Violent Conflict and Gender Inequality:
               An Overview


   Mayra Buvinic † Monica Das Gupta † Ursula Casabonne † Philip Verwimp


Violent conflict, a pervasive feature of the recent global landscape, has lasting impacts on
human capital, and these impacts are seldom gender neutral. Death and destruction alter
the structure and dynamics of households, including their demographic profiles and tradi-
tional gender roles. To date, attention to the gender impacts of conflict has focused almost
exclusively on sexual and gender-based violence. We show that a far wider set of gender
issues must be considered to better document the human consequences of war and to
design effective postconflict policies. The emerging empirical evidence is organized using a
framework that identifies both the differential impacts of violent conflict on males and
females (first-round impacts) and the role of gender inequality in framing adaptive re-
sponses to conflict (second-round impacts). War’s mortality burden is disproportionately
borne by males, whereas women and children constitute a majority of refugees and the
displaced. Indirect war impacts on health are more equally distributed between the
genders. Conflicts create households headed by widows who can be especially vulnerable
to intergenerational poverty. Second-round impacts can provide opportunities for women
in work and politics triggered by the absence of men. Households adapt to conflict with
changes in marriage and fertility, migration, investments in children’s health and school-
ing, and the distribution of labor between the genders. The impacts of conflict are hetero-
geneous and can either increase or decrease preexisting gender inequalities. Describing
these gender differential effects is a first step toward developing evidence-based conflict
prevention and postconflict policy. JEL codes: D0, D1, I0, I1, J0, J1




In the last 50 years, more than half of all nations have been affected by internal
civil conflicts, defined by Blattman and Miguel (2010) as conflicts with 25 or
more deaths per year, or full-fledged civil wars, with 1,000 or more battle deaths

The World Bank Research Observer
# The Author 2012. Published by Oxford University Press on behalf of the International Bank for Reconstruction and
Development / THE WORLD BANK. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
doi:10.1093/wbro/lks011           Advance Access publication November 29, 2012                               28:110–138
per year. Following a peak in the late 1980s and early 1990s, the incidence of
wars and battle deaths has declined in recent decades, but a substantial number
of people, especially in poorer countries and regions, continue to suffer and are
forced to cope with the consequences of destruction and death caused by violent
conflict (World Bank 2011).
   Researching the effects of violent conflict on economic development is inherent-
ly difficult. Conflict itself is difficult to measure and most often occurs in countries
that are poor and that have weak institutions and infrastructure; these countries
also have sparse preconflict statistics and little data-gathering capacity. However,
there have been growing efforts to empirically document the impacts of conflict.
These efforts have found more lasting impacts on human capital than on physical
capital. Economies may recover, but people often do not (Justino Forthcoming).
These human effects of conflict are seldom gender neutral. Death and destruction
alter the structure and dynamics of households, including their demographic pro-
files and traditional gender roles.
   This overview examines the emerging empirical evidence on both the differential
impacts of violent conflict on males and females and the role of gender inequality
in framing adaptive responses to conflict. The descriptive findings are a first step in
developing evidence-based policy. To date, attention to the gender impacts of con-
flict and war has focused almost exclusively on sexual and gender-based violence.
We show that a far wider set of gender issues must be considered to better docu-
ment the human consequences of war and to design effective postconflict policies.
   We borrow a framework used to identify transmission channels for the impact
of economic shocks on women (see Sabarwal, Sinha, and Buvinic 2010) and use
a modified version to organize our review of the evidence on the gender impacts
of violent conflict. This organizing framework, detailed in figure 1, captures both
the gender-differentiated effects of conflict and the adaptive responses of house-
holds and individuals to conflict. Unfortunately, this framework does not provide
a direct window into the household. The evidence we present is limited by the in-
ability of most research to observe and measure the household processes and dy-
namics that mediate the impacts of violent conflict on individuals and families.
   The conflict literature distinguishes the direct effects of conflicts, such as the
killing, wounding, and physical destruction that result from violence, and indirect
effects on economic performance and human welfare. Similarly, we separate direct
and indirect effects, when relevant, and underscore both first- and second-round
impacts on gender differentials. First-round impacts operate through the following
channels: (a) an increase in mortality and morbidity (especially of men and chil-
dren) as direct and indirect consequences of violence and destruction along with
a higher incidence of widowhood; (b) forced displacement and migration; (c) asset
and income loss due to the disruption of markets, infrastructure destruction and
damage, and deaths of household members; and (d) sexual and gender-based

Buvinic et al.                                                                    111
Figure 1. Possible Transmission Channels for the Gender Impacts of Conflict




violence. Second-round gender impacts resulting from excess male mortality
include the induction of women into political and civic participation and changes
in marriage and fertility behavior. Another set of second-round impacts emerge
as conflict-affected households respond to the loss of family members and declines
in household income and consumption with coping strategies that involve
changes in women’s traditional household roles. Women respond to decreases in
household income by increasing their hours of work, entering the labor force, or
adjusting their time and effort in the home. Women can further cope by altering
their fertility or by migrating, and households can curtail (or increase) their in-
vestments in children’s health or education. We cover these second-round impacts
on children because of the interdependence between women’s choices and child-
ren’s well-being that is especially salient in poor countries and among poor
households. We also reference evidence from the economic crisis and gender liter-
ature that illuminates the gender differentials observed in response to conflict.



Background
Once wars begin, they are “development in reverse.” Wars destroy physical and
human capital, disrupt service delivery, divert public expenditures to the military,

112                                    The World Bank Research Observer, vol. 28, no. 1 (February 2013)
disrupt the efficient functioning of markets and transport infrastructure, and lead
to dissaving, capital flight, and the departure of skilled workers (Collier 2007).
The World Bank (2011, 5– 6) estimated that the average cost of civil war is equiv-
alent to more than 30 years of GDP growth for a medium-size developing country
and that after major episodes of violence, trade levels take 20 years to recover.
Between 1981 and 2005, the poverty headcount rose in countries affected by
major violence, and it fell substantially in countries with minor violence and
sharply in those with negligible violence (World Bank 2011, figure F1.3). Collier
and others (2003) have concluded that conflicts account for part of the growing
income gap between the poorest countries in the world and other countries.
   Conflicts often recur: 90 percent of the last decade’s civil wars occurred in
countries that had experienced a civil war since 1945 (World Bank 2011, table
F1.1). The destruction of conflict contributes to “conflict traps,” in which nations
that are already poor and exhibit lagging human development are caught in
cycles of protracted struggles and entrenched poverty. The problems of fragile con-
flict-affected states can also spread to their neighbors. An estimated 75 percent of
refugees are hosted by neighboring states. Refugees put pressure on the resources
of the host state and can bring disease. The World Bank (2011) estimated that
Tanzania, a country that is making development gains, loses an estimated 0.7
percent of GDP every year for each neighbor in conflict.
   In places where institutions are quickly repaired after a war to ensure political
stability and facilitate the resumption of normal economic activity, there can be
rapid postwar recovery. In Germany and Japan after the Second World War, the
Allied powers helped to build strong institutions of governance, including partici-
patory political institutions. In Vietnam, a commitment to maintain primary
health and education during the war, particularly in the north, and intensive in-
vestment in repairing infrastructure in damaged areas in the postwar period miti-
gated the war’s impact (Miguel and Roland 2005). However, the conditions that
facilitated rapid reconstruction are not present in many modern fragile states.



Methodological Issues
Until recently, there has been relatively little rigorous work on the effects of con-
flict on individuals and households, including its effects on gender roles and in-
equalities, because large-scale, high-quality household surveys are generally not
available for countries affected by violent conflict. Where these surveys are avail-
able, they do not include questions on violent behavior. Even if they have informa-
tion on violence, researchers are confronted with difficult methodological
problems. The foremost difficulty is the rigorous attribution of causality; war-torn
countries and war-torn households often differ from peaceful countries and

Buvinic et al.                                                                    113
peaceful households, but it is virtually impossible to design scientific experiments
to test the consequences of conflict on people randomly assigned to variations in
conflict situations. Furthermore, there may be reverse causality for many vari-
ables of interest: violent conflict may be both the cause and the consequence, for
instance, of increased poverty. Panel studies, which can track reverse causality
and document violent cycles, have been largely unavailable.
   An equally serious methodological hurdle is the lack of reliable baseline infor-
mation on many of the variables of interest in conflict research, such as the levels
of violence before a conflict, recall errors, and possible survival bias when using
ex post measures to approximate baseline information and, especially in the case
of sexual and gender-based violence measures, the possibility that increased inci-
dence may simply be an artifact of improved reporting over time. In addition to
these hurdles, there is a general lack of empirical information on gender variables
at the individual and household levels, difficulty in measuring intrahousehold
issues when investigating gender inequalities, and logistical difficulties and risks
involved in both conducting research and acting as a research subject in conflict
and postconflict situations. Safety and ethical issues arise, especially in investiga-
tions of sexual and gender-based violence and interviews with combatants.
Finally, conflict research generally conceptualizes conflict as a discrete event or
shock, although conflict is a process that evolves over time and recurs in repeated
cycles of violence (Bru ¨ ck and others 2010).
   Despite these limitations, recent research on the consequences of conflict has
advanced and has benefitted from more and better micro-level data, increased use
of innovative approaches, and quasi-experimental variation. A growing number of
longitudinal household-level data sets and follow-up household surveys in post-
conflict settings that integrate prewar data are facilitating new microstudies on
the impacts of war (Blattman and Miguel 2010). Confronted with a lack of data
from household surveys on violence, researchers have merged household data sets
with secondary sources that register violent events and death tolls at the village
or district level. An example is the work of Shemyakina (2011) in Tajikistan,
which used variation in the number of incidents reported in newspapers at the
“raion” (district) level to differentiate violent districts from nonviolent ones.
Seeking a more satisfactory solution, Bru  ¨ ck and others (2010) proposed adding a
generic violence module to standard household surveys. The inclusion of such a
module would remediate or offset the inability of survey-based research to infer
the effects of violent conflict on schooling, health, and labor market outcomes dis-
aggregated by gender.
   To address the issue of attribution of causality, researchers have sought to
control for unobserved heterogeneity correlated with both conflict and outcomes
of interest as much as possible (e.g., using area or household fixed effects).
Researchers have also used instrumental variables to control for endogeneity.

