Economic sanctions and the duration of civil conflicts
Author(s): Abel Escribà-Folch
Source: Journal of Peace Research , march 2010, Vol. 47, No. 2 (march 2010), pp. 129-141
Published by: Sage Publications, Ltd.
Stable URL: https://www.jstor.org/stable/25654550
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
Sage Publications, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to
Journal of Peace Research
This content downloaded from
134.117.10.200 on Tue, 23 Feb 2021 10:39:01 UTC
All use subject to https://about.jstor.org/terms
Research article
journal of
RESEARC II
Journal of Peace Research
47(2) 129-141
Economic sanctions and the duration or ?71,6 Aud;or<s) 2010Reprints and permission:
? I ? sagepub.co.uk/journalsPermissions.nav
civil connicts dol 10.1177/0022343309356489
jpr.sagepub.com
Abel Escriba-Folch
Universit?t Pompeu Fabra
Abstract
This article studies the impact of economic sanctions on the duration and outcome of intrastate conflicts. Sanctions are argued to
foster the convergence of beliefs over parties' capacity, to reduce the utility of victory and to increase the costs of continuing
fighting. Using a sample of 87 wars and new data on sanctions and sanction types, the author shows that sanctions and their
durations are statistically associated with shorter intrastate conflicts. It is also shown that total economic embargoes are the most
effective type of coercive measure in these cases and that sanctions imposed either by international organizations or by other
actors have similar negative effects on war duration. In the second part of the article, the dependent variable is disaggregated,
and I demonstrate that sanctions imposed by international institutions increase the likelihood of conflict resolution, whereas
those sanctions not imposed by such institutions tend to increase the probability of a military victory. Moreover, if the targeted
state is a member of the international institution imposing the sanctions, the effect of such coercion is even greater. Economic
embargoes are also proven to increase the likelihoods of a military and a negotiated end, whereas international arms embargoes
reduce the likelihood of a military victory.
Keywords
civil war, duration, sanctions, sanctions types, war outcome
Introduction
Although economic sanctions are among the most commonly
used instruments of international relations, their usefulness is
constantly under debate and, quite often, doubted (Pape,
1997, 1998; N?rnberger, 2003). Sanction episodes have been
motivated by a multiplicity of factors, among which civil con
flicts and political violence are two of the most prominent. To
cite some examples, sanctions have targeted countries experi
encing civil war, such as Liberia, Yugoslavia, Sudan, Rwanda,
Lebanon and Cambodia. As the controversy continues, the
number of sanctions imposed has sharply increased over the
last two decades, especially those imposed by the United
Nations. Assessing whether these instruments of international
relations have had any significant effect in bringing about war
termination is thus especially relevant.
The existing literature on civil war duration has already
emphasized the important role that external actors may play.
Sanctions constitute a rather specific method of intervention
based on coercive measures imposed by one country, an inter
national organization or a coalition of countries against
another country - the government or any group within the
country - with the aim of bringing about a change in a specific
policy or behaviour. The debate about sanctions revolves
around not only their efficacy, but also whether some types
of sanction are more successful than others.
The aim of this article is to analyze the effect of economic
sanctions on the likelihood of civil war termination, in order to
aid our understanding of the usefulness of third-party inter
ventions and their modalities. We also study the effects of dif
ferent types of sanction and how they affect the war outcomes
(i.e. military victory versus negotiated settlement). To do so,
I use a dataset including 87 intrastate conflicts occurring
between 1959 and 1999. The results show that international
sanctions and their duration are negatively and significantly
associated with civil war length. Further refinements of both
the main independent and dependent variables serve to show
also that the most effective measure is a total economic
embargo against the target country. Regarding multilateral
sanctions, we show that sanctions imposed by multilateral
international institutions increase the likelihood of a negoti
ated settlement, while those not conducted through such
institutions enhance the likelihood of military victories.
The article proceeds as follows. First, we summarize the lit
erature on civil war duration and external interventions. Next,
we deal with the potential effects of economic sanctions and
their types on the durability of civil wars and the mode of con
flict termination, paying attention to the mechanisms that may
Corresponding author:
abel.escriba@upf.edu
This content downloaded from
134.117.10.200 on Tue, 23 Feb 2021 10:39:01 UTC
All use subject to https://about.jstor.org/terms
130_journal ?/Peace Research 47(2)
link both variables. In the following sections, we present the
data and the methods employed and discuss the empirical
results. The last section summarizes the main findings.
External interventions and civil war duration:
A review
The growing literature studying the expected duration of civil
wars bases its explanations fundamentally on the costs of war,
forecast errors and actors' capabilities, which can be altered by
the involvement of third parties. Just as in the case of interna
tional sanctions, the efficacy and convenience of external inter
ventions are subject to constant debate. Given that sanctions
are a particular form of outside intervention, it is worth briefly
reviewing the main insights drawn from the literature relating
external interventions and war duration.
In practice, third-party interventions predominantly rely on
the manipulation of the expectations of victory and the costs
associated with continued warfare in order to influence the war
process (Regan, 2002). Interventions may be directed toward
assisting the government, supporting the rebellion or remaining
neutral. The existing evidence is rather mixed and inconclusive,
owing to the different samples and techniques employed.
According to Elbadawi &c Sambanis (2000), outside inter
ventions tend to reduce the cost of coordinating a rebellion.
Taking Regan's data on interventions and covering the period
1960-1999, they estimate the predicted probability that an
external intervention will take place in a given civil conflict and
find that this measure is associated with longer wars.
Balch-Lindsay & Enterline (2000) use Correlates of War
(COW) Intrastate War Data on outside interventions and
conflict and show that biased interventions increase war dura
tion and that balanced interventions result in extremely long
civil wars. Similarly, after incorporating some refinements to
his data, which incorporate 150 conflicts between 1945 and
1999, Regan (2002) reports that both economic and military
interventions have a strong positive impact on intrastate war
length. Buhaug, Gates & Lujala (2002) find that interventions
on the government's side increase the time until a civil war
ends. All these findings are consistent with Cunningham's
(2006) finding that multiple veto players are related to longer
conflicts. Conversely, Collier, Hoeffler &c S?derbom (2004)
use Regan's data on outside interventions but instead of consider
ing contemporaneous effects, they use a measure that cumulates
the number of months of each type of external intervention.
Their results, using a sample of 55 wars between 1960 and
2000, show that economic interventions have a positive effect
on war length but are wholly insignificant, and that only military
interventions on the side of the rebels result in shorter civil wars.
The study of the likelihood of war end may overlook the
alternative ways in which a conflict can actually terminate.
Consequently, simple duration models have the disadvantage
that they do not differentiate the different effects of a given
covariate on the alternative war outcomes, because the
coefficients would be reflecting an average effect of how factors
affect different outcomes (such as a military victory or a
negotiated settlement). So, for instance, one may find that a
covariate is not significant because it has opposing effects on
a government military victory and an opposition military
victory. The distinction between war outcomes has permitted
scholars to yield new insights concerning the precise role of
third-party interventions. Mason, Weingarten & Fett (1999)
use COW data (including 57 civil conflicts between 1945 and
1992) and report that third-party interventions make a nego
tiated settlement more unlikely. Yet, they also show that as the
conflict becomes protracted, interventions increase the
likelihood of a settlement. Similarly, analyzing a sample contain
ing 213 wars and outside interventions compiled from the
COW database covering the period 1816-1997, Balch
Lindsay, Enterline & Joyce (2008) find that a third-party inter
vention supporting one of the sides reduces the time until that
group achieves military victory, but it makes a negotiated setde
ment more unlikely. In contrast, DeRouen &c Sobek (2004)
report that UN interventions increase the probability of a treaty
or a truce and decrease the likelihood of one side's victory.
