American Political Science Association The American Political Science Review

Ethnicity, Insurgency, and Civil War
Author(s): James D. Fearon and David D. Laitin

Source: The American Political Science Review, Vol. 97, No. 1 (Feb., 2003), pp. 75-90
Published by: American Political Science Association
Stable URL: https://www.jstor.org/stable/3118222
Accessed: 24-08-2020 02:40 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
American Political Science Association is collaborating with JSTOR to digitize, preserve and
extend access to The American Political Science Review
This content downloaded from 175.159.123.171 on Mon, 24 Aug 2020 02:40:31 UTC
All use subject to https://about.jstor.org/terms
American
American Political Science Review
Political Science Review Vol. 97, No. 1Vol.February
97, No. 1 February 2003
2003
Ethnicity, Insurgency, and

JAMES D. FEARON and DAVID
An influential conventional wisdom h
the Cold War and that the root cause
antagonisms. We show that the current p
accumulation of protracted conflicts sinc
with a new, post-Cold War international sys
more ethnically or religiously diverse coun
violence in this period. We argue for under
rural guerrilla warfare, a particular form
agendas. The factors that explain which cou
religious characteristics but rather the con
marks financially and bureaucratically wea
rough terrain, and large populations.
B etween 1945 and 1999, about 3.33 million battle The data cast doubt on three influential conventional
deaths occurred in the 25 interstate wars that wisdoms concerning political conflict before and after
killed at least 1,000 and had at least 100 dead the Cold War. First, contrary to common opinion, the
on each side. These wars involved just 25 states that prevalence of civil war in the 1990s was not due to the
suffered casualties of at least 1,000 and had a median end of the Cold War and associated changes in the inter-
duration of not quite 3 months. In contrast, in the same national system. The current level of about one in
period there were roughly 127 civil wars that killedsix at countries had already been reached prior to the
least 1,000, 25 of which were ongoing in 1999. A con- breakup of the Soviet Union and resulted from a steady,
servative estimate of the total dead as a direct result of gradual accumulation of civil conflicts that began im-
these conflicts is 16.2 million, five times the interstate mediately after World War II.
toll. These civil wars occurred in 73 states-more than a Second, it appears not to be true that a greater degree
third of the United Nations system-and had a median of ethnic or religious diversity-or indeed any partic-
duration of roughly six years.1 The civil conflicts in thisular cultural demography-by itself makes a country
period surely produced refugee flows far greater than more prone to civil war. This finding runs contrary to
their death toll and far greater than the refugee flowsa common view among journalists, policy makers, and
associated with interstate wars since 1945. Cases such academics, which holds "plural" societies to be espe-
as Afghanistan, Somalia, and Lebanon testify to cially the conflict-prone due to ethnic or religious tensions
economic devastation that civil wars can produce.and By antagonisms.
these crude measures, civil war has been a far greater Third, we find little evidence that one can predict
scourge than interstate war in this period, though itwhere has a civil war will break out by looking for where
been studied far less. ethnic or other broad political grievances are strongest.
What explains the recent prevalence of violent civil Were this so, one would expect political democracies
conflict around the world? Is it due to the end of the and states that observe civil liberties to be less civil
Cold War and associated changes in the international war-prone than dictatorships. One would further antici-
system, or is it the result of longer-term trends? Why
pate that state discrimination against minority religions
have some countries had civil wars while others have or languages would imply higher risks of civil war. We
not? and Why did the wars break out when they show did? that when comparing states at similar levels of per
We address these questions using data for the period
capita income, these expectations are not borne out.
1945 to 1999 on the 161 countries that had a populationThe main factors determining both the secular trend
of at least half a million in 1990. and the cross-sectional variation in civil violence in this
period are not ethnic or religious differences or broadly
held grievances but, rather, conditions that favor insur-
James D. Fearon and David D. Laitin are Professors, Department gency. Insurgency is a technology of military conflict
of Political Science, Stanford University, Stanford, CA 94305-6044characterized by small, lightly armed bands practicing
(jfearon@stanford.edu and dlaitin@stanford.edu). guerrilla warfare from rural base areas. As a form of
We wish to thank the many people who provided comments on warfare insurgency can be harnessed to diverse politi-
earlier versions of this paper in a series of seminar presentations.
The authors also gratefully acknowledge the support of the Nationalcal agendas, motivations, and grievances. The concept is
Science Foundation (Grants SES-9876477 and SES-9876530); sup-most closely associated with communist insurgency, but
port from the Center for Advanced Study in the Behavioral Sciences the methods have equally served Islamic fundamental-
with funds from the William and Flora Hewlett Foundation; valu-ists, ethnic nationalists, or "rebels" who focus mainly
able research assistance from Ebru Erdem, Nikolay Marinov, Quinn on traffic in coca or diamonds.
Mecham, David Patel, and TQ Shang; sharing of data by Paul Collier.
1 The interstate war data derive from Singer and Small 1994, updated We hypothesize that financially, organizationally, and
to include the Kargil and Eritrean wars. The bases for the civil war politically weak central governments render insur-
estimates are discussed below. gency more feasible and attractive due to weak local
75
Ethnicity, Insurgency, and Civil WarFebruary
February 2003
2003
policing or
that we presently ineptbelieve to meet the following primary and
tices. These often
criteria. (1) They involved fighting between agents ofin
indiscriminate (or claimants to) a state and organized, retal
nonstate groups
ant locals whointo sought either to take control rebelof a government, to f
weakness, take power we in a region, or to use argue,
violence to change gov-
come. Shocks ernment policies. (2) The conflict to killed atcouleast 1,000
from political over its course, with a yearly average insta of at least 100.
loss of a (3) At least 100 were killed on both sides (including
foreign pat
is favored by
civilians attacked by rebels). The rough
last condition is in-
edge of the
tended to rule out massacrespopulat where there is no orga-
and a large nized or effectivepopulati
opposition.4
from superior These criteria are broadly similar govern to those stated
financial by the Correlates of War (COW) project, Doyle an
support, and
Our data Sambanis show (2000), and several others. that We developed our m
grievances own list (workingfail from these and other to sources) mainly po
sures of because we wanted data for the whole 1945-99 period th
conditions
Surely ethnic and because of doubts about particular inclusions and
antagon
grievances exclusions inofteneach list.5 mo
But such Inbroad
one respect our data differ significantly fact from most
the cases others: where We see no reason in principle to exclude civi anti-
surgency colonialcanwars, such as the French be versus the suc National
bers of Liberation Front (FLN) in Algeria.
rebels under We count these as
require occuring within thea
only colonial small
empire. Thus, the French n
get going. state/empire looks highly civil war-prone by our list,
Using data on about 45 civil wars since 1960, Collier with six colonial wars occuring in the 1950s. But
and Hoeffler (1999, 2001) find similarly that measures to drop such cases would be like dropping the cur-
of "objective grievance" fare worse as predictors than rent conflict in Chechnya as a civil war in Russia
economic variables, which they initially interpreted as if the Chechens succeed in gaining independence. Al-
measures of rebel "greed" (i.e., economic motivation).2 ternatively, it would make even less sense to include
More recently, they argue that rebellion is better them as wars within "states" that did not exist (such as
explained by "opportunity" than by grievance (cf. "Algeria" in 1954).
Eisinger 1973 and Tilly 1978) and that the main de- There are both practical and theoretical consider-
terminant of opportunity is the availability of finance ations pointing the other way, however. In practical
and recruits for rebels. They proxy these with measures terms, to include the anticolonial wars in the analysis
of primary commodity exports and rates of secondary- requires that we form estimates of possible explana-
school enrollment for males. We agree that financing tory factors for whole empires, such as gross domes-
is one determinant of the viability of insurgency. We tic product (GDP) per capita, ethnic fractionalization,
argue, however, that economic variables such as per and democracy scores. Further, these estimates must
capita income matter primarily because they proxy change almost by year, as the colonial empires gradu-
for state administrative, military, and police capabil- ally diminished in size. We are able to use country-level
ities. We find no impact for primary commodity ex- data to produce such estimates for ethnic fractionaliza-
ports, and none for secondary schooling rates distinct tion, but our estimates for per capita income are more
from income. Our theoretical interpretation is more
Hobbesian than economic. Where states are relatively 4 We used the following secondary criteria to deal with several other
weak and capricious, both fears and opportunities en- coding issues. (4) The start year is the first year in which 100 were
killed or in which a violent event occurred that was followed by a
courage the rise of would-be rulers who supply a rough sequence of actions that came to satisfy the primary criteria. (5) If a
local justice while arrogating the power to "tax" for main party to the conflict drops out, we code a new war start if the
themselves and, often, for a larger cause. fighting continues (e.g., Somalia gets a new civil war after Siad Barre is
defeated in 1991). (6) War ends are coded by observation of a victory,
wholesale demobilization, truce, or peace agreement followed by at
CIVIL WAR SINCE 1945 least two years of peace. (7) Involvement by foreign troops does not
disqualify a case as a civil war for us, provided the other criteria are
Building on similar efforts by other civil war resear-
satisfied. (8) We code multiple wars in a country when distinct rebel
chers,3 we constructed a list of violent civil conflicts
groups with distinct objectives are fighting a coherent central state
on distinct fronts with little or no explicit coordination. (9) If a state
seeks to incorporate and govern territory that is not a recognized
2 There are 79 wars in their sample, but they lose about 34 due state,towe consider it a "civil war" only if the fighting continues after
missing values on explanatory variables, which are mainly economic. the state begins to govern the territory (thus, Indonesia/East Timor
Standard economic data tend to be missing for countries that 1975,
areyes, and India/Hyderabad 1947, no).
poor and civil war-torn. This highly nonrandom listwise deletion 5 Sambanis
may (2002) discusses a number of conceptual and operational
account for some of the differences between our results. ambiguities and problems with the COW civil war data. Collier and
3 In particular, Doyle and Sambanis (2000), Esty et al. (1998), Hoeffler (2001) base their list on COW data, breaking a number of
Gleditsch et al. (2002), the Institute for International and Strate-COW civil wars into multiple wars according to unspecified criteria
gic Studies (2000), Licklider (1995), Singer and Small (1994), Sivard
and including some colonial wars (coded in nonexistent "states," such
(1996), and Valentino (2002). as Angola 1961) but not others.