114                                  The World Bank Research Observer, vol. 28, no. 1 (February 2013)
Examples of instrumental variables include proxy measures for the intensity of
conflict, such as Shemyakina (2011) used for Tajikistan. The difficulty in identify-
ing convincing instruments and the impossibility of implementing a randomized
controlled trial forces conflict researchers to confront the limits of any identifica-
tion strategy. Panel data, when available, minimize problems related to recall and
ex post measures and enable researchers to trace the dynamics of conflict over
time (e.g., Guerrero-Serda ´ n 2009). Some studies have constructed panel data by
resurveying households that had been surveyed before the conflict, such as
Bundervoet, Verwimp, and Akresh’s (2009) study in Burundi and Andre and
Platteau’s (1998) study in Rwanda.
   Researchers have often empirically addressed the issue of attribution of causali-
ty using difference-in-differences, a nonexperimental technique that compares a
conflict-affected group (or region) with its preconflict situation and with a control
group (or region) that did not experience the conflict. The validity of this method
is contingent on no other changes between the two groups (regions) at the same
time as the exposure to conflict. Even if there are no other changes, there may be
spillover effects from the conflict-affected group to the nonconflict control group,
leading to an underestimation of the effects of conflict. The more that these studies
contribute alternative explanations and perform careful checks for robustness and
the more that studies by different researchers in different settings are able to
observe regularities in the legacy of conflict on human development and gender in-
equality, the more confident we can be that the result is a valid assessment of con-
flict consequences rather than a spurious finding. Similar findings from related
empirical literature on economic shocks add confidence to the findings.



First-Round Impacts of Violent Conflict
First-round impacts of violent conflict include excess male mortality and morbidi-
ty as an obvious direct and indirect consequence of violent conflict, resulting in
widowhood, sexual and gender-based violence, asset and income loss, forced dis-
placement, or migration. These first-round impacts often result in reductions in
household income and consumption, triggering coping strategies that have
gender implications.


Mortality and Morbidity
Young adult men typically suffer the highest mortality in conflicts, creating a
shortage of working-age males and a high share of females and widows in the
population (figure 2 shows missing men in postwar age pyramids for Cambodia,
Germany, and Russia). The World Bank (2011) estimated that men constitute 90

Buvinic et al.                                                                   115
Figure 2. Estimated Population Distributions by Age and Sex




 Source: Authors’ analysis based on data from the United Nations Population Division (2006).



percent of the missing, whereas women and children constitute 80 percent of ref-
ugees and those internally displaced by violence. According to reports by the
Truth and Reconciliation Commission, males accounted for 74 percent of reported
fatalities in Sierra Leone, 87 percent of reported fatalities in East Timor and 84
percent of reported killings and disappearances in Guatemala (Cohen 2011).
Similarly, Obermeyer and others (2008) reported that males accounted for 81
percent of violent war deaths in 13 countries over the period from 1955 to 2002.
De Walque and Verwimp (2009) used age at marriage to partially control for
being a Tutsi and estimated that adult males and educated people were most
likely to die in 1994, the year of the Rwandan genocide that killed at least
500,000 people. In Kosovo, Spiegel and Salama (2000) found that men were 8.9
times more likely than women to die from war-related trauma.
   The World Health Organization estimated that in the year 2000, there were
310,000 deaths due to wars. These deaths were concentrated primarily among
men aged 15 to 44. Nearly half of these deaths occurred in Africa, where there
were an estimated 32 male deaths per 100,000 population due to war-related in-
juries compared with 15 female war deaths per 100,000 population. The disabil-
ity burden from war was similarly skewed toward men and toward Sub-Saharan
Africa (Krug and others 2002).


116                                            The World Bank Research Observer, vol. 28, no. 1 (February 2013)
   The extent to which males suffer higher mortality than females varies somewhat
with the nature of the conflict. Foreign armies sent to participate in a conflict tend
to consist largely of young men, so excess deaths are highly concentrated among
this group (Buzzell and Preston 2007). When the conflict is on home territory,
women may also suffer elevated mortality, and they do so particularly as an indi-
rect consequence of war. Using census and DHS survey data, Neupert and Prum
(2005) estimated that 65 percent of the approximately 2 million deaths during the
Khmer Rouge occupation in Cambodia were men. As many as 45 percent of all
deaths occurred among people younger than 14 years or older than 60, suggesting
that most of these deaths did not occur in combat but were an indirect result of the
war. These deaths were distributed more equally between males and females.


Indirect Impacts on Health
Violent conflicts affect population health in ways that extend beyond the direct
effects of violence through a combination of increased exposure to infectious disease,
acute malnutrition, poor sanitation, and a lack of health services. The evidence sug-
gests that women and children have more exposure to these indirect effects of war
on health than men do. Indirect effects occur because health and other infrastruc-
ture, such as roads needed for effective health system functioning, may be damaged,
and resources may be diverted away from health (Ghobarah, Huth, and Russett
2003). Vulnerability to disease may be further increased by loss of income and
assets, population displacement, or orphanhood. The spread of HIV may be in-
creased through sexual violence or otherwise increased interaction between civilians
and military groups with higher HIV prevalence rates as well as through increases
in refugee movements and commercial sex, and it may be reduced by the better
social services often found among refugee camps (Hankins and others 2002; Mock
and others 2004). Women are often at higher risk of contracting HIV/AIDS as a
result of displacement and have an increased likelihood of being victims of sexual vi-
olence. In Rwanda, 17 percent of women who were survivors of genocide, and 67
percent of rape survivors were HIV positive (McGinn 2000).
   Refugees can also pose significant challenges to receiving nations. A study of
135 countries from 1962 to 1997 found that for every refugee who left a malaria
endemic zone, there were 2– 2.7 new cases in the receiving area (Montalvo and
Reynal-Querol 2007). Baez (2011) found that the flood of refugees from the geno-
cides of Burundi and Rwanda into a neighboring region of Tanzania had adverse
impacts on the health of local children, including a 7 percentage point increase
in child mortality and an increase of 15 –20 percentage points in infectious dis-
eases. These findings are in line with other studies that caution that neglecting
vaccination and disease control efforts in postconflict settings can lead to devas-
tating epidemics and further fatalities (Connolly and others 2004).

Buvinic et al.                                                                    117
   Reduced access to health care may also increase mortality from chronic diseas-
es among older people, as it did in Kosovo in 1998 –99 (Spiegel and Salama
2000). This reduced access may also have severe negative impacts on women’s re-
productive health, resulting in gynecological problems, unwanted pregnancies,
maternal mortality, obstetric fistula, and preterm babies (McGinn 2009).
   Estimations of the contribution of these indirect effects has been hampered by a
lack of data and difficulty establishing a satisfactory counterfactual of health out-
comes in the absence of war (Murray and others 2002). In a cross-national anal-
ysis of 1999 WHO data on Disability Adjusted Life Years, Ghobarah, Huth, and
Russett (2003) found that the additional burden of death and disability incurred
in 1999 alone from the indirect and lingering effects of civil wars in 1991 –97
was almost double the burden incurred directly and immediately from all wars in
1999. Their results are strongest for infectious diseases and traffic accidents.
Overall, the authors conclude that women and children are the most common
long-term victims of civil war.
   Equally critical is the impact of the experiences of conflict on mental health.
Several studies note poor psychosocial health after exposure to conflict, with chil-
dren and adults demonstrating high levels of traumatic symptoms, stress, anxiety,
aggressive behavior, and depression. However, these studies often do not identify
who was exposed to trauma or compare findings with a control group that was
not exposed to conflict (Ajdukovic and Ajdukovic 1999; Bayer, Klasen, and Adam
2007; Cardozo and others 2004; Husain and others 1998; Johnson and others
2008; MacMullin and Loughry 2004; Mollica and others 1997; Pham, Vinck,
and Stover 2009). A meta-analysis of 25 years of research on sex differences in
trauma and posttraumatic stress disorder showed that females are at greater risk
of developing posttraumatic and depressive symptoms after traumatic events,
whereas males report more exposure to trauma, even when comparing the fre-
quency and severity of war trauma experienced by civilian male and female
victims of war or terrorism (Tolin and Foa 2006). In one of the few studies that
uses a tragic natural occurrence (i.e., random abductions by the Lord’s Resistance
Army) to randomize exposure to violence, Annan and others (2011) reported
that women in Uganda experience greater emotional distress than men do from
war violence, although men and women report similar levels of violence (approxi-
mately 25 percent of both female and male Lord’s Resistance Army abductees
were forced to beat, cut, or murder people). Unexpectedly, most women returning
from these armed groups reintegrate socially and are psychologically resilient.