International sanctions and conflict resolution
The first comprehensive empirical study revealed a low success
rate of imposed sanctions (Hufbauer, Schott & Elliott, 1990;
Hufbauer et al., 2007).1 Later research tended to increase scep
ticism contending that sanctions basically do not work and
that any policy effect observed is due to the use of force that
frequently accompanies them (Pape, 1997, 1998). Further
studies then moved to investigate what kind of sanctions were
more effective, under what conditions could sanctions be
expected to work, and the impact of sanction threats.
Concretely and concerning threats, it is argued that, as eco
nomic coercion is the result of strategic interaction, those sanc
tions that are likely to be effective will succeed at the mere
threat stage (Drezner, 2003; Lacy & Niou, 2004). As for
imposed sanctions, some conditions have been identified that
make sanctions more likely to work: the sender's perceptions
of the salience of the issue (Ang & Peksen, 2007), the initial
stability of the target and the cost to the target country
(Dashti-Gibson, Davis & Radcliff, 1997), the political regime
of the targeted country (Nooruddin, 2002; Lektzian & Souva,
2007), and future conflict expectations (Drezner, 1998).
Finally, scholars have also sought to measure and explain coer
cion effectiveness from the perspective of the type of measures
applied or according to the senders' characteristics. One part
of the debate revolves around whether unilateral or multilateral
sanctions are more effective (Bapat & Morgan, 2007; Drezner,
1999; Hufbauer, Schott & Elliott, 1990; Kaempfer &
Lowenberg, 1999), as well as around the role of international
organizations (Drezner, 2000) and the involvement of
1 The updated version of the book reports the same rate of success, about one
third. See Hufbauer et al. (2007).
This content downloaded from
134.117.10.200 on Tue, 23 Feb 2021 10:39:01 UTC
All use subject to https://about.jstor.org/terms
Escriba-Fokh_131
superpowers (Davis & Engerman, 2003). The second branch of
the debate seeks to establish the efficacy of targeted measures vis
a-vis comprehensive sanctions. Some evidence suggests that
financial sanctions increase the success rate of sanctions seeking
to yield a policy change. On the contrary, Cortright & Lopez
(2002), albeit being in favour of smart sanctions, find evidence
supporting the greater effectiveness of comprehensive sanctions,
as they maximize the economic and social impact on the target.
Economic sanctions constitute a concrete type of external
intervention relying on coercive measures consisting of differ
ent types of intended economic damage with the aim of trig
gering a policy change in the target government. The main
motivations for such measures are intrastate conflicts, human
rights violations and political violence. Therefore, an appropri
ate way to assess their success is to analyze whether sanctions
have been of any help in shortening civil wars (and fostering
conflict resolution). However, the studies of external interven
tions reviewed above have not considered sanctions separately;
rather, they consider only 'economic interventions', which
may include aid, grants, loans and restrictions. Thus, we know
little about the specific impact - if any - of economic sanctions
in the specific context of civil war. So far, sanctions, inter
preted as a type of costly signal from an external actor, have
been shown to exert no significant impact on the probability
of civil war onset (Thyne, 2006), while concerning war dura
tion, the existing evidence is only partial and based on an
extremely limited number of cases (Strandow, 2006).
The potential impact of sanctions on war length can be
simply approached from the existing third-party intervention
perspective. Sanctions can be thought of as constituting one
particular form - economic - of external intervention into a
civil conflict and, thus, involving the problems associated with
such interventions. Accordingly, external interventions may
make it difficult to reach an agreement or a military victory
by entailing the involvement in the conflict of a new veto
player, which might reduce the range of acceptable agreements
(Cunningham, 2006), by reducing the costs of rebellion coor
dination (Elbadawi & Sambanis, 2000), or by shifting the bal
ance of parties' capabilities toward parity (Regan, 2002).
Consequently, and according to the existing evidence on exter
nal interventions reviewed in the previous section, economic
sanctions would tend to lengthen the duration of civil wars,
as indeed suggested by many of the empirical findings, such
as Regan's (2002) study, which shows that economic interven
tions tend to prolong civil wars.
In contrast, sanctions can be argued to decrease civil war
duration through various, more specific, mechanisms. In simple
formal terms, and somewhat following Mason, Weingarten &
Fett's (1999) setting, each party's expected utility of continued
fighting at a given time can be expressed as follows:
EUF=peUv + (\-pe)UD-ECF, (1)
where pe is that party's subjective estimate of the probability of
winning, Uv\s the utility of victory, Up is the utility of being
defeated, and ECF stands for the estimated costs of fighting.2
The other possible outcome is that both parties reach an agree
ment from which utility EU$ is derived.3 Hence, a party will
continue fighting unless
EUS > EUp, (2)
that is, if the utility of a settlement exceeds the expe
of fighting; or unless the expected costs of fighting e
overall expected utility of victory, so one party gives u
and accepts defeat:
peUv + {l-pe)UD<ECF.
Following this simple scheme, sanctions may operat
context of civil war (i) by altering the parties' beliefs
relative distribution of power (which affect the com
estimations of the probability of victory, pe), (ii) by
the structure of the incentives of the contending
(so reducing the expected payoff from victory and th
of continuing fighting, Uy), or (iii) by reducing the a
financial resources and arms necessary to sustain w
thereby increasing the costs of fighting, ECp.
On the one hand, sanctions can be placed in the c
the general uncertainty that wars involve, which af
estimated probability of victory, pe. Hence, one set
ments contends that the parties in a conflict will c
fight if they are uncertain about the actual distri
power, which makes the demands of the other side
unacceptable (Filson & Werner, 2002). In other wor
duration can be understood as the result of foreca
usually over-optimistic - on the part of one of the
both (Elbadawi & Sambanis, 2000). This view coinci
the rebellion-as-mistake approach, which stresses t
misperceptions about the military capability and, he
the probability of victory (pe) as a key determinant o
sistence (Hirshleifer, 2001; Collier, Hoeffier & S?d
2004). Indeed, Cunningham, Gleditsch & Salehya
show that civil wars tend to last longer when rebel g
weaker vis-?-vis the government. Under such cond
settlement is clearly implausible. All in all, as
(2006) claims, conflict resolution is the result of con
of the parties' beliefs over their relative power dist
which sanctions can help to promote by influencing
information, thereby making the each side's estim
ability of winning, pe, progressively approach the rea
ity of victory, pe ? pr. This makes possible that o
2 According to Mason, Weingarten & Fett (1999: 242), these cost
summed from the present until that point in time in the futur
party in question estimates that it will be able to achieve victor
t= victory
could also be modelled just as Cp.
t=0
3 We treat it, albeit without disaggregating it, as an expected utility, because
settlements involve uncertainty due to implementation difficulties, potential
cheating and lack of confidence between parties.
This content downloaded from
134.117.10.200 on Tue, 23 Feb 2021 10:39:01 UTC
All use subject to https://about.jstor.org/terms
132 journal of "Peace Research 47(2)
demands can be accepted by the other as they reflect its real
power. For instance, sanctions are argued to have been success
ful in stabilizing the Liberian conflict. The arms embargo
(imposed in 1992) made the acquisition and delivery of weap
ons more difficult and less predictable (Wallensteen, Eriksson
& Strandow, 2006). Charles Taylor even declared we are
hands-tied as a result of the arms embargo' (cited in Strandow,
2006: 11). A comprehensive peace agreement was reached in
August 2003. The settlement prompted the resignation of for
mer president Charles Taylor. The National Transitional Gov
ernment of Liberia (NTGL) - which included both rebel and
government groups - took over two months later.
