the duration and termination of civil war

Upload: alvaro-andres-alfero-maspes

Post on 04-Feb-2018

215 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/21/2019 The Duration and Termination of Civil War

    1/11

    http://jpr.sagepub.com/Journal of Peace Research

    http://jpr.sagepub.com/content/41/3/243Theonline version of this article can be found at:

    DOI: 10.1177/0022343304043768

    2004 41: 243Journal of Peace ResearchHvard Hegre

    The Duration and Termination of Civil War

    Published by:

    http://www.sagepublications.com

    On behalf of:

    Peace Research Institute Oslo

    Journal of Peace Research Replication Data

    can be found at:Journal of Peace ResearchAdditional services and information for

    http://jpr.sagepub.com/cgi/alertsEmail Alerts:

    http://jpr.sagepub.com/subscriptionsSubscriptions:

    http://www.sagepub.com/journalsReprints.navReprints:

    http://www.sagepub.com/journalsPermissions.navPermissions:

    http://jpr.sagepub.com/content/41/3/243.refs.htmlCitations:

    What is This?

    - May 1, 2004Version of Record>>

    at UNIV OF NOTRE DAME on August 19, 2013jpr.sagepub.comDownloaded from

    http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/content/41/3/243http://jpr.sagepub.com/content/41/3/243http://jpr.sagepub.com/content/41/3/243http://www.sagepublications.com/http://www.sagepublications.com/http://www.prio.no/http://www.prio.no/Data/http://jpr.sagepub.com/cgi/alertshttp://jpr.sagepub.com/cgi/alertshttp://jpr.sagepub.com/subscriptionshttp://jpr.sagepub.com/subscriptionshttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsPermissions.navhttp://www.sagepub.com/journalsPermissions.navhttp://jpr.sagepub.com/content/41/3/243.refs.htmlhttp://jpr.sagepub.com/content/41/3/243.refs.htmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://jpr.sagepub.com/content/41/3/243.full.pdfhttp://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/http://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://jpr.sagepub.com/content/41/3/243.full.pdfhttp://jpr.sagepub.com/content/41/3/243.refs.htmlhttp://www.sagepub.com/journalsPermissions.navhttp://www.sagepub.com/journalsReprints.navhttp://jpr.sagepub.com/subscriptionshttp://jpr.sagepub.com/cgi/alertshttp://www.prio.no/Data/http://www.prio.no/http://www.sagepublications.com/http://jpr.sagepub.com/content/41/3/243http://jpr.sagepub.com/
  • 7/21/2019 The Duration and Termination of Civil War

    2/11

    243

    Why Study the Duration of CivilWar?

    Figure 1 shows how the global incidence ofcivil war has changed over the last half-century.1 The incidence is the share of theworlds countries where at least one war isactive during a year. In the absence of goodmeasures of the number of people killed inwar every year and the amount of physicaldestruction, the incidence is the best measure

    to gauge the scope of civil war as a globalproblem. The figure shows that the incidencewas increasing dramatically up to 1990. Atthe peak in 1990, the incidence was almostfour times higher than in the 1950s. Sincethen, the incidence declined and then stabil-ized around 12% from 1995.

    The proportion of countries that started anew war (or took part in one that re-eruptedafter at least a years peace) is representedwith a dark shade in Figure 1. What is

    striking about the figure is that the change inincidence is not due to any change over timein the frequency of war onsets. The fractionof the worlds countries that experienced newconflicts has been stable at between 1% and2% during the entire period. The increase upto 1990 was not caused by an explosion inthe number of new conflicts, nor was thesubsequent decline due to a dearth of newconflicts. This confirms Fearon & Laitins

    (2003: 78) observation that the increasemust be due to a steady accumulation of con-flicts that start at a higher rate than they end

    2004 Journal of Peace Research,vol. 41, no. 3, 2004, pp. 243252Sage Publications (London, Thousand Oaks, CAand New Delhi) www.sagepublications.comDOI 10.1177/0022343304043768 ISSN 0022-3433

    The Duration and Termination of Civil War*

    HVARD HE GRE

    Department of Political Science, University of Oslo & Centre for the Study ofCivil War, PRIO

    An important key to reducing the suffering due to civil war is to shorten conflicts. The marked decrease

    in the incidence of conflicts in the 1990s was mostly due to a high number of conflict terminations,not to a decrease in the number of new wars. The articles in this special issue treat theoretically andempirically the determinants of civil war onset, duration, and termination, with particular emphasis onduration and termination. This introduction gives an overview of the articles in the special issue anddiscusses a few central topics covered by the different contributions: rebel group motivations, theimportance of financing, military factors, misperception, and commitment problems. Finally, the articlesums up some policy recommendations that may be derived from the articles in the issue.

