1. Cost of Conflict and Food
Security in the ESCWA Region
Vito Intini
Section for Emerging and Conflict Related Issues
6 February 2012

2. • Violent conflicts strengthen the popular conviction that
conflict is a development curse. But exactly how much does
it cost?
• Several studies have tried to provide an answer to this
question, but no consensual answer has emerged. In
particular, point estimates tend to vary considerably across
different studies.
• Moreover, the empirical assessment of the cost of conflict
poses a number of relevant methodological challenges that
are often overlooked in the existing literature.
• Finally, very little is known on the relationship between
conflict, on the one hand, and FS and development, on the
other, in the ESCWA region.

3. Outline
• Concepts and Stylized Facts
• Methodological Framework
• Transmission Channels
• Measurement of Impact of Conflict
• Conflict-FS Nexus and Challenges
• Recommendations

4. Concepts & Stylized Facts
• The most cited study in this area is Collier (1999). He finds
that civil war reduces the growth rate of per-capita GDP by
2.2% a year (10-year war would reduce pc GDP by ~25%).
• Collier also shows that after long civil wars the economy
recovers rapidly, whereas after short wars (more intense) it
continues to decline for at least five years.
• Rodrik (1999) argues that outbreaks of social conflict are a
primary reason why national economic growth lacks
persistence.
• Cerra and Saxena (2008) estimate that output declines 6% in
the immediate aftermath of a civil war. However, output
rebounds quickly, recovering half of the fall within a few years.
• Glick and Taylor (2010) find large and persistent impact of
interstate war on trade, national income, and global economic
welfare. The negative effect on trade is strong.

5. Concepts & Stylized Facts
• The conventional wisdom that interstate war disrupts
economic activity and hence reduces growth does not go
unchallenged.
• An increase in military expenditure might boost aggregate
demand and hence increase income through a multiplier
effect. However, this would likely be a short-term effect,
while in the long-term the growth potential of the economy
should decrease because of crowding-out effect on private
investment.
• A more general argument is that the occurrence of a war
would not alter the long-term trend of per-capita income, but
only cause a transitionary deviation from the steady state.

6. Methodology Issues
• Most of the existing literature estimates reduced form
models: models where per-capita income (or growth) is
regressed on a measure of conflict duration (or occurrence)
and a set of controls. However, these controls are often
variables that are themselves affected by war. This in turn
causes possible multicollinearity between conflict and the
controls and biases the estimates of the conflict effect.
• Some authors recognise this problem and exclude controls
that are most likely to be affected by conflict (for instance
Collier, 1999). However, if the excluded controls are
significant determinants of development (as measured by
income or any other variable), then an omitted variables
problem arises and estimates of the effect of conflict are
again likely to be biased (i.e. underestimated and/or not
significant).

7. Methodological Framework
• In our study, a structural model of conflict and development is
estimated for a large sample of countries over a period of >40
years (divided in sub-periods of 5 years each). In this way,
conflict is allowed to affect development both directly and
indirectly via its effect on the controls (institutions, policy,
investment).
• The structural model consists of a set of equations that capture
the effect of conflict on per-capita income and various other
(MDG) variables for ESCWA and other war-torn economies
while separating interstate from intrastate conflicts.
• Equations are estimated by taking into account dynamic effects
as well as endogeneity between conflict and development.
• Source of information on conflicts is the Correlates of War
project (COW): i) conflict duration, ii) number of deaths (more
challenging).