Widowhood
As result of the excess mortality of men, wars create widows. There is little empiri-
cal evidence on the economic status of these war widows, but the little existing

118                                  The World Bank Research Observer, vol. 28, no. 1 (February 2013)
evidence suggests that widows and the households that many of them head are
                                                                ¨ ck and Schindler
especially vulnerable to the consequences of violent conflict. Bru
(2009) noted the need to empirically study the changes that war widowhood
brings to women and households, including the formation of female-headed
households and subfamilies when war widows and their children take refuge in
larger, male-headed households. These authors specifically examined the case of
widows in the Rwandan genocide. Most of the more than 500,000 deaths from
genocide were men, resulting in unbalanced sex ratios after the genocide. These
authors found that genocide widows and their children face different and often
more severe constraints in earning incomes because of the loss of economic re-
sources tied to men, including male labor and land ownership, and the destruc-
tion of social networks. They hypothesized that widows may constitute a group of
households that are affected differently by conflict and that require specific
attention.
   In follow-up work, Schindler (2010) compared male-headed households with
households headed by widows in postgenocide rural Rwanda, controlling for the
unequal distribution of size and composition between these household types. That
study found that 23.3 percent of rural households are female-headed. Widow-
headed households have a higher incidence of poverty and extreme poverty when
compared to male-headed households, but they show a less pronounced gender
division of labor, suggesting that the conflict triggered a change in traditional
gender roles within the home. Schindler (2010) also showed that the work inten-
sity of teenage girls, adult women, and widows is significantly higher in conflict
regions with fewer males in the population, holding household wealth and com-
munity infrastructure constant (Schindler 2010).
   A recent analysis of widows in Mali corroborates findings on the poverty and
disadvantage faced by widows independent of a conflict situation, presenting evi-
dence on potentially long-term negative consequences for children, particularly
girl children (in terms of nutritional status), even when a widow remarries (van
de Walle 2011). Widowhood is also strongly associated in developing countries
with poor mental health (Das, Friedman, and Mckenzie 2008). The welfare costs
of male deaths due to conflict may be reproduced intergenerationally through the
constraints faced by widows.


Sexual and Gender-based Violence
Sexual violence (rape and sexual abuse) and other forms of gender-based violence
(domestic abuse and beatings) have become distressingly common features associ-
ated with violent conflict, although it is extremely difficult to obtain reliable esti-
mates of their incidence and prevalence. These types of violence can be a direct
weapon of war used for ethnic cleansing and for punishing opponents, although

Buvinic et al.                                                                    119
carefully collected evidence questions the extent to which such violence has oc-
curred. More commonly, this type of violence may be a crime of opportunity facili-
tated by the general breakdown of social order, a climate of impunity, and the
contagion effect of war violence. Such violence has been widely reported, for
example, in the former Yugoslavia in the mid-1990s, where it is estimated that
20,000 women were raped; the Rwandan genocide in 1994, with estimates that
approximately 300,000 to 400,000 women suffered rape; Somalia in the early
1990s; the conflict in Kashmir; the 15-year-long civil war in Peru; and the recent
civil war in Sudan (See, for example, McGinn 2000; El Jack 2003; Human Rights
Watch 1995, 1996; McGinn 2000; Swiss and Giller 1993). A global review of 50
countries found significant increases in gender-based violence following major
wars (World Bank 2011).
   Estimates of sexual and gender-based violence can suffer in both wartime and
peacetime from serious underreporting (i.e., because people are unwilling or
afraid to report gender-based violence, especially when the perpetrator is a family
member) or overreporting, when incidence statistics are inflated because reporting
improves with time (Norda   ˚ s and Cohen 2011). This situation also occurs in
peacetime, making it very difficult to accurately assess the increases in sexual and
gender-based violence that are associated with conflict.
   Recent evidence highlights variations in the prevalence of sexual and gender-
based violence in war situations and relates this variation to combatant norms
and group cohesion. This evidence shows that Bosnia and Rwanda are anomalous
cases of wartime rape being used as a war weapon for ethnic cleansing. In most
other cases, sexual and gender-based violence is a crime of opportunity that is
often committed by relatives rather than strangers (Wood 2009, 2006). A study
of sexual violence in a pediatric ward in Goma, Democratic Republic of the
Congo, showed the predominance of domestic sexual violence over militarized
rape. Of 500 pediatric cases treated for sexual violence at the hospital (2006 –
2008), nearly all were females between the ages of 10 and 18 (Kalisya and others
2011). Also in the Democratic Republic of the Congo, the results of a population-
based household survey with a randomly assigned module on sexual violence (a
subsample of 3,436 women) yielded a very high prevalence of rape—an estimate
of more than 400,000 women were raped in the 12 months prior to the 2007
survey—and showed that the most pervasive form of sexual violence was from in-
timate partners and that the most conflict-affected provinces were at higher risk
of sexual violence (Peterman, Palermo, and Bredenkamp 2011).
   A population-based random cluster survey of adults in Liberia (conducted in
2008) is one of the few quantitative studies on the legacy of sexual violence in
conflict situations. The study showed that both female and male combatants who
experienced sexual violence (42 percent of all female combatants and 33 percent
of all male combatants) had worse mental health outcomes than noncombatants

120                                 The World Bank Research Observer, vol. 28, no. 1 (February 2013)
and former combatants who did not suffer sexual violence (Johnson and others
2008). This study made the important observation that men are also victims of
sexual violence in conflict situations, but their victimization is largely underre-
ported. The empirical literature on the consequences of domestic violence in
peacetime has shown lower productivity and earnings among female victims of
intimate partner violence (Morrison and Orlando 1999). Sexual and gender-
based violence triggered by conflict may have similar economic outcomes, trap-
ping victims and their families in poverty. It may also have psychosocial and
reproductive health consequences. Research that tracks sexual and gender-based
violence victims over time alongside appropriate controls is needed to assess the
full consequences of violent conflict and to design appropriate interventions.


Migration and Displacement
Violent conflict sometimes disrupts normal life to the extent that people are forced
to move elsewhere in the hope of finding more secure conditions for themselves
and their families. These movements are sometimes massive, as was the case in
Mozambique, where residents of approximately half of all households were inter-
nally displaced or were refugees by the end of the war (Bru  ¨ ck 2006). By the end
of 2009, it is estimated that 42 million people had been forced to leave or flee
their homes due to violence—15 million refugees outside their country of resi-
dence and 27 million internally displaced persons, with women and children
comprising 80 percent of all refugees and internally displaced people (World Bank
2011). Such circumstances leave women almost entirely alone to care for their
families under very difficult circumstances.
   Displaced people typically face significant asset losses and major economic and
social difficulties in the resettlement process. By cutting off large numbers of
people from economic opportunities, internal conflict can lead to a vicious cycle
of displacement and household poverty from which it is difficult to escape. This
situation is made worse by the destruction of social networks and the consequent
depletion of important elements of people’s social, economic, and political capital,
including the previously mentioned constraints that accompany widowhood and
female headship.
   In Colombia, internal displacement has both income- and gender-differential
labor effects. Annual labor income among displaced persons fell by 80 percent
and recovered to only half of its predisplacement level after more than a year
(Iban˜ ez and Moya 2010). Displaced females work longer hours, earn similar
wages, and contribute in larger proportions to household earnings than do rural
women who do not move, but this increased contribution does not seem to
strengthen their household bargaining power (Calderon and others 2011).
Researchers have observed increased domestic violence against women and

Buvinic et al.                                                                   121
violent punishment of children, both of which may be related to the traumatic
events that led to displacement.
   Kondylis (2010) found that in Bosnia, more able individuals are more likely to
be displaced but are less likely to find work after displacement than the remaining
population. In a study of the economic performance of people who returned to
Rwanda after its 1997 resettlement policy for refugees from previous conflicts,
Kondylis (2008) found that returns to farm labor are higher for returnees relative
to stayers and suggested that such higher returns occur because upon their
return from conflict-induced exile, returnees are more motivated to increase their
economic performance. There may also be a positive selection into refugee status
among more able individuals than observed in Bosnia.
   Despite generally bleak conditions, refugee camps can sometimes offer better ser-
vices than those available in the refugees’ places of origin. However, people may
linger in refugee camps as internally displaced people for years, if not decades. One
review of global displacement trends estimated the average length of displacement
at 14 years (Norwegian Refugee Council 2004). In some settings (e.g., Burma,
southern Sudan, Burundi), higher proportions of refugee children were in school
than those who stayed behind. Examining the 10 settings that produced the largest
number of refugees as of 2002, the Women’s Commission for Refugee Women and
Children (2004) found that among refugees, enrollment levels decrease after grade
4 for both boys and girls, with gender differentials steadily increasing; enrollment
rates were 20 percent higher among boys than girls in first grade and were nearly
twice as high at the secondary level. Refugees and the internally displaced reflect
highly gender-specific consequences of violent conflict and civil wars.