On the other hand, sanctions may shorten conflict through
their impact on the parties' incentives to continue the fight,
that is, the expected utility a party would derive from victory
(Uy) and the utility of sustained conflict. If some civil wars
effectively contain a greed component (Collier & Hoeffler,
2004), then, as Collier Hoeffler & S?derbom (2004: 254) pro
pose, 'a key prediction is that the higher the payoff from vic
tory, the longer would be the warranted rebellion'. Rebellion
is seen as an investment in this case, and economic sanctions,
by restricting the expected benefits of resource control and
export, may alter parties' incentives to prolong a conflict. Like
wise, in the conceptualization of rebellion as business, groups
have an incentive to continue fighting when the general state
of lawlessness makes contraband and other illegal activities
possible and highly profitable (Collier, 2000). Indeed, Fearon
(2004) shows that contraband by rebels increases war dura
tion, especially given the presence of lootable goods like dia
monds and gemstones (Buhaug, Gates & Lujala, 2002).
Perhaps the clearest examples of measures seeking to limit such
incentives are diamond sanctions and other export restrictions.
These seek to control illicit trade and limit the funding of rebel
groups and, thus, the expected utility of controlling the extrac
tion and trade of natural resources and other commodities. For
example, the export ban and the oil embargo imposed by the
UN against the Khmer Rouge seriously limited many gem
mining operations around Pailin (Cambodia). Then, if
S stands for sanctions, we have that < 0. So, as a result,
both the right-hand side in (2) and the left-hand side in (3)
decrease, as the expected utility of continued fighting shrinks.
Finally, independently of the underlying motivations for
continuing a war, its durability is going to be crucially influ
enced in the short term by both financial and military feasibil
ity (Buhaug, Gates & Lujala, 2002; Fearon, 2004;
Humphreys, 2005). This is the main mechanism through
which sanctions are supposed to operate, as they are especially
designed to try to limit such viability by curtailing contending
parties' arms procurements, illicit flows and the collection of
public revenues from trade and foreign aid, so increasing the
expected costs of conflict, > 0. Therefore, according to
our setting, this makes the right-hand side in (2) decrease and
the right-hand side in (3) increase, making war termination
more likely. For instance, a trade embargo may decrease the
exports of goods, thereby affecting the resources available to
the government. As for the rebel side, sanctions are usually
aimed at limiting the funding that some of these groups obtain
through contraband. For example, it is argued that the Lancas
ter House agreement (1979) that put an end to the conflict in
Rhodesia was in part made possible by the extensive UN sanc
tions, imposed in 1966, that forbade trade (the insuring of
commodities or goods, exports and imports from Rhodesia)
as well as financial exchange with the target country, and so
weakened Ian Smith's government and forces. The diamond
embargo recently imposed on Ivory Coast (2005) is intended
to reduce the revenues raised from export and production by
the rebels of the New Forces (Wallensteen, Eriksson &
Strandow, 2006). Diamond sanctions have been imposed
against UNITA in Angola (along with petroleum sanctions),
Sierra Leone and Liberia. It is argued that the 2001 diamond
sanctions on Liberia were effective in reducing government
revenues and, consequently shortening the conflict (Wallens
teen, Eriksson & Strandow, 2006). If properly enforced, all
of them are principally designed to cut to some extent the flow
of monetary and military resources into the hands of the con
tending parties (especially the rebels).
So let us frame the following guiding hypotheses, relying on
both the arguments against the usefulness of sanctions and on
the three basic mechanisms just mentioned above in favour of
sanctions' role in shortening conflicts:
Hla: Sanctions, as a sort of external intervention, are negatively
related to the probability of war end.
lb: The duration of civil war is decreased by the imposition of
sanctions.
The debate over sanctions does not only revolve around
whether they are effective or not. The apparent reliance for
sanctions' success on the costs they impose on targets com
pelled governments and scholars to discuss and assess the type
of cost that should be inflicted, and on what actor or group
those costs should be concentrated in order to maximize effec
tiveness. In this case, comprehensive sanctions must be distin
guished from the so-called smart - or targeted - sanctions. The
former are intended to maximize general costs on target. Advo
cates of smart sanctions assert that instead, targeted measures
maximize the cost on the specific group whose obedience is
sought while avoiding causing the general population to suffer.
Kaempfer & Lowenberg state that 'the sanctions which are
most likely to precipitate the desired political change in the tar
get country are those which concentrate income losses on
groups benefiting from the target government's policy'
(1988: 792). For instance, Dashti-Gibson, Davis & Radcliff
(1997) find that when senders seek to achieve policy changes,
the imposition of financial sanctions is an important determi
nant of success. On the contrary, Gershenson (2002) formally
shows that it is the strength of sanctions that affects the alloca
tion of resources to conflict and the utility of the contending
parties. In fact, the existing evidence suggests that comprehen
sive sanctions are more effective (Cortright & Lopez, 2002),
This content downloaded from
134.117.10.200 on Tue, 23 Feb 2021 10:39:01 UTC
All use subject to https://about.jstor.org/terms
Escriba-Folch 133
and many studies have identified th
imposed on the target, the higher th
coercion episodes (Hufbauer, Sch
Hufbauer et ah, 2007; Dashti-Gibs
1997; Drury, 1998; Nooruddin, 2002).
partial evidence on some specific types
to support the more pessimistic view a
tiveness. For instance, following a com
some UN arms embargoes, such meas
largely irrelevant (Tierney, 2005). The
failure include the exemption of Secu
from the restrictions, the incentives t
weak enforcement, late application a
power asymmetries between parti
2002; Tierney, 2005). In clear contras
tends that arms embargoes, given th
actors' military power, should be the m
geted measure - if properly impleme
likelihood of conflict resolution by, a
the convergence of beliefs over each p
offers quantitative evidence of the p
tions on two cases, Liberia and Ivory
I attempt to evaluate the relationship b
of measures - such as total economic e
goes, trade restrictions and aid cuts (m
in sanction episodes) - and the durat
above statements are summarized in th
H2: Sanctions that maximize costs on the
goes) should decrease the duration of civ
Another way of studying and classify
focusing on the senders. Internation
multilateral or unilateral or, more sp
either by international institutions or
small coalition of them). Some eviden
view that unilateral sanctions tend to
multilateral ones (Drezner, 1999; Hufb
1990; Kaempfer & Lowenberg, 1999), p
above-mentioned problems in targeted
embargoes - namely, free-riding and en
Nevertheless, a more recent study, re
finds that multilateral sanctions have a
tive (Bapat & Morgan, 2007).4 Drezne
possibility that the key variable is no
multilateral or not but whether those m
imposed by an international organiza
backsliding and ensures the maintenan
to cooperate with the sanctioning col
increases the senders' capability of i
4 Most of the analyses of sanction duration an
Hufbauer, Schott & Elliott's (1990) dataset.
5 Bapat & Morgan (2007) also find that the succe
if they are conducted through an international in
on the target countr
cause of the success o
Apart from this cap
ment of an internat
has other implicatio
a signal that an outsi
the terms of a pote
been true of some
which have been fol
peacekeeping operat
the peace agreement
the UNOMSIL in S
dia. The reduction of
of a settlement (BUs
failure to fulfil the
hold. In the light o
hypothesize:
H3: Sanctions that ar
tional organization s
internal conflict dur
Furthermore, I will
sanction senders on t
victory vs. negotiat
tially assess the role
peace. The same will
Data and meth
In order to answer
on civil wars cover
on intrastate conf
Laitin (2003), Hum
and include 87 civil
Since the data have
better control for the
is under economic s
With regard to the d
analysis I use a binar
ends in a given yea
coded 1 if the war e
year. For the second
used by Humphrey
flict termination an
is ongoing in that y
6 The list constructed by
meet the following prim
agents of the state and o
least 1,000 over its cours
100 were killed on both s
(2003).