76
American
AmericanPolitical SciencePolitical
Review Science Review Vol. 97, Vol.
No. 97, No.11
FIGURE 1. Number and Percent

o _ _ o
O) CO
# civil wars /
o _, - - - % with civil war ,
CClIU' \
'o O r/ \ :--
0) C~ )
~~~~~0 'r~~~~~~~~~ \
-c ~ ~ ~ ~ ~ C 0
0e', ?e-
o - - o
1945 1950 1960 1970 1980 1990 1999
year
Russia are the of

problematic and the question most how to code
civil war-prone countriesthe
in the e
pires on a democracy index iswith
sample, onsets each.7Regarding theory
vexed.
six
the colonial empires differed so radically from oth
independent states, and Trends faced such an inhospitable
over Time
ternational environment after the war (with press
from the United States and the new United Nations
Figure 1 shows the number of countries with ongoing
system), that we need to be cautious about any empir-
civil wars by year from 1945 to 1999. Since the number
ical results that depend wholly on these cases. Thus
of we
independent states grew sharply in this period, it
analyze the data both with and without the anticolonial
also shows the proportion of countries with at least one
wars.
ongoing war in each year.
The graph indicates that, contrary to popular belief,
the prevalence of civil wars in the 1990s is not due
Descriptive Statistics to effects of the end of the Cold War. The 1999 level
We identified 127 conflicts that meet theof 25 ongoing

above wars had already been reached by the
criteria,
of which 13 were anticolonial wars.6 This mid 1980s.
makes forConflicts
127 associated with the Soviet collapse
were
civil war starts in a sample of 6,610 country partlya responsible
years, rate for the sharp increase in the
early 1990s,
of 1.92 per 100. The periods following major but a marked decline has followed.8
interna-
Onerates.
tional systemic change had the highest onset might Civil
conjecture that more and more civil wars
wars broke out in the late 1940s and the 1950s at 4.6 and are breaking out over time, thus producing the secu-
lar increase. This is incorrect. The rate of outbreak is
2.2 per 100 country-years, respectively, followed by the
1990s, at 2.0. In absolute terms, the largest number of2.31 per year since 1945, highly variable but showing
civil wars began in the 1990s (31), followed by the 1960s no significant trend up or down. The secular increase
stems from the fact that civil wars have ended at a
and 1970s (19 and 25, respectively).
Omitting the anticolonial conflicts, most civil wars rate of only about 1.85 per year. The result has been a
occurred in sub-Saharan Africa (34) and Asia (33),steady, almost-linear accumulation of unresolved con-
flicts since 1945.
followed by North Africa/the Middle East (17), Latin
America (15), Eastern Europe and the former Soviet Put differently, states in the international system
Union (13), and "the West" (2). The rate of outbreak have been subject to a more or less constant risk
was highest in Asia, at three per 100 country-years;
Africa, North Africa/the Middle East, and Latin 7 Four outbreaks are coded in the Soviet Union in 1946, in the Baltics
America all had rates of about two per 100 country- and Ukraine, plus Russia's two Chechen wars in the 1990s.
years. "France," Indonesia, and the Soviet Union/
8 Gurr (2000) notes the late-1990s decline in ethnic war and argues
that the trend reflects improved management strategies by states and
international organizations. The basic pattern in Figure 1 is not an
artifact of the way we have coded "civil war"; it is observed in a broad
range of other data sets on violent domestic conflict for this period
6 See Fearon and Laitin 2003 for a list of the conflicts. (e.g., Gleditsch et al. 2002).
77
Ethnicity,
Ethnicity,
Insurgency, and Civil War
Insurgency, and Civil War February
February
2003
2003
of violent the civil conflic

dominant group. (When preexisting differences are
flicts theyslight,
suffer
assimilation is more likely.) have
age duration of thus
The two approaches the civi
imply the same hypothesis
steadily from two
about the relationship between y
cultural diversity and
in 1999. civil conflict.
From a policy
tion about seeing as a
military H1: Measures
and of a country's ethnic or religious di-
political pr
makers versity should be associated
have faced with a higher risk of re
Bosnia, Somalia, Haiti, East Timor, Colombia, and civil war.
elsewhere.
Seeking to explain anticolonial nationalist move-
ments in countries barely affected by economic mod-
ernization, Gellner (1983, 108, nl) argued that even
ETHNICITY, DISCRIMINATION, "the advance shadow" of modernization was sufficient
AND GRIEVANCES
to start nationalist dynamics in motion. For Anderson,
nationalism quickly became a "modular form" that
During the Cold War, political scientists and sociolo-
could
gists often sought to trace rebellion to economic in- be easily "pirated," even under conditions quite
different
equality (Muller 1985; Paige 1975; Russett 1964), to from those during its origination. Nonethe-
less,
rapid economic growth said to destabilize traditionalif one took the modernist approach literally one
might
rural social systems (Huntington 1968; Scott 1976), or infer that more modernization should imply
more discrimination and thus more nationalist con-
to frustrations arising from the failure to gain expected
tention in culturally divided countries.
benefits of economic modernization (Gurr 1971). A
few scholars argued that the real source of rebellion
H2: The effect of ethnic diversity on the probability
was often ethnic nationalism (Connor 1994), and a rich
of civil war should increase at higher levels of per
literature on the sources nationalist mobilization de-
capita income (a proxy for economic moderniza-
veloped in comparative politics (e.g., Anderson 1983,
tion).
Deutsch 1953, and Gellner 1983). With the collapse
of the Soviet Union and Yugoslavia, such culturalist
Horowitz (1985) and several others argue that the
perspectives became a dominant frame for interpreting relationship between ethnic diversity and severe ethnic
inter- and intranational conflict (e.g., Huntington 1996).
conflict is nonmonotonic, with less violence for highly
Using a broad brush, we can distinguish between
homogeneous and highly heterogeneous countries. The
perennialist and modernist (or constructivist) positions
politics of a highly diverse country might be strate-
on the nature and sources of ethnic nationalism. Given
gically similar to those of a homogeneous one, since
in purest form by nationalist politicians and journalistssmall ethnic groups must join coalitions to have political
reporting on nationalist conflicts in progress, perenni-influence, and there are many possibilities for cross-
alist arguments stress long-standing cultural practices
cutting and shifts in coalition membership. Horowitz
said to define and distinguish ethnic groups. Differ-suggests that the most severe ethnic conflicts will arise
ences between these practices are argued to have madein countries where a substantial ethnic minority faces
conflict more likely. Academics rarely make such ar- an ethnic majority that can, given ethnic voting, win for
guments as baldly as do nationalist leaders. But au-sure in any national election. Such arguments yield the
thors who stress the long-standing, "deep" nature of following hypothesis.
ethnic differences and suggest that these make domes-
tic peace difficult include Huntington (1996), Ignatieff H3: Countries with an ethnic majority and a sig-
(1993), Moynihan (1993), Rabushka and Shepsle nificant ethnic minority are at greater risk for civil
(1972), and Smith (1986). Arguably, the main message war.
of Horowitz's (1985) influential book on ethnic conflict
is that plural societies face a host of pathologies that We consider several measures for these concepts.
render them especially prone to conflict and, at the H1 we use (1) the commonly employed ethnoling
extreme, violence. tic fractionalization (ELF) index based on data f
In contrast, modernist theories see the thorough- Atlas Narodov Mira 1964, which gives the probab
going politicization of cultural difference that ethnic that two randomly drawn individuals in a countr
nationalism represents as a development of the last 200 from different ethnolinguistic groups;9 (2) a me
to 500 years. The core argument is that economic mod- of the share of population belonging to the largest
ernization and the development of the modern statenic group that we constructed from the CIA Fact
and other sources (Fearon 2002); (3) the number
make upward social mobility possible, but contingent
on sharing the culture of the group that dominatesdistinct languages spoken by groups exceeding 1
state or society. When the state or society poses ascrip-
tive barriers to upward mobility for minority groups,
they may develop separatist nationalist movements: All9 For this and several other variables we filled in values for m
country years whenever possible based on our own research
the more so, according to Anderson (1983), Deutsch the sources were the CIA Factbook, Encyclopedia Brittanica
(1953), and Gellner (1983), the greater the preexisting
the Library of Congress Country Studies, though we used cou
cultural differences between the minority group and specific sources when necessary.
