25 billion that has coincided with a substantial increase in the world today. relect levels of overall violence in a given
USD in 2014 as the number of international aid both aid worker attacks and battle deaths. country. As such, aid workers operating in more
workers has also tripled. In addition, given the changing role of humani- dangerous regions may simply be attacked with
Greater Humanitarian Insecurity? tarian engagement in conlict, aid organizations greater frequency due to the underlying regional
Yet this international goodwill has also are working much more extensively in remote insecurity – either in the form of criminal or
produced dangerous consequences. Aid delivery These igures support the assertion that attacks ield settings, doing more sophisticated work, political violence.
areas tend to be in conlict or crisis zones, on aid workers are increasing. Two factors in and undertaking a much wider variety of de-
increasing operational insecurity and at times particular are responsible for the increase. First velopment tasks than just a decade ago. This To better understand the causes and drivers
blurring motivations for aid disbursement. 2014 is the increasing number aid workers in the has required INGOs to employ new strategies behind these attacks, we combined existing
was also the irst year in almost three decades ield, with the 1997-2014 time period covered to attempt to reduce operational risk by leaning datasets on aid worker deaths, violent conlict,
with more than 100,000 battle-related deaths here having seen a dramatic increase in the more heavily on national partners, presenting security and development to more systematically
globally, and an additional 58 million people number of aid workers deployed. The dual a less-visible local proile, and increasing staf assess why aid worker at-tacks occur, and what
were displaced in 2014 – the highest total ever increases lead to the inding that per capita security mechanisms. In response, some have national-level conditions place workers most at
recorded. Aid worker attacks have increased in attacks on aid workers are in fact stable. Second, raised concerns about how the links between risk. These indings both undercut and corrobo
tandem with the increases in violence. In 2013, the increased number of total attacks are driven aid actors and donor governments in the rate existing conventional wisdoms.
461 aid workers were attacked, representing the by a small number of countries that register Global South and the business-like actions of
most violent year on record against aid workers. many more aid worker attacks than what is humanitarian aid agencies may be potentially Explaining attacks on aid workers Figure 2: Trends in armed conlict and attacks on aid workers, 1997–2014
commonly seen, led by, in particular, Afghani- eroding the value and neutrality of humanitar-
The number of aid worker attacks from stan and Syria. ian space. Other scholars believe that this places Drawing on a cross-national sample between attacks – from 20 attacks in a year to more than available for analyzing this hypothesis than for
1997-2014 are shown in Figure 1. The size of INGO staf at greater risk, and that increasing 1997 and 2014, we used a quantitative approach 40 attacks per year. This relationship holds most of the others.
the bubbles correspond to the number of aid Why are attacks on the rise? Scholars and prac- institutional politicization and deeper ties with to test over one dozen factors to better when we diferentiate between diferent types
worker attacks recorded in the country over the titioners are increasingly trying to understand governments and/or militaries is considered a understand the spatial and temporal distribu of attacks on aid workers, with one exception: An international military presence does
period. For reference, all countries that have ex- where, why, and how aid workers become targets key factor motivating violence against INGO tion of attacks on aid workers. Our research Kidnappings are much less common in low not add to aid worker risk
perienced armed conlict over the same period of violent acts. For many, the key message is workers. examined the nature of conlicts that humani intensity then in high intensity conlicts. For aid
are in grey. Aid worker attacks occur dispropor- simply that it has become more dangerous to tarian agencies operate within; the structure of organizations the lesson here is fairly clear: the Contrary to much existing analysis suggesting
tionally in some countries such as Afghanistan, be an aid worker in the ield, particularly in But do these critiques have merit? It is possible humanitarian operations; and political factors more intense the conlict, the more attacks are an international military presence may increase
Pakistan, and Syria, but many African countries conlict-afected regions. However, we lack that aid workers are indeed targeted much economic conditions that may motivate attacks. to be expected. Unfortunately, these are often risks for aid workers, we ind no efect of a
have also seen high rates of attacks on aid systematic evidence explaining where and why more than before, but the improvement of We ofer six key indings that help to understand precisely the conlicts where aid workers are NATO presence on the estimated number of
workers. aid workers are attacked. This knowledge gap security by INGOs for their ield staf may have attacks on aid workers. most needed. attacks on aid workers. Comparative areas in
which NATO deploys forces see just as many (or
The conlict context of the country General levels of insecurity don’t appear as few) attacks on aid workers as those where
matters – to an extent to inluence the risk of aid worker NATO is not present. The estimated efect
attacks cannot be reliably distinguished from zero, but
We ind, not surprisingly, that aid workers are if anything, it is negative – countries with NATO
much more likely to be attacked in countries Yet unlike the clear risk of civil conlict to aid forces see fewer attacks on aid workers.