    * This introductory essay and several of the articlesincluded in the special issue are outputs from the Econ-omics of Civil War, Crime, and Violence project in the

    World Bank Development Research Group. The projectwas initiated and directed by Paul Collier, and receivedfunding from the Norwegian, Swiss, and Greek govern-ments, the World Bank Post-Conflict Fund, and the WorldBank Research Committee. I am grateful for comments onthe introduction by Jim Fearon, Heather Congdon Fors,Nils Petter Gleditsch, Anke Hoeffler, and Michael Ross.1 The figure is based on the armed conflict data in Gled-itsch et al. (2002) and Eriksson, Wallensteen & Sollenberg(2003). The figure includes all conflicts that have led to at

    least 1,000 deaths over the course of the conflict. Thegraphs were smoothed with a five-year moving average. SeeCollier et al. (2003) for regional breakdowns of the inci-dence of civil war.

    at UNIV OF NOTRE DAME on August 19, 2013jpr.sagepub.comDownloaded from

    http://www.sagepublications.com/http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/http://www.sagepublications.com/
  • 7/21/2019 The Duration and Termination of Civil War

    3/11

    (also see Collier et al., 2003: ch. 4) and anincrease in average duration.2 Moreover, thedecline in the early 1990s is to a large extentdue to changes in the duration of civil wars.

    This underscores the importance ofresearch on civil war duration to guide thepolicies of the international community.Policies targeted at shortening conflict andreducing the risk of recurrence are a neces-sary supplement to policies aimed at

    reducing the risk of war initiation. First, atthe global level, the impact of effectiveconflict prevention will be felt only aftermany years, when older conflicts finally dieout. Second, over time there is more vari-ation in war duration than in the rate of waronset, possibly implying that means to alterthe prospects for war termination are more

    readily available than measures for conflictprevention.

    The first papers from the Economics ofCivil War, Crime, and Violence projectinvestigated which factors are systematicallyassociated with a high risk of onset ofconflict.3 Some of the articles in this issueadd to this research, but the majority of thearticles seek to increase our understanding ofwhy civil wars endure and how they end.

    Topics in the Civil War Literature

    This introduction discusses some selectedtopics as they are covered by the articles inthe issue, with particular emphasis on issuesrelevant to explaining the duration of civil

    journal o fPEA CE RES EA R CH volume 41 / number 3 / may 2004244

    0.00

    0.02

    0.04

    0.06

    0.08

    0.10

    0.12

    0.14

    0.16

    0.18

    0.20

    1 95 0 1 95 5 1 96 0 1 96 5 1 97 0 1 97 5 1 98 0 1 98 5 1 99 0 1 99 5 2 00 0

    Proportion

    ofcountries

    in

    war

    Ongoing wars New wars

    Figure 1. The Global Incidence of Civil War, 19502002

    2 Fearon (2004: 276) presents a similar figure based on hiscompilation of civil wars. The figure shows the same

    pattern, and also shows that the average duration of warsin progress increased up to 1990. Collier, Hoeffler &Sderbom (2004: 260) include period dummies in theirmodel that confirm this impression.

    3 See Collier et al. (2003) for an overview of the project.Prominent publications from the project are Collier &Hoeffler (2002), Elbadawi & Sambanis (2002), Hegre et

    al. (2001), and the special issues edited by Collier &Sambanis (2002) and Hegre & Sandler (2002). For workon similar topics, see Esty et al. (1998), Fearon & Laitin(2003), Sandler (2000), and Murshed (2002).

    at UNIV OF NOTRE DAME on August 19, 2013jpr.sagepub.comDownloaded from

    http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/
  • 7/21/2019 The Duration and Termination of Civil War

    4/11

    war. The articles formulate various models ofhow civil wars erupt and evolve. These dis-cussions highlight the heterogeneity of civilwars, and the articles suggest different ways

    to categorize civil wars in order to explaintheir onset, duration, or risk of recurrence.Below, I review some of the recurrent topics.