78
American
No. 97, No.1
1
the country's population,

by Deininger and Squire based
(1996); using theon Grim
country-
1996; and (4) a measure of
average of Gini values religious
produced the same results. For fr
(analogous to the ELF) that
policies on religion, we
we coded, construc
by decade, indicators
from the CIA Factbook and
for whether the state other
had an official religion, gavesourc
re-
interact these measures with
sources to one religion not availableperto others, capit
regu-
H3, we use a dummy variable
lated missionary marking
activities, required religious groups th
whose largest and to second-largest
get official approval for religious activities, ethnic
or sin-
49% and 7% of the gled population,
out for or permitted harassment respecti
of a particular
The mechanism that gives
religious group. For policies onrise
language, weto coded by nat
tention in modernist arguments
year whether a language spoken by atis state
least 5% of the o
crimination along the
populationlines of recognition
received no official cultural at any level dif
is thought to create the grievances tha
of government.12
bellions. Grievances One are
might argue,difficult
contrary to H, above, that to we me
dently of our knowledge of between
should expect an association the actions
measures of eth-
to explain (rebellions,
nic diversitycivil war),
and the occurence but
of ethnic strife, but not me
age levels of discrimination
between ethnic diversity andare feasible
civil war more broadly.
being equal, political
In the limitdemocracy
this argument borders on tautology, should
since
with less discrimination and
it would not be surprising repression
if ethnic wars are rare in
countriesdemocracy
or other lines, since whose citizens think of themselvesendows
as ethni-
political power (thecally
vote)
homogeneous.they
However, one do mightnot hav
ask if "ethnic
ships. Even more directly,
wars" become more likelymeasures of st
as ethnic diversity increases
of civil rights suchamongas freedom
those countries of
that have at least one assoc
nontrivial
sion, and due process should be associa
ethnic minority.
repression and thus lower grievances. Stat
H7: Amongof
discriminate in favor countries
awith an ethnic minority com-
particular gr
prising at least 5%
or religion should be associated with of the population, greater eth- gr
grievances. Finally, nic diversity
it should associate with
is often a higher risk of tha
argued
nomic inequality ethnic civil war.
creates broad grievanc
civil conflict (e.g., Muller 1985). Thus, we
lowing hypotheses. The measurement problem here is to say what an
"ethnic" civil war is, and we suspect that under any
H4: Measures of political plausible statement of the concept there will be numer- an
democracy
erties should be associated with lower risks of civil ous cases that are mixed or ambiguous. We coded as
war onset. "ethnic" conflicts in which the fighters were mobilized
primarily along ethnic lines, marking off 58 (51%) as
ethnic, 20 (18%) as mixed or ambiguous, and 36 (32%)
H5: Policies that discriminate in favor of a partic-
ular language or religion should raise the risk as of
not ethnic (excluding the anticolonial conflicts).
civil war onset in states with religious or linguistic
minorities.
INSURGENCY
HS6: Greater income inequality should be associ-
If many post-1945 civil wars have been "ethnic" or
ated with higher risks of civil war onset.
"nationalist" as these terms are usually understood
We consider both the Polity IV and the Przeworski then even more have been fought as insurgencies. I
et al. 2000 democracy measures, along with the Free- surgency is a technology of military conflict character
dom House indicator of the observance of civil ized by small, lightly armed bands practicing guerri
liberties.11 For income inequality we interpolated warfare
andfrom rural base areas. To explain why som
countries have experienced civil wars in this period one
extended as necessary the Gini coefficients assembled
needs to understand the conditions that favor insur-
gency, which are largely independent of cultural dif-
ferences between groups and even group grievances.
10 Alternative thresholds for the second-largest group, such as 10%,
make no difference in the results. These conditions are best summarized by way of a brief
11 Freedom House codes countries annually from 1972 on a seven-statement of the logic of insurgency.13
point scale based on a "checklist" that awards points for government The fundamental fact about insurgency is that insur-
observance of a long list of civil liberties and rule of law, including gents are weak relative to the governments they are
freedom of the press, religion, and association, independent judiciary,
equal treatment under the law, civilian control of police, "protec-
tion from political terror," secure property rights, and equality of 12 "Official recognition" implies use by some public, officially sanc-
opportunity. "Freedom from war and insurgency" is one element of tioned entity. For language we relied mainly on Asher 1994 and
their checklist, so we are careful to lag this independent variable.Gunnemark 1992; and for religion, the U.S. State Department's 1999
For Polity IV, we use the difference between the 11-point democracyAnnual Report on Religious Freedom.
and autocracy scales. Following the Polity coders' suggestions, we13 Though our formulations differ, we have been influenced here
interpolate values for "transition period" years (-88), treat foreignby Stathis Kalyvas's work on the Greek civil war. The literature on
occupation years (-66) as missing, and treat "interruptions" (-77) guerrilla warfare is extensive; see, for examples, Desai and Eckstein
as zeros. See http://www.bsos.umd.edu/cidcm/polity/. 1990, Griffith 1961, and Laqueur 1976.
79
Ethnicity,
Ethnicity,
February
2003
2003
fighting, at least
life, and they at
may also need information the
and instruction
ment forces knew who the rebels were and how to find in the practical details of running an insurgency.15
them, they would be fairly easily destroyed or captured. Most important for the prospects of a nascent insur-
This is true even in states whose military and policegency, however, are the government's police and military
capacities are low. The total number of active rebels in capabilities and the reach of government institutions into
many wars in which thousands of civilians have been rural areas. Insurgents are better able to survive and
killed (through the actions of both governments and prosper if the government and military they oppose
rebels) is often in the hundreds or low thousands. are relatively weak-badly financed, organizationally
The numerical weakness of the insurgents implies inept, corrupt, politically divided, and poorly informed
that, to survive, the rebels must be able to hide from
about goings-on at the local level.
government forces. Several hypotheses follow. Effective counterinsurgency requires government
forces to distinguish active rebels from noncombatants
H8: The presence of (a) rough terrain, poorly without destroying the lives and living conditions of the
served by roads, at a distance from the centers latter. This is an extremely difficult political, military,
of state power, should favor insurgency and civil and organizational problem even for well-equipped
war. So should the availability of (b) foreign, cross- and well-paid modern militaries; witness the U.S. mil-
border sanctuaries and (c) a local population that itary's failures in Vietnam (Avant 1994; Krepinevich
can be induced not to denounce the insurgents to 1986), early British efforts in Northern Ireland
government agents. (Kennedy-Pipe 1997), or Soviet efforts in Afghanistan.
For less well-financed and bureaucratically competent
Much scholarly writing holds that ethnic or class sol-states, the problem appears to be nearly insoluble. Such
idarity and grievances are necessary for Hc, the lo-states either cannot prevent the abuse of local powers
cal population's support of active rebels. In line withby field commanders or may even permit these abuses
Kriger (1992) and some analysts of communist insur- as a sort of tax farming to the military. That is, they
gencies (e.g., Clutterbuck 1967, Leites and Wolf 1970, "pay" the soldiers with the opportunity to loot and pil-
and Thompson 1966), we argue that while grievanceslage, a practice that tends to sustain rather than end
and ethnic solidarity can be helpful in this regard, theyinsurgencies (see Keen 1998 for examples). Thus, we
are not necessary. Instead, the key to inducing the lo- have the following hypothesis.
cal population not to denounce the active rebels is
local knowledge, or information about who is doing Hg: Proxies for the relative weakness or strength of
what at the village level. Local knowledge allows the the insurgents-their odds of being killed or cap-
active rebels to threaten retribution for denunciation tured for a given level of counterinsurgent effort
credibly.14 Ethnic insurgents use this informational ad- by the government-should be associated with the
vantage to great effect, often threatening and inflicting likelihood that a country develops a civil war. In
unimaginably harsh sanctions on "their own" people particular, a higherper capita income should be as-
(Kalyvas 1999; Kriger 1992). The presence of an eth- sociated with a lower risk of civil war onset because
nic insurgency does not imply that the members of the (a) it is a proxy for a state's overall financial, ad-
ethnic group are of one mind in their determination to
ministrative, police, and military capabilities, and
fight the state till they realize a nationalist dream. The
(b) it will mark more developed countries with ter-
immediate concern is how to survive in between gov-
rain more "disciplined" by roads and rural society
ernment forces using violence to gain information or
punish alleged rebel supporters and rebel forces usingmore penetrated by central administration.
violence to punish alleged informants, "moderates," orThere is an additional reason why a lower per capita
government sympathizers. income should favor the technology of insurgency:
An empirical implication of the importance of lo-(c) Recruiting young men to the life of a guerrilla is eas-
cal knowledge is hypothesis Hsd: Having a rural base
ier when the economic alternatives are worse. Though
should greatly favor insurgency. In the city, anony-
we try below, it is difficult to find measures to distinguish
mous denunciation is easier to get away with, giving
among these three mechanisms associating a low per
the government an advantage in its counterinsurgentcapita income with civil war onset. We believe that the
efforts.
strong results for per capita income reported below are
Given the basic constraints posed by numerical
due largely to its acting as a proxy for state military and
weakness-the need to hide and not be denounced-
police strength relative to potential insurgents (a and b
various factors determine insurgents' ability to wage
in H9). The fact that measures such as the percentage of
war. To survive, rebels need arms and materiel, money
young males and male secondary schooling rates pre-
to buy them, or smugglable goods to trade for them.
dict less well than per capita income is consistent with
They need a supply of recruits to the insurgent way of conjecture, though not definitive.
this
14 A "second-order" mechanism by which ethnicity may favor 15 In the case literature one frequently finds either that rebels leaders
insur-
gency is that ethnic minorities are sometimes marked by densehave spent time at guerrilla training camps in, for example, Libya,
social
Afghanistan, Lebanon, or Mozambique (in the 1970s) or that they
networks that are isolated from dominant group networks, thus giving
an informational advantage to local rebels (Fearon and Laitin gained
1996). guerrilla experience in one insurgency that they apply in pur-
But such an advantage does not require ethnic distinctiveness. suing another.
80
American
No. 97, No.11
Additional popular grievances

factors that or ethnic solidarities to operate
would be ef- e
(or proxy) the fectively. The latter point
strength of suggests
an a contrast to H4-Hk.
insurg
a state follow.