7 Humphreys compiled it from Walter (2002).
This content downloaded from
134.117.10.200 on Tue, 23 Feb 2021 10:39:01 UTC
All use subject to https://about.jstor.org/terms
134 journal ?/Peace Research 47(2)
military means, and 2 if the war ends through a negotiated
settlement.
Data on sanction episodes are taken from Marinov's (2005)
dataset, which recast Hufbauer, Schott & Elliott's (1990) data
set in country-year format and updated it. These data have
been contrasted and widened using the Threat and Imposition
of Sanctions (TIES henceforth) dataset, which collected data
on threats and sanction implementation during the period
1971-2000 (Morgan, Krustev & Bapat, 2006). Hufbauer,
Schott & Elliott's (1990) dataset covers 116 cases of sanctions
(between 1914 and 1990), whereas the TIES dataset includes
529 instances of sanctions imposition (Bapat & Morgan,
2007). Moreover, the TIES dataset classifies episodes accord
ing the type of measure applied. Thus, I have constructed two
more variables: The first, 'Institutional sanction', takes the
value 0 if no sanction is imposed, 1 if a country is under sanc
tions not imposed by an international institution, and 2 if a
country is under sanctions imposed by an international insti
tution in a given year. The second variable distinguishes
between the main types of measure adopted: total economic
embargoes, multilateral arms embargoes (UN, EU and other
multilateral embargoes),8 restrictions on imports and exports,
the termination of foreign aid and other measures (such as
asset freezes, travel bans, suspension of agreements and block
ades). Finally, I have also constructed a variable that distin
guishes between threats of sanctions and imposed sanctions.
As mentioned above, the analyses on sanctions effectiveness
may suffer from selection bias, which lies in the fact that
threats of sanctions might be more successful than imposed
ones but they are generally not observed. So, following Drez
ner (2003), to test the selection argument, we need to study
those cases in which sanctions are threatened but not imple
mented. Fortunately, the new TIES data set contains informa
tion on threat episodes that did not end up with the
imposition of sanctions. Therefore, our variable 'sanction
threat' is coded 0 if no threat or sanction is applied, 1 if a coun
try is threatened but eventually sanctions are not imposed, and
2 if sanctions are imposed. We consider a number of control
variables that refer to commonly identified factors included
in recent studies on civil war duration.
Geographical characteristics of the country are argued to
influence war viability and actors' capacity (DeRouen &
Sobek, 2004; Buhaug, Gates & Lujala, 2005). Both the per
centage of the terrain that is mountainous and forested and the
number of bordering states are frequently argued to increase
the ability of the rebels to resist. As for the government's capac
ity, we include the size of the army (per 1,000 inhabitants),
8 The TIES dataset does not include a category for arms embargoes, so given
the policy relevance of international arms embargoes, I have constructed a
dummy for arms embargoes imposed by international organizations. Data
have been taken from Fruchart et al. (2007) and the SIPRI website:
www.sipri.org/contents/armstrad/embargoes.html.
9 These variables have all been compiled from DeRouen & Sobek (2004).
obtained by dividing the size of the army by the total
population.9
Some researchers point to the potential polarizing effect of
ethnic fractionalization (Montalvo & Reynal-Querol, 2007),
and a high degree of fractionalization may hamper cooperation
on the rebel side. Collier, Hoeffler & S?derbom (2004) and
Elbadawi & Sambanis (2000) find a non-linear relationship
between ethnic diversity and war duration. Therefore, I con
trol for the degree of ethnic diversity and this potential curvi
linear pattern using Fearon's (2003) variable, which measures
the probability that two randomly selected persons from a
given country will not belong to the same ethnic group.
On the socio-economic side, I include the logarithm of the
country's population and per capita GDP. The presence of
natural resources and primary commodities has been shown
to have an effect on both the onset and the duration of civil
war (Collier, Hoeffler & S?derbom, 2004; Ross, 2004). I use
several measures: first, 'mineral resources', which takes the
value 1 if the average ratio of ore and mineral exports in any
year for which a country has data exceeded 50% of total
merchandise exports, and 0 otherwise.10 The second variable,
'oil-producing country', is coded 1 for those country-years in
which fuel exports exceeded one-third of total export revenues,
and 0 otherwise.11 I have also used the variables on oil produc
tion and reserves and diamond production (expressed both in
total production and in per capita terms) developed by
Humphreys (2005).
To capture the possibility of financing through contraband,
the dummy variable constructed by Fearon is included. It is
coded 1 if there is 'evidence of major reliance by the rebels
on income from production or trafficking in contraband'
(2004: 284), and 0 otherwise.
As stressed above, Pape (1997, 1998) argues that sanctions
do not work, since any policy effect observed so far is actually
due to the use of force that sometimes accompanies sanctions.
To control for this possibility, I include the dummy variable
'military intervention', which is coded 1 for each country
year in which some sort of military intervention, as categorized
by Regan (2000), takes place.
Other relevant controls are: the logarithm of the average
number of deaths per year as compiled by Fearon (2004). And,
also compiled from Fearon (2004), some dummies capturing
the type of ongoing civil war are considered. The first is 'ethnic
war', which takes value 1 if the ongoing war is of an ethnic
nature.12 The second dummy takes value 1 if the war is clas
sified as a 'sons of the soil' conflict. These are wars in which
the state is dominated by an ethnic group facing population
pressure. When members of this group migrate to less
populated areas, often with the support of the state, the ethnic
10 Variable compiled from Gandhi & Przeworski (2006).
11 Compiled from Fearon & Laitin (2003).
12 We have coded as 1 only those cases with value 3, excluding those
considered by Fearon to be mixed or ambiguous.
This content downloaded from
134.117.10.200 on Tue, 23 Feb 2021 10:39:01 UTC
All use subject to https://about.jstor.org/terms
Escriba-Folch_135
10 20 30 40~
Years in progress
-No sanctions Under sanctions I
Figure 1. Proportion of civil wars ongoing by year
minorities living in those regions often take up arms against
the migrants and the state.
The methodology employed consists basically in logistical
regressions, both binary, to analyze the likelihood of conflict ter
mination, and multinomial, to analyse the mode in which the
war ends (military victory or negotiated settlement). As usual
when duration is analysed using discrete-time data, the potential
time dependence in the data is corrected by including natural
cubic splines on the right-hand side of the equation to be esti
mated (Beck, Katz & Tucker, 1998). Errors have been clustered.
Analyses and results
Civil war termination and economic sanctions
I begin by enquiring whether sanctions have any effect on the
length of intrastate conflicts in a simple manner. Figure 1 portrays,
using the non-parametric Kaplan?Meier estimate, the survival
curves of those conflicts under sanctions versus those not under
sanctions. The differences between the groups are substantial and
suggest a significant correlation between the two variables. As the
plot reveals, the proportion of ongoing wars is clearly smaller for
those cases targeted by international economic sanctions.13 This
preliminary evidence suggests that international sanctions are
related to the decline of the survival rate of intrastate wars.
Moving on to the multivariate analyses, Table I reports the
estimated coefficients of the binary logistic regressions used to
estimate the impact of economic sanctions with the aim of
testing our first hypotheses. In Columns 1-3, I include the
sanctions dummy, while in Columns 4 and 5, I include the
variable 'sanctions duration', which is a measure that
13 The variable 'sanction' includes only sanctions actually imposed, not
threats.