experiencing conlict than in peaceful countries. workers, insecurity and criminal violence is
Somewhat surprisingly, though, the diference not a factor related to the number of aid worker We do ind, however that countries that have
between minor and major conlict setting is not attacks. Attacks do not appear to be inluenced large UN peacekeeping operations (PKO)
that large measured in the absolute expected by the homicide rate in the country, a measure deployed, i.e. PKOs with large budgets which we
count of attacks. An average country with a that captures the generalized level of insecurity assume to be highly correlated with the size of
major conlict is likely to see double the aid and violence, the relative presence of criminal the force, see more attacks against aid workers.
worker attacks than a similar country with a groups and other ‘everyday’ threats to humani
minor armed conlict. This relationship holds tarian workers. Instead, countries with high This somewhat surprising inding is tempered
when we measured conlict intensity through homicide rates see just as many aid worker by the result that the type or mandate of the
the number of battle related deaths as well. We attacks as those with low crime rates. This could PKO matters crucially for understanding
found a strong and signiicant efect of conlict potentially be good news for aid organizations, attacks on aid workers. PKOs with a traditional
intensity on the expected number of attacks. An perhaps meaning that their extra precautions mandate (observing the terms of truce or peace
increase of battle deaths from around 400 deaths when deploying to countries that have high agreements, or policing a bufer zone and
per year, a medium intensity conlict, to 2000 levels of homicides have been efective. Note, assisting in negotiating a peace agreement) are
deaths a year, a high intensity conlict, roughly however, that the data for homicides is quite associated with more attacks on aid workers,
Figure 1: Aid worker attacks by country, 1997–2014 doubles the expected amount of aid worker poor; consequently we have much less data whereas PKOs with transformational mandates
C O N F L I C T T R E N D S 0 6 2 015 w w w. p r i o.o r g /C o n f l i c tTre n d s
Figure 2 shows the total number of aid worker has important implications for security and risk reduced the number of attacks to below what
100000
A Golden Age of Humanitarian Aid?
500
attacks globally from 1997 to 2014, and the protocols in aid organizations; for donors and re they would otherwise have been without such
The humanitarian aid sector has grown expo number of battle-related deaths globally over searchers seeking to understand humanitarian improvements. Moreover, it is also possible
nentially over the past two decades. Global aid the same period. Here, the dashed vertical line insecurity; and in understanding how humani that aid worker attacks are less of a ‘special’
80000
spending by governments and private actors marks the start of the current conlict in Syria tarian agencies could and should engage with phenomenon that may be assumed, and merely
400
has increased 400% since 2000 to $25 billion that has coincided with a substantial increase in the world today. relect levels of overall violence in a given
USD in 2014 as the number of international aid both aid worker attacks and battle deaths. country. As such, aid workers operating in more
60000
workers has also tripled. In addition, given the changing role of humani dangerous regions may simply be attacked with
Battle deaths
Greater Humanitarian Insecurity? tarian engagement in conlict, aid organizations greater frequency due to the underlying regional
300
Attacks
Yet this international goodwill has also are working much more extensively in remote insecurity – either in the form of criminal or
produced dangerous consequences. Aid delivery These igures support the assertion that attacks ield settings, doing more sophisticated work, political violence.