H1l: After controlling for per capita income (or
H1o: The political and military technology of in- other measures of state strength), neither political
surgency will be favored, and thus civil war made democracy, the presence of civil liberties, higher in-
more likely, when potential rebels face or have come inequality, nor nondiscriminatory linguistic
available the following. or religious policies should associate strongly with
lower odds of civil war. Given the right environ-
(a) A newly independent state, which suddenly
mental conditions, insurgencies can thrive on the
loses the coercive backing of the former im-
basis of small numbers of rebels without strong,
perial power and whose military capabilities
widespread, popular support rooted in grievances
are new and untested (Fearon 1998).
and, hence, even in democracies.
(b) Political instability at the center, which may
indicate disorganization and weakness and As for measures, for "rough terrain" we use the pro-
thus an opportunity for a separatist or center- portion of the country that is "mountainous" according
seeking rebellion. to the codings of geographer A. J. Gerard.16 This does
(c) A regime that mixes democratic with auto- not pick up other sorts of rough terrain that can be
favorable to guerrillas such as swamps and jungle, and
cratic features, as this is likely to indicate polit-
it takes no account of population distributions or food
ical contestation among competing forces and,
availability in relation to mountains; but it is the best we
in consequence, state incapacity. (In contrast, have been able to do for H8a. For H9 we use Penn World
pure autocracy tends to reflect the successful Tables and World Bank data on per capita income, es-
monopolization of state coercive and adminis- timating missing values using data on per capita en-
trative power by an individual or group.) ergy consumption.17 For H1oa (new states) we mark
(d) A large country population, which makes it countries in their first and second years of indepen-
necessary for the center to multiply layers of dence; for H1ob (political instability) we use a dummy
agents to keep tabs on who is doing what at the variable indicating whether the country had a three-or-
local level and, also, increases the number of greater change on the Polity IV regime index in any of
potential recruits to an insurgency for a given the three years prior to the country-year in question.18
For countries that mix democratic and autocratic fea-
level of income.
tures (called "anocracies" or "semidemocracies" in
(e) A territorial base separated from the state's
the international relations literature and "praetorian
center by water or distance-for example, regimes" by Huntington 1968) we mark regimes that
East Pakistan (now Bangladesh) from West score between -5 and 5 on the difference between
Pakistan or Angola from Portugal. Polity IV's democracy and autocracy measures (the
(f) Foreign governments or diasporas willing to difference ranges from -10 to 10). Country popula-
supply weapons, money, or training. tion (Hiod) is based largely on World Bank figures. For
(g) Land that supports the production of high- oil exporters we marked country-years in which fuel
value, low-weight goods such as coca, opium, exports exceeded one-third of export revenues, using
diamonds, and other contraband, which can be World Bank data.19 We coded a dummy variable for
used to finance an insurgency. states with noncontiguous territory ourselves (H1ie).20
(h) A state whose revenues derive primarily from
oil exports. Oil producers tend to have weaker 16 Gerard produced this measure for the DECRG project on civil
state apparatuses than one would expect given wars at the World Bank. Our sample of countries differs slightly,
so we estimated values for 21 missing countries using the difference
their level of income because the rulers have
between the highest and the lowest point of elevation in each country,
less need for a socially intrusive and elaboratewhich is well correlated with the mountains measure (0.78 in logs).
17 We used income growth rates from the World Development Indi-
bureaucratic system to raise revenues-a po- cators 2001 to extend the estimates in the Penn World Tables 5.6 and
litical "Dutch disease" (Chaudhry 1989; Karl then used the per capita energy consumption estimates provided by
1997; Wantchekon 2000). At the same time, the COW project to estimate additional missing values. For details
see Fearon and Laitin 2003.
oil revenues raise the value of the "prize" of
18 For this variable, "transition periods" and "interruptions" (which
controlling state power. indicate a "complete collapse of central authority") are coded as
instability; foreign occupations are treated as missing.
Partially excepting f, none of these conditions cru-
19 The data are for five-year intervals beginning in 1960; we inter-
cially involves cultural differences, ethnic minority
polated for years after 1960, set the value to that in 1960 for years
status, or group grievances. We do not claim that these prior to 1960, and used country-specific sources for a few countries
factors provide no help to would-be insurgents in spe- without World Bank coverage.
cific cases. But, to reiterate, grievances and ethnic 20 Countries with territory holding at least 10,000 people and sep-
arated from the land area containing the capital city either by land
differences are too common to help distinguish the or by 100 km of water were coded as "noncontiguous." Ignoring the
countries andy tat ee ii a nd years that see civil wars, and in any event
colonial empires, 25 of our 161 countries meet this criterion at some
the technology of insurgency does not require strong
time since 1945.
81
Ethnicity, Insurgency, and Civil War February
February
2003
2003
The remaining hypotheses

chance of (H
civil war outbreak in the next five years. In
contrast, countries
Hiog) present more at the eightieth percentile on mea
difficult ethnic
Whether availability
homogeneity and at theof twentietha rural
percentile on income
had a 15% chance of
(HId) is better tested inwar inathe next five years.24
research
groups are the unitNote that for
of any level of ethnic diversity, as one
analysis, so
ferent moves up the
geographic income scale (to the right in Figure 2),
concentrations
the odds of civil war decrease,
Although it is possible to by substantial factors
code reb
for whether the in all cases and dramatically
rebels among the most homoge-
receive she
neous countries.
foreign countries (H3b, The richest fifth is practicallythe
Hlof), immune
of these aids to rebel
regardless strength
of ethnic composition. In contrast, for given is
levels of country
prior to the onset of income, no consistent effect is associ-
fighting. In t
ated with variation in
potential availability ofethnic homogeneity
support (i.e., moving f
to governments is
up or down observable-i
the figure). Among the poorest countries
"Brezhnev doctrine") in
where we observe the highest ratesEastern
of civil war, the data
indicate a tendency
policy with regard to for itsmore homogeneous
former countries to c
be more civil expect
Africa.22 We would war-prone. Among the richest countries
such su
there may be of
relative advantage a weak tendency for the most homoge-
governmen
neous countries
tential insurgents and to have fewer
thuscivil wars, butassoc
the size of
of civil war onset. We also consider a more tenuous the effect, if any, is small.
measure of potential support to rebels-the number The empirical pattern is thus inconsistent with H1,
of civil wars ongoing in neighboring countries-which the common expectation that ethnic diversity is a major
might yield more easily available weapons, training, and
or direct cause of civil violence. Nor is there strong
the presence of experienced guerrillas.23 evidence in favor of H2, which expects ethnic strife to
be activated as modernization advances. Ethnic diver-
sity could still cause civil war indirectly, if it causes a
EMPIRICAL ANALYSIS
low per capita income (Easterly and Levine 1997) or
Our central hypotheses concern the relationshipabe- weak state. But then the mechanisms that actually
tween ethnic and religious diversity or structure,produce
on the violence would more likely be those of the
insurgency perspective than the culturalist arguments
the one hand, and the susceptibility of a country to civil
in either perennialist or modernist forms.
war, on the other. Several multivariate analyses of the
country-year data are presented below, but the main The data for Figure 2 omit the anticolonial wars, and
the analysis does not control for other possible influ-
story emerging from them is made clear by the contour
ences on civil strife enumerated earlier. We consider
plot in Figure 2.
these issues next.
Are More Diverse Countries Prone

to Civil War? Multivariate Results
Figure 2 shows how probabilities of civil war onset We coded a variable onset as "1" for all country-years
in which a civil war started and "0" for all others.25
vary at different percentiles for country income (on the
x axis, measured in lagged 1985 dollars) and ethnic Model 1 in Table 1 shows the results of a logit analysis
homogeneity (on the y axis, measured by the popu-using onset as the dependent variable and a fairly full
lation share of the largest ethnic group). The lines inspecification of independent variables discussed above.
the plot show the probability of war onset in the next Prior war is a control variable indexing whether the
five years for a country at the given level of income
and ethnic homogeneity. For example, countries at the 24 The figure was produced using R's locfit package, with a smoothing
twentieth percentile in terms of the size of their largestparameter of 0.9, and transforming annual probabilities of outbreak
ethnic group-thus quite ethnically diverse-but at the
to five-year equivalents. The figure looks highly similar if we use other
eightieth percentile on income have had about a 5%measures of ethnic diversity, such as fractionalization.
25 We do not code as ones years in which a civil war continues, which
would be relevant if our focus were causes of war duration rather
21 Using the Phase III Minorities at Risk (MAR) data, Fearon andthan onset. Nor do we drop country-years that have an ongoing
Laitin (1999) found that groups without a rural base area were far less war (as in Collier and Hoeffler 2001), since this would omit the
likely to be engaged in violent conflict with the state, even after con- 14 wars in our data that start while another war is in progress. Other
trolling for various country- and group-specific factors. Toft (1996) approaches to the dependent variable and estimation-such as mak-
was the first to note and examine the strong bivariate relationship in ing the dependent variable "1" for all war years and using dynamic
the MAR data. probit or Poisson models that interact independent variables with a
22 U.S. support to rightist regimes in Latin America during thelagged Cold dependent variable (Jackman 2001; Przeworski et al. 2001)-
War might also qualify, although this was perhaps more offset by virtually identical results. Only one country-year in the data
produce
support for armed insurgency in this area from the Soviet Union hasand
more than one onset (Soviet Union 1946, with one Ukrainian
Cuba. and three Baltic rebellions). We treat this as a one in the analyses
reported
23 The presence of valuable minerals or the suitability of land for here, though there are no significant changes if we use
Poisson regression instead (treating the dependent variable as a
the cultivation of narcotics is also codable in principle, but at present
we lack such measures (H10e). Nor do we have measures for count)
the or weight this observation four times in the logit likelihood.
