Tilman Brück, Negar Habibi, Charles Martin-Shields,
Astrid Sneyers, Wolfgang Stojetz and Stijn van Weezel
ISDC – International Security and Development Center, Berlin
www.isd-center.org
FAO, Rome,
7 March 2017
The Relationship between Food Security and Violent Conflict: Summary of a Report to FAO
1. The Relationship between
Food Security and Violent Conflict:
Summary of a Report to FAO
Tilman Brück, Negar Habibi, Charles Martin-Shields,
Astrid Sneyers, Wolfgang Stojetz and Stijn van Weezel
ISDC – International Security and Development Center, Berlin
tilman.brueck@isd-center.org
www.isd-center.org
FAO, Rome
7 March 2017
ISDC
policies for
peace and
prosperity
2. Overview
1. Introduction
2. Conflict-affected and Fragile Countries
3. Correlating Food Security, Conflict and Fragility
4. From Conflict to Food Security
5. From Food Security to Conflict
6. Need for New Data
7. Policy Implications
ISDC
policies for
peace and
prosperity
3. Project Overview
Objective
– to explore and map the analytical and empirical relationships
between food security and conflict
Team
– ISDC (with advice from FAO)
Duration
– August 2016 – February 2017
Outputs
– joint FAO-Households in Conflict Network workshop in 10/2016
– 5 working papers and 1 policy report by ISDC (not yet online)
– various products by FAO
Further details
– www.isd-center.org and www.hicn.org
ISDC
policies for
peace and
prosperity
4.
5. Call for Papers on Conflict, Aid and Humanitarian Assistance.
Papers due 15 May 2017. HiCN Workshop on 15-16 Nov 2017.
7. 1: War leads to hunger.***
2: Rising food prices cause political instability.**
3: Improved food security may strengthen peace.*
4: Across all settings: institutions matter!**
food
security
hunger
peacewar
1
3
4
target
Key
Findings
2
8. Overview
1. Introduction
2. Conflict-affected and Fragile Countries
3. Correlating Food Security, Conflict and Fragility
4. From Conflict to Food Security
5. From Food Security to Conflict
6. Need for New Data
7. Policy Implications
ISDC
policies for
peace and
prosperity
9. Key Questions
Which conflict parties?
– interstate, intrastate and internationalized intrastate conflicts
Which conflict intensity and duration?
– high (above 1000 battle deaths) or low (25-1000 battle deaths)
– some conflicts are extremely sticky
Which direction of violence?
– one-sided violence
Beyond violence: fragility
– strength of formal and informal institutions
Overlap of categories possible
10. Global Conflict Trends, 1996-2014
0
5
10
15
20
25
30
35
40
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Intrastate conflict Interstate conflict Internationalized Intrastate conflict
This series includes a high share of countries
experiencing very long intrastate conflicts.
11. 11
Conflict
Types 2011-17
Inter Intra Int. Intra One-sided Years Conflict
2011-14
# Types
Afghanistan Yes Yes 4 2
Burundi 0 3
CAR Yes Yes Yes 4 3
Chad 0 0
North Korea 0 0
DRC Yes Yes Yes 3 3
Djibouti 0 0
Eritrea 0 0
Ethiopia Yes Yes 4 2
Haiti 0 0
Kenya 0 0
Liberia 0 0
Niger 0 0
Somalia Yes Yes 4 2
South Sudan Yes Yes Yes 4 3
Sudan Yes Yes Yes 4 3
Syria Yes Yes 4 2
Yemen Yes Yes 4 2
Zimbabwe 0 0
Total/Mean 2 6 6 8 1,84 1,32
FAO Protracted Crisis List, 2017 Update
12. Deconstructing Fragility: The Macro Level
• Traditional focus on ‘the state’ and the aggregate
level of observation in discussing ‘fragility’
• Yet, no clear definition of what constitutes a ‘failed’ or
‘fragile’ state
• And, obviously, the experience of fragility varies
across groups or even individuals
• OECD (2015: 40) argues for precisely such a
multidimensional approach, based on a range of
different indicators
• 3 key domains, from our perspective:
• human security
• economic inclusion
• social cohesion
including subjective indicators
13. Deconstructing Fragility: The Micro Level
Source: Brück et al (2017), Microeconomic Foundations of Fragility, AERC, Nairobi.