40000
areas tend to be in conlict or crisis zones, on aid workers are increasing. Two factors in and undertaking a much wider variety of de
200
increasing operational insecurity and at times particular are responsible for the increase. First velopment tasks than just a decade ago. This To better understand the causes and drivers
blurring motivations for aid disbursement. 2014 is the increasing number aid workers in the has required INGOs to employ new strategies behind these attacks, we combined existing
20000
was also the irst year in almost three decades ield, with the 1997-2014 time period covered to attempt to reduce operational risk by leaning datasets on aid worker deaths, violent conlict,
with more than 100,000 battle-related deaths here having seen a dramatic increase in the more heavily on national partners, presenting security and development to more systematically
100
globally, and an additional 58 million people number of aid workers deployed. The dual a less-visible local proile, and increasing staf assess why aid worker at-tacks occur, and what
were displaced in 2014 – the highest total ever increases lead to the inding that per capita security mechanisms. In response, some have national-level conditions place workers most at
0
recorded. Aid worker attacks have increased in attacks on aid workers are in fact stable. Second, raised concerns about how the links between risk. These indings both undercut and corrobo-
1997 2000 2003 2006 2009 2012 2014
tandem with the increases in violence. In 2013, the increased number of total attacks are driven aid actors and donor governments in the rate existing conventional wisdoms.
461 aid workers were attacked, representing the by a small number of countries that register Global South and the business-like actions of Aid worker attacks Battle deaths
most violent year on record against aid workers. many more aid worker attacks than what is humanitarian aid agencies may be potentially Explaining attacks on aid workers Figure 2: Trends in armed conlict and attacks on aid workers, 1997–2014
commonly seen, led by, in particular, Afghani eroding the value and neutrality of humanitar
The number of aid worker attacks from stan and Syria. ian space. Other scholars believe that this places Drawing on a cross-national sample between attacks – from 20 attacks in a year to more than available for analyzing this hypothesis than for
1997-2014 are shown in Figure 1. The size of INGO staf at greater risk, and that increasing 1997 and 2014, we used a quantitative approach 40 attacks per year. This relationship holds most of the others.
the bubbles correspond to the number of aid Why are attacks on the rise? Scholars and prac institutional politicization and deeper ties with to test over one dozen factors to better when we diferentiate between diferent types
worker attacks recorded in the country over the titioners are increasingly trying to understand governments and/or militaries is considered a understand the spatial and temporal distribu- of attacks on aid workers, with one exception: An international military presence does
period. For reference, all countries that have ex where, why, and how aid workers become targets key factor motivating violence against INGO tion of attacks on aid workers. Our research Kidnappings are much less common in low not add to aid worker risk
perienced armed conlict over the same period of violent acts. For many, the key message is workers. examined the nature of conlicts that humani- intensity then in high intensity conlicts. For aid
are in grey. Aid worker attacks occur dispropor simply that it has become more dangerous to tarian agencies operate within; the structure of organizations the lesson here is fairly clear: the Contrary to much existing analysis suggesting
tionally in some countries such as Afghanistan, be an aid worker in the ield, particularly in But do these critiques have merit? It is possible humanitarian operations; and political factors more intense the conlict, the more attacks are an international military presence may increase
Pakistan, and Syria, but many African countries conlict-afected regions. However, we lack that aid workers are indeed targeted much economic conditions that may motivate attacks. to be expected. Unfortunately, these are often risks for aid workers, we ind no efect of a
have also seen high rates of attacks on aid systematic evidence explaining where and why more than before, but the improvement of We ofer six key indings that help to understand precisely the conlicts where aid workers are NATO presence on the estimated number of
workers. aid workers are attacked. This knowledge gap security by INGOs for their ield staf may have attacks on aid workers. most needed. attacks on aid workers. Comparative areas in
which NATO deploys forces see just as many (or
The conlict context of the country General levels of insecurity don’t appear as few) attacks on aid workers as those where
matters – to an extent to inluence the risk of aid worker NATO is not present. The estimated efect
attacks cannot be reliably distinguished from zero, but
We ind, not surprisingly, that aid workers are if anything, it is negative – countries with NATO
much more likely to be attacked in countries Yet unlike the clear risk of civil conlict to aid forces see fewer attacks on aid workers.