1 Therefore, in this case, sanctions are not considered just as a constant
variable, but as a steadily increasing value over time, as it is assumed that
the higher the number of years a given country is targeted, the higher are
the accumulated costs. See Collier, Hoeffler & S?derbom (2004) for a
similar methodology applied to outside interventions.
cumulates the number of years a country has been under sanc
tions in a given year.14 The results serve to reject Hypothesis
la and confirm Hypothesis lb: both sanctions and their dura
tion (in years) are significantly associated with a higher likeli
hood of civil war termination. These results are robust to the
inclusion of the variable 'military intervention' and to the use
of alternative measures of natural resource availability. Note
that the effect of a military intervention is negative, as previous
research had already indicated, but not significant. Being
under sanctions involves an increase of 0.044 in the probabil
ity of conflict termination (according to the estimates in
Column 1), while a one-unit change in the time a country has
been targeted by sanctions increases the probability of war
termination by 0.0041 points (Column 4).15 The computed
probability of civil war termination when a country has been
just one year under coercive sanctions is 0.0384, whereas the
probability of termination for a country that has been targeted
for five years is 0.0572.
Column 6 includes the distinction between sanction
threats and effectively imposed sanctions. The number of
observations is reduced in this case, as the sample is restricted
to the TIES data, so it starts in 1971. Threats that did not end
with the imposition of sanctions have an important positive
effect on the likelihood of conflict end, but the relationship
is not statistically significant, probably due to the reduced
number of such instances in our sample.
I now move on to examine the effects of different types of
sanctions. In the first column of Table II, the estimates use the
categorization of sanctions according to the kind of measure
adopted against the target. This model is restricted to the TIES
sample (from 1971), so the number of observations is lower.
The results tend to confirm Hypothesis 2 and make clear that
it is basically comprehensive sanctions directed towards cut
ting the total flow of funds to the rival parties that have signif
icant and negative effects on the duration of intrastate
conflicts. The examination of the marginal effects of each of
the types included in the regression reveals that comprehensive
sanctions (total embargoes) are the most effective measure in
shortening civil wars, followed by trade restrictions (of exports
and imports), which include commodity sanctions.16 Concre
tely, the increases in the probability of war termination when
those dummies change from 0 to 1 are 0.30 and 0.075, respec
tively. Multilateral arms embargoes do not appear to have any
significant effect on civil war duration.
With regard to whether sanctions are imposed by an inter
national institution, the results in Column 2 reveal that, when
no further distinction is introduced in the dependent variable,
the impacts of both kinds of sanctions are almost identical in
size. The changes in the likelihood of war termination are
0.050 and 0.057 when these two dummy variables increase
15 The rest of the variables are held constant at their means.
16 Trade restrictions are almost significant; the />-value of this variable is just
0.105.
This content downloaded from
134.117.10.200 on Tue, 23 Feb 2021 10:39:01 UTC
All use subject to https://about.jstor.org/terms
136 journal ?/Peace Research 47(2)
Table I. Sanctions and civil war duration (logistic regression)
Event: Civil war end = 1
Independent variables (1) (2) (3) (4) (5) (6)
Intercept
Mountains
Forests
Log population
Log GDP per capita
Mineral exporting
Oil exporting
Oil production
Diamond production
Ethnic fractionalization
(Ethnic fractionalization)2
Contraband
Number of borders
Army size (log)
Deaths/year
Ethnic war
Sons of soil war
Military intervention
Economic sanctions
Sanction duration
Threat
Imposed sanction
Duration splines
Observations
Log-pseudolikelihood
Pseudo R-squared
2.32
(1.86)
-0.007
(0.006)
-0.001
(0.009)
-0.119
(0.231)
-0.546**
(0.250)
2.13***
(0.789)
0.285
(0.419)
7.09***
(2.52)
-8.02***
(2.65)
-1.65***
(0.527)
-0.066
(0.078)
-0.353*
(0.184)
4.13e-06*
(2.18e-06)
0.563
(0.470)
_2 \ 1 ***
(0.730)
0.847***
(0.293)
yes
663
-151.90
0.1309
2.59
(1.65)
-0.003
(0.007)
0.006
(0.009)
-0.070
(0.236)
-0.635**
(0.287)
6.32
(6.33)
1.67**
(0.845)
5.45*
(2.94)
-6.28**
(3.14)
-1.44**
(0.583)
-0.094
(0.085)
-0.338*
(0.195)
3.93e-06*
(2.17e-06)
0.540
(0.463)
-2.08***
(0.740)
0.885***
(0.291)
yes
619
-148.01
0.1232
2.23
(2.15)
-0.009
(0.007)
0.003
(0.011)
-0.202
(0.263)
-0.398
(0.292)
2.73**
(1.19)
0.245
(0.436)
6.44*
(3.40)
7.42**
(3.36)
-1.68***
(0.614)
-0.033
(0.082)
-0.393*
(0.226)
6.21e-06**
(2.54e-06)
0.512
(0.580)
-2.06***
(0.750)
-0.606
(0.541)
0.883***
(0.316)
yes
638
-138.34
0.1359
3.87*
(2.33)
-0.008
(0.007)
-0.003
(0.010)
-0.120
(0.244)
-0.794**
(0.315)
2.97***
(1.01)
0.603
(0.484)
10.48***
(0.343)
-11.99***
(3.68)
-1.77***
(0.624)
-0.112
(0.090)
-0.334*
(0.202)
4.10e-06*
(2.26e-06)
0.598
(0.503)
-2.32***
(0.662)
0.104***
(0.028)
yes
663
-150.34
0.1398
3.53
(2.65)
-0.009
(0.008)
0.002
(0.011)
-0.225
(0.285)
-0.608*
(0.328)
3.65***
(1.41)
0.557
(0.499)
10.24**
(4.81)
-11.77**
(4.83)
-1.79**
(0.717)
-0.069
(0.095)
-0.405*
(0.238)
6.24e-06**
(2.67e-06)
0.636
(0.595)
-2.24***
(0.679)
-0.566
(0.565)
0.096***
(0.029)
yes
638
-137.46
0.1414
2.34
(2.18)
-0.017*
(0.009)
-0.004
(0.012)
-0.383
(0.289)
-0.304
(0.328)
2.83**
(1.30)
-0.161
(0.521)
9.19**
(4.22)
-10.48**
(4.27)
-1.45**
(0.604)
0.099
(0.084)
-0.336
(0.290)
0.00001***
(4.l4e-06)
0.619
(0.614)
-1.85***
(0.533)
-0.563
(0.529)
1.34
(1.06)
0.921***
(0.341)
yes
565
-123.27
0.1615
Robust standard errors in parentheses.
***/><.01; *><.05; ><?10.
from 0 to 1, respectively. Although it is true that the effect of
those sanctions imposed by international institutions is bigger,
the negligible differences in the sizes of both effects give little
support to Hypothesis 3. In the next section, I examine
whether these sanctions are associated with different war
outcomes.
This content downloaded from
134.117.10.200 on Tue, 23 Feb 2021 10:39:01 UTC
All use subject to https://about.jstor.org/terms
Escriba-Folch 137
Table II. Sanction types and the duration
Event: Civil war end = 1
Independent variabks(1) (2)
Intercept
Mountains
Forests
Log population
Log GDP per capita
Mineral exporting
Oil exporting
Ethnic fractionalization
(Ethnic fractionalization)2
Contraband
Number of borders
Army size (log)
Deaths/year
Ethnic war
Sons of soil war
Military intervention
Non-institutional sanction
International institution sanction
Total embargo
Aid termination
Trade restrictions
Multilateral arms embargo
Othet sanctions
Duration splines
Observations
Log-pseudolikelihood
Pseudo R-squared
3.04
(2.33)
-0.013
(0.011)
-0.011
(0.015)
-0.212
(0.303)
-0.487
(0.317)
3.65***
(1.33)
0.593
(0.580)
7.84*
(4.36)
-9.54**
(4.67)
-1.84***
(0.631)
0.144
(0.095)
-0.537*
(0.307)
9.04e-06*
(4.87e-06)
0.418
(0.579)
-2.28***
(0.675)
-0.645
(0.540)
2.70***
(0.730)
0.546
(0.479)
1.25
(0.773)
-0.627
(1.07)
0.584
(0.922)
yes
565
-119.23
0.1596
2.73
(2.07)
-0.010
(0.007)
0.001
(0.012)
-0.197
(0.268)
-0.438
(0.315)
2.64**
(1.20)
0.437
(0.432)
6.38*
(3.58)
?7.45**
(3.56)
-1.76***
(0.628)
-0.022
(0.088)
-0.403*
(0.232)
4.26e-06
(3.11e-06)
0.372
(0.580)
-2.12***
(0.828)
-0.608
(0.532)
0.940**
(0.390)
1.01*
(0.607)
yes
631
-136.25
0.1322
Robust standard errors in parentheses.