8. Preliminary General Findings
Civil Conflict Interstate Conflict
Av. Duration Av. Deaths Av. Duration Av. Deaths
World 0.27 106 0.07 271
ESCWA 0.39 105 0.13 903
• Most of the negative association between conflict time and growth/income
level is driven by civil conflicts. Interstate conflict duration appears to be
uncorrelated with growth and/or per-capita income level.
• One additional year of civil conflict is directly associated with a decline in
per-capita GDP growth of ~0.6% a year.
• However, we have not yet determined direction of causality.
11 12 11 12
10 10 8
Rate of growth of per-capita GDP
8
Rate of growth of per-capita GDP
Per-capita GDP (in log)
9 9
Per-capita GDP (log)
4 4
8 8
0 0
7 7
-4 -4
6 6
-8 -8
5 5
4 -12 4 -12
0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5
Years of intrastate war (in a quinquennium) Years of intrastate war (in a quinquennium) Years of interstate war (in a quinquennium) Years of interstate war (in a quinquennium)

9. Effect of Conflict on Institutions
Variable Estimate
Lagged Institution Index Quality 0.79 (0.01)***
Lagged GDP p.c. 0.43 (0.55)
Civil War -0.24 (0.56)
Inter-state War -2.57 (0.89)**
Civil War (ESCWA) -0.80 (0.12)***
Inter-state War (ESCWA) 0.50 (1.14)
ESCWA -0.30 (0.34)
Fuels 0.00 (0.00)
Legor 0.19 (0.31)
Ethnic 0.29 (0.50)
Malaria -0.08 (0.02)***
Constant 3.46 (0.55)***
Observations 503
Points on a scale 1-10. Robust standard errors in parenthesis; ***estimate is significant at 99% , **estimate is
significant at 95%, *estimate is significant at 90%; Endogenous variables: Civil War, Inter-state War, Civil
War (ESCWA), Interstate War (ESCWA), Fuels.

10. Effect of Conflict on Trade
Variable Estimate
Lagged Trade 0.89 (0.05)***
Civil War -1.51 (0.70)**
Inter-state War 0.96 (0.83)
Civil War (ESCWA) 1.43 (2.93)
Inter-state War (ESCWA) 14.19 (28.00)
ESCWA 4.30 (5.00)
GDP 6.22 (2.52)**
GDP p.c. -6.88 (3.52)*
Landlocked -23.84 (16.62)**
Country Surface -5.12 (1.64)***
Constant 54.78 (22.79)**
Observations 696
Figures in %. Robust standard errors in parenthesis; ***estimate is significant at 99% , **estimate is
significant at (95%), *estimate is significant at 90%; Endogenous variables: Inter-state War, Inter-state War
(ESCWA), GDP, GDP p.c.

11. Effect of Conflict on Investment
Variable Estimate
Lagged Investment 0.53 (0.07)***
Civil War -7.25 (3.55)**
Inter-state War 0.61 (0.41)
Civil War (ESCWA) -2.55 (0.69)***
Inter-state War (ESCWA) -1.13 (7.99)
ESCWA -0.61 (1.04)
Volatility -0.00 (0.00)***
Financial Openness 0.03 (0.24)
M2 -0.08 (0.02)***
Institutions 0.54 (0.60)
Trade 0.42 (0.15)***
Latitude 4.24 (3.33)
Landlocked -1.24 (1.69)
Malaria -0.34 (0.12)***
Constant 7.91 (2.98)***
Observations 551
Figures in %. Endogenous variables: war_intl, war_intl_escwa, war_civil, war_civil_escwa, M2,
Institutions, Trade, Fin_open, Volatility.

12. Effect of Conflict on MDG Dimensions)
HDI Social Immuniz. Life Average Access Child Inequal.
Expect. School. to Sanit. Mortal.
Civil War -0.00 -0.09*** -1.53* 0.30 -0.01 -0.41 1.32 -0.32
Inter-state -0.92*** -0.02 -4.12 -0.55*** -0.21** 0.32*** 3.33** -4.11
War
Civil War -0.13 0.13 1.94 -1.38*** -0.03 -0.88 2.43** 15.23***
(ESCWA)
Inter-state -1.63 -2.73*** -26.47 -1.78 -0.26 -0.03*** 0.02 -0.06
War
(ESCWA)
GDP p.c. 10.63*** 1.96*** 30.19*** 8.13*** 2.32*** -0.81 -0.01 2.36*
Human = av. years of schooling * life expectancy; Immunization = % children aged 12-
23 months that have been immunized against measles.; Social: weighted av. of (i) av.
years of education in the population, (ii) life expectancy, and (iii) rate of children
immunization. Access to water, maternal mortality, child malnutrition are not significant
after controlling for PC GDP.

