Re-assessing the ‘Where’ of Violence: Subnational

Ethnic Diversity and Wartime Violence in Liberia

Sekou Jabateh∗
[email protected]
New York University
August 2023

Abstract
While various country-level studies imply a higher inclination towards violence in di-
verse countries, the ’where’ of violence remains largely unexplored. This research delves
into this gap by examining the effect of subnational ethnic diversity on the escalation of
wartime violence. Utilizing a unique clan-level ethnic composition data and the Armed
Conflict Location & Event Dataset (ACLED), the study investigates the association
between ethnic heterogeneity and the escalation of violence during the Liberian civil
war. The findings reveal a complex relationship between ethnic polarization and dif-
ferent forms of wartime violence. While ethnic polarization positively correlates with
protests, it negatively associates with riots. Ethnic fractionalization, however, does not
significantly predict any of the violence measures. These findings highlight the multi-
faceted nature of ethnic conflicts and the importance of considering additional factors
when assessing the consequences of ethnic diversity in conflict situations.

∗
Research Fellow at New York University

1
1 Introduction

How does ethnic diversity influence the initiation and intensification of civil wars? This paper
interrogates whether conflict severity is exacerbated in regions characterized by homogeneity
or heterogeneity within ethnically pluralistic nations. Ethnic identity has garnered substan-
tial focus in civil war scholarship (Hegre et al., 2009; Collier & Hoeffler, 2004; Fearon &
Laitin, 2003; Reynal-Querol, 2002). Prevailing narratives attribute the genesis of numerous
civil wars to religious or ethnic antagonisms. A broad consensus among political scientists
and economists posits a correlation between ethnic diversity and diminished social capital
(Easterly & Levine, 1997), hindered economic progression (Alesina et al., 2003; Alesina & La
Ferrara, 2005), and adverse societal outcomes (Kustov & Pardelli, 2018), which encompass
an elevated propensity for civil discord and a reduction in public good provision. Existing
literature frequently associates ethnic fractionalization with poorer public outcomes due to
diminished preference aggregation and collective action potential (Cruz et al., 2020). Re-
search on civil conflict causation and proliferation employs three primary frameworks: onset,
incidence (the number of events), and duration of civil conflicts (Montalvo & Reynal-Querol,
2017). These phases, while complementary, address distinct aspects of civil war dynamics.
Discerning the drivers of each phase is crucial for developing effective policies to mitigate or
resolve civil wars and to promote post-conflict peace.
Fearon and Laitin (2003) propose that the emergence of civil wars is shaped by conditions
conducive to insurgency, such as poverty, challenging terrains, political instability, and large
populations. They argue that the post-Cold War surge in civil wars resulted from "a steady
accumulation of protracted conflicts since the 1950s and 1960s" and assert that, controlling
for per capita income, countries with greater ethnic or religious diversity do not face a
significantly higher risk of civil war outbreak (Fearon & Laitin, 2003, p. 1). The incidence
and duration of civil wars have been the subjects of numerous studies. Montalvo and Reynal-
Querol (2005) identified ethnic polarization as a key determinant of civil war frequency. Gates
and Strand (2004), and Montalvo and Reynal-Querol (2007) developed models estimating

2
the duration of civil wars. While some research posited that uncertainty about contenders’
relative power extends war duration, Gates and Strand (2004) and Montalvo and Reynal-
Querol (2007) concluded that the length of civil wars expands with societal polarization.
Empirically, these findings suggest that ethnically polarized countries experience prolonged
civil conflicts.
However, these inquiries predominantly occurred at regional or national levels. Histori-
cally, the bulk of civil conflict research prioritized regional or national ethnic group distri-
butions, deeming them more critical than subnational scrutiny (Toft, 2003). Recent studies
have indicated that regional and country-level conflict analyses may incur inaccuracies, as
many conventional and popular conflict explanations require an understanding of the geo-
graphic variance within countries in ethnic differences, discrimination, and inequality (Hegre
et al., 2009; Østby & Rød, 2006). This recognition has spurred demands for more granular
civil war analyses, focusing on subnational causes, incidence, and duration (Chaudoin et
al., 2017; Montalvo & Reynal-Querol, 2007; Buhaug & Lujala, 2005; Buhaug & Rod, 2006;
Hegre & Raleigh, 2006; Raleigh & Urdal, 2007; Koren & Sarbahi, 2018). Notably, while
state-level studies imply heightened violence susceptibility in ethnically diverse nations, it
remains largely unexplored and unverified whether such violence predominantly occurs in ho-
mogeneous or heterogeneous areas within these nations. The previous studies’ emphasis on
regional and country-year aggregates, coupled with their lack of intra-national disaggregated
analysis, has impeded empirical testing of their implications at the subnational level. This
paper aims to bridge this gap, focusing on two main objectives: (a) to offer a more detailed
and nuanced perspective on the local-level determinants of wartime violence, particularly
within Liberia’s protracted civil war; and most importantly, (b) to empirically test the link
between subnational ethnic heterogeneity and the escalation of violence.
The Liberian civil war, which raged from 1989 to 2003, offers a valuable opportunity to
empirically examine this relationship. Liberia is a highly ethnically diverse country, home
to over 16 recognized ethnic groups. These groups were often pitted against each other in a

3
brutal struggle for power during the civil war. This paper leverages original pre-civil war clan-
level ethnic composition data, along with the Armed Conflict Location & Event Dataset’s
(ACLED) disaggregated conflict event data on Liberia, to investigate the association.
The findings of the paper reveal a complex relationship between ethnic polarization and
different forms of wartime violence. While ethnic polarization positively correlates with
protests, it negatively associates with riots. These findings suggest that ethnic polarization
may have differential effects on different types of violence. Ethnic fractionalization, however,
does not significantly predict any of the violence measures. This suggests that the relation-
ship between ethnic diversity and conflict may be more complex than previously thought
and may be mediated by other factors, such as economic inequality or political instability.
The findings of this paper have important implications for our understanding of the rela-
tionship between ethnic diversity and wartime violence. They suggest that the relationship
is not always straightforward and that ethnic polarization may have differential effects on
different types of violence. Additionally, the findings highlight the importance of considering
other factors in addition to ethnic diversity when assessing the risk of conflict.