**> < .01; *> < .05; > < .10.
As for the rest of the variables, their estimated patterns con
form to some of the evidence already provided by previous
research. In line with Humphreys's (2005) results, the produc
tion of diamonds and the export of minerals tend to shorten
civil wars. Confirming Fearon's (2004) findings, I also find
that contraband hinders conflict resolution and that 'sons of
the soil' wars tend to last longer than other types of conflict.
Ethnic wars are slightly shorter, but the effect is not significant.
Furthermore, the size of the army is related to longer wars (as
already observed by DeRouen & Sobek, 2004), while the
number of fatalities, as well as the GDP per capita, tend to
shorten wars. I find a curvilinear relationship between ethnic
fractionalization and conflict duration, too. The geographic
characteristics of the country do not have any significant effect
in these pooled regressions (only mountainous terrain in
Column 6).
Civil war outcomes and the effect of sanctions
Not all civil wars end in the same way, so outcomes may need
to be treated as competing risks. According to our sample, and
following the codification developed by Walter (2002), 49 of
the 66 civil conflicts that finished within the period under
study ended because of a military victory of one of the sides,
while 17 ended through a negotiated settlement. I want to
investigate whether sanctions imposed by an international
institution are more conductive to negotiated settlements,
whereas those that do not involve an international organiza
tion are more prone to lead to military victories. The method
employed to test this proposal is, in this case, multinomial
logit.
The estimates are reported in Table III. The likelihood of
each war outcome is strongly influenced by the existence of
sanctions in the direction pointed out (see Models 1 and 2).
As is clearly revealed, although sanctions generally help to
reduce conflict length, they do so with varying consequences.
Sanctions imposed by international institutions significantly
increase the probability of reaching a negotiated settlement
that brings the conflict to an end. Increased cooperation
between senders augments the costs and efficacy of sanctions
episodes. Moreover, the intervention of an international orga
nization signals the parties that an outside actor may intervene
to guarantee the terms of a potential settlement (Walter,
1997), so the utility of a pact increases as the likelihood of a
unilateral defection is diminished by a third actor. On the
other hand, those sanctions unilaterally imposed by individual
countries or a small coalition have an important impact on the
probability of a civil war ending through military means. One
possible reason (needing further study) may be that those sanc
tions not under the direction of an international multilateral
institution may be biased and possibly inspired by the domes
tic interests of the primary sender.
In Model 3, a further refinement to our variable on multi
lateral sanctions has been introduced. It can be argued that
sanctions imposed by international institutions will tend to
be more effective if the target country is itself a member of the
multilateral institution. Greater diplomatic contact, military
This content downloaded from
134.117.10.200 on Tue, 23 Feb 2021 10:39:01 UTC
All use subject to https://about.jstor.org/terms
138 journal of Peace Research 47(2)
Table III. Civil war outcome and international sanctions (multinomial logit)
y = 1, conflict resolved through military y = 2, conflict resolved through negotiated
means settlement
Independent variables (1) (2) (3) (1) (2) (3)
Intercept 2.03 1.22
(1.76) (2.12)
Mountains 0.008 0.004
(0.010) (0.010)
Forests 0.020* 0.023*
(0.012) (0.012)
Log population 0.006 -0.073
(0.270) (0.292)
Log GDP per capita -0.705** -0.539
(0.331) (0.368)
Oil production 14.78** 12.72*
(6.51) (6.86)
Diamond production 1.55 1.80
(1.05) (4.44)
Ethnic fractionalization 4.84 6.25
(4.61) (8.35)
(Ethnic fractionalization)2 -4.96 -6.19
(4.69) (7.88)
Contraband -1.42** -1.37**
(0.633) (0.684)
Number of borders -0.235* -0.186
(0.122) (0.126)
Army size (log) -0.283 -0.319
(0.273) (0.324)
Deaths/year 2.08e-06 3.89e-06
(3.11e-06) (3.96e-06)
Ethnic war 0.516 0.598
(0.570) (0.674)
Sons of soil war -2.56** -2.45**
(1.20) (1.19)
Military intervention -0.293
(0.588)
Non-institutional sanction 1.17*** 1.19***
(0.367) (0.385)
International institution sanction 0.654 0.582
(0.761) (0.935)
Int. institution sanction
(no member)
Int. institution sanction (membet)
Model 1
Duration splines yes
Observations 612
Log-pseudolikelihood -159.91
Pseudo R-squared 0.1692
1.27 5.25 7.19 6.00
(2.13) (4.43) (6.05) (6.32)
0.004 -0.044 -0.046 -0.045
(0.010) (0.038) (0.046) (0.048)
0.023* -0.030 -0.047** -0.056*
(0.012) (0.022) (0.019) (0.029)
-0.074 -0.816 -1.02 -1.17
(0.289) (0.504) (0.749) (0.827)
-0.542 -0.222 -0.060 0.331
(0.369) (0.559) (0.686) (0.696)
12.79* -25.45 -34.33 -80.32
(6.85) (23.13) (34.49) (54.75)
1.89 3.16* 2.67 10.93
(4.59) (1.66) (5.29) (7.17)
6.25 6.01 2.57 3.51
(8.31) (7.30) (11.49) (14.31)
-6.20 -10.16 -7.86 -9.96
(7.84) (8.29) (13.37) (16.75)
-1.38** -0.954 -0.701 -1.01
(0.689) (0.841) (0.904) (0.978)
-0.187 0.382* 0.489* 0.708**
(0.128) (0.232) (0.294) (0.351)
-0.322 -0.744** -0.836* -1.20*
(0.322) (0.380) (0.440) (0.687)
3.88e-06 -6.13e-06 -2.36e-06 -8.90e-06
(3.96e-06) (0.00002) (0.00003) (0.00004)
0.601 0.784 0.366 0.197
(0.672) (0.893) (1.18) (1.34)
-2.45** -1.97 -1.93 -2.49
(1.17) (2.02) (2.51) (3.11)
-0.292 -1.63* -1.67*
(0.587) (0.865) (0.896)
1.19*** 0.374 0.566 0.645
(0.381) (0.600) (0.557) (0.451)
1.84** 2.32*
(0.922) (1.23)
0.553 0.116
(1.49) (0.883)
0.561 3.45**
(0.981) (1.59)
2 3
yes yes
589 589
-142.65 -141.64
0.1873 0.1931
Robust standard errors in parentheses.
**>< .01; **/>< .05; >< .10.
Cooperation and commercial interdependence are likely to
found among the members of some multilateral organizations
To test this straightforward proposition, I have recoded th
variable used in Model 1 by dividing the sanctions impose
by an international institution into two extra dumm
variables: the first is coded 1 if t
by an international institution b
tution, and the second is coded
tioned by an international instit
coefficients of the newly created
This content downloaded from
134.117.10.200 on Tue, 23 Feb 2021 10:39:01 UTC
All use subject to https://about.jstor.org/terms
Escriba-Folch_139
expectations. Note first that the strong positive impact of
institutional' sanctions on military victory is not altere
contrast, for those countries belonging to the institut
imposing the sanctions, the impact of such coercive mea
is great and significant. The estimated probability of a ne
ated settlement is 0.0008 when no sanction is present; i
country is targeted by sanctions of an international instit
it belongs to, the probability of a negotiated terminati
0.024. On the other hand, a country under
institutional sanctions faces a probability of 0.074 of ha
the conflict resolved through military means, while if no
tion is imposed, the likelihood is just 0.023.