Individual-level Fragility Indicators in Kenya, 2016
governancegap
centralgovernment
14. Overview
1. Introduction
2. Conflict-affected and Fragile Countries
3. Correlating Food Security, Conflict and Fragility
4. From Conflict to Food Security
5. From Food Security to Conflict
6. Need for New Data
7. Policy Implications
ISDC
policies for
peace and
prosperity
15. Key Messages
Over time, countries in Protracted Crisis shift between different
categories of conflict, but rarely achieve full peace:
– Shifts could be in intensity of the conflict or in the actors involved.*
As the scale of a conflict increases, the impacts on food security tend
to be felt at different levels:
– Low intensity, localized conflict is related to low Utilization and Availability.*
– Larger scale conflict is related to Access as well as Utilization and Availability.*
– Fragility correlates with low Availability.*
There is a polarisation around some very conflicted, food insecure
countries and other protracted crises.*
Variables that matter are conflict type, duration and intensity,
conflict-fragility interactions, level of observation, and food
security dimension.*
17. • Ability to distribute food
nationwide.
• Organize imports and
exports.
• Capacity to balance
import versus domestic
production.
• Domestic infrastructure for food
distribution.
• Health and state services to
address malnutrition.
• Government extension
services for household food
security.
• Inability to distribute
food.
• Exposure to global price
shocks.
• Lack of coherent food
production or import
policies.
• Inability to manage food
distribution.
• Low/limited infrastructure.
• Lack of capacity to meet
nutritional public health needs.
• No extension services.
High
Fragility
High
Administrative
Capacity
Systemic
Food Security
Individual
Food Security
Correlating Food Security, Fragility and
the Level of Observation
18. Low-intensity
Conflict
Internationalized
Intrastate Conflict
Fragility
Exposure to Global
Price Shocks
Food Insecurity +
Domestic Price
ShocksFood Insecurity
Type 3:
Country is functional in
some other ways, but at high
risk of price shocks beyond
its control that could lead to
conflict
Type 2:
Conflict is intensive enough to
include external actors; impacts
on population (under-
nourishment) and domestic
food prices are noticeable.
Type 1:
Conflict is localized enough
that impacts are not felt in
nationwide food prices; only
impact is on people local to
the violence, expressed in
the form of under-
nourishment.
Scenario is likely if a civil
conflict is increasingly drawing
in external actors, exposing the
country to wider global price
issues, plus destruction of
domestic crop capacity
Conflict is lower intensity and any
external participation is at a small
scale; total scope of violence is
not larger enough to impact
national food markets.
Administrative capacity is very
low, and aside from distribution
problems, policy or economic
issues have led to food shortages.
Risks include food riots, and urban
violence.
Administrative capacity is low, and
food security is a problem of
distribution instead of supply. Any
potential violence would be
localized and due to relative
deprivation.
Country is fragile, and has
noticeable levels of localized or
low intensity violence, but food
security issues are only noticeable
in terms of exposure to global
markets.
Localized conflict leads to
undernourishment in affected
areas, while food price issues
are more likely due to policy or
administrative issues as
opposed to the low intensity
conflict.
Typology of Conflict, Fragility and Food Security
19. Low-intensity
Conflict
Internationalized
Intrastate Conflict
Fragility
Exposure to Global
Price Shocks
Food Insecurity +
Domestic Price
ShocksFood Insecurity
Type 3:
Country is functional in all
other ways, but at high risk of
price shocks beyond its control
that could lead to conflict
Type 2:
Conflict is intensive enough to
include external actors; impacts
on population
(undernourishment) and
domestic food prices are
noticeable.
Type 1:
Conflict is localized enough
that impacts are not felt in
nationwide food prices; only
impact in on people local to
the violence, felt in the form
of undernourishment.