2 Ethnic Diversity and Conflict: Theory and Evidence

The theoretical underpinnings of this study are anchored in the premise that ethnic diversity
critically influences the likelihood and intensity of civil conflicts. Ethnic diversity, defined
as the presence of a wide range of ethnic differences among individuals, has been shown to
significantly affect the dynamics and severity of wartime violence.
Empirical evidence supporting this notion can be found in studies such as Cederman et
al. (2013), who argue that the presence of diverse ethnic groups, especially minority groups
fighting for territorial rights or larger groups advocating for governmental reform, escalates
the probability of ethnic conflicts in settings like competitive elections.
Easterly and Levine (1997) highlight that ethnically diverse societies often encounter

4
difficulties in aggregating preferences and achieving collective action, leading to reduced
social capital and public goods provision. Cruz et al. (2020) extend this observation by
linking lower social capital in such societies to hindered conflict resolution mechanisms,
thereby increasing the potential for intensified violence during civil conflicts.
Furthermore, the geographic variance of ethnic diversity within countries can lead to
different levels of tension and conflict across various regions. Hegre et al. (2009) and Østby
and Rød (2006) emphasize the importance of understanding these subnational differences in
ethnic composition to fully grasp the dynamics of civil conflicts.
Ethnic polarization, or the degree of division and animosity between ethnic groups within
a society, is another crucial aspect. Montalvo and Reynal-Querol (2007) argue that societies
with significant ethnic polarization are prone to more intense violence due to entrenched
divisions and a lack of willingness among rival groups to compromise. This situation often
leads to prolonged conflicts, where the depth of ethnic divisions exacerbates the violence.
Adida et al. (2022) contribute to this discourse by showing that changes in ethnic di-
versity, particularly in areas that were previously homogeneous, can trigger conflict. This
finding underscores the complexity of ethnic dynamics and how shifts in ethnic composition
can disrupt existing social equilibriums, potentially leading to violence.
This paper aims to analyze the nexus between the presence of diverse and polarized ethnic
groups and the intensification of civil conflicts. The paper’s central hypothesis postulates
a relationship between ethnic diversity and the increased propensity for wartime violence
within a clan. This overarching hypothesis encompasses several sub-hypotheses, potentially
leading to a variety of distinct intra-national outcomes or findings, outlined as follows:

H1:

Ethnic polarization intensifies the escalation of civil conflicts. This hypothesis advances a
robust positive correlation between the degree of ethnic polarization and the amplification
of wartime violence in a clan.

5
H2:

Ethnic diversity contributes to the escalation of civil conflicts. This hypothesizes a strong
positive correlation between the level of ethnic diversity within a clan and the augmentation
of wartime violence.

H3:

Neither ethnic diversity nor ethnic polarization contributes to the escalation of civil con-
flicts. This hypothesis contends that there is no significant correlation between either ethnic
diversity or polarization and the escalation of wartime violence in a clan.

3 Research Design

3.1 The Significance of Liberia

The Liberian civil war commenced as a vehement ethnic conflict (Young, 2022). Given the
extensive geographical variation in political violence during the Liberian civil war, as recorded
in the Armed Conflict Location and Events Data, Liberia presents itself as a unique case
study for quantitative analysis of wartime violence. The availability of pre-civil war ethnic
composition and distribution data, obtained through fieldwork in Liberia, facilitates the
examination of potential correlations between ethnic heterogeneity and the escalation of
violence. A comprehensive population survey conducted in 2011 on attitudes toward dispute
resolution, security, and post-conflict reconstruction revealed that 40% of Liberians perceived
ethnic division as a fundamental cause of the civil war (Vinck, Pham, and Kreutzer, 2011).
Furthermore, the final report of the Liberian Truth and Reconciliation Commission (2009)
identified tribalism or ethnic division as one of the three primary causes of the civil war.
Scholars such as Reno (1998) and Outram (1999) contend that warlords exploited ethnicity
for political and economic ends during Liberia’s conflict. While additional post-conflict

6
qualitative studies and surveys, including those by Pugel (2007) and Bøås and Hatløy (2008),
offer significant insights into the origins and nature of the Liberian civil war, they fall short
in detailing the specific intra-national locations of wartime violence, thus underscoring the
need for a systematically disaggregated quantitative analysis.
The Liberian Truth and Reconciliation Commission discerned that a fundamental un-
derlying cause of the civil war was rooted in the historical decision to establish Liberia as
a nation divided between settlers—Americo-Liberians, or freed slaves from America who
settled in Liberia in 1822—and the indigenous Liberians, reinforced by the use of force to
maintain Americo-Liberians’ dominance (TRC Final Report 2009, p. 49). Despite the gen-
eral clash between the indigenous Liberians, comprising at least 95% of the then 2.5 million
population, and the elite minority Americo-Liberians, accounting for just under 5% of the
population, Liberia remains an ethnically diverse country. In Liberia, ethnicity refers to
tribal identities and sociocultural groups, with the Americo-Liberians representing just one
of seventeen "generally recognized sociocultural groups" (TRC Final Report, 2009, p. 51).
The 95% indigenous Liberian population is composed of sixteen recognized ethnic and socio-
cultural groups, including the Bassa, Gbandi, Krahn, Mano, Loma, Vai, Mende, Mandingo,
Belle, Kru, Grebo, Gola, Dei, Gio, Kpelle, and Kissi. Each of these groups possesses distinct
linguistic and cultural identities, despite some similarities among them, and were distributed
across thirteen counties (Grimes, 2000).
Ethnicity in Liberia, as in many other regions globally, is "a complex and often artificial
notion" (TRC Final Report, 2009, p. 52). Historically, ethnic polarization and divergences in
ethnic-religious ideologies and practices have engendered prejudice and tension in daily life,
underpinned various forms of discrimination against certain minorities or underrepresented
tribes, and exacerbated land disputes across various counties in Liberia (The International
Peacebuilding Alliance - Interpeace, 2008).
As the Liberian civil war evolved, it became increasingly characterized by factional af-
filiations, often defined along ethnic lines (Young, 2008). For instance, the majority of

7
combatants in the National Patriotic Front of Liberia (NPFL), the initial rebel group of the
Liberian civil war led by warlord Charles Taylor, belonged to the Gio and Mano ethnic groups
from Nimba County in Northeastern Liberia. These groups formed the foundational sup-
port for Taylor’s forces. The Liberians United for Reconciliation and Democracy (LURD),
another major rebel faction opposing Taylor, predominantly consisted of the Mandino and
Krahn ethnic groups, with the Krahn originating from Grand Gedeh in the Southeast and
the Mandingos from Lofa County in the North and Nimba County in the Northeast (El-
lis, 1999). LURD subsequently fragmented, leading to the emergence of the Movement for
Democracy in Liberia (MODEL) as a breakaway faction composed of the Krahn, while the
Mandingoes continued their involvement in LURD.