Table IV shows the results of the effects of the types of
sure sanctions involve on war outcome. These results sh
be carefully interpreted as they are quite sensitive to sa
size, which in this case is restricted to the TIES data (f
1971). To increase the number of observations, the vari
'log GDP per capita' has been ruled out, since it was not
nificant anyway, and we use the natural resources dumm
The evidence suggests that the only measures effectiv
shortening conflicts are total embargoes, which increas
likelihood of both a military victory and a settlement. In
trast, international arms embargoes have a negative and si
icant impact on the probability of a military victory. T
result gives credit to those arguing that this type of san
has been irrelevant at best, and even counterprodu
(Tierney, 2005). Multilateral arms embargoes may
balance the power between parties, thereby making victory
likely. On the other hand, multilateral arms embargoes
had a positive, albeit not significant, effect on the likel
of a settlement.17 This latter finding supports to some
the proposition that arms embargoes decrease uncertainty
each party's power.
Other results of these tables merit comment: military in
ventions significantly reduce the likelihood of a settlem
The inclusion of this variable reduces the sample some
and slightly alters some of the results. Again, contrab
reduces the prospects of conflict termination, especially t
of a military victory, as it allows the rebels to finance
activities. The number of borders has contradictory eff
On the one hand, it serves to hinder a military victory
ably, by government forces), since it eases the maintenanc
rebel bases outside the boundaries of the state; but, on
other hand, it increases the likelihood of a negotiated r
tion, as in DeRouen & Sobek (2004). The size of the a
also represents a significant obstacle to a negotiated resolu
of conflicts, as it may tend to make the government ove
mate its relative strength, pe, and the probability of wi
(Mason & Fett, 1996). Natural resources, both diamond
oil, are again related to shorter wars, although through di
ent mechanisms. Abundant forest areas increase the likel
of a military victory and hinder negotiated settlements.
17 The coefficient is almost significant at the 0.10 level; the Rvalue is
Table IV. Types of sanction and
logit)
Event: Civil war outcome = j
Military Negotiated
Independent variables victory settlement
Intercept -0.167 9.09
(0.278) (6.15)
Mountains -0.012 -0.041
(0.012) (0.043)
Forests 0.003 -0.081*
(0.015) (0.043)
Log population -0.481* -1.12**
(0.261) (0.568)
Mineral exporting 4 71*** 6.45**
(1.81) (2.60)
Oil exporting 0.257 0.441
(0.566) (1.57)
Ethnic fractionalization 14.68 0.739
(9.53) (14.06)
(Ethnic -15.84* -5.58
fractionalization)2
(9.27) (15.22)
Contraband -1.53** -3.08
(0.748) (2.14)
Number of borders 0.141 0.665**
(0.114) (0.302)
Army size (log) -0.663* -1.56***
(0.372) (0.498)
Deaths/year 0.00001** 1.15e-06
(5.64e-06) (0.00004)
Ethnic war 1.16 -0.373
(0.761) (1.25)
Sons of soil war -2.36*** -2.56
(0.666) (2.34)
Military intervention -0.656 -1.34***
(0.629) (0.490)
Total embargo 2.04*** 6.03***
(0.634) (1.98)
Aid termination 0.781 0.692
(0.479) (1.42)
Trade restrictions 0.508 1.60
(0.671) (2.08)
Multilateral arms -1.48** 1.71
embargo
(0.673) (1.05)
Other sanctions 0.538 1.02
(1.10) (1.45)
Duration splines Yes
Observations 592
Log-pseudolikelihood -141.65
Pseudo R-squared 0.2152
Robust standard errors in parentheses.
***/>< .01;**/>< .05; >< .10.
Concluding remarks
Studies of the efficacy of economic sanctions have t
analyze it in a rather general way and have used con
variables that often assess the success of a sanction
This content downloaded from
134.117.10.200 on Tue, 23 Feb 2021 10:39:01 UTC
All use subject to https://about.jstor.org/terms
140 journal a/Peace Research 47(2)
on an ordinal scale. In this work, I sought to explore the effect
of sanctions in a very specific context, civil war duration, using
a time-series cross-sectional dataset and a direct measure of
event occurrence, civil war termination and outcome.
I hypothesized that sanctions would have a significant neg
ative effect on intrastate conflict duration, mainly due to three
basic mechanisms: convergence of parties' beliefs over power,
reduced utility of victory and financial pressure that reduce
parties' viability of continued fighting. Our second proposi
tion contended that those sanctions maximizing the costs to
the target would be more effective than other types of targeted
measures in bringing war to an end. Finally, I further proposed
that those sanctions imposed by an international institution
would be more effective in augmenting the probability of
intrastate conflict termination.
Our empirical evidence using new data on sanctions shows
that, effectively, sanctions have had a significant and negative
impact on intrastate conflict duration. Moreover, this effect
grows the more years a given country is targeted. These results
are robust to the inclusion of a variable controlling for external
military interventions. Moreover, the empirical evidence
seems to suggest that the most successful measures so far have
been total economic embargoes. Such measures are shown to
increase the likelihood of both military victory and negotiated
settlement, while international arms embargoes are found only
to significantly decrease the likelihood of a military victory by
one of the parties.
Regarding the debate about the efficacy of sanctions
imposed by international institutions or not, this article has
shed light on the distinctive effect of both types of sanction.
Although the coefficients for sanctions backed by international
institutions and those which are not are extremely similar in
size in the general models of war duration, I find that sanctions
imposed by international institutions significantly increase the
likelihood of conflict resolution, especially if the targeted
country is a member of the organization. In contrast, sanctions
applied by other bodies are much more conducive to military
victories.
In sum, the article has several policy implications. First, it
suggests that, overall, sanctions have been relatively useful in
helping to shorten civil wars, as the statistical association sug
gests. Yet, this role needs to be improved. Concretely, our
results suggest that any coercive measures should preferably
be conducted by international organizations, especially if we
are interested in promoting conflict resolution. Concerning
the measures imposed, the evidence suggests that maximizing
costs via embargoes is more effective than other sanction types.
Multilateral arms embargoes can also result in increased
chances of negotiated settlement, although implementation
problems so far have limited their effectiveness, as remarked
by many scholars.
Replication data
The data used in this article can be found at http://
www. prio. no/j pr/datasets.
Acknowledgements
I thank Clifton Morgan for sharing the data on sanctions and
the editor, T?nia Verge and three anonymous referees for very
helpful comments. I am also grateful to the Institut Barcelona
d'Estudis Internationals, where this research was started, and
all the people there.
References
Ang, Adrian U-Jin & Dursun Peksen (2007) When do economic
sanctions work? Asymmetric perceptions, issue salience, and out
comes. Political Research Quarterly 60(1): 135-145.
Balch-Lindsay, Dylan & Andrew J Enterline (2000) Killing time:
The world politics of civil war duration, 1820-1992. International
Studies Quarterly 44(4): 615-642.
Balch-Lindsay, Dylan; Andrew J Enterline & Kyle A Joyce (2008)
Third-party intervention and the civil war process. Journal of Peace
Research 45(3): 345-363.
Bapat, Navin A & Clifton Morgan (2007) Multilateral versus uni
lateral sanctions reconsidered: A test using new data. Manuscript,
University of North Carolina.
Beck, Nathaniel; Jonathan Katz & Richard Tucker (1998) Taking
time seriously: Time-series-cross-section analysis with a binary
dependent variable. American Journal of Political Science 42(4):
1260-1288.