Asset and Income Loss
Conflicts impose shocks on the population. Recent empirical literature has begun
to measure the substantial costs of violent conflict on economies and communi-
ties. These costs encompass the most immediate and observable consequences of
war, such as damage to the national productive structure and the redirection of
resources from productive to military uses, as well as more indirect effects on
households’ assets and income and other attributes of economic well-being.
   Violent conflict damages public resources and services (World Bank 2011). It
destroys school infrastructure, displaces teachers, and interrupts schooling, often
for years. In Sudan, the destruction of infrastructure included the destruction and
closure of schools and hospitals. Immunization, preventive medicine, and malaria
eradication services stopped completely in the south, and malnutrition affected
most of the children in the region (Teodosijevic 2003). Violent conflict also
damages transportation networks and other infrastructure, crippling production
in the primary and secondary sectors. In Mozambique, rail and other rural

122                                  The World Bank Research Observer, vol. 28, no. 1 (February 2013)
infrastructure were severely affected (Bru ¨ ck 2006). As the government shifts ex-
penditure toward the military, public investment and expenditure on maintenance
experience pressure that further deteriorates public infrastructure. The postcon-
flict legacy of these effects can be extensive.
   Private assets are also lost or destroyed during fighting and looting. These
assets include houses, land, utensils, livestock, and other productive assets.
Livestock is a key form of savings and an insurance substitute in many developing
countries that helps to smooth consumption in adverse times. Agricultural pro-
duction was badly affected during the war in Mozambique, and 80 percent of
cattle stock was lost (Bru ¨ ck 2006). In northern Uganda, many people lost all of
their cattle, homes, and assets (Gersony 1997; Annan and others 2006). During
the 1994 Rwandan genocide, the average cattle stock was halved, and 12 percent
of households lost their homes (Berlage and others (2003) In Tajikistan, the civil
war damaged the homes and livelihoods of 7 percent of households between
1992 and 1998 (Shemyakina 2011). The Burundi conflict in the 1990s was ac-
companied by severe asset depletion (Bundervoet, Verwimp, and Akresh 2009). In
Colombia, violence significantly affected the efficiency of farm holdings as a result
of the disruption of rural labor markets and limits imposed on the operation of
larger farms (Gonzalez and Lopez 2007).
   These losses can impoverish societies and create or entrench household poverty,
leading to persistent cycles of war and poverty (Justino 2006, Forthcoming).
However, armed conflict can also result in new opportunities, such as firms and
households that take advantage of the redistribution of assets that often follows
conflict (Justino Forthcoming).
   The ability of households to respond and adapt to the asset and income shocks
of war depends on a number of factors, including the often overlooked issue of ci-
vilian agency (Justino Forthcoming). Women are often severely asset and land
constrained, making it difficult for them to manage farms and households in the
absence or death of men. However, women are resourceful and can play important
roles in helping households adapt to the shocks of war. Gender roles and inequali-
ty are clearly important in terms of how individuals and households experience
the loss of assets and income during conflict and how they accommodate these
losses. These factors help to explain the interaction between violent conflict and
poverty and the channels through which violent conflict can perpetuate house-
hold poverty, as described in the section below on second-round impacts.



Second-Round Impacts of Violent Conflict
The massive human losses from violent conflict can exacerbate gender inequali-
ties, as described above, but other impacts of violent conflict can reduce these

Buvinic et al.                                                                  123
inequalities by altering the landscape of opportunities available to women. Several
studies have found that households recover some of their earlier economic indices
(often roughly measured) within years or decades and manage to limit the
damage to human capital accumulation and losses to family members. In Sierra
Leone, households’ experiences of war victimization during the 12-year period of
civil war from 1991 to 2002 had no strong or consistent association with house-
holds’ assets, although the study only examined ownership of a stove or radio in
2005 (Bellows and Miguel 2006). In Rwanda, McKay and Loveridge (2005) and
Justino and Verwimp (2006) found that per capita GDP reached its pregenocide
level less than 10 years after the genocide but with increasing levels of inequality.
   However, other studies have found longer-term impoverishment following
violent conflict, especially for low-income households. In a small panel of rural
households in Rwanda, Justino and Verwimp (2006) found that households
whose dwelling was destroyed or who lost land experienced a significant decrease
in average incomes. They found a 15 percentage point increase in households
that were classified as extremely poor between 1990 and 2002. In a large cross-
sectional study, McKay and Loveridge (2005) found that rural incomes in Rwanda
had recovered to their (very low) 1990 preconflict levels by 2000. However, this
recovery was due largely to the improved position of the richer households, who
could take skilled labor jobs. In contrast, income levels did not recover for nearly
70 percent of the rural population (especially for the poorest 30 percent), who
had small landholdings and little access to wage income.
   The way that households recover from the shocks produced by violent conflict
is part of their adaptive responses, and women often assume the primary respon-
sibility for ensuring the survival of families. Women take over this pivotal eco-
nomic role especially when working-aged males have died or have joined (or been
forced to join) fighting units and when families are forced to move internally or to
another country, and women and children form the majority of those displaced
(World Bank 2011).
   Below, we review the empirical evidence on some of the second-round impacts
of violent conflict on individuals and households. These impacts are associated
with responses that differ by gender, including adaptive responses by households
to the violent shock. The demographic changes triggered by the sex-unbalanced
mortality and morbidity of conflict alter or change marriage and fertility patterns
and create opportunities for political participation among those who have been
formally excluded. These changes are reviewed next. The destruction of assets and
the disruption of state and market institutions due to conflict require households
to accommodate sudden sharp declines in household income and consumption.
Households reallocate labor between the genders and reallocate resources as-
signed to children’s well-being to cope with the aftermath of these conflicts.



124                                  The World Bank Research Observer, vol. 28, no. 1 (February 2013)
Changes in Marriage and Fertility
People in conflict zones alter marriage and childbearing patterns to minimize the
disruptive effects of conflict on their household economy. Households that experi-
ence a decrease in income often defer marriage expenditures and childbearing
until times are better. This phenomenon has been found in studies of the effects
of armed conflict and in many studies of economic shocks. For example,
Jayaraman, Gebreselassie, and Chandrasekhar (2009) found that in Rwanda,
women living in areas that were more exposed to violent conflict during the 1994
genocide (as measured by the proportion of sibling deaths in 1994) were more
likely to marry and have children later than those living in clusters that experi-
enced less violence. Fertility can be depressed as a result of lower coital frequency,
as couples are separated by male out-migration and male combat duties, and
poor nutritional status and stress lower fecundity and increase spontaneous abor-
tions (Blanc 2004).
    Conflict has an additional dimension beyond other kinds of shocks in that con-
flict-related excess male mortality creates shortages of potential grooms. This sit-
uation may increase the search and dowry costs incurred by women’s families, as
Shemyakina (2009) found in Tajikistan, where women of marriageable age who
lived in conflict-affected areas were one-third less likely to be married than
women in less affected areas. In Cambodia, a rebound in marriage occurred after
the war but could not be sustained due to a shortage of young men of marriage-
able age (Heuveline and Poch 2007). The shortage of grooms may lead to
changes in marriage practices, such as an increase in polygamous marriages and
informal unions. For instance, in the case of Colombia, the impact of male mortal-
ity due to internal conflict is partially responsible for the increasingly high fre-
quency of consensual unions and, potentially, for female rural to urban migration
(Holland and Ferguson 2006).
    Fertility is frequently found to rebound once a crisis recedes. In Cambodia, fer-
tility fell by 30 percent during the Khmer Rouge period, but it nearly doubled two
years after the fall of the regime and remained above prewar levels for several
years (Heuveline and Poch 2007). In Ethiopia, annual conception probabilities
fell sharply in crisis years during the 1970 – 1982 conflict and resumed a more
gradual secular decline after 1982, as the country’s demographic transition pro-
ceeded (Lindstrom and Berhanu 1999). In Angola, Agadjanian and Prata (2001)
found that fertility dropped when hostilities peaked and rebounded in periods of
peacefulness, and these fluctuations were stronger in regions that were more af-
fected by the fighting.
    Alternately, conflict may lead to increased fertility when besieged ethnic or reli-
gious groups feel the need to increase their numbers. For example, among the
Palestinian population, fertility is substantially higher than would be expected


Buvinic et al.                                                                    125
from their level of socio-economic development (DellaPergola 2001). Families may
seek to replace children lost during conflict-related disruption. Some years after
the conflict in Rwanda, the number of surviving children among displaced popu-
lations was similar to that of the nondisplaced populations (Verwimp and van
Bavel 2004). The displaced populations had higher fertility, but this increased fer-
tility was offset by their children’s lower chances of survival. Similarly, 30 years
after the end of the Khmer Rouge period in Cambodia, Heuveline and Poch
(2007) found virtually no difference in the number of surviving children between
women who lost children during that period and those who did not, despite the
conflict-related shortage of men.


Political and Civic Participation
Evidence suggests that violent conflict can trigger unexpectedly positive civic and
political behaviors by women and other groups in the population who are largely
excluded from participating in civic and political life during peacetime.
Experiences of war violence are highly correlated with greater levels of social
capital, community engagement, and peaceful political engagement. Carmil and
Breznitz (1991) found that exposure to war led to greater political activism
among groups such as Jewish Holocaust survivors and Palestinian victims of
bombardment (Punama      ¨ ki, Qouta, and El Sarraj 1997). Regarding El Salvador,
Wood (2003) argued that government violence prompted its victims to support
and even join opposition forces out of moral outrage. Similarly, in Sierra Leone,
Bellows and Miguel (2006, 2009) found that individuals living in households
that experienced mortality, injury, or displacement due to war are more likely to
be politically active and to participate in local collective action, as evidenced by
voting, attending community meetings, being more politically knowledgeable and
engaging in community maintenance projects.
   These findings extend to ex-combatants. Employing survey data from northern
Uganda, where rebel recruitment generated quasi-experimental variation in people
who were conscripted, Blattman (2009) found that abduction leads to greater
postwar political participation, with a 27 percent increase in the likelihood of
voting and a doubling of the likelihood of being a community leader among former
abductees. Ex-combatants were also nearly twice as likely to participate in youth
peace clubs. Many of the abductions were brief, especially among those younger
than 11 or older than 20 years. However, the average duration of abduction was
15.3 months, and the average abductee reported receiving, witnessing, or perpe-
trating 11 violent acts (Blattman 2009, 233). Abductees who witnessed the most
violence were also most likely to participate politically later in life. However, abduc-
tion does not generally affect nonpolitical forms of social activity, suggesting that
the effects of war on participation may be uniquely political.