3.1.1 Context of the Liberian Civil War

Between 1989 and 2003, Liberia was engulfed in a devastating civil war, resulting in the
loss of over 300,000 lives and the displacement of numerous civilians, including women and
children (Ballah & Abrokwaa, 2003). This period of intense conflict followed a phase of
relative political stability (Kieh, 2004) and was marked by exacerbated ethnic tensions,
political instability, and socioeconomic decline. Various Liberian scholars have attributed
the war to multiple factors, including the predatory abuse of power, corrupt political systems,
economic disparities, and discrimination stemming from the elite dominance of a small group
of Americo-Liberians (Dolo, 1996; Ballah & Abrokwaa, 2003).
Identifying the root causes of civil wars presents a complex challenge, often involving the
obscure nature of these causes and the difficulty in distinguishing them from their effects.
The Liberian civil war, in particular, has been analyzed through various lenses, including
religious conflicts (Ellis, 1999), dynamics of “power, plunder, and revenge” (Ellis, 2001, p.
224), warlord commerce (Reno, 1998), the concept of "a new war" or gradual state weakening
(Kaldor, 2007, p. 1), issues of identity and resource distribution (Bøås, 2005), and youth
dissatisfaction (Utas, 2003). Utas, based on ethnographic fieldwork and interviews conducted

8
in 1988, advocates for an ’opportunity’ explanation of the Liberian civil war, encompassing
perspectives from Monrovia and surrounding rural areas.1
The Liberian civil conflict unfolded in two distinct phases. The first, spanning from 1989
to 1997, commenced on Christmas Eve with Charles Taylor’s rebel forces entering Butuo,
Nimba County. Taylor’s insurgency eventually led to the assassination of President Samuel
K. Doe, who had seized power in a 1980 coup d’état, ending Americo-Liberian oligarchic rule.
The initial attempts to overthrow President Doe included a failed coup in 1985, primarily
orchestrated by the Gio and Mano ethnic groups, followed by the successful offensive led by
Charles Taylor’s forces (Dunn, 1995). This attack precipitated the formation of the National
Patriotic Front of Liberia (NPFL) (Reno, 1997), intensifying the conflict and resulting in
seven years of brutal violence, which claimed over two hundred thousand lives and displaced
more than half of Liberia’s then 2.5 million population (Humphreys & Richards, 2005). The
aftermath of Doe’s assassination in 1997 saw Charles Taylor’s election as president.
The second civil war phase, from 1999 to 2003, began with Taylor’s presidency. The con-
flict was marked by the emergence of the Liberians United for Reconciliation and Democracy
(LURD) in the Northwestern part of Liberia and the Movement for Democracy in Liberia
(MODEL) in the Southeast, both opposing Taylor’s NPFL (Kamara, 2003). LURD’s coun-
terstrategies culminated in Taylor’s resignation and the peace talks in Accra, Ghana, in
August 2003, effectively ending this phase of the war (Pugel, 2007). Subsequently, members
of the Krahn ethnic group, belonging to former President Samuel Doe’s community, gained
control over significant political and military institutions (Nilsson & Kovacs, 2005). Collier
and Hoeffler (2004) note that, unlike other developing countries’ civil conflicts, the Liberian
civil war was predominantly driven by greed rather than grievance.
The Liberian civil crisis resulted in dire consequences, including heightened malnourish-
ment, soaring unemployment, and other socioeconomic conditions leading to severe emotional
distress (Dolo, 1996). Only recently has the impact of civil wars on economic development
1
Utas’s fieldwork and interviews were carried out in Liberia’s capital, Monrovia, as well as in nearby
provinces and rural areas.

9
been fully recognized. They not only destroy infrastructure and human resources but also
dismantle institutions and the rule of law, creating an uncertain environment for future for-
eign investment. These deleterious effects collectively contribute to a “conflict trap,” keeping
countries mired in low per capita income equilibria (Collier, 2003, p. 106).

3.2 Unit of Analysis and Sample Size

The primary unit of analysis in this paper is the clan. Liberia is organized into fifteen first-
order political subdivisions known as counties, which are further divided into 136 second-
order administrative divisions, termed districts. These districts are subsequently subdivided
into third-order local administrative systems, referred to as clans. The clan system, estab-
lished in 1904, was initially implemented to organize indigenous communities and chiefdoms
(Sawyer, 1995; Nyei, 2014). Historically, these clans have not only loosely corresponded to
local political entities but have also, to some extent, represented "parts of a limited number
of officially recognized tribes" (Nelson, 1985, p. 14). As per the 1974 National Population
and Housing Census—Liberia’s third national census since its independence—the country
was comprised of 351 clans, which constitute the third-level administrative subdivisions.
These 351 clans form my sample size for this paper.

3.2.1 Data

This paper integrates a unique sub-national administrative dataset, personally collected

during extensive fieldwork in Liberia, with geo-coded data on political violence within the
country. The wartime violence data is sourced from the Armed Conflict Location & Event
Dataset (ACLED), which comprehensively collects, analyzes, and maps a range of attributes
of political violence and protest events, including locations, actors, dates, fatalities, and event
types, in countries facing instability (Raleigh, 2010, p. 652). As the most extensive sub-
national dataset on political violence (Raleigh, 2012), ACLED’s detailed records of violent
civil war events facilitate quantitative research into various local-level factors, especially those

10
influencing the escalation of civil conflicts. The local-level subnational ethnic composition
data is extracted from the 1974 Liberian National Population and Housing Census, conducted
before the onset of the Liberian civil war, thus providing pre-treatment data to control for
potential wartime impacts on ethnic distribution.
The process of accessing the ACLED data involved straightforward online steps, includ-
ing account creation and dataset downloading. In stark contrast, acquiring the clan-level
subnational ethnic composition data from the 1974 National Population and Housing Cen-
sus of Liberia demanded extensive on-the-ground data collection and substantial financial
resources, all undertaken personally. This critical dataset for my research necessitated direct
interaction and collaboration with key governmental agencies in Liberia, particularly the
Liberia Institute of Statistics and Geo-Information Services, to ensure accurate and compre-
hensive data gathering.