Scenario is likely if a civil
conflict is increasingly drawing
in external actors, exposing the
country to wider global price
issues, plus destruction of
domestic crop capacity
Conflict is lower intensity and any
external participation is at a small
scale; total scope of violence is
not larger enough to impact
national food markets.
Administrative capacity is very
low, and aside from distribution
problems, policy or economic
issues have led to food shortages.
Risks include food riots, and urban
violence.
Administrative capacity is low, and
food security is a problem of
distribution instead of supply. Any
potential violence would be
localized and due to relative
deprivation.
Country is fragile, and has
noticeable levels of localized or
low intensity violence, but food
security issues are only noticeable
in terms of exposure to global
markets.
Localized conflict leads to
undernourishment in affected
areas, while food price issues
are more likely due to policy or
administrative issues as
opposed to the low intensity
conflict.
Conflict, Fragility and Food Security
For example: Rwanda,
Angola, Haiti, Myanmar,
Djibouti, Ethiopia, Chad,
Georgia, Afghanistan,
Rep. of Congo
For example: Angola,
Iraq, Uganda, Rep. of
Congo, Rwanda, Sierra
Leone, Yemen
For example: Djibouti,
Papua New Guinea,
Solomon Islands,
Vanuatu, Yemen,
Mauritania, Comoros,
Somalia, Liberia
20. Overview
1. Introduction
2. Conflict-affected and Fragile Countries
3. Correlating Food Security, Conflict and Fragility
4. From Conflict to Food Security
5. From Food Security to Conflict
6. Need for New Data
7. Policy Implications
ISDC
policies for
peace and
prosperity
21. Key Messages
Conflict has a significant negative impact on food security, whether
measured in terms of access, utilization, stability or availability.***
The reason for fighting has a statistically significant impact on the
severity of conflict-induced food insecurity.**
– Conflicts over control of the government tend to show wider impact on food
insecurity than conflicts over control of territory.
– Conflicts over control of the entire country will have larger effects on food
systems than conflicts over a specific geographic region.
Intrastate conflict in drought-affected settings has a significant
impact on increases in the number of underweight individuals.**
Conflict may have negative impact on access to food.*
– For example, fighting itself often destroys existing infrastructure. While already
weak states may also lack infrastructure and are more likely to suffer conflict,
these trends are accentuated by fighting, which in turn worsens food security.
23. Case Study: Macro-level Trends
Data on food supply
– calorie intake per day per capita (FAO Food Balance Sheets)
Data on conflict
– country-years with at least 25 battle-related fatalities
(UCDP/PRIO)
Coverage
– 106 countries in the Global South (Africa, Asia, Latin America)
in the period 1961-2011
Methods
– regression analysis to establish causality
24. Case Study: Weak Positive Association
Peace Duration and Food Supply Levels
25. Case Study: Findings
• On average, conflict is negatively associated with food
supply levels.
• Conflict onset is associated with 65% drop in growth of
food supply.
• This effect is larger for conflicts with high intensity and
about power.
• Small differences in food supplies during conflict and
non-conflict years.
• But there is some empirical evidence that longer peace
durations are associated with higher food supply levels.
• And not all countries seem to suffer the same negative
effect of conflict on food supply levels.
26. Case Study: Disclaimers
• Using data aggregate at the national level provides some
useful insights into the macro-level trends, but this comes
at the cost of missing sub-national variation both in food
security and conflict.
• This makes it hard to pin down the exact channels and
mechanisms.
• Moreover, the data used might be biased towards urban
consumers.