experiencing conlict than in peaceful countries. workers, insecurity and criminal violence is
Somewhat surprisingly, though, the diference not a factor related to the number of aid worker We do ind, however that countries that have
between minor and major conlict setting is not attacks. Attacks do not appear to be inluenced large UN peacekeeping operations (PKO)
that large measured in the absolute expected by the homicide rate in the country, a measure deployed, i.e. PKOs with large budgets which we
count of attacks. An average country with a that captures the generalized level of insecurity assume to be highly correlated with the size of
major conlict is likely to see double the aid and violence, the relative presence of criminal the force, see more attacks against aid workers.
worker attacks than a similar country with a groups and other ‘everyday’ threats to humani-
minor armed conlict. This relationship holds tarian workers. Instead, countries with high This somewhat surprising inding is tempered
when we measured conlict intensity through homicide rates see just as many aid worker by the result that the type or mandate of the
the number of battle related deaths as well. We attacks as those with low crime rates. This could PKO matters crucially for understanding
found a strong and signiicant efect of conlict potentially be good news for aid organizations, attacks on aid workers. PKOs with a traditional
intensity on the expected number of attacks. An perhaps meaning that their extra precautions mandate (observing the terms of truce or peace
increase of battle deaths from around 400 deaths when deploying to countries that have high agreements, or policing a bufer zone and
per year, a medium intensity conlict, to 2000 levels of homicides have been efective. Note, assisting in negotiating a peace agreement) are
deaths a year, a high intensity conlict, roughly however, that the data for homicides is quite associated with more attacks on aid workers,
Figure 1: Aid worker attacks by country, 1997–2014 doubles the expected amount of aid worker poor; consequently we have much less data whereas PKOs with transformational mandates
w w w. p r i o.o r g /C o n f l i c tTre n d s C O N F L I C T T R E N D S 0 6 2 015
– those designed to address the roots of the from conlict, surely they will not be able to that have functioning state institutions, and
Understanding Attacks on
conlict, such as economic reconstruction and protect aid workers either. But in fact, this does therefore have a lower likelihood of experienc-
institutional transformation (i.e. reform of not hold. We ind a tendency towards more ing government instability, are also associated
police, army, judicial system, elections) – are aid worker attacks in the most authoritarian with lower levels of risk to aid workers.
Humanitarian Aid Workers
not associated with a greater number of attacks countries, yet surprisingly, we do not ind any
on aid workers. This result is in line with recent evidence to support our hypothesis that in- Limitations and forward research
research that shows that only PKOs with trans- consistent states see more aid worker attacks.
formational mandates are efective at reducing This is especially interesting considering the Our study is among the irst to analyze the
the intensity of conlict in a country. In contrast, abundance of evidence showing these countries determinants of attacks against aid workers. As
traditional peacekeeping forces are not equipped in general to be more conlict-prone and instable such, there is scope for reinement.
to use lethal force to protect either themselves, or experience greater rates of criminal or social
civilians, or aid workers. violence. First, data improvements could improve the
quality and extent of reporting of attacks,
The type of conlict – and degree of While some have argued that particularly agreeing on what constitutes an incident, and
civilian targeting – matters less than ‘brutal’ or ‘humane’ types of insurgencies may more accurately deining and determining the
expected inluence attacks on humanitarian actors, we number of workers in the ield. With these im-
www.prio.org/ConlictTrends
Conlict Trends Project
found no evidence indicating that the aims of provements, a better understanding of the risks
We found no evidence indicating that countries the rebels signiicantly inluence the degree involved at the sub-national level and of the There is an increasing demand for
that experience one-sided violence (i.e. countries of aid worker attacks. Conlicts where rebels speciic types of humanitarian initiatives and
deploying humanitarian workers
where either the government or insurgents are
actively targeting civilians) have higher rates of
attacks on aid workers. The efect of one-sided
seek secession or regional autonomy, which in
turn are often conlicts in which rebels have
a strong regional presence and compete with
deployments could be made. As aid agencies
continue to expand across the globe – and to
new areas of conlict and insecurity, such as in to conlict-afected areas. However, • Violent conlicts see more aid worker
attacks
violence on aid worker attacks is essentially zero.
