The Economic Crisis, Violent Conflict, and Human Development

UNITED NATIONS DEVELOPMENT PROGRAMME













The Economic Crisis, Violent Conflict, and Human Development



A UNDP/ODS Working Paper


By Namsuk Kim and Pedro Conceição











Office of Development Studies
United Nations Development Programme, New York
—May 2009—


Note: The views expressed in this paper are those of the authors and do not necessarily reflect
those of UNDP. The authors thank Emmanuel Letouzé and Bartholomew Armah for helpful
comments. Nina Thelen provided excellent research assistance. Please send comments and
suggestions to the following e-mail addresses: Namsuk.Kim@undp.org,
Pedro.Conceicao@undp.org.

---

The Economic Crisis, Violent Conflict, and Human Development


1. Introduction

The unfolding global economic crisis is expected to bring the world economy into recession in
2009. Figure 1 shows the population weighted real GDP growth from 1991 to 2009 (estimates
for 2008 and projection for 2009) for the world economy and for different groups of countries.
The annual real GDP growth rate of the global economy was 5.1% in 2007 but the world
economy is projected to shrink by -1.3% in 2009 (IMF, 2009). Emerging and developing
economies are also projected to suffer a sharp slowdown as a result of the crisis, with a projected
growth rate of 1.6% in 2009 compared to 8.3% in 2007. For many developing countries, the
sharp economic slowdowns will translate into deep recessions. The UN projects that 15
developing countries will have negative per capita growth in 20091, while projections from the
World Bank adjusting for terms-of-trade changes increase this to 502.

A recent strand of literature, reviewed in some detail in this paper, suggests that economic
conditions are important determinants of the outbreak and recurrence of conflict. In particular,
wars often start following growth collapses (Collier et al, 2009, p.15). Sharp economic
slowdowns and low levels of income per capita appear to increase the likelihood of conflicts. In
this context, it is opportune to explore insights from this literature, linking it also with the human
development implications of both growth slowdowns and conflict. In particular, the paper
highlights the risks of the emergence of low-human-development/conflict traps.

Given that the probability of conflict recurrence is high, as elaborated upon below, post-conflict
countries – those that have experienced armed conflicts until recent years – may be particularly
vulnerable.3As Figure 1 shows, post-conflict countries are projected to have a substantial
decrease in the economic growth, from 7.4% in 2007 to 3.1% in 2009. Advanced economies may
have a sharper slowdown (2.7% in 2007 and -3.8% in 2009), but they have well-developed social
protection, and stable political systems that may facilitate the recovery and absorb the pressures
for social instability and conflict. In contrast, post-conflict countries, may be more vulnerable to
a more protracted and slower recovery from the slowdown, given the higher risks of conflict
recurrence.

                                                            
1 http://www.un.org/esa/policy/wess/wesp2009files/wesp2009.pdf 
2 http://siteresources.worldbank.org/INTGLOMONREP2009/Resources/5924349-
1239742507025/GMR09_overview.pdf 
3 A post-conflict country refers to a country with armed intra-state conflicts that ended, or significantly diminished,
after the end of the Cold war (see UNDP 2008a, p.7). 
1

---

Figure 1. Real GDP growth (annual percentage change)

Note: 2009* is a projected value.
Source: Real GDP growth rate is obtained from IMF (2009); The list of post-conflict countries is from UNDP
(2008a, Table 1.2); The population weight is from World Bank (2009).

Drawing on a review of both theoretical and empirical literature, this paper frames the the
connection between economic factors and conflict within a conceptual framework in which
levels of human development and the risk of conflict are linked. Violent conflict is one of the
most extreme forms of suppressing choices and advancing rights, and therefore a major threat to
human development (UNDP, 2005, p.151). Since 1990, more than 3 million people have died in
armed conflicts in developing countries (Marshall, 2005). The total war deaths are far more than
the battle deaths. For example, the total war deaths are estimated as 1.2 million in Ethiopia
during 1976-1991, but only 2% of them were directly engaged in the battles. (Bethany and
Gleditsch, 2004) Conflict has also non-lethal consequences that may last across generations
(UNDP, 2008a).

As far as drivers of conflict are concerned, one of the most robust findings in the literature is that
many economic conditions (low income, slow growth, and especially severe economic
downturns) are correlated with the outbreak of conflict, with some evidence strongly suggesting
that the causal direction runs from economic conditions to conflict (Collier and Hoeffler, 2004).
There is also a rich literature on the impact of horizontal inequality and dependence on natural
resources as drivers of increases in the risk of conflict. This paper however focuses only on the
economic factors, reviewing the findings in light of the current economic crisis and the severe
economic downturn that it now occurring.