126                                   The World Bank Research Observer, vol. 28, no. 1 (February 2013)
   Another positive outcome of peace processes and political transitions has been
women’s increased participation in civil and political life. As survivors of conflict,
the expansion of women’s roles in postconflict reconstruction often leads to the
emergence of women’s organizations and networks. Through these organizations
and networks, women mobilize to integrate a gendered perspective and women’s
representation into peace negotiations and throughout the post-conflict period
(World Bank 2011). In Haiti, Liberia, Nicaragua, and Sierra Leone, for instance,
transitional governments introduced female staffing and gender-specific service in
the police force (World Bank 2011). In Timor-Leste, the transitional administra-
tion supported by the United Nations engaged women in rebuilding public institu-
tions (UNIFEM 2009, 30 – 31). The new constitutions in Uganda, Burundi, the
Democratic Republic of the Congo, and Nepal adopted affirmative action mecha-
nisms, especially quotas and cooptation systems, to help empower women eco-
nomically and politically.


Labor Reallocation
The loss of men in conflict and declines in household income trigger changes in
the household allocation of labor. These changes include women’s increased par-
ticipation in the labor market, the so-called “added worker” effect in which
women join the workforce to help families weather income shocks and compen-
sate for the absence of an earning spouse or partner. Studies in industrialized
countries suggest that the added worker effect was strong during the World Wars
and the Great Depression, but it has become less important over time as women’s
status in the labor market has improved (Acemoglu, Autor, and Lyle 2004). In de-
veloping countries, however, recent evidence suggests that aggregate economic
shocks yield added worker effects for women in low-income countries and low-
income households, whereas “discouraged worker” effects seem to prevail for
women in high-income countries and high-income households (Sabarwal, Sinha
and Buvinic 2010). For instance, analyzing DHS data for 66 countries over 21
years (1985 –2006), Bhalotra and Umana (2009) found that women with more
education often reduce their labor force participation in response to income
shocks, whereas women with less education increase their participation.
   In line with the above findings on economic shocks, a study in Rwanda
(Schindler 2010) observed increased labor intensity among teenage girls and
adult women in districts with low sex ratios, indicating the absence of males as
                                 ´n
                       ´ ndez, Iba
result of the war. Ferna                       ˜ a (2011) found that better-off agri-
                                   ˜ ez, and Pen
cultural households that are targets of violence in Colombia expand their labor
market supply and shift to off-farm employment to compensate for the decline in
household income produced by the violent shock. Men in these wealthier
Colombian households are more likely to participate in off-farm nonagricultural

Buvinic et al.                                                                    127
work, although this work does not fully compensate for drops in consumption.
Women attempt to find off-farm work, but with little apparent success. A related
study of civilian displacement as result of violent conflict in Colombia ( previously
cited in this review) found increased labor force participation among forcefully
displaced females compared to rural women who remain in rural areas and
reduced participation among displaced males, but the study did not test for added
worker effects (Calderon, Ga              ´n
                            ´ faro, and Iba ˜ ez 2011). If added worker effects were
significant and these displaced households were low-income households, the
results of these two studies of Colombia would be consistent with the findings in
the literature on differential labor market allocations by gender in response to
shocks for low-income versus high-income groups.


Children’s Human Capital
Violent conflict negatively affects children’s health and schooling. Although
households attempt to mitigate this impact (often with remarkable success, under
the circumstances), children’s life opportunities are typically negatively affected
because poor health and schooling outcomes for children can translate into signif-
icant differences in their lifetime earnings. Child stunting may be the most persis-
tent negative economic effect of violent conflict. Bundervoet, Verwimp, and
Akresh (2009, 558) estimated that the impact of child stunting from the Burundi
conflict could translate into as much as a 20.5 percent reduction in adult wages.
In terms of schooling, Ichino and Winter-Ebmer (2004) estimated that German
and Austrian children lost approximately 0.2 years of schooling, on average,
during the Second World War, with stronger effects for those with less-educated
parents. This loss of schooling translates into a long-term loss of earnings of 2 –3
percent, which, even in the 1980s, may have amounted to 0.8 percent of the GDP     .


Health
Children are especially vulnerable to worsening health conditions during and
after conflict periods. Hoeffler and Reynal-Queral (2003) found that a five-year
war is associated with a 13 percent increase in infant mortality, an effect that per-
sists during the first five years of peace. After decreasing during the decade of the
1980s, the infant mortality rate in south-central Iraq nearly doubled in 1991
and remained at this level throughout much of the 1990s (Ali and others 2003).
Parental education can help, but only to a point. In Ethiopia, parental education
is associated with lower child mortality in low-conflict areas but is less protective
in high-conflict areas, perhaps because of acute distress caused by conflict
(Kiros and Hogan 2001). Toole (2000) noted a lack of evidence suggesting any


128                                  The World Bank Research Observer, vol. 28, no. 1 (February 2013)
consistent female (or male) disadvantage in mortality among displaced children.
However, the data have been quite limited. Many studies have not included data
disaggregated by sex and age in emergency situations (Reed 2008).
   Many studies have shown that children’s growth is affected by violent conflict,
and some studies have reported gender differentials. Children living in high-con-
flict areas during the 2003 Iraq war were found to be 0.8 cm shorter than those
living in low-conflict areas (Guerrero-Serda   ´ n 2009). One reason for stunting is
the higher incidence of diarrhea in high-intensity conflict districts. Another possi-
ble factor, according to the author, is maternal stress. Valente (2011) observed an
increased probability of miscarriage in early to mid-pregnancy in high-conflict
Nepali districts. She hypothesized that maternal stress, rather than poor access to
health facilities, is responsible for this increased neonatal mortality. She also
found some evidence for selection effects; that is, healthier women are more likely
to become pregnant and to give birth in times of violent conflict.
   In an interesting twist, Akresh, Verwimp, and Bundervoet (2011) examined
the effects of both civil war and crop failure on children’s height-for-age z-scores.
They found evidence of child stunting associated with civil war in Rwanda and
Burundi but little evidence of gender or wealth differentials. By contrast, they
found that a crop failure in southern Rwanda only affected girls, especially those
from poor households. They suggested that this child stunting may have resulted
from the sudden occurrence of the civil war, leaving parents unable to protect
their children. Faced with an anticipated crop failure, poor households may have
chosen to protect their sons. This finding corresponds with Baird, Friedman, and
Schady (2011), who examined data from 59 developing countries and found that
infant mortality rises with negative economic shocks and that female infants’ sur-
vival is especially sensitive to such shocks. These authors noted the importance of
implementing policies that protect the health status of female infants during eco-
nomic downturns.
   Conflict and other traumas, such as orphanhood, can have a lasting impact on
child development through biological reactions that have been called “toxic
stress,” with adverse consequences for health and economic outcomes in adult-
hood. These reactions can potentially affect the intergenerational transmission of
human capital and can have a long-term impact on societies through reduced
future adult productivity. In a review of the literature, de Walque (2011) found
that although many studies worldwide document the adverse impact of conflict on
children and teenagers, there is wide variation in the resilience to conflict shocks.
For example, follow-up studies found that among Cambodian refugee adolescents,
the intensity of posttraumatic stress disorder and depression declined over time
(Sack and others 1993). De Walque (2011) noted factors that contribute to post-
traumatic resilience, such as placing orphans with family members, receiving



Buvinic et al.                                                                    129
posttraumatic professional care, and reinstituting normal school life, but much
more research is required on the factors that facilitate resilience to traumatic
events.


Schooling
Studies in many settings have found that violent conflict has a limited impact on
child schooling, even under difficult conditions. It seems that households attempt
to keep their children in school. Blattman and Annan (2010) reported that the
conflict in Northern Uganda had little impact on schooling for children who had
not been abducted, despite the violence experienced by communities. However, ab-
ducted male youths lost nearly a year of schooling on average, were less likely to
be functionally literate, and obtained lower-skilled work with lower earnings after
the war. The length of abduction was strongly correlated with losses in education
and literacy: each year of abduction was associated with a loss of half a year of
schooling and a 9 percentage point reduction in literacy (Blattman and Annan
2010, 882, 894).
   In Rwanda, Akresh and de Walque (2008) found that school-age boys and
girls exposed to genocide have 0.5 and 0.3 fewer years of schooling, respectively,
and are 15 percent less likely to complete third or fourth grade. The authors
argued that the impact of genocide could be characterized as a negative shock
that produces low levels of schooling outcomes and that disproportionately affects
boys and nonpoor children who had previously enjoyed an advantage in terms of
education. During the Cambodian genocide period from 1975 to 1978, the
Khmer Rouge closed schools, and individuals with an urban, educated back-
ground were more likely to have died. De Walque (2006) found that a higher per-
centage of Cambodian males of secondary school age in 1975 had less secondary
education compared with the preceding and succeeding birth cohorts. Girls were
less affected, at least partly because their levels of secondary schooling were far
lower than that of males before 1975 (de Walque 2006, figure 9).
   Traditional gender roles can work against both boys and girls. Boys’ schooling
may be more affected by conflict than the schooling of girls because there are
fewer expectations for the latter to participate in the labor force. In Colombia,
Angrist and Kugler (2008) and Rodriguez and Sanchez (2012) found that conflict
had a negative effect on teenage boys’ school enrollment as a result of boys’ in-
creased labor supply. In contrast, the expectation of higher economic returns from
boys’ schooling may tilt the balance of resource allocation to boys in conflict-
stressed conditions. In Guatemala, Chamarbagwala and Mora      ´ n (2011) found that
in rural areas that experienced a higher intensity of war, girls were significantly
more likely than boys to suffer loss of schooling. After the war, the government’s
girl scholarship program helped to reduce this gap. These findings suggest that

130                                 The World Bank Research Observer, vol. 28, no. 1 (February 2013)
gender-differentiated responses to child schooling are highly conflict and country
specific.