    What motivates rebel groups is one of these.Related to this is the issue of rebel groupfinancing. Military factors are also discussed,as well as problems related to misperceptionsand credible commitments. In the lastsection, I discuss some possible policy impli-cations of the research.

    Rebel Group MotivationsCollier & Hoefflers work on why civil warstarts (1998, 2002; see also Collier, 2000)sparked a lively debate on whether rebelgroups most commonly are motivated bygrievance or greed whether the aim of therebels is to generate public goods such asequal political rights or redistribution, or toyield private gains for the rebels. They arguethat the latter more often is the case, partlybecause of the collective action problemsinvolved in bringing about public goods (seeOlson, 1965; Lichbach, 1995), and partlybecause their empirical analysis indicates thatcivil wars occur where the opportunities forarmed predation are present rather thanwhere injustice is particularly prevalent.

    Collier, Hoeffler & Sderbom (2004)

    spell out the implications of this argumentfor the duration of civil war. They restate thetwo conceptualizations as rebellion-as-investment and rebellion-as-business. If mostrebellions are best seen as investments wherethe payoff is attainment of some politicalgoal in the future, wars are expected to lastlonger the larger are the contested stakes, aslong as there is no funding or credit con-straint. If for instance the goal is release from

    repression, wars should be longer the moresevere is the prewar repression. Stakes willalso be high if control of the government

    gives access to large revenues from naturalresource extraction (see Olsson & Fors,2004). A different empirical implicationfollows from conceptualizing rebellion as

    business, where only the profitability of therebellion as it proceeds should affectduration not the gains that can be reapedonly upon victory. Collier, Hoeffler &Sderbom (2004) conclude from the empiri-cal analysis that the rebellion-as-investmentconceptualization sits less well with the datathan the rebellion-as-business one: neitherprewar repression nor high primary com-modity dependence is associated with longer

    wars. Declines in commodity prices that mayhurt the profitability of the rebellion, on theother hand, are found to increase the chancesof war termination.

    Collier, Hoeffler & Sderbom infer rebelmotivations from how observed duration isempirically related to characteristics of theconflict and the country in which it takesplace. Fearon (2004) instead chooses to codeall the wars in his sample into different types.These types are to some extent based on rebelgroup motivations: he finds that warsemerging from coups, revolutions, and anti-colonial struggles are relatively brief. On theother hand, sons of the soil conflicts landor natural resource conflicts between aperipheral ethnic minority and state-sup-ported migrants of a dominant ethnic group(Fearon, 2004: 277) and wars where the

    rebel group is funded by contraband such asdrugs or diamonds tend to be long.However, Fearon explains the variance incivil war duration less with reference to themotivations of the rebels than to the com-mitment problems involved in reaching anegotiated solution.

    The formal model developed by Olsson& Fors (2004) and their case study of theDemocratic Republic of Congo in the 1990s

    refer back to Collier & Hoefflers (2002) dis-tinction between greed and grievance. Thearticle develops a formal model of conflict

    Hvard Hegre INT R OD U CT I ON 245

    at UNIV OF NOTRE DAME on August 19, 2013jpr.sagepub.comDownloaded from

    http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/
  • 7/21/2019 The Duration and Termination of Civil War

    5/11

    which distinguishes between greed, concep-tualized as the desire to gain control ofnatural resource rents, and grievance,thought of as the rulers deliberate insti-

    tutional distinction between the formal andthe informal sector. Their terminology isslightly different from Collier & Hoefflers,since the grievance term partly models theopportunity costs for potential rebels. Rebelschoose to devote resources to predation if theexpected gain is large relative to what theyearn through production in the non-formalsector. On the basis of the model and the casestudy, the article argues that the onsetsof the

    two wars in Congo are best explained bygrievance and the effectiveness of defense.