11
Figure 1: Source: The clan shapefiles (digital maps) were acquired during fieldwork from the
Department of Geographic Information System (GIS) at the Liberia Institute of Statistics
and Geo-Information Services (LISGIS). These shapefiles were produced based on informa-
tion pertaining to clans from the 1974 Liberia National Population and Housing Census. In
these digital maps, red crosses within the polygons denote wartime events—interpreted in a
broad sense—in each clan. The geo-referenced data for these wartime events were sourced
from the Armed Conflict Location & Event Dataset (ACLED). Spatial analysis was con-
ducted using packages within the R statistical software to accurately map ACLED events to
their respective clans.

12
3.3 Operationalization of Variables

3.3.1 Independent Variables

My independent variable is ethnic diversity. I measure the level of ethnic diversity using
the index of ethnolinguistic fractionalization (ELF). Numerous scholars have explained the
observed negative correlation between ethnic diversity and economic growth as a result of
the elevated likelihood of conflict in highly fractionalized societies. Consequently, many
studies employ the ELF index as a measure of ethnic heterogeneity. Calculated by Taylor
and Hudson (1972), the fractionalization index measures the probability that two randomly
chosen individuals from a given unit of analysis—in this case, a given clan—will belong to
different ethnic groups. Simply put, the ethnic diversity aspect of my independent variable
looks at how many different ethnic groups are present in each clan.
Formula 1 below represents the equation of the ethnolinguistic fractionalization index.
In the formula, πi indicates the share or proportion of ethnic groups and N represents the
amount or sum of ethnic groups in each clan. In plain words, ELF is equal to one minus the
sum of the squares of ethnic groups in a clan.2
The Ethno-Linguistic Fractionalization Index (ELF) is delineated by the following for-
mula:
N
X
ELF = 1 − (πi )2 (1)
i=1

Where:

ELF : Ethno-linguistic fractionalization index

πi : Share or proportion of ethnic group i

N : Total number of ethnic groups

For robustness and because ethnic fractionalization alone may arguably not be sufficient
2
The original source of the ELF is Taylor and Hudson (1972). However, the current formula I am
referencing has been extracted from the work of Montalvo and Reynal-Querol (2002, 2005).

13
to explain the outbreak and escalation of violence, I also consider the level of polarization
or the distribution of the population across group lines. To measure this, I use the index
of ethnic polarization (RQ). Originally constructed by Reynal-Querol (1998), the RQ index
captures the distance of ethnic groups from a bipolar distribution. Following the argument
that ethnic violence requires a big minority, not multiple subdivided minorities, facing an
ethnic majority (Horowitz, 1985), the RQ index considers the shares or percentage distribu-
tion of ethnic groups. Unlike the index of fractionalization, the RQ index captures the idea
“that a large minority is the worst possible situation” for a high likelihood of violence because
the index is close to maximum in this case (Montalvo and Reynal-Querol, 2005, p. 798). I
use the RQ index not only because it is regarded as an influential predictor of civil wars,
but also because its explanatory power is shown to increase with the escalation of conflicts’
intensity (Montalvo and Reynal-Querol, 2005; Montalvo and Reynal-Querol, 2008).
The mathematical formula for RQ is slightly more complex than that of the ELF. RQ is
represented by Equation 2. As in the ELF, πi in RQ also indicates the share or proportion
of ethnic groups, and N represents the number or sum of ethnic groups in a clan. In plain
words, RQ is equal to the sum of the squared differences between each ethnic group’s share
and 0.5 (1/2), weighted by the share of the ethnic group within a clan. 3

The index of ethnic polarization (RQ) is represented by the formula:

N  2
X 0.5 − πi
RQ = 1 − πi (2)
i=1
0.5

Where:

RQ : Index of ethnic polarization

πi : Share or proportion of ethnic group i

N : Total number of ethnic groups in a clan

3
The original source of the RQ index is Reynal-Querol (1998, 2001). However, I have extracted the
current formula from Montalvo and Reynal-Querol (2002, 2005) and Reynal-Querol (2002).

14
3.3.2 Dependent Variables

My dependent variable is the escalation of civil conflicts —in other words, the presence,
number and intensity of violence in each clan during the civil war. As indicated above in the
data section, I draw on the Armed Conflict Location and Event Dataset, which codes various
types of conflict events—from strategic developments to fatalities—at the sub-country level.
However, this study cares mainly about violent events. From ACLED’s event and sub-
event types, I use seven measures of wartime violence—violence against civilians, violence
against civilian fatalities, total battles, total riots, total protest, total fatalities, and total
events. Several quantitative studies utilizing ACLED have looked at one key dependent
variable or aspect of violence. For example, Hegre et al. (2009) examined the broader
wartime events, and Raleigh (2012) investigated violence against civilians. By inspecting
seven distinct types of violence from ACLED in this study and regressing each of my seven
dependent variables on measures of ethnic diversity, I hope to better highlight diversity’s
relationship with violence, thus providing more nuanced accounts of the potential effect of
ethnic differences and polarization on the degree of conflict within countries.

3.4 Statistical Method

I conduct multiple ordinary least squares regression (OLS) models to understand the rela-
tionship between ethnic diversity and various measures of wartime violence. I utilize two
different measures of ethnic diversity, which are commonly used in the literature. Table 1
is based on the statistical models where I used the index of ethnic polarization (RQ) as
the independent variable, and Table 2 is based on the models where I operationalize ethnic
diversity using the index of ethnic linguistic fractionalization (ELF). Both indices of ethnic
diversity range from 0 to 1, with 1 representing maximum ethnic diversity and 0 representing
complete ethnic homogeneity.
I include seven measures of wartime violence in the analysis, which are violence against
civilians, violence against civilian fatalities, total battles, total riots, total protest, total

15
fatalities, and total events. These measures are presented across the two tables. Additionally,
I controlled for population size and density, distance to the capital, distance to diamond
mines, and distance from neighboring conflict and border areas. These control variables
were included in the statistical models based on strong theoretical justifications as explained
below. I used the same control variables in both tables.