This is encouraging, as periods of one-sided
violence are situations where civilian popula-
the government for service provision, do not
have more attacks on aid workers than conlicts
where the rebels are primarily seeking to
urban settings – the evolving and contested role
of humanitarian space is of high priority for
further study.
this need has expanded the risk
of violent attacks against staf in
• Countries with greater criminal
violence do not have more attacks on
tions are especially vulnerable and needing aid overthrow the government.
the most. That these situations do not appear to The next step in the study of aid worker attacks insecure ield settings. In this aid workers
be more dangerous to aid workers may further
encourage the international community to
provide more extensive support for vulnerable
More developed states have fewer
attacks, and country risk is important to
consider
is to dig deeper into the micro dynamics of
attacks. We have the data on conlict dynamics
to do such an analysis, but the crucial missing
brief, we identify six country-level • International military forces do not add
to aid worker risk, but International
populations. ingredient is ine-grained information from aid factors that can inluence attacks Peacekeeping Operations do
•
In general, we ind that the more developed the organizations themselves about who, where,
Regime type and rebel organizational country, measured in terms of GDP per capita, and what aid workers are doing. Regardless, on aid workers. These six factors Groups who actively target civilians do
structure matter less than expected the fewer aid worker attacks. The same pattern
holds for other measures of state capacity and
by further reining the conditions for violence
against humanitarians, aid organizations of
help us to better understand the not attack aid workers more
A long-standing literature has found that au-
tocracies and democracies are roughly as good
at containing violence within their territories
state consolidation, namely ‘time since regime
change’ and ‘time in peace’. We ind that higher
capacity states see dramatically fewer attacks on
all types can better prepare for and prevent hu-
manitarian attacks.
causes of aid worker attacks, and
may ultimately guide towards their
• More democratic and economically
developed countries pose fewer risks
– they have roughly equal levels of armed aid workers, presumably as they are better able Further Reading for aid workers
conlict. Semi-democracies, often called hybrid to protect aid workers present in the country. prevention in the future. This
or inconsistent states, in contrast have much
higher rates of conlict. We expected this to The same holds where countries are assessed
Hoelscher, Miklian & Nygård (2015).
Understanding Violent Attacks against Humani- policy brief is the irst from an
hold for attacks on aid workers as well. If semi-
democracies are unable to protect themselves
by international organisations as being of lower
political risk. This indicates that countries
tarian Aid Workers. PRIO Working Paper.
emerging PRIO research program
on Humanitarian Security.
THE AUTHORS THE PROJECT PRIO
Kristian Hoelscher, Jason Miklian and Håvard The Conlict Trends project aims to answer The Peace Research Institute Oslo (PRIO) is a
Mokleiv Nygård are Senior Researchers at questions related to the causes of, conse- non-proit peace research institute (estab-
PRIO. Their research examines how political quences of and trends in conlict. The project lished in 1959) whose overarching purpose
institutions, commerce, aid and peacekeeping will contribute to new conlict analyses within is to conduct research on the conditions for
affect peace and conlict processes. areas of public interest, and works to produce peaceful relations between states, groups and
krihoe@prio.org thorough and quality-based analysis for the people. The institute is independent, interna-
jason @prio.org Norwegian Ministry of Foreign Affairs. tional and interdisciplinary, and explores is-
havnyg@prio.org sues related to all facets of peace and conlict.
C O N F L I C T T R E N D S 0 6 2 015 w w w. p r i o.o r g /C o n f l i c tTre n d s