When it comes to the consequences of conflict, there is no doubt that it is harmful to human
welfare, but it becomes even more hazardous if conflict results in a persistent low human
2

---

development/conflict trap. A typical country reaching the end of a civil war faces a 44 percent
risk of returning to conflict within five years (Collier et al, 2003, p. 83). Whether or not a country
will experience a new civil war can be best predicted by whether the country experienced wars in
the past (Collier, et al, 2004).

The high rates of recurrence of conflict, along with the economic determinants of conflict,
suggest the possibility of the existence of poverty-conflict traps (Collier et al, 2003; Bloomberg
and others 2000). Given that poverty and low per capita income are also correlated with worse
health and education outcomes, and also that these outcomes suffer as a result of conflict, the
conflict trap can be conceptualized in the framework of a low human development – conflict trap
(Collier and Hoeffler, 2004; Justino and Verwimp, 2006; Alderman, Hoddinott and Kinsey,
2004).

A self-reinforcing circle from conflict to low human development, and vice versa, is suggestively
illustrated below (Chart 1). Conflict destroys accumulated physical and human capital, forces
replacement of labor, deteriorates institutional capacity. A country experiencing conflict cannot
secure long term returns for investments in both in physical and human capital, resulting in low
investment in health and education. All of these factors lead to low levels of human development.
A country with low levels of human development has more difficulty in improving institutions,
and in increasing productivity and potential growth. In turn, lower growth rates heighten the risk
of conflict, potentially trapping a country in the loop.
3

---

Chart 1. Low Human Development – Conflict Trap
 
Conflict
• Loss of life 
• Low institutional 
•
capacity 
Destruction of assets 
•
• Low productivity 
Forced migration 
•
• Low potential growth 
Low investment in 
health and education 
• Slow growth 
Low Human 
Development 


The remainder of the paper discusses the empirical findings and theoretical background for
linkage between the low-human development and conflict. Section 2 considers how low levels of
human development can affect the risk of violent conflict. Section 3 shows how the conflict can
result in low human development, completing the vicious circle. Section 4 concludes the paper
with a brief discussion on the policy responses.
2. From Low Human Development to Conflict
While there are number of factors that could cause conflict, empirical studies find that poor
economic performance is associated with higher incidence of conflict. Being a poor country is
correlated with most forms of violence (UNDP, 2008a). Figure 2 shows that economic development
and conflicts are observed to be clearly related. The level of GDP is negatively correlated with
observing a new conflict. Collier et al (2009) find that the predicted risk for a hypothetical
country with characteristics set at the study’s sample mean was 4.6 per cent. If the level of per
capita income were to be halved from this level, the risk would be increased to 5.3 per cent.

4

---

Figure 2. GDP per capita and the probability of observing a new conflict
 
Source: Humphreys (2003, p.2), as reported d in UNDP (2008a).

Growth rates are also strongly associated with risks of conflict in developing countries. If the
growth rate in developing countries is increased by one percentage point from the mean, the risk
of conflict decreases by 0.6 percentage points to 4.0 per cent (Collier et al, 2009). Kang and
Meernik (2005) show that the growth rate in conflict countries in the five years prior to conflict,
including cases of conflict recurrence, was on average 0.5 percent compared to 2 percent in
countries that remained peaceful.

Empirical analysis of growth and conflict has inherent data limitations, but some recent studies
using more careful methodology shows a strong causal link running from poor economic
performance to conflict. One problem is that the direction of impact between the income per
capita and conflict can run both ways. Assuming a priori one-way causality – that is, ignoring
endogeneity – in regression analysis can result in biased estimates. Other information used in the
empirical studies, such as income inequality, population, ethnic distribution, are also subject to
difficulties of econometric identification and data quality (Hegre and Sambanis, 2006; Sambanis,
2004).To address the endogeneity problem, some studies adopt instrumental variable analysis,
using a strictly exogenous variable that moves with income per capita, but not with conflict. For
instance, Miguel, Satyannath and Sergenti (2004) use annual changes in rainfall data as an
instrument for income growth. The rainfall data predicts growth fluctuation in agricultural
economies in Africa. They find that income shocks are drive conflict. Besley and Persson (2008)
and Bazzi and Blattman (2008) use international commodity price and trade shocks as the
exogenous variables, but they find that the evidence on the relationship between economic
shocks as drivers of conflict is mixed.