3.4.1 Control Variables

To improve the robustness of the study’s models and better account for alternative factors
influencing conflict escalation, I incorporate a collection of control variables that might be
linked to the geographical occurrence of conflicts within country. I provide explanation for
these control variables as follow:
Distance from Monrovia (Liberia’s capital): Regions closer to the capital are typi-
cally better governed than those farther away. The distance from the capital city serves as a
useful indicator of the likelihood of violence outbreak. Governments often allocate significant
resources to maintain peace in areas near the capital. Thus, regions close to the capital city
are expected to experience less wartime violence, even if they are highly diverse.
Distance from diamond mines: Natural resources are a critical source of tension
between ethnic groups in low-income countries. Ethnic groups may resort to violence to
control valuable natural resources in their vicinity, such as diamond mines. Consequently,
the proximity of ethnic groups to the nearest diamond mine can escalate ethnic tensions or
even lead to full-fledged civil wars. I anticipate that more ethnically diverse clans located
closer to diamond mines face a higher risk of conflict. Natural resources can be both a
blessing and a curse, as they have the potential to trigger conflict and instability.
Distance from neighboring countries: The distance of clans from neighboring coun-
tries can be another significant source of conflict. I have included this variable in the model
because neighboring countries also fueled tensions in Liberia. They provided financial and
armed support to various wartime factions, and rebel recruitment took place in Liberia’s

16
neighboring Mano River Union countries during the civil war. The conviction of Liberian
war criminal Charles Taylor by The Hague court, partly for extending the war to neighboring
Sierra Leone (Bowcott, 2013), illustrates the cross-border impact of the conflict. I expect
that clans closer to border regions would be at a greater risk of violent conflict, and thus,
I control for the distance of clans from Sierra Leone, Guinea, and the Ivory Coast in my
models.
Population Density: Increased population density, particularly in ethnically diverse
regions, is posited to elevate the risk of civil unrest (Raleigh and Hegre, 2009). This phe-
nomenon arises as intensified competition for essential public resources—such as education,
healthcare, and employment—fuels grievances among different ethnic groups. In areas where
diverse populations coexist, these struggles for resource allocation can become acute, poten-
tially exacerbating ethnic tensions and conflicts. This dynamic is exemplified in the Liberian
civil war, specifically through the protracted land disputes between the Gio and Mandingo
ethnic communities in Nimba County, northeastern Liberia.4
Population Size: As with population density, the size of the population can theoreti-
cally increase the likelihood of conflict escalation due to increased competition for resources
(Buhaug and Urdal, 2013), potential for recruitment by militant groups (Weidmann 2009),
relative ethnic group sizes (Cederman, Weidmann, and Gleditsch, 2011), and challenges
related to governance (Goldstone, 2002).

4 Results

4.1 Ethnic polarization and wartime violence

In the regression outputs in Table 1 below, my variable of interest—Ethnic Polarization

(RQ)—does not show a statistically significant relationship with fatalities (Model 1), civilian
4
Notably, the animosity between the Gios and the Mandingos was further compounded by their allegiances
to opposing factions during the Liberian conflict.

17
fatalities (Model 2), battles (Model 3), and events (Model 4). However, it appears to have
a positive and significant impact on protests (Model 6) at the 5% level, suggesting that
higher ethnic polarization increases the likelihood of protests. Interestingly, the coefficient
for riots (Model 7) is negative and significant at the 5% level, indicating that higher ethnic
polarization might be associated with a decrease in riots.
Control Variables:

• Population Size shows a positive correlation with battles (Model 3), events (Model
4), and civilian violence (Model 5), all significant at the 1% level. This suggests that
larger populations may experience more of these types of violence.

• Population Density is positively correlated with fatalities, battles, and events (Models
1, 3, and 4 respectively), each significant at the 1% level, implying that denser areas
are more likely to witness these forms of violence.

• Distance from Monrovia is positively associated with battles (Model 3) at the 10%
significance level, potentially indicating that proximity to the capital may relate to the
occurrence of battles.

• Distance to Diamond Mines presents a strong positive relationship with fatalities,

civilian fatalities, and battles (Models 1, 2, and 3) at the 1% significance level, and
with events (Model 4) at the 5% level, which could suggest that areas closer to diamond
mines are more prone to such violence.

• Distance from Guinea shows a significant negative relationship with fatalities (Model
1) at the 5% level, indicating fewer fatalities further from this border.

• Distance from Sierra Leone is negatively associated with civilian fatalities, battles, and
events (Models 2, 3, and 4), all significant at the 1% level, suggesting that proximity to
the Sierra Leone border may be related to reduced instances of these types of violence.

18
 Table 1: Regressing my seven measures of wartime violence on ethnic
polarization index (RQ)

Table 1: Table 1 shows results from OLS regressions with ethnic polarization index (RQ) as
the independent variable in each model and violence against civilians, total fatalities, total
events, total protest, total riots, violence against civilian fatalities, and total battles5 in a
clan as dependent variables.

Model Fit and Significance: The models’ R-squared values vary, with the highest
explanatory power seen in models for riots (Model 7) and protests (Model 6), suggesting
these models explain a significant portion of the variability in the dependent variables. The
F statistics are significant across all models, indicating that the sets of independent variables
are jointly significant in explaining variations in the dependent variables.
5
Battles are perhaps the most recurrent ACLED event type in the case of the Liberian civil war.

19
4.2 Ethnic fractionalization and wartime violence

The regression results in Table 2 indicates that my variable of interest, Ethnic Fractionaliza-
tion (ELF), shows no statistically significant association with fatalities (Model 1), civilian
fatalities (Model 2), battles (Model 3), events (Model 4), civilian violence (Model 5), protests
(Model 6), or riots (Model 7). This suggests that ethnic fractionalization, as measured by
the ELF index, does not have a discernible direct effect on these specific measures of wartime
violence within the scope of this data.
Control Variables:

• Population Size is found to have a positive and statistically significant effect on battles
(Model 3), events (Model 4), and civilian violence (Model 5), with a significance level
at 1%. This indicates that larger population sizes are correlated with an increase in
these forms of violence.