Using "rare events logit" (King and Zeng 2001) had no appreciable
comparative disadvantage of governments in access to village-level
information (Hs). impact on any estimates reported.
82
American Political Science Review Vol. 97, No. 1
FIGURE 2. Probability of Civil War Onset per Five-Year Period
GDP/capita in lagged 1985 U.S. dollars

48 808 1433 2724 5689 17280
03 q00
i i i i i
d ao
e0 /
O O. O
0)
% . 5- 0)
Q /
U) C
o
. -/o - \
eCL
i U/
0
0o_~~
o \ \ \
0 20 40 60 80 99
Percentile
country had a distinct civil war ongoing in the previousmight have low rates of civil war for reasons of culture
year.26 or history that have little to do with income. The esti-
mated coefficient drops only to -0.28 when a dummy
Per Capita Income. Per capita income (measured as for the West is included and remains strongly significant
thousands of 1985 U.S. dollars and lagged one year) is
despite the high correlation between the two variables.
strongly significant in both a statistical and a substantive
The relationship holds even within the poorest regions.
sense: $1,000 less in per capita income is associated with
Among the (mainly) former colonies of Africa, the
41% greater annual odds of civil war onset, on average. Middle East, and Asia, we estimate that $1,000 less
Holding other variables at their median values, a coun-
in income corresponds to 34% greater annual odds of
try in the tenth percentile on income has an 18% chance outbreak.27
of a civil war outbreak over a decade, compared to an
11% chance for a country at the median income and Ethnic and Religious Composition. The estimates
a 1% chance for a country at the ninetieth percentile for the effect of ethnic and religiousfractionalization are
($573, $1,995, and $9,505, respectively). The income substantively and statistically insignificant. Alternative
variable is not just a proxy for "the West," whose states measures of ethnic and religious diversity-such as the
26 With onset as the dependent variable, the data are grouped dura-
tion data and we need to consider the possibility of temporal depen- 27 Contrary to Hibbs 1973 and Hegre et al. 2001, adding the square of
dence between observations. One approach is dynamic probit; see per capita income does not significantly improve the fit of the model.
footnote 25. We also tried Beck et al.'s (1998) method of including Regarding H2, the interaction of income and ELF has the expected
dummies for each successive "peace year" prior to an onset, or fitting sign but is not significant. For lack of space, we report the details of
natural cubic splines. This had no appreciable effect on the estimates, these and other "nonresults" mentioned below in the tables included
and the splines were jointly insignificant. in Fearon and Laitin 2003.
83
Ethnicity,
Ethnicity,
February
2003
2003
TABLE 1. Logit Analy

Model
(4) (5)
(1) (2) (3) Civil War Civil War
Civil War "Ethnic" War Civil War (Plus Empires) (COW)
Prior war -0.954** -0.849* -0.916** -0.688** -0.551
(0.314) (0.388) (0.312) (0.264) (0.374)
Per capita incomeab -0.344*** -0.379*** -0.318*** -0.305*** -0.309**
(0.072) (0.100) (0.071) (0.063) (0.079)
log(population)ab 0.263*** 0.389*** 0.272*** 0.267*** 0.223**
(0.073) (0.110) (0.074) (0.069) (0.079)
log(% mountainous) 0.219** 0.120 0.199* 0.192* 0.418***
(0.085) (0.106) (0.085) (0.082) (0.103)
Noncontiguous state 0.443 0.481 0.426 0.798** -0.171
(0.274) (0.398) (0.272) (0.241) (0.328)
Oil exporter 0.858** 0.809* 0.751** 0.548* 1.269***
(0.279) (0.352) (0.278) (0.262) (0.297)
New state 1.709*** 1.777*** 1.658*** 1.523*** 1.147**
(0.339) (0.415) (0.342) (0.332) (0.413)
Instabilitya 0.618** 0.385 0.513* 0.548* 0.584*
(0.235) (0.316) (0.242) (0.225) (0.268)
Democracya,c 0.021 0.013
(0.017) (0.022)
Ethnic fractionalization 0.166 0.146 0.164 0.490 -0.119
(0.373) (0.584) (0.368) (0.345) (0.396)
Religious fractionalization 0.285 1.533* 0.326 1.176*
(0.509) (0.724) (0.506) (0.563)
Anocracya 0.521 * 0.597*
(0.237) (0.261)
Democracya.d 0.127 0.219
(0.304) (0.354)
Constant -6.731*** -8.450*** -7.019*** -6.801*** -7.503***
(0.736) (1.092) (0.751) (0.681) (0.854)
N 6327 5186 6327 6360 5378
Note: The dependent variable is
parentheses. Estimations perform
a Lagged one year.
bIn 1000's.
c Polity IV; varies from -10 to 10.
d Dichotomous.
proportion of the largest group and the log of the num- while evidence of an independent effect of cultural di-
ber of languages spoken by at least 1%-prove to be versity is surprisingly weak. In Model 2 (Table 1), the
just as unrelated. The ethnic diversity measures show dependent variable marks the onset of wars that we
a strong bivariate relationship with civil war onset (not coded as "ethnic" or partially "ethnic," and we consider
so for the religion measures), but this evaporates when only countries with at least a 5% ethnic minority. The
we control for income.28 coefficient for ethnic fractionalization barely changes.
Nor are countries that are ethnically or religiously The same is true for our other measures of ethnic diver-
polarized in the sense of H3 more likely to experience sity. Our two measures for religious diversity give weak
major civil violence. When we add dummy variables for and inconsistent results, and support for H7 diminishes
countries that have an ethnic or religious majority and further if we code the "partially" or "ambiguously" eth-
a minority of at least 8% of the country's population, nic wars as nonethnic.29
both are incorrectly signed and neither comes close to
statistical significance. This finding does not depend on
Democracy and Civil Liberties. Broad social and po-
which other variables are included in the model. litical grievances should be lower, on average, in politi-
cal democracies. But contrary to H4 and consistent with
Ethnic War. The strong effect of per capita income H11,
re- civil war onsets are no less frequent in democracies
mains even when we restrict attention to "ethnic wars,"
after controlling for income, as shown by the positive
and statistically insignificant coefficient for democracy,
28 Collier and Hoeffler (2001) find the interaction of ethnic and reli-
gious fractionalization to be negatively related to onset. Others 29find
Coefficients for political instability and mountains diminish in
that fractionalization has a nonmonotonic relation to conflict. Nei- this subsample, partly because the associations appear to have been
ther relationship appears in these data; we suspect that one reason weaker for ethnic wars and partly due to the omission of highly ho-
may be listwise deletion on missing economic data in other datamogeneous
sets. countries.
84
American
No. 97, No.11
the (lagged) PolityNew IV States and Political Instability.

measure. The Consistent with
resu
if we use Przeworski et al.'s dichotomous measure in- H1Ja, the odds of civil war onset are estimated as
stead (available for 1950-91) or the lag of Freedom
5.25 times greater in the first two years of a state's in-
House's measure of observance of civil liberties (avail-
dependent existence than in other years, a huge effect
able for 1972-99). In the former case the sign is (Model 3). For the "median country," this translates to
"wrong," while in the latter the sign is consistent withan 11% chance in its first two years. Political instability
H4 but the coefficient on civil liberties is thoroughly
at the center matters in other years as well. The odds
insignificant. of onset in a given year are estimated to increase by
Some past studies of civil strife found an "inverted- 67% if there was instability in governing arrangements
U" relationship with democracy and sought to explain in any of the previous three years.32
this by the observation that the most autocratic regimes
can repress dissent and thus avoid civil violence despite Mountains and Noncontiguous Territory. Mountain-
facing the highest levels of popular grievance (Hegreous terrain is significantly related to higher rates of
et al. 2001; Muller and Weede 1990). This observation civil war. A country that is about half "mountainous"
does not explain why the leaders of a partially demo- (ninetieth percentile) and otherwise at the median has
an estimated 13.2% chance of civil war over the course
cratic regime are not able to implement full autocracy
or democracy to avoid conflict and opposition. As sug- of a decade. A similar country that is not mountainous
gested in H1oc, we suspect that the answer is often thatat all (tenth percentile) has a 6.5% risk. The estimated
"anocracies" are weak regimes, lacking the resourcessign for states with noncontiguous territory is consistent
to be successful autocrats or containing an unstablewith H10e, but its 95% confidence interval includes the
mix of political forces that makes them unable to movepossibility of no effect.33
to crush nascent rebel groups. Consistent with this hy- Population. Holding other variables at medians, the
pothesis and with the prior studies, Model 3 (Table 1) estimated risk of civil war over the course of a decade
shows that a dummy variable marking anocracies takes for a country at the tenth percentile in population is
a positive coefficient. In substantive terms the estimate
6.4%, versus 16.4% for a country at the ninetieth per-
suggests that these regimes have about 68% greatercentile (1.4 versus 56.3 million, e.g., Botswana versus
odds of civil war outbreak in any given year than would
Iran). This effect is not due to large states being more
a full autocracy. This is so despite the fact that we are ethnically diverse. We are controlling for ethnic diver-
controlling for recent political instability, which is muchsity, and in any event it turns out that there is essen-
more common in anocracies.30
tially no correlation between diversity measures and
log of size, even if we omit the relatively homogeneous
Linguistic and Religious Discrimination. Added to
China.34
Model 3, our measures of state discrimination against
regional languages or against minority religions are not Oil, Islam, Young Males. Consistent with H1oh, de-
associated with systematically higher risks of civil war riving at least one-third of export revenues from fossil
onset.31 This nonresult persists when we restrict thefuels is estimated to more than double a country's odds.
sample to those countries with at least a 5% religious
Based on Model 3, the "median country" had a 10%
or ethnic minority. chance of civil war over a decade, whereas the same
country as an oil exporter would have an estimated
Inequality. Whether in a bivariate model or added
21% chance.
to Model 3, the Gini coefficient estimates of income
inequality do not come close to either statistical or sub-
Oil producers are disproportionately Middle
Eastern, which raises the question of whether this
stantive significance. The poor quality of the inequality
variable might proxy some other relevant attribute of
data, available for only 108 countries, does not allow
these countries. But a dummy for the North Africa/
us to go beyond the claim that there appears to be no
Middle East region is insignificant when added to
powerful cross-national relationship between inequal-
Model 3, while the estimate for oil is barely affected.
ity and onset, consistent with H1.