13. Estimated Impact of Conflict on Development
Civil Conflict (%) Inter-state Confl (%)
All ESCWA All ESCWA
Countries Countries
GDP p.c. (without Investment) -14.70 -17.40 -8.03 -8.03
GDP p.c. (with Investment) -15.15 -18.50 -7.06 -7.06
GDP p.c. (direct war effect) -1.70a -1.70a -1.90b -1.90b
HDI -7.80 -9.20 -8.85 -8.85
Social Indicators -24.67 -27.44 -10.36 -19.32
Immunisation Rate -8.24 -9.35 -3.31 -3.31
Life Expectancy -1.98 -2.93 -1.96 -1.96
Average Years of Schooling -4.56 -4.56 -6.34 -6.34
Access to Sanitation Facilities -6.86 -6.86 -9.01 -12.12
Child Mortality 12.11 14.02 15.02 15.02
Income Inequality -0.80 37.5 -0.40 -0.40
Notes: Estimated effect of one additional year of war. For human development, immunization rate, life
expectancy, average years of schooling, access to sanitation facilities, child mortality, and income inequality,
the percentage is computed on the sample mean value of the indicator. For social development, the
percentage is computed on the sample standard deviation of the indicator.

14. Conflict – FS Nexus
LOG10(foodimpshr) vs FSI
ceryld vs FSI
2.0
10000.0
LOG10(foodimpshr)
1.2
6666.7
YEM
ceryld
0.3 SDN
3333.3
PAL
IRQ
YEM SDN -0.5
0.0 0.0 40.0 80.0 120.0
0.0 40.0 80.0 120.0 FSI
FSI
PubHealth vs FSI
LOG10(kcal) vs FSI
12.0
3.6
LOG10(kcal)
PubHealth
8.0
3.4
SDN
PAL
YEM
3.3 4.0
IRQ
YEM
SDN
3.1 0.0
0.0 40.0 80.0 120.0 0.0 40.0 80.0 120.0
FSI FSI

15. Post-conflict Challenges (1)
• Service provision in post-conflict is particularly challenging
• Populations needs are most acute and particularly concentrated in
this phase
• Public institutions and governance are particularly weak and
fragmented (center-local divides and horizontal cleavages)
• Elites capture and patronage risk is higher
• Different political interests, approaches, and parallel diplomatic,
military, and development-sponsored activities, reflected in
fragmented institutional mechanisms
• Limited capacity of military aid and its lack of coordination with ODA
• The short-term aid governance approach often prevails over the
long-term one (PIUs vis-à-vis governmental staff, shopping lists,
subcontracting of essential services, fragmented short-term small
scale projects)
• Risk of emergency/aid dependence

16. Post-conflict Challenges (2)
Crowding-in or out Effect of Military Expenditures vs Social Expenditures? (% of GDP)
Military Public education Public health
expenditure expenditure expenditure
Country average average average
2000- 2005- 2001- 2005- 2000- 2005- Conflict-affected
2004 2009 2004 2008 2004 2009 countries tend to
Bahrain 4.40 3.36 - 3.10 2.79 2.66
Egypt 3.24 2.50 4.80 4.06 2.32 2.12
have over-inflated
Iraq 2.38 4.54 - - 1.01 2.58 wage bills that can
Jordan 5.74 5.22 - - 4.86 5.30 reach up to 1/3 of
Kuwait 6.92 3.79 6.30 4.24 2.49 1.92 government budget
Lebanon 4.91 4.29 2.64 2.46 3.44 3.93
(OPT, Iraq, Yemen)
Oman 11.89 9.86 4.05 3.72 2.56 1.99
Qatar 3.88 2.25 2.14 - 2.37 2.07
Saudi Arabia 9.80 8.93 7.27 5.97 2.99 2.78
Sudan 3.77 4.24 - - 1.07 1.88
Syria 5.49 4.22 - 5.09 2.21 1.45
UAE 8.40 5.73 1.84 1.09 2.44 1.81
Yemen 6.16 4.62 9.63 5.15 2.42 1.62
ESCWA average 5.92> 4.89> 4.83> 3.88~ 2.71< 2.67<
Arab Region 6.40 5.28 5.07 3.95 2.57 2.44
EAP 1.49 1.58 3.98 3.79 4.69 4.28
LAC 1.35 1.36 4.18 3.96 3.19 3.51
World average 2.39 2.48 4.33 4.45 5.64 5.76