• Population Density has a mixed influence, with a significant positive effect on fatalities
and events (Models 1 and 4 respectively) at the 1% level, but it does not significantly
affect the other forms of violence.

• Distance to Monrovia shows a significant positive relationship with battles (Model 3)

at the 10% level, hinting at more frequent occurrences of battles closer to the capital.

• Distance to Diamond Mines is significantly and positively associated with fatalities,

civilian fatalities, battles (Models 1, 2, and 3), and events (Model 4), suggesting a
linkage between proximity to diamond mines and higher instances of violence.

• Distance from Guinea has a negative effect on fatalities (Model 1) at the 5% level,
implying fewer fatalities occur further away from the Guinean border.

• Distance from Sierra Leone shows a consistently negative relationship with civilian
fatalities, battles, and events (Models 2, 3, and 4), each significant at the 1% level.

20
 This could suggest that areas further from the Sierra Leone border are associated with
lower instances of these forms of violence.

Table 2: Regressing my seven measures of wartime violence on ethno-linguistic

fractionalization index (ELF)
n

Figure 2: Table 2: Table 2 shows results from OLS regressions with ethno-linguistic frac-
tionalization index (ELF) as the independent variable in each model and violence against
civilians, total fatalities, total events, total protest, total riots, violence against civilian fa-
talities, and total battles in a clan as dependent variables

Model Fit and Significance: The R-squared values across the models indicate the
proportion of variance for a dependent variable that’s explained by the independent variables
in the model. The values suggest that the models explain a considerable amount of variance
in battles (Model 3), events (Model 4), and riots (Model 7), with R-squared values above 0.5,

21
which is relatively high for social science data. The F statistics in each model are significant,
which indicates that the models are a good fit for the data and that the independent variables,
as a set, provide a statistically significant prediction of the dependent variables.

5 Discussion

This paper’s empirical analysis offers a nuanced understanding of the relationship between
ethnic diversity and wartime violence in Liberia, particularly focusing on ethnic polarization
and fractionalization. The findings illuminate the intricate dynamics of violence in ethnically
diverse settings and highlight the importance of considering multiple factors when assessing
the consequences of ethnic diversity in conflict situations.

5.1 Ethnic Polarization and Its Varied Impact on Violence

The results reveal that ethnic polarization, as operationalized in this study, exhibits a com-
plex relationship with different forms of wartime violence. Notably, the significant impact
of ethnic polarization on protests, as demonstrated in the regression models, suggests that
higher levels of ethnic polarization may indeed escalate protest activities. This finding is
consistent with theoretical perspectives that posit heightened ethnic tensions leading to in-
creased public expressions of dissent. Conversely, the negative and significant association
with riots is counterintuitive, implying that higher ethnic polarization might be linked with
a decrease in such violent expressions. This paradoxical outcome invites further investiga-
tion into the specific contexts and mechanisms through which ethnic polarization influences
collective actions in conflict settings.

5.1.1 Control Variables: Demographic and Geographic Influences

The statistical significance of various control variables such as population size, density, and
geographic location relative to Monrovia, diamond mines, and borders with Guinea and

22
Sierra Leone, underscores the multifaceted nature of wartime violence. Larger population
sizes and denser areas, for instance, are correlated with increased violence, aligning with
literature that emphasizes demographic pressures in conflict-prone regions. The proximity
to diamond mines and national borders also appears to be a crucial factor, reflecting the
role of economic incentives and cross-border dynamics in the escalation of violence. These
findings contribute to a more comprehensive understanding of the interplay between ethnic
diversity and other critical factors in shaping the landscape of wartime violence.

5.1.2 Ethnic Fractionalization: A Non-significant Predictor

Interestingly, the analysis shows that ethnic fractionalization, measured by the ELF index,
does not significantly predict the various forms of wartime violence examined. This outcome
challenges some conventional assumptions in the field about the direct impact of ethnic
diversity on conflict. It suggests that the relationship between ethnic fractionalization and
violence may be more complex than previously thought, possibly mediated by other social,
economic, and political factors. This insight calls for a broader approach to studying ethnic
conflicts, one that incorporates a variety of determinants beyond mere ethnic composition.

5.1.3 Implications and Future Research Directions

The findings of this paper, particularly the complex and sometimes counterintuitive results
regarding ethnic polarization and fractionalization, demonstrate the challenges in drawing
generalized conclusions about the effects of ethnic diversity on violence. This complexity is
further compounded by the potential limitations of the study, such as the relatively small
sample size and the constraints inherent in using only subnational data. These limitations
suggest a need for future research to expand the data scope both in terms of sample size and
temporal coverage.
Future studies should aim to collect more comprehensive subnational data over extended
time periods to provide a more detailed and nuanced understanding of how ethnic diversity

23
interacts with various forms of violence. Additionally, exploring alternative methodologies
or analytical frameworks may yield further insights into the complex dynamics at play in
ethnically diverse settings experiencing conflict.
In conclusion, while this paper contributes to the understanding of ethnic diversity’s role
in conflict, it also highlights the need for continued exploration and nuanced analysis in this
area. The multifaceted nature of ethnic conflicts, as evidenced by the findings, underscores
the importance of a holistic and context-specific approach in examining the causes and
manifestations of violence in ethnically diverse societies.