Huntington (1996) argues that "Islam has bloody
borders" and "bloody innards" because Islamic
societies have cultural and demographic features that
30 In Model 3 Democracy is a dichotomous variable marking regimes
make them violence-prone. When we add a variable
that scored higher than 5 on the Polity scale; the excluded category is
autocracies. For Model 3 we followed Polity IV's suggestion in coding
regime "interruptions" as anocracies; if these values are instead in-
terpolated, the resulting anocracy indicator gets a smaller estimated 32 Snyder (2000) argues that democratization puts states at greater
coefficient that is not quite significant at the 5% level, while that for
risk for civil war, ethnic in particular. Coding variables for three-or-
political instability increases. Another issue, raised by Hegre et al. greater moves toward and away from democracy on the Polity index,
(2001), is that the observation of some violence may lead Polity IV we find in these data that moves away from democracy are much
to code democracies as less "democratic" and autocracies as less more strongly associated with civil war onset in the next year than
autocratic (since there is mobilized opposition). So if theremoves is low-
toward it, though both forms of instability elevate the risk.
level violence preceding the reaching of our thresholds, some of the
Hegre et al. (2001) found the same thing using COW civil war data.
effect of "anocracy" may be that it is picking up nascent civil 33war.
Dropping this or other "insignificant" variables from Model 3
31 We tried the several religion indicators both separatelybarely and asaffects
a the remaining estimated coefficients.
single measure based on a factor analysis of the components. 34 Discrim-
Huntington (1996) argues that rapid population growth puts states
ination against regional languages is in fact associated with slightly at greater risk for civil war. Using the average growth rate for the
lower odds of civil war onset, consistent with Laitin 2000. three prior years, we find no significant impact.
85
Ethnicity,
Ethnicity,Insurgency, and Civil War
Insurgency, and Civil War February February
2003 2003
measuring the the other estimates remain stable). A dummy

percentage of variable
Mus
to our main specification,35 for all former French colonies is unrelated to civil
it war take
is not statistically onset, which suggests that the observed effect in sub-The
significant.
strong. Saharan Africa is due to the mechanism behind Hlof
Huntington also argues that societies with a surfeit of rather than the effects of French colonial administra-
young males will be prone to civil violence, suggesting tion or law.
this as a part of the explanation for Islam's "bloody
innards." Given that young males have physical and Other Variables and Robustness Checks
perhaps psychological characteristics that make them
apt guerrillas, our arguments on insurgency point in the Anticolonial Wars. Figure 2 and the multivariate
same general direction. If we include a lagged measure analyses above omitted the 13 anticolonial wars in five
of the proportion of males aged 15 to 24 in the popula- colonial empires (Britain, France, Belgium, Portugal,
tion in Model 3, it has the expected positive sign,36 but and The Netherlands). Including them is problematic
the estimate is highly uncertain (p = .21); one problem for both theoretical and empirical reasons, but so is
is that percentage of young males has a strong negative ignoring them. The main practical obstacle is the lack
correlation with income, and income gets the better of of comparable data. However, by using the ethnic
it when both are included. The same is true for male sec- fractionalization, population, mountainousness, and in-
ondary schooling rates (cf. Collier and Hoeffler 2001). come data for each colony in its first year of inde-
pendence, we were able to produce annual estimates
Foreign Support. One determinant of the prospects
of these variables for each empire. The income esti-
for insurgency is the availability of third-party support
mates are systematically too high-we use, for example,
to either the rebels or the government of the state in
Nigerian GDP per capita in 1962 as an estimate of its
question. "Availability" is difficult to observe ex ante,
income as a British colony in 1945. This creates some
however. A possible proxy is the number of civil wars
bias against the income variable and in favor of the ELF,
ongoing in neighboring countries in the previous year,
which has a median value of 0.83 for the 114 "empire
which might increase the availability of arms, support,
years," compared to 0.34 for nonempire years.
and seasoned guerrillas. Adding the lagged version of
Model 4 reports a logit analysis parallel to our speci-
this variable to Model 3, its sign is as predicted, but it
fication in Model 1 (Table 1), omitting only the democ-
is not significant (p = .33). Civil war next door is too
racy variable and religious fractionalization (which is
common, occuring in 44% of the country-years in the
hard to estimate for the empires due to shared religions
sample.
between colonies and metropoles). Given that the em-
Another possible proxy is foreign support to govern-
pires were quite civil war-prone and highly diverse, it
ments against domestic challengers. A dummy variable
is not surprising that the estimated coefficient for ELF
for "Soviet bloc" prior to 1991 cannot be included in the
logit analysis because it perfectly predicts peace. But by increases. Surprisingly, though, it remains statistically
indistinguishable from zero, while, excepting nonconti-
itself this observation is consistent with our argument.3
guity, the others change little. Noncontiguous states are
In the late 1980s, Mitterand's government departed now estimated to have had 2.2 times the odds of civil
from long-standing French foreign policy by support-
war in this period, other things being equal. Because the
ing, to a limited extent, democratization in some of its
empires are all noncontiguous and because noncontigu-
former sub-Saharan colonies. This involved encourag-
ity otherwise does a better job than the ELF at picking
ing "national conferences," elections, and some politi-
out civil wars onsets, it is given greater substantive and
cal opposition, which all suggested that the prior policy
statistical significance by the logit. If we drop noncon-
of immediate military support for French-speaking dic-
tiguity, the estimate for ELF increases to 0.62-which
tators might have changed. Civil war rates for for-
mer French African colonies had been much lower corresponds to a 59% increase in risk moving from the
than those for other sub-Saharan states. The 25 others tenth to the ninetieth percentile-but still fails statis-
tical significance at the 5% level (p = .07). Thus even
saw 17 civil war onsets before 1990, a rate of 2.4 per
if we include the anticolonial wars, the evidence that
100 country-years. The 16 former French colonies had
ethnic diversity directly causes states to be more civil
only three onsets in the same period, for a rate of
war prone remains slim at best.38
0.65 per 100 country-years (Chad in 1965, Senegal
Stepping back from the statistical analysis, we ob-
[Casamance] in 1989, and Mali in 1989). If a variable serve that these five imperial states managed their
marking former French colonies in sub-Saharan Africa
empires for many years before World War II with
prior to 1990 is included in Model 3, its estimated co-
fairly little active rebellion despite the same levels of
efficient is -1.0, which implies an almost-threefold re-
ethnic fractionalization, noncontiguity, and mountain-
duction in the annual odds of civil war onset (p = .11;
ousness. Nor did discrimination and oppression sud-
denly increase in the empires after the war, causing
35 Coded from the CIA Factbook and a variety of country-specific
sources.
36 Data from the World Bank, linearly interpolated 38 Wefor

created a democracy
missing years estimate for the empires by taking a
within countries. weighted average of the metropole's lagged Polity democracy score
37 Foreign support may be at least as important as communist party and -10, weighting by the metropole's proportion of total imperial
structure in explaining Huntington's (1968) observation that the com-population. Adding this or dummies for "anocracy" and democracy
munist countries rarely faced insurrections. to Model 5 has no substantive impact.
86
American
No. 97, No.11
anticolonial movements as
efficient that log a response.
of income takes in the standard logit Rat
greatly weakened theon the full
mainsample (-0.78), significant at p =.02.39state
imperial In
addition, the effect
and the new international legalestimates for the other variables
order (the U
gave support to anticolonial movements.
are, with one exception, virtually identical to the es-
to the relative power
timates inof the full,the metropoles
"pooled" model, the significance
"civil war onset" in these
remaining high for allcases (Fearon
but anocracy and oil (which has 20
little temporal variation). The exception is population,
Regional Effects. which
Different regions
varies vastly more across countries than within of
share a variety of them
historical, cultural,
over time. While bigger countries are more civil an
traits. It is reasonable tothere
war-prone, wonder
is no evidence that if any
population growth of t
considered in the multivariate
within a country raises the risk. analysis ju
such factors. Further, if regional dummy
Primary Commodity
not add to the model's Exports. Using an "expanded"
explanatory powe
version of the COW
have managed to account for data, Collier
the and Hoeffler (2001)
influenc
find that the risk of civil war onset is maximized when
characteristics with our more general indep
primary
ables. We find that if we commodity
add exports comprise
all about 32% one)
(but of the r
country's GDP, with risk
mies to Model 3, the coefficients and signif declining on either side of this
for our other variables are little affected. A likelihood-figure. They propose that, up to a point, higher levels
ratio test fails to reject the null hypothesis that "re- of primary commodity exports provide more motiva-
gions do not matter" beyond the included variablestion and opportunity for rebels to support themselves
(p = .22). Including region dummies individually re-through "looting." Beyond this point, the revenue gains
available to the state are said to be large enough to
veals that none has a rate of civil war onset significantly
different from what one would expect on the basismake civil war less likely.
of the country characteristics already included in the We find little evidence of such a relationship in our
model. data. Neither the share of primary commodity exports
in GDP nor its square is remotely significant when
Period and Fixed Effects. If added to Model 3, added to Model 3.