Discussion
Violent conflicts have been a pervasive feature of the recent global landscape, es-
pecially in countries and regions that are poor and have weak institutions and
low levels of human development. Once they begin, wars lead to further increases
in poverty and leave nations vulnerable to additional cycles of violence. However,
the study of war and its social and economic legacies is difficult. Those who par-
ticipate in or simply live through wars often suffer from persistent injuries, receive
less education, and experience a permanent decline in their productivity and
earnings. However, it remains unclear which impacts are most profound and per-
sistent, which disproportionately strike the poor, and how these effects can be
contained by local institutions and economic policies (Blattman and Miguel
2010). In particular, there is very little knowledge of the factors that make some
individuals and households more resilient than others to the impacts of conflict.
    The direct and indirect effects of violent conflict are seldom distributed random-
ly in the population and are seldom gender neutral. They usually affect one
gender more than the other (for example, the burden of mortality due to war vio-
lence appears to be disproportionately borne by males due to greater daily expo-
sure to violence in combat), and they may exacerbate or shrink preexisting gender
inequalities. Gender inequalities also mediate households’ coping responses to the
shocks inflicted by violent conflict. Until very recently, however, gender issues
were not part of the empirical literature on violent conflict, except for a focus on
gender when documenting sexual and gender-based violence in the context of
war. As this overview of recent studies shows, gender inequalities shape and are
both shaped by the responses of individuals and households to violent conflict.
These inequalities are a legitimate and important focus for the policy-oriented lit-
erature on conflict.
    The literature review reveals, first, the heterogeneity of impacts across contexts,
conflicts, and countries for girls and boys, women and men. In Cambodia and
Colombia, boys’ schooling suffered more from conflict than did girls’ schooling. In
rural Guatemala, however, girls’ schooling was more affected. In Burundi, conflict
seems to have especially affected the health of boys, whereas some evidence sug-
gests that girls’ health was more affected than boys’ health in Iraq. Second, the
literature also shows regularities in the effects of conflict on the labor market by
gender for countries and households that mirror those observed in response to
economic shocks. In some settings, women cope with the loss of men’s income by


Buvinic et al.                                                                    131
joining paid employment, whereas in wealthier agricultural households in
Colombia, conflict leads men to increase their participation in off-farm paid work.
    Third, the evidence shows that conflict changes households’ demographic pro-
files and that families’ coping responses include adjustments in marriage and fer-
tility behavior, such as in Rwanda. Fourth, the evidence reveals surprising
resilience and adaptability in the responses of many households to the human
and economic shocks of violent conflict and underlines women’s contributions
through alterations to their traditional labor allocation and roles in the family.
This evidence suggests that unseen benefits of the costs and destruction of violent
conflict may be changes in traditional gender roles and greater gender equality in
the household. A follow-up empirical question is how lasting these changes are.
Fifth, although many households rebound from the shock inflicted by conflict,
women left alone to provide for their families may be particularly vulnerable to
poverty that can persist across generations. Targeting widows and their families
with postconflict assistance and resources may break this transmission channel
and halt the reproduction of conflict-related poverty. Finally, the empirical evi-
dence counters commonly held views, such as the pervasive view that sexual and
gender-based violence is a phenomenon that affects only girls and women, that
rape is a weapon of war rather than a crime of opportunity, or that girls and
women are always worse off as result of conflict (when, in fact, they may be com-
paratively better off because they remain in school or are able to substitute for
men in jobs and political positions). This evidence also highlights the many gaps
in knowledge about the gender-differentiated effects of and adaptive responses to
conflict.
    There are several questions for future research. What happens to families in
terms of loss of land, physical and financial capital, and livelihood when conflict
occurs? How are these losses distributed within the family? How do gender roles
and inequalities affect families’ coping responses to these losses? How do these re-
sponses differ by gender for the poor and the nonpoor? What are the implications
of civil conflict for girls and boys and for women and men in terms of psychologi-
cal trauma, sexual violence and sexually transmitted infections, and schooling?
What are the long-term consequences of sexual violence for female and male
victims? What are the factors that foster resilience by individuals and families to
the devastation produced by conflict? When, under what circumstances, and for
whom does conflict create opportunities?
    The concept of “violent conflict” cannot be understood by a single indicator,
and different indicators may yield different effects on labor, education, and health.
It is therefore of the utmost importance to be as precise as possible about the
channels through which violent conflict affects the outcomes of interest. Not all
channels are equally harmful; some may be gender neutral, and some may even
favor girls and women. This overview has offered an organizing framework to

132                                  The World Bank Research Observer, vol. 28, no. 1 (February 2013)
help identify the transmission channels for the effects of conflict on households
and genders and their adaptive responses to it, but there is room for greater preci-
sion in future research.
   Research on conflict-affected populations will encourage social scientists to
move in new directions in several ways. Interdisciplinary collaboration, particular-
ly with the aim of measuring psychosocial outcomes and informing public health
practice, is crucial. The usual steps of research design may differ as a result of a
lack of sampling frames, population mobility, and the overall difficult conditions
in which conflict-affected groups live, but the goal should be to produce, to the
greatest extent possible, useful and methodologically rigorous insights that can
inform the design of effective conflict prevention and postconflict policies—the
foremost objective of any conflict research. Gender is an important variable in this
research that highlights particular conflict-triggered vulnerabilities, such as infant
girls’ nutritional status, boys’ schooling deficits and widows’ burdens, and shapes
specific resilient responses to inform the design of these policies.



Notes
Mayra Buvinic is a Senior Fellow with both the UN Foundation and Vital Voices; email address:
mayra.buvinic@gmail.com. Monica Das Gupta is a Senior Demographer at the World Bank. Ursula
Casabonne is an Operations Analyst at the World Bank. Philip Verwimp is Associate Professor of
Development Economics at the Solvay Brussels School of Economics and Management, Universite  ´
Libre de Bruxelles. This work was supported by a grant from the Government of Norway to the
World Bank.




References
Acemoglu, Daron, David Autor, and David Lyle. 2004. “Women, War, and Wages: The Effect of
  Female Labor Supply on the Wage Structure at Midcentury.” Journal of Political Economy 112(3):
  497–551.
Agadjanian, Victor, and Ndola Prata. 2001. “War and Reproduction: Angola’s Fertility in
  Comparative Perspective.” Journal of Southern African Studies 27(2): 229–347.
Ajdukovic, Marina, and Dean Ajdukovic. 1998. “Impact of Displacement on the Psychological Well-
   Being of Refugee Children.” International Review of Psychiatry 10:186 –95.
Akresh, Richard, and Damien de Walque. 2008. “Armed Conflict and Schooling: Evidence from the
  1994 Rwanda Genocide.” World Bank Policy Research Paper No 4606. World Bank, Policy
  Research Department, Washington, DC.
Akresh, Richard, Philip Verwimp, and Tom Bundervoet. 2011. “Civil War, Crop Failure, and the
  Health of Young Children in Rwanda.” Economic Development and Cultural Change 59(4):
  777–810.
Ali, Mohammed M., John Blacker, and Gareth Jones. 2003. “Annual Mortality Rates and Excess
   Deaths of Children Under Five in Iraq, 1991–98.” Population Studies 57(2): 217–26.


Buvinic et al.                                                                              133
Andre, Catherine, and Jean-Philippe Platteau. 1998. “Land Relations under Unbearable Stress:
  Rwanda Caught in the Malthusian Trap.” Journal of Economic Behavior and Organization 34:
  1 –37.
Angrist, Joshua, and Adriana Kugler. 2008. “Rural Windfall or Resource Curse? Coca, Income and
  Civil Conflict in Colombia.” Review of Economics and Statistics 90: 191–215.
Annan, Jeannie, Christopher Blattman, Dyan Mazurana, and Khristopher Carlson. 2011. “Civil
  War, Reintegration and Gender in Northern Uganda.”Journal of Conflict Resolution 55(6):
  877 –908.
Annan, Jeannie, Christopher Blattman, and Roger Horton. 2006. “The State of Youth and Youth
  Protection in Northern Uganda: Findings from the Survey of War Affected Youth.” UNICEF
  Uganda.
Baez, Javier E. 2011. “Civil Wars Beyond Their Borders: The Human Capital and Health
   Consequences of Hosting Refugees.” Journal of Development Economics 96(2): 391–408.
Baird, Sarah, Jed Friedman, and Norbert Schady. 2011. “Aggregate Income Shocks and Infant
   Mortality in the Developing World.” Review of Economics and Statistics 93(3): 847 –56.
Bayer, Christophe Pierre, Fionna Klasen, and Hubertus Adam. 2007. “Association of Trauma and
   PTSD Symptoms with Openness to Reconciliation and Feelings of Revenge among former
   Ugandan and Congolese Child Soldiers.” Journal of the American Medical Association 298(5): 555.
Bellows, John, and Edward Miguel. 2009. “War and Local Collective Action in Sierra Leone.” Journal
   of Public Economics.” 93(11-12): 1144–57.
Berlage, Lode, Marijke Verpoorten, and Philip Verwimp. 2003. “Survival Strategies of Poor
   Households in Post-genocide Rwanda.” Report for the Flemish Interuniversity Council and the
   Belgian Department of International Cooperation, Policy Research Program, Leuven, Belgium.
Bhalotra, Sonia, and Marcela Umana-Aponte. 2010. “The Dynamics of Women’s Labour Supply in
  Developing Countries.” IZA Discussion Paper 4879. Institute for the Study of Labor, Bonn,
  Germany.
Blanc, Ann. 2004. “The Role of Conflict in the Rapid Fertility Decline in Eritrea and Prospects for
   the Future.” Studies in Family Planning 35(4): 236 –24.
Blattman, Christopher. 2009. “From Violence to Voting: War and Political Participation in Uganda.”
   The American Political Science Review 103(2): 231 –47.
Blattman, Christopher, and Jeannie Annan. 2010. “The Consequences of Child Soldiering.” Review
   of Economics and Statistics 92(4): 882–98.
Blattman, Christopher, and Edward Miguel. 2010. “Civil War. ” Journal of Economic Literature 48(1):
   3 –57.
  ¨ ck, Tilman. 2006. “War and Reconstruction in Northern Mozambique.” The Economics of Peace
Bru
   and Security Journal 1(1): 30 –9.
  ¨ ck, Tilman, and Kati Schindler. 2009. “The Impact of Violent Conflicts on Households: What We
Bru
   Know and What Should We Know about War Widows?” Oxford Development Studies 37(3):
   289 –309.
  ¨ ck, Tilman, Patricia Justino, Philip Verwimp, and Alexandra Avdeenko. 2010. “Identifying
Bru
   Conflict and Violence in Micro-Level Surveys.” Working Paper 79. Households in Conflict
   Network (HiCN), Brighton, UK.
Bundervoet, Tom, Philip Verwimp, and Richard Akresh. 2009. “Health and Civil War in Rural
  Burundi.” Journal of Human Resources 44(2): 536–63.
Buzzell, Emily, and Samuel H. Preston. 2007. “Mortality of American Troops in the Iraq War.”
  Population and Development Review 2007 33(3): 555–66.