24
References
[1] Alesina, Alberto, Arnaud Devleeschauwer, William Easterly, Sergio Kurlat, and Romain
Wacziarg. “Fractionalization,” 2003. https://doi.org/10.3386/w9411.
[2] Alptekin, Hüseyin. “A Theory of Ethnic Violence: Ethnic Incorporation and Ethno-
Political Mobilization in Bulgaria and Cyprus.” Ethnic and Racial Studies 40, no. 15
(2016): 2672–90. https://doi.org/10.1080/01419870.2016.1266008.
[3] Asiedu, Kwabena. Comprehending the Ethnic Dynamic of the Liberian Civil Conflict.
Burnaby, B.C.: Simon Fraser University, 2004.
[4] Ballah, Heneryatta, and Clemente K. Abrokwaa. “Ethnicity , Politics and Social Conflict
: The Quest for Peace in Liberia.” Semantic Scholar, 2003.
[5] Bøås, Morten, and Anne Hatløy. “’Getting in, Getting out’: Militia Membership and
Prospects for Re-Integration in Post-War Liberia.” The Journal of Modern African Stud-
ies 46, no. 1 (2008): 33–55. https://doi.org/10.1017/s0022278x07003060.
[6] Bøås, Morten. “The Liberian Civil War: New War/Old War?” Global Society 19, no. 1
(2005): 73–88. https://doi.org/10.1080/1360082042000316059.
[7] Buhaug, Halvard, and Henrik Urdal. “An Urbanization Bomb? Population Growth
and Social Disorder in Cities.” Global Environmental Change 23, no. 1 (2013): 1–10.
https://doi.org/10.1016/j.gloenvcha.2012.10.016.
[8] Buhaug, Halvard, and Jan Ketil Rød. “Local Determinants of African Civil Wars,
1970–2001.” Political Geography 25, no. 3 (2006): 315–35. https://doi.org/10.1016/
j.polgeo.2006.02.005.
[9] Buhaug, Halvard, Lars-Erik Cederman, and Jan Ketil Rød. “Disaggregating Ethno-
Nationalist Civil Wars: A Dyadic Test of Exclusion Theory.” International Organization
62, no. 3 (2008): 531–51. https://doi.org/10.1017/s0020818308080181.
[10] Buhaug, Halvard, Scott Gates, and Päivi Lujala. “Geography, Rebel Capability, and the
Duration of Civil Conflict.” Journal of Conflict Resolution 53, no. 4 (2009): 544–69.
https://doi.org/10.1177/0022002709336457.
[11] Cederman, Lars-Erik, and Kristian Skrede Gleditsch. “Introduction to Special Issue on
‘Disaggregating Civil War.” ’ Journal of Conflict Resolution 53, no. 4 (2009): 487–95.
https://doi.org/10.1177/0022002709336454.
[12] Cederman, Lars-Erik, Kristian Skrede Gleditsch, and Simon Hug. “Elections and Ethnic
Civil War.” Comparative Political Studies 46, no. 3 (2012): 387–417. https://doi.org/
10.1177/0010414012453697.
[13] CEDERMAN, LARS-ERIK, NILS B. WEIDMANN, and KRISTIAN SKREDE GLED-
ITSCH. “Horizontal Inequalities and Ethnonationalist Civil War: A Global Compari-
son.” American Political Science Review 105, no. 3 (2011): 478–95. https://doi.org/
10.1017/s0003055411000207.

25
[14] Chaudoin, Stephen, Zachary Peskowitz, and Christopher Stanton. “Beyond Zeroes and
Ones: The Intensity and Dynamics of Civil Conflict,” 2014. https://doi.org/10.
3386/w20258.

[15] Collier, P., and Anke Hoeffler. “Greed and Grievance in Civil War.” Oxford Economic
Papers 56, no. 4 (2004): 563–95. https://doi.org/10.1093/oep/gpf064.

[16] Collier, Paul. Breaking the Conflict Trap: Civil War and Development Policy. Washing-
ton, DC: World Bank, 2003.

[17] CRUZ, CESI, JULIEN LABONNE, and PABLO QUERUBÍN. “Social Network Struc-
tures and the Politics of Public Goods Provision: Evidence from the Philippines.” Amer-
ican Political Science Review 114, no. 2 (2020): 486–501. https://doi.org/10.1017/
s0003055419000789.

[18] Dolo, Emmanuel. Democracy versus Dictatorship: The Quest for Freedom and Justice
in Africa’s Oldest Republic–Liberia. Lanham: University Press of America, 1996.

[19] Dunn, D. Elwood. Liberia. Vol. 23. World Bibliographical Series. Oxford, England: Clio
Press, 1995.

[20] Easterly, W., and R. Levine. “Africa’s Growth Tragedy: Policies and Ethnic Divisions.”
The Quarterly Journal of Economics 112, no. 4 (1997): 1203–50. https://doi.org/
10.1162/003355300555466.

[21] Ellis, Stephen. “Mystical Weapons: Some Evidence from the Liberian War.” Journal of
Religion in Africa 31, no. 2 (2001): 222. https://doi.org/10.2307/1581517.

[22] Ellis, Stephen. “The Mask of Anarchy. the Destruction of Liberia and the Religious
Dimension of an African Civil War.” Critique internationale n° 7 2 (1999): 73–73.
https://doi.org/10.3917/crii.p2000.7n1.0073.

[23] FEARON, JAMES D., and DAVID D. LAITIN. “Ethnicity, Insurgency, and Civil War.”
American Political Science Review 97, no. 01 (2003): 75–90. https://doi.org/10.
1017/s0003055403000534.

[24] Fjelde, Hanne, and Gudrun Østby. “Socioeconomic Inequality and Communal Conflict:
A Disaggregated Analysis of Sub-Saharan Africa, 1990–2008.” International Interactions
40, no. 5 (2014): 737–62. https://doi.org/10.1080/03050629.2014.917373.

[25] Gates, Scott & Håvard Strand. “Modeling the Duration of Civil Wars: Measurement
and Estimation Issues.” 2004. Presented at the Meeting of the Standing Group on In-
ternational Relations, the Hague September 9-11.

[26] Goldstone, Jack A. “Population and Security: How Demographic Change Can Lead to
Violent Conflict.” Journal of International Affairs 56, no. 1 (2002): 3–21.

[27] Grimes, Barbara. “Ethnologue. Languages of the World.” African Studies Companion
Online, 2000. https://doi.org/10.1163/_afco_asc_441.

26
[28] Hegre, Håvard, Gudrun Østby, and Clionadh Raleigh. “Poverty and Civil War Events.”
Journal of Conflict Resolution 53, no. 4 (2009): 598–623. https://doi.org/10.1177/
0022002709336459.

[29] Humphreys, Macartan, and Raul Richards. “Prospects and opportunities for achieving
the MDGs in post-conflict countries: A case study of Sierra Leone and Liberia.” CGSD
Working Paper No. 27, Center on Globalization and Sustainable Development, The
Earth Institute at Columbia University. 2005.

[30] Kaldor, Mary. New and Old Wars: Organized Violence in a Global Era. 2nd ed. Stanford,
CA: Stanford University Press, 2007.