dummy variables marking each decade (but one) are We agree with the general argument behind Collier
and Hoeffler's hypothesis. In our terms, insurgency
jointly significant in a likelihood-ratio test (p= .04),
should be more feasible if sources of financing are read-
showing evidence of a general upward trend in civil war
risk after the 1940s. Adding a dummy for the 1940s andily available. But we doubt that primary commodity
a variable marking the year indicates that from 1950 exports are a good measure of financing potential for
rebels.
the odds of civil war outbreak rose about 2% per year, This measure combines oil, agricultural prod-
ucts,
controlling for the other variables. As noted earlier, and metals and minerals, which, except for some
there is no secular trend when we do not control for minerals, are hard to exploit without control of a na-
tional distribution system and ports. Oil exports may be
other factors. Part of the explanation is that the median
of country incomes roughly doubled from 1950 to 1999, relevant not so much because they finance rebel groups
which "should have"-based on Model 3-reduced the but, as we argued, because they mark relative state
onset odds for the "median country" by about weakness35%. at a given level of income.41 A better measure
The fact that the rate of outbreak remains fairly for financing potential would focus on the presence of
con-
minerals or contraband that can reward control of a
stant while incomes generally increased in effect pe-
nalizes income in Model 3. When year is added,small the enclave with huge profits (Leonard and Strauss
2001).
effect estimate for income increases to -0.36 (the other
effect estimates stay steady, except for noncontiguity,
Trade Openness. Using a measure of "state failure"
which becomes statistically significant at p = .035). We
that includes both civil war and "disruptive regime tran-
do not know what accounts for steady rates of outbreak
sitions," Esty et al. (1998) found that trade share of
despite increasing incomes. Increased international in-
GDP strongly predicted civil peace. Using the Penn
equality (so that there is a set of persistently poor coun-
World Tables measure of trade as a share of the GDP,
tries at risk for civil war), greater availability of small
we find no such relationship in our data. Trade is related
arms, and diffusion of insurgency "know-how" come to
to civil peace in a bivariate logit, but this is because
mind, but are just conjectures. smaller countries have more trade and less civil war.
This observation also raises the question of how
much the estimated impact of income is due to cross-
country comparisons and how much to increasing
39 Whether we use income or logged income in the standard logit
income within given countries associating withmakes
civilfor no significant differences.
peace. Applying conditional fixed effects logit
40 Theto
data series on primary commodity exports begins in 1960 and is
Model 3 without the variables that have little or no tem- available only at five-year intervals. We interpolated missing values
poral variation (mountains, noncontiguity, ethnic and and extended the 1995 values through 1999. We also tried Collier
religious fractionalization) and with a control for year, and Hoeffler's procedure of grouping the data by five-year periods
beginning in 1960, again finding no relationship.
we find that the estimated coefficient for income drops
41 Our oil variable is moderately correlated with the primary com-
only to -0.24 but is no longer significant (p=.16). modities measure, at r =.46, but the latter and its square remain
Using log of income, however, returns the same co- insignificant when oil is dropped from the model.
87
Ethnicity,
Ethnicity,
February
2003
2003
Alternative How could democracy and cultural or and

Definitions religious ho- C
Our coding mogeneityadmit
rules fail to associate with civil peace across coun-
civil con
1,000 have been killed.
tries? Viewing "ethnic wars" as a speciesSome
of insurgency
threshold is may help explain
too low this paradoxical
toresult. If, under the
disting
"civil war" right say,
and, environmental conditions,
terrorism. just 500 to 2,000 ac-
conflicts in tive guerrillas
our samplecan make for a long-running,
estima destruc-
than 5,000 as tive internal war, then the
zeros, the average level of grievance
result
identical.42 in a group may not matter that much. What matters is
We also checked to see how our results differed if whether active rebels can hide from government forces
and whether economic opportunities are so poor that
we based the coding of civil war onset on the lists
the life of a rebel is attractive to 500 or 2,000 young
given by Collier and Hoeffler (2001), the COW project,
and Doyle and Sambanis (2000). As shown in Model men.
5 Grievance may favor rebellion by leading nonac-
(Table 1), which bases onset on the COW data (1945- tive rebels to help in hiding the active rebels. But all
the guerrillas really need is superior local knowledge,
92), the coefficients and significance levels for the "in-
surgency" variables are remarkably stable.43 We obtain which enables them to threaten reprisal for denun-
quite similar results for the other two coding schemes. ciation.
Although these four civil war lists differ on many spe-If our analysis is correct, then policy makers should
not assume that civil wars and the "failed states" they
cific start years and, to some extent, on what cases con-
sometimes produce are temporary phenomena of the
stitute "civil wars" at all, there proves to be substantial
shared variation. For example, 53 of the 93 COW civilimmediate post-Cold War world. Nor should policy
wars in the sample start in the same year as an onsetmakers or academics infer that ethnic diversity is the
root cause of civil conflict when they observe insur-
in our data, and 68 start within two years of one of our
gents in a poor country who mobilize fighters along
onsets. At the country level, the bivariate correlation
between our estimate of the total number of warsethnicby lines. Instead, the civil wars of the period have
structural roots, in the combination of a simple, robust
country and the equivalent for the others ranges from
military technology and decolonization, which created
0.71 with Collier and Hoeffler to 0.82 with Doyle and
Sambanis. an international system numerically dominated by frag-
ile states with limited administrative control of their
CONCLUSION
peripheries.
Regarding policy implications, the spread of democ-
racy and tolerance for ethnic and religious minorities
The prevalence of internal war in the 1990s is mainly
should be major foreign policy goals because they are
the result of an accumulation of protracted conflicts
desirable for their own sake, but not with the expec-
since the 1950s rather than a sudden change associated
tation that they are "magic bullets" for the prevention
with a new, post-Cold War international system. Decol-
or resolution of civil war. Sometimes recommended as
onization from the 1940s through the 1970s gave birth
a general
to a large number of financially, bureaucratically, and international policy for resolving ethnic civil
wars
militarily weak states. These states have been at (e.g., Kaufmann 1996), ethnic partitions should be
risk
viewed as having large international implications and
for civil violence for the whole period, almost entirely
high costs. International support for partition would
in the form of insurgency, or rural guerrilla warfare.
Insurgency is a mode of military practice that increase
can be the expected benefits for rebels, who, we have
harnessed to various political agendas, be it commu-argued, may be able to get a nasty civil war going on
nism in Southeast Asia and Latin America, Islamic the basis of small numbers when the conditions for in-
fundamentalism in Afghanistan, Algeria, or Kashmir, surgency are right.
right-wing "reaction" in Nicaragua, or ethnic national- Policies to redress grievances, or, in the limit, parti-
ism in a great many states. The conditions that favor tion, could be important to resolve ongoing conflicts.
insurgency-in particular, state weakness marked by We cannot say on the basis of this research, which fo-
cused on civil war onset rather than termination. We
poverty, a large population, and instability-are better
predictors of which countries are at risk for civil war find little evidence that civil war is predicted by large
than are indicators of ethnic and religious diversity or cultural divisions or broadly held grievances. But it
measures of grievances such as economic inequality, seems quite clear that intense grievances are produced
lack of democracy or civil liberties, or state discrimina- by civil war-indeed, this is often a central objective
tion against minority religions or languages. of rebel strategy. These could well pose obstacles to
settlement.
Regarding prevention, our analysis suggests that
42 The coefficient on income increases 21%, noncontiguity becomes
while economic growth may correlate with fewer civil
significant, and anocracy weakens slightly. We coded total deaths
ourselves, based on Brogan 1998, Institute for International and wars, the causal mechanism is more likely a well-
Strategic Studies 2000, Sivard 1996, Valentino 2002, and country- financed and administratively competent government.
specific sources. In specific terms, international and nongovernmen-
43 The one exception is religious fractionalization, which just man- tal organizations should develop programs that im-
ages significance at the 5% level here. This does not occur in the
other data sets, and even with the COW data the estimate for reli- prove legal accountability within developing world
gious fractionalization is highly sensitive to which other variables are militaries and police, and make aid to governments
included in the model. fighting civil wars conditional on the state observing
88
American
No. 97, No.1
1
Fearon, James D., and David D.practices

counterinsurgency Laitin. 1999. "Weak States, Rough
cruit militias.Terrain, and Large-Scale Ethnic Violence since 1945." Presented
Government
at the Annual Meetings of the American Political Science Associ-
perpetuating ation,
counterinsurge
Atlanta, GA.
rupt as to be hopeless
Fearon, James shou
D., and David D. Laitin. 2003. "Additional Tables for
when there are major
"Ethnicity, Insurgency, implic
and Civil War."" http://www.stanford.edu/
or international
group/ethnic/. terrorism,
Gellner, Ernest. 1983. Nations
"neotrusteeship" under and Nationalism. Ithaca, NY: Cornell
the
University Press.
military and political orga
Gleditsch, Nils, Havard Strand, Mikael Eriksson, Margareta
the European Union.
Sollenberg, The
and Peter Wallensteen. 2002. "Armed Conflict 1946- l
ready see operating, in
2001: A New Dataset." Journal of Peace Research eff
39 (5): 615-37.
East Timor,Griffith,
should be
Samuel B., ed. 1961. Mao Tse-Tung rat
on Guerrilla Warfare.
internal New York: Praeger.
coordination among
Grimes, Joseph E., and Barbara F Grimes. 1996. Ethnologue: Lan-
in such operations.
guages of the World. 13th ed. Dallas, TX: Summer Institute of
Linguistics.
REFERENCES Gunnemark, Erik V. 1992. Countries, Peoples and Their Languages.
Gothenburg, Sweden: Geolingua.