Buhaug, Harvard; Scott Gates & P?ivi Lujala (2002) Lootable natural
resources and the duration of armed civil conflict, 1946-2001.
Paper presented at the Annual Meeting of the Peace Science Soci
ety, Tucson, AZ, 1-3 November.
Buhaug, Harvard; Scott Gates & P?ivi Lujala (2005) Geography, stra
tegic ambition, and the duration of civil war. Paper presented at
the Annual Meeting of the American Political Science Associa
tion, Washington, DC, 1-4 September.
Collier, Paul (2000) Rebellion as a quasi-criminal activity. Journal of
Conflict Resolution 44(6): 839-853.
Collier, Paul & Anke Hoeffler (2004) Greed and grievance in civil
war. Oxford Economic Papers 56(4): 563-595.
Collier, Paul; Anke Hoeffler & Mans S?derbom (2004) On the dura
tion of civil war. Journal of Peace Research 41(3): 253?273.
Cortright, David & George A Lopez (2002) Smart Sanctions: Target
ing Economic Statecraft. New York: Rowman & Littlefield.
Cunningham, David E (2006) Veto players and civil war duration.
American Journal of Political Science 50(4): 875?892.
Cunningham, David; Kristian S Gleditsch & Idean Salehyan (2005)
Dyadic interactions and civil war duration. Paper prepared for the
46th Annual Convention of the International Studies Association,
Honolulu, HI, 1-5 March.
Dashti-Gibson, Jaleh; Patricia Davis & Benjamin Radcliff (1997) On
the determinants of the success of economic sanctions: An empiri
cal analysis. American Journal of Political Science 41(2): 608-618.
Davis, Lance & Stanley Engerman (2003) History lessons: Sanctions:
Neither war nor peace. Journal of Economic Perspectives 17(2):
187-197.
DeRouen Jr., Karl R & David Sobek (2004) The dynamics of civil
war duration and outcome. Journal of Peace Research 41(3):
303-320.
This content downloaded from
134.117.10.200 on Tue, 23 Feb 2021 10:39:01 UTC
All use subject to https://about.jstor.org/terms
Escriba-Folch 141
Drezner, Daniel (1998) Conflict expectations
nomic coercion. International Studies Quar
Drezner, Daniel (1999) The Sanctions Paradox
International Relations. Cambridge: Cambrid
Drezner, Daniel (2000) Bargaining, enforc
sanctions: When is cooperation counterpro
Organization 54(1): 73-102.
Drezner, Daniel (2003) The hidden hand
International Organization 57(3): 643-659
Drury, A Cooper (1998) Revisiting economi
Journal of Peace Research 35(4): 497-509.
Elbadawi, Ibrahim A & Nicholas Sambanis
tions and the duration of civil war. World
Working Paper 2433.
Fearon, James D (2003) Ethnic and cultural
Journal of Economic Growth 8(2): 195-222
Fearon, James D (2004) Why do some civil w
than others? Journal of Peace Research 41
Fearon, James D (2005) Primary commodit
Journal of Conflict Resolution 49(4): 483-5
Fearon, James D & David D Laitin (2003)
and civil war. American Political Science R
Filson, Darren & Suzanne Werner (2002) A
and peace: Anticipating the onset, duration
American Journal of Political Science 46(4
Fruchart, Damien; Paul Holtom, Siemon W
dow & Peter Wallensteen (2007) United Nat
Their Impact on Arms Flows and Targ
Uppsala Universitet.
Gandhi, Jennifer & Adam Przeworski (20
tion, and rebellion under dictatotships.
18(1): 1-26.
Gershenson, Dmitriy (2002) Sanctions and c
69(274): 185-206.
Hirshleifer, Jack (2001) The Dark Side of th
tions of Conflict Theory. Cambridge: Cambr
Hufbauer, Gary C; Jeffrey J Schott & Ki
Economic Sanctions Reconsidered: Histor
Washington, DC: Peterson Institute for In
Hufbauer, Gary C; Jeffrey J Schott, Kimb
Oegg (2007) Economic Sanctions Recons
DC: Peterson Institute for International E
Humphreys, Macartan (2005) Natural reso
flict resolution. Journal of Conflict Resolu
Kaempfer, William & Anton D Lowenbe
international economic sanctions: A pub
American Economic Review 78(4): 786-79
Kaempfer, William & Anton D Lowenbe
sus multilateral international sanctions: A
tive. International Studies Quarterly 43(1)
Lacy, Dean & Emerson S Niou (2004) A th
tions and issue linkage: The roles of prefere
threats. Journal of Politics 66(1): 25-42.
Lektzian, David & Mar
sanctions onset and su
848-871.
Marinov, Nikolay (2005) Do economic sanctions destabilize country
leaders? American Journal of Political Science 49(3): 564-576.
Mason, David & Patrick J Fett (1996) How civil wars end:
A rational choice approach. Journal of Conflict Resolution 40(4):
546-568.
Mason, David; Joseph Weingarten & Patrick J Fett (1999) Win,
lose, or draw: Predicting the outcome of civil wars. Political
Research Quarterly 52(2): 239-268.
Montalvo, Jose G & Marta Reynal-Querol (2007) Ethnic polariza
tion and the duration of civil wars. World Bank Policy Research
Working Paper 4192.
Morgan, Clifton; Valentin Krustev & Navin A Bapat (2006)
Threat and Imposition of Sanctions (TIES) Data user's manual. Case
level data. University of North Carolina (http://www.unc.edu/
~ bapat/TIES.htm).
Nooruddin, Irfan (2002) Modeling selection bias in studies of sanc
tions efficacy. International Interactions 28(1): 59-75.
N?rnberger, Ralph (2003) Why sanctions (almost) never work. The
International Economy 17(4): 71-72.
Pape, Robert A (1997) Why economic sanctions do not work. Inter
national Security 22(2): 90-136.
Pape, Robert A (1998) Why economic sanctions still do not work.
International Security 23(1): 66-77.
Regan, Patrick (2000) Users' manual for Pat Regan's data on inter
ventions in civil conflicts. Binghamton, NY: Binghamton
University.
Regan, Patrick (2002) Third-party interventions and the duration
of intrastate conflicts. Journal of Conflict Resolution 46(1): 55-73.
Ross, Michael L (2004) What do we know about natural resources
and civil war? Journal of Peace Research 41(3): 337-356.
Strandow, Daniel (2006) Sanctions and civil war: Targeted measures
for conflict resolution. Uppsala: Department of Peace and Con
flict Research, Uppsala University.
Thyne, Clayton L (2006) Cheap signals with costly consequences:
The effect of interstate relations on civil war. Journal of Conflict
Resolution 50(6): 937-961.
Tierney, Dominic (2005) Irrelevant or malevolent? UN arms embar
goes in civil wars. Review oj International Studies 31 (4): 645-664.
Tostensen, Arne & Beate Bull (2002) Are smart sanctions feasible?
World Politics 54(3): 373-403.
Wallensteen, Peter; Mikael Eriksson & Daniel Strandow (2006)
Sanctions for conflict prevention and peace building. Uppsala:
Department of Peace and Conflict Research, Uppsala University.
Walter, Barbara F (1997) The critical barrier to civil war settlement.
International Organization 51(3): 335-364.
Walter, Barbara F (2002) Committing to Peace. Princeton, NJ: Prince
ton University Press.
ABEL ESCRIB?-FOLCH, b. 1978, PhD in Political Science
(Universidad Autonoma de Madrid, 2007, and Juan March Insti
tute); Visiting Professor, Universit?t Pompeu Fabra. Research
interests: economic sanctions and authoritarian politics.
This content downloaded from
134.117.10.200 on Tue, 23 Feb 2021 10:39:01 UTC
All use subject to https://about.jstor.org/terms