134                                        The World Bank Research Observer, vol. 28, no. 1 (February 2013)
      ´ n, Valentina, Margarita Ga
Caldero                           ´ faro, and Ana Marı ´a Iba´n
                                                              ˜ ez. 2011. “Forced Migration, Female
   Labor Force Participation, and Intra-household Bargaining: Does Conflict Empower Women?”
   Centro de Estudios sobre Desarrollo Econo´ mico, Universidad de los Andes, Colombia.
Cardozo, Barbara Lopes, Oleg O. Bilukha, A. Carol, Gotway Crawford, Irshad Shaikh,
   Mitchell I. Wolfe, Michael L. Gerber, and Mark Anderson. 2004. “Mental Health, Social
   Functioning and Disability in Postwar Afghanistan.” Journal of the American Medical Association
   292(5): 575 –84.
Carmil, Devora, and Shlomo Breznitz. 1991. “Personal Trauma and World View—Are Extremely
   Stressful Experiences Related to Political Attitudes, Religious Beliefs, and Future Orientation?”
   Journal of Traumatic Stress 4(3): 393 –405.
Chamarbagwala, Rubiana, and Hilcı
                                ´as E. Mora ´ n. 2011. “The Human Capital Consequences of Civil
  War: Evidence from Guatemala.” Journal of Development Economics 94:41 –61.
Cohen, Dara Kay. 2011. “Causes of Sexual Violence During Civil War: Cross-National Evidence
  (1980-2009).” Humphrey School of Public Affairs University of Minnesota, Prepared for the
  Minnesota International Relations Colloquium, March 28, 2011.
Collier, Paul. 1999. “On the Economic Consequences of Civil War.” Oxford Economic Papers 51:
   168–83.
        . 2007. The Bottom Billion: Why the Poorest Countries are Failing and What Can Be Done About
   It. Oxford: Oxford University Press.
Collier, Paul, V               ˚ vard Hegre, Anke Hoeffler, Marta Reynal-Querol, and Nicholas
                . L. Elliott, Ha
   Sambanis. 2003. Breaking the Conflict Trap: Civil War and Development Policy. Washington, D.C.:
   World Bank; Oxford and New York: Oxford University Press.
           ´ ire A., Michelle Gayer, Michael J. Ryan, Peter Salama, Paul Spiegel, and David L.
Connolly, Ma
  Heymann. 2004. “Communicable Diseases in Complex Emergencies: Impact and Challenges.”
  Lancet 364(9449): 1974–83.
Das, Jishnu, Quy-Toan Do, Jed Friedman, and David McKenzie. 2008. “Mental Health Patterns and
  Consequences: Results from Survey Data in Five Developing Countries.” World Bank Economic
  Review 23(1): 31 –55.
de Walque, Damien. 2006. “The Socio-Demographic Legacy of the Khmer Rouge Period in
   Cambodia.” Population Studies 60(2): 223 –31.
      . 2011. “Conflicts, Epidemics, and Orphanhood: The Impact of Extreme Events on the Health
   and Educational Achievements of Children.” In Harold Alderman, ed., No Small Matter: The
   Impact of Poverty, Shocks, and Human Capital Investments in Early Childhood Development.
   Washington, DC: The World Bank.
de Walque, Damien, and Philip Verwimp. 2010. “The Demographic and Socio-economic
  Distribution of Excess Mortality during the 1994 Genocide in Rwanda.” Journal of African
  Economies 19(2): 141– 62.
DellaPergola, Sergio. 2001. “Population Change and Political Transitions Demography in Israel/
   Palestine: Trends, Prospects, Policy Implications.” Presented at the IUSSP XXIV General
   Population Conference Salvador de Bahia.
El Jack, Amani. 2003. “Gender and Armed Conflict: Overview Report.” BRIDGE, Institute of
   Development Studies, University of Sussex, Brighton, UK.
    ´ ndez, Manuel, Ana Marı
Ferna                             ´n
                             ´a Iba                    ˜ a. 2011. “Adjusting the Labor Supply to
                                    ˜ ez, and Ximena Pen
   Mitigate Violent Shocks: Evidence from Rural Colombia.” World Bank Policy Research Working
   Paper 5684. World Bank, Policy Research Department, Washington, DC.
Ghobarah, Hazem Adam, Paul Huth, and Bruce Russett. 2003. “Civil Wars Kill and Maim People –
  Long After the Shooting Stops.” American Political Science Review 97(2): 189 –202.


Buvinic et al.                                                                                  135
    ´ lez, Marı
Gonza         ´a A., and Rigoberto A. Lopez. 2007. “Political Violence and Farm Household
  Efficiency in Colombia.” Economic Development and Cultural Change 55(2): 367– 92.
             ´n, Gabriela. 2009. “The Effects of the War in Iraq on Nutrition and Health: An
Guerrero-Serda
  Analysis Using Anthropometric Outcomes of Children.” Working Paper 55. Households in
  Conflict Network (HiCN), Brighton, UK.
Hankins, Catherine A., Samuel R Friedman, Tariq Zafar, and Steffanie A. Strathdee. 2002.
  “Transmission and Prevention of HIV and Sexually Transmitted Infections in War Settings:
  Implications for Current and Future Armed Conflicts.” AIDS 16(17): 2245– 52.
Heuveline, Patrick, and Bunnak Poch. 2007. “Demographic Crisis and Rebound in Cambodia.”
  Demography 44(2): 405 –26.
Holland Jones, James, and Brodie Ferguson, 2006. “The Marriage Squeeze in Colombia, 1973–
  2005: The Role of Excess Male Death.” Social Biology 53(3-4): 140 –51.
Human, Rights Watch. 1995. Global Report on Women’s Human Rights. New York: Human Rights
  Watch.
      . 1996. Shattered Lives: Sexual Violence during the Rwandan Genocide and its Aftermath.
   New York: Human Rights Watch.
Husain, Syed Arshad, Jyotsna Nair, William Holcomb, John C. Reid, Victor Vargas, and Satish S.
  Nair. 1998. “Stress Reactions of Children and Adolescents in War and Siege Conditions.”
  American Journal of Psychiatry 155(12): 1718–19.
Iban
   ˜ ez, Ana Marı
                ´a, and Andre ´ s Moya. 2010. “Do Conflicts Create Poverty Traps? Asset Losses and
   Recovery for Displaced Households in Colombia.” In Rafael di Tella, Sebastian Edwards, and
   Ernesto Schargrodsky, eds., The Economics of Crime: Lessons for and from Latin America, 137–72.
   Cambridge, MA: National Bureau of Economic Research; Chicago, IL: University of Chicago Press.
Ichino, Andrea, and Rudolf Winter-Ebmer. 2004. “The Long-Run Educational Cost of World War II.”
   Journal of Labor Economics 22(1): 57 –86.
Jayaraman, Anuja, Tesfayi Gebreselassie, and S. Chandrasekhar. 2009. “Effect of Conflict on Age at
   Marriage and Age at First Birth in Rwanda.” Population Research Policy Review 28:551 –67.
Johnson, Kirsten, Jana Asher, Stephanie Rosborough, Amisha Raja, Rajesh Panjabi, Charles Beadling,
   and Lynn Lawry. 2008. “  Association of Combatant Status and Sexual Violence with Health and
   Mental Health Outcomes in Post-conflict Liberia.” Journal of the American Medical Association
   300(6): 676–90.
Justino, Patricia. 2006. “On the Links between Violent Conflict and Chronic Poverty: How Much Do
   We Really Know?” Working Paper 18. Households in Conflict Network (HiCN), Brighton, UK.
       . Forthcoming. “War and Poverty.” In M Garfinkel, and S. Skaperdas, eds., Oxford Handbook of
   the Economics of Peace and Conflict. Oxford: Oxford University Press.
Justino, Patricia, and Philip Verwimp. 2006. “Poverty Dynamics, Violent Conflict and Convergence
   in Rwanda.” Working Paper 16. Households in Conflict Network (HiCN), Brighton, UK.
Kalisya, Luc Malemo, Paluku Lussy Justin, Christophe Kimona, Kavira Nyavandu, Kamabu
   Mukekulu Eugenie, Kasereka Muhindo Lusi Jonathan, Kasereka Masumbuko Claude,
   and Michael Hawkes. 2011. “Sexual Violence toward Children and Youth in War-Torn Eastern
   Democratic Republic of Congo.” PLoS One (6): 1.
Kiros, Gebre-Egziabher, and Hogan Dennis P     . 2001. “War, Famine and Excess Child Mortality in
   Africa: the Role of Parental Education.” International Journal of Epidemiology 30(3): 447 –55.
Kondylis, Florence. 2008. “Agricultural Outputs and Conflict Displacement: Evidence from a Policy
  Intervention in Rwanda.” Economic Development and Cultural Change 57(1):31–66.
      . 2010. “Conflict displacement and labor market outcomes in post-war Bosnia and
   Herzegovina.” Journal of Development Economics 93(2): 235 –48.