Gurr, Ted
Anderson, Benedict. 1983. Imagined Communities. London: R. 1971. Why Men Rebel. Princeton, NJ: Princeton Uni-
Verso.
versity
Asher, R. E., ed. 1994. The Encyclopedia of Language and Press.
Linguis-
tics. Oxford: Pergamon Press. Gurr, Ted R. 2000. Peoples versus States. Washington, DC: United
Atlas Narodov Mira. 1964. Moscow: Glavnoe upravlenie States Institute
geodezii i of Peace Press.
kartografii. Hegre, Havard, Tanja Ellingsen, Scott Gates, and Nils Petter
Gleditsch.
Avant, Deborah D. 1994. Political Institutions and Military Change.2001. "Toward a Democratic Civil Peace? Democracy,
Ithaca, NY: Cornell University Press. Political Change and Civil War, 1916-1992." American Political
Beck, Neal, Jonathan N. Katz, and Richard Tucker. 1998. Science Review 95 (March): 33-48.
"Taking
Hibbs,
Time Seriously: Time-Series-Cross-Section Analysis with Douglas A. 1973. Mass Political Violence. New York: Wiley.
a Binary
Dependent Variable." American Journal of Political Science 42 (4):
Horowitz, Donald L. 1985. Ethnic Groups in Conflict. Berkeley: Uni-
1260-88. versity of California Press.
Brogan, Patrick. 1998. World Conflicts. 3rd ed. Lanham, MD:
Huntington, Samuel P 1968. Political Order in Changing Societies.
Scarecrow Press. New Haven, CT: Yale University Press.
Chaudhry, Kiren Aziz. 1989. "The Price of Wealth: Business and State
Huntington, Samuel. 1996. The Clash of Civilizations and the Remak-
in Labor Remittance and Oil Economies." International Organi- ing of World Order. New York: Simon and Shuster.
zation 43 (1): 101-145. Ignatieff, Michael. 1993. Blood and Belonging. London: Noonday
Clutterbuck, Richard L. 1967. The Long, Long War: The Emergency Press.
in Malaya. London: Cassell. Institute for International and Strategic Studies. 2000. The Military
Collier, Paul, and Anke Hoeffler. 1999. Justice Seeking and Balance, 2000. London: IISS.
Loot-seeking in Civil War. World Bank. Typescript. http://econ. Jackman, Simon. 2001. "In and Out of War and Peace: The Statistical
worldbank.org/programs/conflict/library (November 18, 2002). Analysis of Discrete Serial Data on International Conflict." Ms.,
Collier, Paul, and Anke Hoeffler. 2001. "Greed and Grievance Stanford University.
in Civil War." World Bank. Typescript. http://econ.worldbank.org/Kalyvas, Stathis. 1999. "Wanton and Senseless? The Logic of
programs/library (November 18, 2002). Massacres in Algeria." Rationality and Society 11 (3): 243-85.
Connor, Walker. 1994. Ethnonationalism. Princeton, NJ: PrincetonKarl, Terry L. 1997. The Paradox of Plenty. Berkeley: University of
University Press. California Press.
Deininger, Claus, and Lyn Squire. 1996. "A New Data Set MeasuringKaufmann, Chaim. 1996. "Possible and Impossible Solutions to Eth-
Income Inequality." World Bank Economic Review 10 (3): 565-91. nic Civil Wars." International Security 20 (4): 136-75.
Desai, Raj, and Harry Eckstein. 1990. "Insurgency: The Transforma- Keen, David. 1998. The Economic Functions of Violence in Civil
tion of Peasant Rebellion." World Politics 42 (4): 441-66. Wars. London: Oxford University Press.
Deutsch, Karl W. 1953. Nationalism and Social Communication. Kennedy-Pipe, Caroline. 1997. The Origins of the Present Troubles
Cambridge, MA: MIT Press. in Northern Ireland. London: Longman.
Doyle, Michael W., and Nicholas Sambanis. 2000. "Interna- King, Gary, and Langche Zeng. 2001. "Logistic Regression in Rare
tional Peacebuilding: A Theoretical and Quantitative Analysis." Events Data." Political Analysis 9 (2): 137-63.
American Political Science Review 94 (4): 778-801. Krepinevich, Andrew F 1986. The Army and Vietnam. Baltimore,
Easterly, William, and Ross Levine. 1997. "Africa's Growth Tragedy: MD: Johns Hopkins University Press.
Policies and Ethnic Divisions." Quarterly Journal of Economics Kriger, Norma J. 1992. Zimbabwe's Guerrilla War. New York:
112 (4): 1203-50. Cambridge University Press.
Eisinger, Peter. 1973. "The Conditions of Protest Behavior in Laitin, David D. 2000. "Language Conflict and Violence." In Inter-
American Cities." American Political Science Review 67: 11-28. national Conflict Resolution After the Cold War, ed. Paul C. Stern
Esty, Daniel C., Jack A. Goldstone, Ted R. Gurr, Barbara Harff, Marc
and Daniel Druckman. Washington, DC: National Academy Press.
Levy, Geoffrey D. Dabelko, Pamela T. Surko, and Alan N. Unger. Laqueur, Walter. 1976. Guerrilla Warfare. New Brunswick, NJ: Trans-
action Books.
1998. State Failure Task Force Report: Phase II Findings. McLean,
VA: Science Applications International Corporation. Leites, Nathan, and Charles Wolf. 1970. Rebellion and Authority.
Fearon, James D. 1998. "Commitment Problems and the SpreadChicago,of IL: Markham.
Ethnic Conflict." In The International Spread of Ethnic Conflict, ed.
Leonard, David K., and Scott Strauss. 2001. "Power without Respon-
David A. Lake and Donald Rothchild. Princeton, NJ: Princetonsibility: The International Community and Africa's Failed Devel-
University Press. opment." Book ms., University of California, Berkeley.
Fearon, James D. 2001. "Why Do Some Civil Wars Last So Much
Licklider, Roy A. 1995. "The Consequences of Negotiated Settle-
Longer than Others?" Presented at the World Bank DECRG Con- ment in Civil Wars, 1945-93." American Political Science Review
ference on Civil War held, University of California, Irvine. 89 (3): 681-90.
Fearon, James D. 2002. "Ethnic Structure and Cultural Diversity Moynihan, Daniel P. 1993. Pandaemonium: Ethnicity in International
around the World." Presented at the Annual Meetings of the Politics. New York: Oxford University Press.
American Political Science Association, Boston. Muller, Edward N. 1985. "Income Inequality, Regime Repres-
Fearon, James D., and David D. Laitin. 1996. "Explaining Interethnic
siveness, and Political Violence." American Sociological Review
Cooperation." American Political Science Review 4: 715-35. 50: 47-61.
89
Ethnicity,
Ethnicity,
February
2003
2003
Muller, Edward N.,

Sivard, Ruth Leger. 1996. World Military andand
Social Expenditures, Eric
tions in Political Violence:
1996. Washington, DC: Global Priorities. A
of Conflict Smith, Anthony. 1986. The Ethnic Origins of Nations. Oxford:
Resolution 34: 62
Paige, Jeffery Blackwell. M. 1975. Agr
Przeworski, Snyder,
Adam, Michael
Jack. 2000. From Voting to Violence. New York: Norton.
and Fernando Limongi.
Thompson, Robert G. 1966. Defeating Communist Insurgency.20
Cambridge: London: Chatto and Windus.
Cambridge Univ
Rabushka, Alvin, and
Tilly, Charles. 1978. From Kennet
Mobilization to Revolution. Reading, MA:
cieties: A Addison-Wesley.
Theory of Democra
Russett, Toft, MonicaM.
Bruce Duffy. 1996. 1964.
"The Geography of Ethnic Conflict:
"Ine
16: 442-54. Do Settlement Patterns Matter?" Presented at the Annual
Sambanis, Nicholas. 2002. "Defining and Measuring Civil War:Meeting of the Midwest Political Science Association,
Conceptual and Empirical Complexities." Ms., Yale University. Chicago.
Valentino, Benjamin A. 2002. Final Solutions. Ithaca, NY: Cornell
Scott, James C. 1976. The Moral Economy of the Peasant. New Haven,
CT: Yale University Press. University Press.
Wantchekon, Leonard. 2000. "Why Do Resource Dependent
Singer, J. David, and Melvin H. Small. 1994. "Correlates of War
Project: International and Civil War Data, 1816-1992." ICPSRStates Have Authoritarian Governments?" Ms., Yale Uni-
9905, April. versity.
90

American Political Science Association The American Political Science Review

Uploaded by

Copyright:

Available Formats

American Political Science Association The American Political Science Review

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

American Political Science Association The American Political Science Review

Uploaded by

Copyright:

Available Formats

Ethnicity, Insurgency, and Civil War

Author(s): James D. Fearon and David D. Laitin

Ethnicity, Insurgency, and

FIGURE 1. Number and Percent

1945 1950 1960 1970 1980 1990 1999

Russia are the of

We identified 127 conflicts that meet theof 25 ongoing

of violent the civil conflic

the country's population,

Additional popular grievances

The remaining hypotheses

Are More Diverse Countries Prone

FIGURE 2. Probability of Civil War Onset per Five-Year Period

GDP/capita in lagged 1985 U.S. dollars

TABLE 1. Logit Analy

the (lagged) PolityNew IV States and Political Instability.

measuring the the other estimates remain stable). A dummy

36 Data from the World Bank, linearly interpolated 38 Wefor

Alternative How could democracy and cultural or and

Fearon, James D., and David D.practices

Muller, Edward N.,

You might also like