    An abundance of appropriable resourcesincreases the incentives for predation, but italso increases the rulers ability to defend hisriches. Hence, the risk of war onset may notbe affected by the amount of resource rents.The vast amount of resource rent and thewidespread appropriation by both Mobutuand Kabila, on the other hand, provide abetter explanation of the intensityof the con-flicts. When war erupts, the amount of rentsto fight over affects the scale of the conflict.

    Walter (2004) also focuses on the oppor-tunity costs for the potential recruits to rebelorganizations. Civil wars are more likely torecur where living conditions are poor.Collier, Hoeffler & Sderbom (2004: 262)note that wars are longer where average

    income is low or inequality is high. This maybe because opportunity costs are low so thatrebel recruitment is inexpensive.

    Demands for secession or regionalautonomy form another class of motivationsfor war. Ross (2004: 342) notes that oil, inparticular, increases the risk of the outbreak ofseparatist conflicts. Fearons long-standingsons of the soil conflicts regularly involvesuch demands, in particular where the

    government has strong interests in controllingthe region. These interests make it difficult forgovernments to reach agreements, even when

    they lack the military means to settle theconflict. Walter (2003) notes that govern-ments are likely to fear demands from othergroups if they give in to one, and Sambanis

    (2000) and Walter (2004) show that warsending with partition are disproportionatelylikely to experiencing war recurrence.

    Different motivations are likely to lead todifferent outcomes for the wars. DeRouen &Sobek (2004) estimate a multinomial logitmodel of civil war outcomes, distinguishingbetween government victory, rebel victory,truce, treaty, or continued war. If wars areconsistent with the rebellion as business

    conceptualization, rebel victory may not bethe goal. Rebels might prefer a truce, wherethey are allowed de facto control of a region.DeRouen & Sobek support this idea byshowing that truces are most likely in coun-tries with high dependence on primary com-modities. They also find wars in ethnicallyheterogeneous countries to be unlikely tolead to rebel victory, either because of thecoordination problems for rebel groups thatcannot draw on one large ethnic group, orbecause the wars in ethnically heterogeneouscountries typically follow sons of the soildynamics.

    Rebel Group FinancingCentral to the arguments of Collier &Hoeffler (2002) and Fearon & Laitin (2003)is the question of opportunities available for

    rebel groups to organize an army, recruitsoldiers, and obtain the funds to coverrunning costs. Collier & Hoeffler focus inparticular on the importance of naturalresources as sources of rebel group finances.

    The article by Ross (2004) reviews theevidence for the relationship between naturalresources and civil war in a set of recentempirical studies. The article surveys bothcase studies and cross-national statistical

    analyses of civil war onset and duration. Heshows that there is an emerging consensusthat oil increases the risk of the outbreak of

    journal o fPEA CE RES EA R CH volume 41 / number 3 / may 2004246

    at UNIV OF NOTRE DAME on August 19, 2013jpr.sagepub.comDownloaded from

    http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/
  • 7/21/2019 The Duration and Termination of Civil War

    6/11

    conflicts, and that conflicts are longer wherethere exist natural resources that can financerebel group financing (e.g. diamonds ordrugs). The article also points out issues

    where empirical studies reach different con-clusions: for instance, whether all types ofprimary commodity extraction and exportsare related to conflict, or whether such com-modities are related to revolutionary warsand not to ethnic ones. Ross also surveys thedebate on the causal mechanisms that linkresources with conflict.

    Fearon (2004) shows that wars where therebels have access to contraband finance are

    long. Collier, Hoeffler & Sderbom (2004)present results indicating that increases in theprices of primary commodities tend toprolong war, and DeRouen & Sobek (2004)show that dependence on primary com-modities increases the chance that civil warswill end in truces rather than outrightvictories for either side.

    The last article in the issue studies theconflict trap. Walter (2004) seeks an expla-nation of why conflicts re-erupt in the incen-tives for individuals to join or rejoin a rebelgroup. This focus on the incentives for enlist-ment as they evolve in the postwar period isin contrast to earlier studies on why warrecurs. These often explain recurrence bylooking at the nature of the original war: whyit started, how it was fought, or how itended. The empirical analysis in the article

    indicates that some aspects of the previouswar are important, such as the duration ofthe war and whether it ended in partition.Equally important are the living conditionsafter the war: war is most likely to re-eruptin those post-conflict countries that have thepoorest living conditions, or have semi-democratic political systems.