[31] Kamara, Thomas. “Liberia: Civil war and regional conflict.” Writenet paper no. 17 for
United Nations High Commissioner for Refugees. 2003. http://www.globalsecurity.
org/military/library/report/2003/030501-liberia.pdf.

[32] Kieh Jr, George Klay. “Irregular Warfare and Liberia’s First Civil War.” Journal of
International and Area Studies 11 (2004): 57-77.

[33] Koren, Ore, and Anoop K Sarbahi. “State Capacity, Insurgency, and Civil War: A
Disaggregated Analysis.” International Studies Quarterly, 2018. https://doi.org/10.
1093/isq/sqx076.

[34] KUSTOV, ALEXANDER, and GIULIANA PARDELLI. “Ethnoracial Homogeneity and

Public Outcomes: The (Non)Effects of Diversity.” American Political Science Review
112, no. 4 (2018): 1096–1103. https://doi.org/10.1017/s0003055418000308.

[35] Montalvo, José G, and Marta Reynal-Querol. “Ethnic Polarization, Potential Conflict,
and Civil Wars.” American Economic Review 95, no. 3 (2005): 796–816. https://doi.
org/10.1257/0002828054201468.

[36] Nelson, Harold D. Liberia, a Country Study. Washington, D.C.: Headquarters, Dept. of
the Army, 1985.

[37] Nilsson, Desirée, and Mimmi söderberg Kovacs. “Breaking the Cycle of Violence?
Promises and Pitfalls of the Liberian Peace Process.” Civil Wars 7, no. 4 (2005):
396–414. https://doi.org/10.1080/13698280600683138.

[38] “Nimba County Reconciliation Project - Findings and Recommendations.”

Interpeace, September 23, 2008. https://www.interpeace.org/resource/
nimba-county-reconciliation-project-findings-and-recommendations/.

[39] Nyei, Ibrahim Al-bakri. “Decentralizing the State in Liberia: The Issues, Progress and
Challenges.” Stability: International Journal of Security & Development 3, no. 1 (2014).
https://doi.org/10.5334/sta.eg.

[40] Østby, Gudrun, Ragnhild Nordås, and Jan Ketil Rød. “Regional Inequalities and Civil
Conflict in Sub-Saharan Africa.” International Studies Quarterly 53, no. 2 (2009):
301–24. https://doi.org/10.1111/j.1468-2478.2009.00535.x.

27
[41] Outram, Quentin. “Liberia: Roots and Fruits of the Emergency.” Third World Quarterly
20, no. 1 (1999): 163–73. https://doi.org/10.1080/01436599913974.

[42] Pugel, James. “What the fighters say: A survey of ex-combatants in Liberia.” Monrovia,
Liberia: United Nations Development Programme. 2007. http://www.lr.undp.org/
UNDPwhatFightersSayLiberia-2006.pdf.

[43] Raleigh, Clionadh, and Havard Hegre. “Population Size, Concentration, and Civil War
: A Geographically Disaggregated Analysis.” Policy Research Working Papers, 2007.
https://doi.org/10.1596/1813-9450-4243.

[44] Raleigh, Clionadh, and Håvard Hegre. “Population Size, Concentration, and Civil War.
A Geographically Disaggregated Analysis.” Political Geography 28, no. 4 (2009): 224–38.
https://doi.org/10.1016/j.polgeo.2009.05.007.

[45] Raleigh, Clionadh. “Violence against Civilians: A Disaggregated Analysis.” Interna-

tional Interactions 38, no. 4 (2012): 462–81. https://doi.org/10.1080/03050629.
2012.697049.

[46] Reno, William. “War, Markets, and the Reconfiguration of West Africa’s Weak States.”
Comparative Politics 29, no. 4 (1997): 493. https://doi.org/10.2307/422016.

[47] Reno, William. Warlord Politics and African States. Boulder, CO: Lynne Rienner Pub-
lishers, 1998.

[48] Reynal-Querol, Marta, and Jose G. Montalvo. “Ethnic Polarization and the Duration
of Civil Wars.” Policy Research Working Papers, 2007. https://doi.org/10.1596/
1813-9450-4192.

[49] REYNAL-QUEROL, MARTA. “Ethnicity, Political Systems, and Civil Wars.” Jour-
nal of Conflict Resolution 46, no. 1 (2002): 29–54. https://doi.org/10.1177/
0022002702046001003.

[50] Rights, The Advocates for Human, Dulce Foster, Dianne Heins, Mark Kalla, Michele
Garnett McKenzie, James O’Neal, Rosalyn Park, et al. “A House with Two Rooms:
Final Report of the Truth and Reconciliation Commission of Liberia Diaspora Project.”
TexasScholarWorks. DRI Press, January 1, 1970. https://repositories.lib.utexas.
edu/handle/2152/7110.

[51] Sawyer, Amos. “The Failure of the Centralized State: Institutions and Self-Governance
...,” 1995. https://www.jstor.org/stable/45290125.

[52] The Guardian. “Charles Taylor: Liberian President’s Sentence Upheld.” The
Guardian, September 26, 2013. https://www.theguardian.com/world/2013/sep/26/
charles-taylor-liberian-president-sentence-upheld.

[53] Toft, Monica Duffy. The Geography of Ethnic Violence: Identity, Interests, and the
Indivisibility of Territory. Princenton etc.: Princenton University Press, 2006.

28
[54] Utas, Mats. Sweet Battlefields: Youth and the Liberian Civil War. Uppsala, Sweden:
University of Uppsala Press, 2003.

[55] Vinck, Patrick, Phuong Pham, and Tino Kreutzer. “Talking Peace: A Population-Based
Survey on Attitudes about Security, Dispute Resolution, and Post-Conflict Reconstruc-
tion in Liberia.” SSRN Electronic Journal, 2011. https://doi.org/10.2139/ssrn.
1874025.

[56] Weidmann, Nils B. “Geography as Motivation and Opportunity: Group Concentration

and Ethnic Conflict.” Journal of Conflict Resolution 53, no. 4 (2009): 526–543. https:
//doi.org/10.1177/0022002709336458.

[57] Young, Andrew T. “Costly Discrimination and Ethnic Conflict: The Case of the
Liberian Civil Wars.” SSRN Electronic Journal, 2022. https://doi.org/10.2139/
ssrn.4038881.

29