5

---

Other components of human development other than income levels/growth rates may also affect
the risk of conflict. Education outcomes are closely linked with the outbreak of conflict. Collier
and Hoeffler (2004) find strong evidence that higher levels of secondary school attainment are
associated with a lower risk of civil war. If the enrollment rate is 10 percentage points higher
than the average in their sample, the risk of war is reduced by about three percentage points (a
decline in the risk from 11.5 percent to 8.6 percent). This draws on date that refers to the period
between 1960 and 1999 for developing countries.

Very few countries with low Human Development Index (HDI) show high levels of political
stability (the higher the indicator, the higher the level of political stability). Figure 3 plots one
political stability indicator and HDI for 178 countries. High values of this political stability
indicator imply that the country suffers less violence. The figure suggests that high HDI (say,
above 0.5) does not guarantee high political stability. However, low HDI (below 0.5) is clearly
associated with political instability.

Figure 3. Political Stability (2007) and Human Development Index (2006) for all countries

Note: Political Stability is obtained from World Bank (2008); Human Development Index is from UNDP (2008b).
Number of countries is 178.

Going now to the theory behind the outbreak of conflict, a great number of potential channels
and mechanisms have been studied through which social, political and economic factors can
cause conflict. Following Blattman and Miguel (2009), four distinctive models are briefly
reviewed in this section: 1) Contest model; 2) Rational behavior with asymmetric information or
imperfect bargaining; 3) Collective action and selective incentives; 4) Feasibility hypothesis.

The most well-known framework is the contest model where two competing parties allocate
resources to production and appropriation (Garfinkel, 1990; Skaperdas, 1992). Garfinkel and
6

---

Skaperdas (2007) employs conventional optimization techniques and game-theoretic tools to
study the allocation of resources among competing activities - productive and otherwise
appropriative, such as grabbing the product and wealth of others as well as defending one's own
product and wealth.

When the conflict is regarded as a deviation from equilibrium between players, it could occur
when the players either act irrationally, or act rationally with asymmetric information or
incomplete bargaining (Fearon, 1995). Acemoglu and Robinson (2006) demonstrated the
existence of an equilibrium in a bargaining process between the rich and the poor. A number of
studies, including Powell (2007), Esteban and Ray (2008), Chassang and Pedro-i-Miquel (2008),
and Dal Bo and Powell (2007), show how conflict is sometimes unavoidable with asymmetric
information or in multi player settings. Conflict can also occur when the bargaining process is
incomplete, that is, credible commitment to maintain peace cannot be made (Powell, 2006;
Walter, 2006; Garkinfel and Skaperdas, 2000).

Even if contest or rational behavior models are close to reality that there exists underlying
tension between two groups, the formation of coalition and participation in conflict of each
individual is a different problem (Olson, 1971, 1982). Grossman (1999) and Fearon (2007)
consider monetary incentives to motivate participation in conflict. Walter (2004) shows the
absence of alternative to conflict (non-violent change) could form rebel groups, while Gates
(2002) and Mkandawire (2002) argue that organization of selective punishment/treat can be the
critical force of individual participation. Ethnic nationalism is often argued to be the leading
cause of conflict, but sometimes it is strategically used to coordinate and enforce cohesion even
when it is not the cause of conflict (Fearon, 2006).

Economic drivers matter in part because it is economic characteristics that make a
rebellion/conflict feasible. They enable fighting parties to buy the weapons, to maintain a private
army over long periods, and to perpetrate large scale killing without endangering themselves
(Collier, 2006). This theory, so called, the “feasibility hypothesis” is being tested with aggregate
and micro level data.

In summary, many dimensions of human development affect the risk of conflict through
behavior of economic agents and institutional capacity. Blattman and Miguel (2009), after
extensively surveying literature on this topic, concluded that low per capita incomes, slow
economic growth and geographic conditions favoring insurgency are the factors most robustly
linked to civil war.

3. From Conflict to Low Human Development

There is a huge literature on the consequence of conflict for economic growth. Rodrik (1999)
argues that domestic social conflicts are key to understanding why growth rates lack persistence
and why so many countries have experienced a growth collapse after the mid-1970s.
Econometric evidence shows that countries that experienced the sharpest drops in growth after
1975 were those with divided societies (as measured by indicators of inequality, ethnic
7

---

fragmentation etc) and with weak institutions of conflict management (proxied by indicators of
the quality of governmental institutions, rule of law, democratic rights, and social safety nets).
Cerra and Saxena (2008) show that, compared to the average recession, those that are associated
with civil wars are ten percentage points deeper and last for ten more months (Table 1). They
also show that economic contractions are not always followed by offsetting fast recoveries and
adverse shocks may lead to absolute divergence and lower long-run growth. Chen et al (2008)
also provides rich analysis on the effect of war on changes in the level and growth of GDP.