27. Overview
1. Introduction
2. Conflict-affected and Fragile Countries
3. Correlating Food Security, Conflict and Fragility
4. From Conflict to Food Security
5. From Food Security to Conflict
6. Need for New Data
7. Policy Implications
ISDC
policies for
peace and
prosperity
28. Key Messages
Generally speaking, the impact of variations in food security on the
likelihood of violent conflict are difficult to identify statistically.*
National level indicators of food security tend to change relatively slowly
and thus lack the variance necessary to demonstrate statistically how they
can lead to conflict (even if they do in practice).*
Sub-national analyses of food security and violence show promising results
for understanding how anthropometric and production measures of food
security can impact conflict.**
– For example, evidence from subnational analysis of rainfall and production in Ethiopia
indicates that variation in food production and access has a significant effect on the
likelihood of local violence.**
– For example, analysis at the household-level in Somalia indicates that food production and
availability pressures, such as drought, have a statistically significant impact on conflict
likelihood. This relationship is not observed at the district level, indicating that
understanding channels from food security to conflict may be best observed at the
household level.**
30. Tackle Endogeneity Food Security ↔ Conflict
Key strategies
– Cross-sectional analysis
– Panel data analysis (incl. lag specifications)
– Exploit exogenous variation in treatment
• Controlled experiments
• Natural experiments
– Exploit exogenous variation correlated with treatment (IV)
Key issues
– Main concern: omitted variable and simultaneity bias
– For key questions controlled experiments are not available
• Can‘t randomize conflict
• Can‘t randomize climate
30
31. The Case of Ethiopia: Research Questions
• Are higher annual precipitation levels associated with
lower probability of conflict onset?
• Has precipitation affected conflict through affecting
total production?
Ho: ↑ Rainfall ↑ Production ↓ Conflict
32. Case Study: Methodology and Data
Method:
– Variations in precipitation used as a proxy for exogenous
shocks to food security
Data:
– Conflict: Geo-coded data from the Armed Conflict Location and
Event Data (ACLED)
– Annual Precipitation: Geo-coded data from PRIO-GRID at
0.5×0.5 decimal degrees resolution (55km×55km)
33. Case Study: Main Results
Precipitation levels have a negative and statistically
significant effect on the probability of conflict onset
The same result holds across different types
– low intensity, intra-state and non-state conflict events
Precipitation affects probability of conflict onset through
affecting production levels
Other variables affecting conflict:
– Higher percentage of agricultural area in a geographic unit is
associated with higher probability of low intensity conflict
– Political exclusion of ethnic groups has a positive and
statistically significant impact on probability of conflict
– More populated areas are more likely to experience conflict
34. Case Study: Discussion
Lack of precipitation results in lower production levels
which in turn increases the probability of conflict.
Agriculture dependent areas are
– more vulnerable to weather conditions
– more likely to experience low intensity conflict during low
precipitation periods
Diversification of production sources and lower reliance
on rainfall dependent agriculture will decrease the
sensitivity of total production levels to weather
conditions and in turn reduce the risk of violent conflict.
35. Overview
1. Introduction
2. Conflict-affected and Fragile Countries
3. Correlating Food Security, Conflict and Fragility
4. From Conflict to Food Security
5. From Food Security to Conflict
6. Need for New Data
7. Policy Implications
ISDC
policies for
peace and
prosperity
36. Addressing
Data Challenges
Hard to collect
complete and
consistent global
cross-sectional data
for understanding
conflict
Add conflict
variables to
standard survey
tools to correlate C
and FS within
surveys (Brück et al
2016a)
Match and merge
multiple data sets
from different
sources
Strengthen the capacity
of food security experts
on conflict and of conflict
experts working on food
security, also with a view
to strengthening multi-
disciplinary data (Brück et
al 2016b)
Explore and analyze
administrative (or
project) data from
emergency settings,
e.g. from enhanced
M&E frameworks
Statistical
Capacity
Micro
Macro
Meso
Use geo-coded
remote sensing data
to track socio-
economic
developments and
conflict
Conduct RCTs on
interventions for
peace and food
security (Brück et al
2015)
Learn from case
studies to build up
robust evidence base
Use data from
mobile phones to
overcome data
collection
challenges
Strengthen
statistical
capacity will also
reduce fragility
and improve
accountability
37. Overview
1. Introduction
2. Conflict-affected and Fragile Countries
3. Correlating Food Security, Conflict and Fragility
4. From Conflict to Food Security
5. From Food Security to Conflict
6. Need for New Data
7. Policy Implications
ISDC
policies for
peace and
prosperity