17. FS, Vulnerability, and Conflict
- Vulnerability to shocks makes poor households risk averse in their
asset-allocation strategy and this aversion is even more pronounced
in conflict-affected areas often characterized by a cascading series
of a combination of conflict-related, natural shocks, and international
price shocks with the result of ever-decreasing food-security levels
and passing up more risky but more profitable businesses.
- Complementing social protection with government supported forms of
insurance (but eventually market-based) can help the poor to
improve their risk copying strategies
Natural Conflict-driven Commodity
Shock Shock price Shock

18. Review of Food Aids
• Food Aid highly volatile in the last two decades partly due to donors’
policies not always based on needs assessment
• Food Aid not a tool to address long-term FS
• Need for closer coordination among relief agencies and for medium-
term budgeting
• Food Aid works when local economic policy is addressed, it is
temporary and limited to contexts lacking institutions and markets
• Need to coordinate short-term relief ops with medium-term
considerations (WFP P4P)

19. Volatility of Humanitarian and Food Aid
1600
3,000,000
1400
2,500,000
1200
Afghanistan
Angola Yemen
1000 2,000,000
DRC Syria
Ethiopia Sudan
800
Mt
Indonesia 1,500,000 OPT
Iraq Lebanon
600
OPT Jordan
1,000,000
Somalia Iraq
400
Sudan Egypt
500,000
200
0 0
Period 1990-2008
1998 2000 2002 2004 2006 2008 2010

20. But aid is also volatile at the regional level
Deviation of Gulf state ODA and crude oil price from 1970–2007 average
20

21. Governance deficit – conflict trap (I)
The risk of renewed conflict in countries with good governance
drops rapidly after conflict. In countries characterized by poor
governance, this process takes much longer. Hence, improving
governance is an important part in reducing conflict, and good
governance will in turn decrease the likelihood of conflict.
21

22. Governance deficit – conflict trap (II)
Voice and Accountability Government Effectivness Rule of Law
0 0 0
-0.2 -0.2 -0.2
-0.4 -0.4
-0.4
-0.6
-0.6
-0.8 -0.6
-0.8
-1 -0.8
-1.2 -1
-1
-1.4 -1.2
-1.6 -1.2 -1.4
Regulatory Quality Control of Corruption
Political Stability
1996
1998
2000
2002
2003
2004
2005
2006
2007
2008
2009
2010
0
0
0
-0.2
-0.5 -0.2
-0.4
-1 -0.4
-0.6
-0.6
-0.8 -1.5
-0.8
-1 -2 -1
-1.2 -1.2
-2.5

23. Importance of land rights in conflict (I)
Insecure or weakly enforced property rights:
• Increase risk of expropriation, which diminishes incentives to invest
and to produce
• decrease productivity by necessitating the need to defend property
in insecure environments
• do not allow productive assets to be transferred to those who can
use them for their livelihood (subsistence farmers) or most
productively (commercial farmers)
• do not allow a crucial asset supporting other transactions such as
obtaining financing
• can cause both grievance and greed
• exacerbate over-urbanization and socio-economic disparities in the
aftermath of a conflict
• are an important driver of food security and therefore of social
stability given that food price shocks hit the landless hardest and
increase the incidence of riots
23