Table 1. Civil Wars lead to Deeper and Longer Recessions
 
Cumulative loss 
Duration (years) 
Number of recessions 
(percent of GDP) 
All episodes 
‐7.5 
1.62 
637 
Low income 
‐7.1 
1.58 
259 
Low‐middle income 
‐10.0 
1.84 
163 
Upper‐middle income 
‐8.6 
1.67 
97 
High income 
‐4.1 
1.38 
118 
 
 
 
 
Financial crisis 
‐6.8 
1.64 
182 
Banking crisis 
‐11.7 
2.19 
104 
 
 
 
 
Civil wars 
‐17.4 
2.42 
60 
Source: Cerra and Saxena (2008)

Conflict leaves severe impacts on poverty. Civil war and genocide in the 1990-2000 period in
Rwanda caused convergence between provinces following the conflict shocks: previously richer
provinces in the east and in the north of the country experienced lower, even negative, economic
growth compared to the poorer western and southern provinces. This has in turn affected
significantly the dynamics of household poverty in Rwanda in the same period (Justino and
Verwimp, 2006).

Conflict has major consequence in all aspects of human development, not only on income
poverty. A simple illustration in Table 2 shows that life expectancy, infant and under-5 mortality,
and secondary school enrollment are worse during war, and improve during post-war periods
(three subsequent years after the end of war) on average.

Table 2. Selected Human Development Indicators, 1960-2006, Global Average
War/peace 
Life expectancy 
Infant mortality 
Under‐5 mortality  Secondary school 
(at birth) 
(per 1,000 live 
enrollment rate 
birth) 
During War 
59.6 
84.2 
121.2 
57.2 
During Post‐war 
59.2 
75.2 
113.1 
57.8 
During peace 
67.1 
39.8 
52.5 
66.6 
Total 
66.1 
47.6 
59.7 
65.9 
Source: Population weighted sample averages. Identification of war is obtained from Collier and Hoeffler 
(2001, p.21‐22). Human development indicators are from World Bank (2009). 
8

---

Some evidence suggests that the human development impact of conflict is significant in the long-
run. Alderman, Hoddinott and Kinsey (2004) find that the war in Zimbabwe had a permanent
effect of malnutrition on children, resulting in lower labor productivity. Bundervoet, Verwimp,
and Akresh (2008) also show that young children affected by civil wars display much worse
health-scores relative to non-affected children. The study examines the impact of Burundi's civil
war on children's health status and finds that children exposed to the war have on average 0.515
standard deviations lower height-for-age z-scores than non-exposed children.

Conflict has negative impact on education attainment. Shemyakina (2006) finds that from 1992
to 1998, exposure to the conflict, as measured by past damage to household dwellings, had a
significant negative effect on the enrollment of girls of ages 12-15. Girls who were of school age
during the conflict and lived in conflict affected regions were 13% less likely to complete
mandatory schooling as compared to girls who had the opportunity to complete their schooling
before the conflict started, and 7% less likely to complete school than girls of the same age group
who lived in regions relatively unaffected by conflict.

Many studies considered transmission channels through which conflict can affect growth rates.
Grossman and Kim (1996) and Gonzalez (2003) argue that the diversion of resources from
productive to unproductive activities caused by conflict limits potential economic growth. Lloyd-
Ellis and Marceu (2003) points out that the return to investment in physical and human capital is
not secured when conflict is present. Derger and Sen (1983) argue that military spending against
possible conflict crowds-out investment in more productive activities (see also Kalyoncu and
Yucel 2005).

Conflict has severe social and economic consequences. People lose their physical or social assets,
they often flee and become refugees, and lose opportunities to invest in the health and education
of younger generations.

4. Conclusion

An adverse economic shock can be dangerous because its impact may be long-lived if countries
are forced into a vicious cycle of low human development and conflict.

It is important to pursue policies that mitigate the risks of conflict outbreak, especially when
countries are facing, as in the context of the current crisis, sharp economic slowdowns.
International efforts on peacekeeping can play significant role in peace building. Doyle and
Sambanis (2006) review the evidence on United Nations peacekeeping missions and find that
they are associated with a higher likelihood of peace two years after the end of the war. Their
multivariate regressions show that international capacity and hostility variables are very robust
and local capacity variables less so due to their competitive interaction with other covariates.
Fortna (2008) also demonstrates that peacekeeping is an extremely effective policy tool,
dramatically reducing the risk that war will resume. Relatively small and militarily weak
9

---