136                                       The World Bank Research Observer, vol. 28, no. 1 (February 2013)
Krug, Etienne G., Linda L. Dahlberg, James A. Mercy, Anthony B. Zwi, and Rafael Lozano, eds.
   2002 World Report on Violence and Health. Geneva: World Health Organization.
Lindstrom, David P., and Betemariam Berhanu. 1999. “The Impact of War, Famine, and Economic
   Decline on Marital Fertility in Ethiopia.” Demography 36(2): 247 –61.
MacMullin, Colin, and Maryanne Loughry. 2004. “Investigating Psychosocial Adjustment of Former
  Child Soldiers in Sierra Leone and Uganda.” Journal of Refugee Studies 17(4): 460 –72.
McGinn, Theresa. 2000. “Reproductive Health of War-Affected Populations: What Do We Know?”
  International Family Planning Perspectives 26(4): 174–80.
       . 2009. “Barriers to Reproductive Health and Access to Other Medical Services in Situations
   of Conflict and Migration.” In S. Martin and J. Forbes, eds., Women, Migration, and Conflict:
   Breaking a Deadly Cycle. Dordrecht: Springer.
Miguel, Edward, and Gerard Roland. 2005. The Long Run Impact of Bombing Vietnam. Berkeley, CA:
  University of California.
Mock, Nancy B., Sambe Duale, Lisanne F. Brown, Ellen Mathys, Heather O’Maonaigh, Nina KL
  Abul-Husn, and Sterling Elliott. 2004. “Conflict and HIV: A Framework for Risk Assessment to
  Prevent HIV in Conflict-Affected Settings in Africa.” Emerging Themes in Epidemiology 1(6).
Mollica, Richard F., Charles Poole, Linda Son, Caroline C. Murray, and Svang Tor. 1997. “Effects of
  War Trauma on Cambodian Refugee Adolescents’ Functional Health and Mental Health Status.”
  Journal of the Academy of Child and Adolescent Psychiatry 36(8): 1098–106.
Montalvo, Jose G., and Marta Reynal-Querol. 2007. “Fighting against Malaria: Prevent Wars while
  Waiting for the ’Miraculous’ Vaccine.” Review of Economics and Statistics 89(1): 165 –77.
Morrison, Andrew R., and Maria Beatriz Orlando. 1999. “Social and Economic Costs of Domestic
  Violence: Chile and Nicaragua.” In A. R. Morrison and M. L. Biehl, eds., Too Close to Home:
  Domestic Violence in the Americas. Washington, DC: The Inter-American Development Bank and
  Johns Hopkins U. Press.
Murray, Christopher J. L., Gary King, Alan D. Lopez, Niels Tomijima, and Etienne Krug. 2002.
  “Armed Conflict as a Public Health Problem.” British Medical Journal 239: 346 –49.
Neupert, Ricardo, and Virak Prum. 2005. “Cambodia: Reconstructing the Demographic Stab of the
  Past and Forecasting the Demographic Scar of the Future.” European Journal of Population 21:
  217–46.
    ˚ s, Ragnhild, and Dara Kay Cohen. 2011. Wartime Sexual Violence – Challenges and
Norda
  Opportunities for Data Collection and Analysis. Oslo: CSCW.
Obermeyer, Ziad, Christopher J.L. Murray, and Emmanuela Gakidou. 2008. “Fifty Years of Violent
  War Deaths from Vietnam to Bosnia: Analysis of Data from the World Health Survey
  Programme.” British Medical Journal 336: 1482– 86.
Peterman, Amber, Tia Palermo, and Caryn Bredenkamp. 2011. “Estimates and Determinants of
   Sexual Violence in the Democratic Republic of the Congo.” American Journal of Public Health
   101(6): 1060–67.
Pham, Phuong N. P  ., Patrick Vinck, and Eric Stover. 2009. “Returning Home: Forced Conscription,
  Reintegration, and Mental Health Status of Former Abductees of the Lord’s Resistance Army in
  northern Uganda.” BMC Psychiatry 9(1): 23.
Punama ¨ ki, Raija-Leena, Samir Qouta, and Eyad El Sarraj. 1997. “Relationships between Traumatic
  Events, Children’s Gender, and Political Activity, and Perceptions of Parenting Styles.”
  International Journal of Behavioral Development 21(1): 91 – 109.
Reed, Holly. 2008. “Who is Most Vulnerable? Age and Sex-Specific Mortality Patterns in Refugee
   and Displaced Populations.” Unpublished manuscript, Department of Sociology and Population
   Studies and Training Center, Brown University, Providence, RI.


Buvinic et al.                                                                                 137
Rodriguez, Catherine, and Fabio Sanchez Torres. 2012. “Armed Conflict Exposure, Human Capital
  Investments and Child Labor: Evidence from Colombia.” Defense and Peace Economics 23(2):
  161 –84.
Sabarwal, Shwetlena, Nistha Sinha, and Mayra Buvinic. 2010. “How Do Women Weather Economic
   Shocks?: A Review of the Evidence.” Working Paper 5496. World Bank, Policy Research
   Department, Washington DC.
Sack, W .H., G. Clarke, C. Him, D. Dickason, B. Goff, K. Lanham, and J.D. Kinzie. 1993. “A 6-Year
   Follow-Up Study of Cambodian Refugee Adolescents Traumatized as Children.” Journal of the
   American Academy of Child and Adolescent Psychiatry, 32(2): 431 –37.
Schindler, Kati. 2010. “Who Does What in a Household after Genocide? – Evidence from Rwanda.”
   Discussion Paper 1072. DIW, Berlin, Germany.
Shemyakina, Olga N. 2011. “The Effect of Armed Conflict on Accumulation of Schooling: Results
   from Tajikistan.” Journal of Development Economics 95(2): 186 –200.
      . 2009. “The Marriage Market and Tajik Armed Conflict.” Working Paper 66. Households in
   Conflict Network (HiCN), Brighton, UK.
Spiegel, Paul B., and Peter Salama. 2000. “War and Mortality in Kosovo, 1998-99: An
   Epidemiological Testimony.” The Lancet 355(9222): 2204–9.
Swiss, Shana, and Joan E. Giller. 1993. “Rape as a Crime of War: A Medical Perspective.” Journal of
  the American Medical Association 270(5): 612 –15.
           ´ , Slobodanka 2003. “Armed Conflicts and Food Security.” Working Paper No. 03 –11.
Teodosijevic
   Food and Agriculture Organization, ESA, Rome, Italy.
Tolin, David F., and Edna B. Foa. 2006. “Sex Differences in Trauma and Post-Traumatic Stress
   Disorder: A Quantitative Review of 25 Years of Research.” Psychological Bulletin 132(6): 959 –92.
Toole, Michael J. 2000. “Displaced Persons and War”. In S. Barry, and Victor W Sidel, eds., War and
   Public Health, Levy, 197–214. Washington, DC: American Public Health Association.
UNIFEM (United Nations Development Fund for Women). 2009. Progress of the World’s Women
  2008/2009. Who Answers to Women? Gender and Accountability Technical Report. United
  Nations Development Fund for Women, New York.
Valente, Christine. 2011. “What Did the Maoists Ever Do for Us?: Education and Marriage of
   Women Exposed to Civil Conflict in Nepal.” Working Paper 5741. World Bank, Policy Research
   Department, Washington, DC.
van de Walle, Dominique, 2011. “Lasting Welfare Effects of Widowhood in a Poor Country.”
   Working Paper 5734. World Bank, Policy Research Department, Washington, DC.
Verwimp, Philip, and Jan Van Bavel. 2005. “Child Survival and the Fertility of Refugees in Rwanda
   after the Genocide.” European Journal of Population 21:271 –90.
Women’s Commission for Refugee Women and Children. 2004. Global Survey on Education in
  Emergencies. New York: Women’s Commission for Refugee Women and Children.
Wood, Elisabeth J. 2003. Insurgent Collective Action and Civil War in El Salvador. New York:
  Cambridge University Press.
Wood, Elisabeth Jean. 2006. “Variation in Sexual Violence during War.” Politics and Society 34(3):
  307 –41.
       . 2009. “ Armed Groups and Sexual Violence: When Is Wartime Rape Rare?” Politics and
   Society 37(1): 131 –61.
World Bank. 2011. World Development Report 2011: Conflict, Security and Development. Washington
  DC: The World Bank.




138                                        The World Bank Research Observer, vol. 28, no. 1 (February 2013)