24. Importance of land rights in conflict (II)
• Friction or even open conflict between modern state-led
(+ colonial) and customary or tribal property rights
systems
• Property disputes constitute the single largest cause of
backlog cases in the judicial systems of most post-
conflict countries including LDCs (great entry point to
reform the judiciary)
• Post-conflict government capacity is weak in terms of
systematic and consistent law enforcement and
informal governance systems tend to be relatively
stronger
• However, recent experiences have shown that
remedies to this inherent contradiction are possible
(DRC, Mali, Zimbabwe)
24

25. Recommendations (I)
• The breadth and depth of available data require immediate and
significant improvements.
• Conflict - Poverty (and informality) – Food Insecurity compound into
inter-generational transmission of poverty, inequality, and conflict
relapse
• One year of a civil conflict in the region is enough to cancel
development progress made by an average ESCWA country in 5-10
years
• The potential presence of endogeneity suggests an existence of a
‘feedback-loop’ whereby the process of conflict and de-development
feed on each other. In such instances, a ‘circuit-breaker’ is required
to break this symbiotic vicious cycle.

26. Recommendations (II)
• Governmental action can be directed at reactivating the channels through
which public goods are delivered.
• Closer regional integration and better business environment can increase
the peace dividend on these countries.
• Expansion of family (and nutrition support) programs especially in rural
areas to improve maternal health, provision of education with specific focus
on girls education, and reduce fertility rates.
- Support community-based development funds for infrastructure and services
→ labor generation, social capital and peace dividends
- More use of JAM, joint CAS, PRSPs, SWAPs
- Regional action aimed at tapping Sudan’s agricultural potential in a
transparent manner
- Review land policies in conflict-affected environments
- Learn lessons from food aid and relief modalities and how to link them to
livelihood programs (WFP’s P4P)
- Reform CERF and establish a food aid emergency reserve (ASEAN) or
multi-annual regional fund managed by WFP/OCHA

27. Recommendations (III)
• Strengthen governments’ efforts to preserve water resources
and increase their rational utilization.
• Agriculture (and the rural poor) in these countries is trapped
in low value-added activities and mainly characterized by low
productivity farming. It is therefore vital to increase efficiency
in rain-fed agriculture, enhance ag R&D, and promote
market linkages.
• Donors need to move towards multi-year budgeting for their
transfers to the relief agencies working in a conflict-affected
country so that they can improve their planning and increase
efficiency gains in their operations. Due to absorptive
capacity constraints aid disbursements should gradually rise
during first 3-5 years and then slowly revert back.

28. Recommendations (IV)
• With respect to the issues specifically raised in this report,
future research should provide an even more
comprehensive account of cross-regional differences in the
development costs of war.
• More work will be needed in the future in order to provide
theoretical foundations to our findings: why is the negative
effect of conflict sometimes associated with interstate
conflict while other times is associated with civil conflict?
Why is inequality disproportionally affected by conflict in this
region? And what can we do to reduce the overall effect?

29. Enhance Regional Cooperation through FS
Conflict and FS are two regional problems that require a strong
regional approach
• How can cooperation be promoted?
– Create Awareness: Previous Pan-Arab cooperation efforts were
driven by ideology, not necessity. Cooperation in light of common
threats to the conflict-food security nexus is a necessity
– Build Political Commitment: EU and ASEAN were launched to
promote peace and stability. The obstacles to successful regional
integration were not necessarily easier than in the Arab world today
– Provide Incentives: Cooperation must pay a peace dividend, financed
by transfers and the creation of new economic opportunities. It is a
positive sum game.
– Promote specific actions: reinvigorated supranational actions need to
be undertaken with specific responsibilities in the area of social
assistance programs, rural development, gender and family planning,
and the advancement of an Arab common market.
– Manage External Support: The international community can help
increase the opportunity cost from engaging in conflict through
greater regional economic cooperation, including the dismantling of
protectionism (by trade partners) and more efficient development aid.

30. Thank you
intini@un.org