    Misperceptions and Commitment

    ProblemsMisperceptions are an important cause ofwar onset (Blainey, 1973; Fearon, 1995;

    Gartzke, 1999), since wars are costly andmost conflicts could be solved by bargaining.The argument is also relevant for war termin-ation: the literature on war and mispercep-

    tions also shows that war provides a meansof revealing information that is not availablein the standard bargaining models (Wagner,2000: 472; see also Fearon, 1995; Wittman,1979; Goemans, 2000; Filson & Werner,2002). Collier, Hoeffler & Sderbom (2004)note the empirical implication of this: theprobability of reaching an agreement shouldincrease with time after the war started. Boththeir empirical analysis and that of Fearon

    (2004) and Regan (2002: 70) indicate thatthe longer a war has lasted, the better is thechance that it will end. Mason & Fett (1996)find long wars to be more likely to end withnegotiated settlements than short wars,strengthening the interpretation that the riseis due to revelation of information ratherthan military victory due to attrition of theweaker side. The effect is not very strong,however, and Fearon (2004: 291ff.) arguesthat commitment problems are an alterna-tive explanation to misperceptions of whyparties are unable to reach mutually pre-ferred agreements (see also Walter, 2002).

    Fearon develops a game model wheresecessionist war is modeled as a commitmentproblem. Even though peaceful solutionsthat are preferable to both the rebel groupand the government always exist in the

    model, it shows that there are conditions inwhich the parties will never reach such asolution. The problem is that the govern-ment cannot commit credibly to autonomydeals signed in periods when it is relativelyweak. When the government regainsstrength, nothing will stop it from overturn-ing the agreement. As mentioned above, thiscommitment problem is more severe whenthe governments stakes in the conflict are

    high.Walter (2004) finds that the risk of war

    recurrence is higher after long wars. This

    Hvard Hegre INT R OD U CT I ON 247

    at UNIV OF NOTRE DAME on August 19, 2013jpr.sagepub.comDownloaded from

    http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/
  • 7/21/2019 The Duration and Termination of Civil War

    7/11

    result seems more consistent with expla-nations based on commitment problemsthan those based on misperceptions.

    Military Factors and State CapacityFearons (2004: 276) point of departure isthat wars are short when conditions favor adecisive military victory. Moreover, bothFearons and Olsson & Forss models pointout the importance offluctuationsin relativepower or the effectiveness of defense. As wehave seen, gaining access to finance is crucialfor rebel group strength. But the balance ofpower obviously also depends on the

    government side. Herbst (2004) focuses onthe government side of conflicts in Africa.The behavior of the armed forces is crucialto understand the development and out-come of civil wars. Herbst studies this bydiscussing how the threats to governmentstypically develop over the course of aconflict: in the initial stage, the insurgencyis generally small and should be easy todefeat. As the insurgency grows, therequired organization, political support, andfinancing of the military campaign becomeever more challenging to the government.Hence, effective counter-insurgency strat-egies require swift mobilization. Herbstshows that governments very often are notable to do this, and argues this is due to poorintelligence capabilities, limited resources,and political systems that often are authori-

    tarian and non-responsive to local demands.Especially in large countries, this often leadsto long conflicts where neither the govern-ment nor the rebel group is able to win. Incontrast, African countries have been moresuccessful in mobilization against foreigncountries. DeRouen & Sobek (2004)confirm Herbsts description of Africangovernments when showing that wars in

    African countries are disproportionately

    long and much more likely to end with rebelgroup victories than wars on other conti-nents. They also find that the size of the

    army has little impact on the chance ofgovernment victory.

    External military intervention is anothermeans to alter the military balance. Olsson

    & Fors (2004) note the importance ofMobutus and Kabilas changing allies for thedevelopment of the conflict in Congo. Howintervention affects the duration of wardepends on which side the intervener linesup with, and on the initial distribution ofpower between the parties. An interventionthat enables one side to achieve outrightmilitary victory is likely to shorten wars. It isdifficult to test this with cross-national

    studies, however, since it is hard to measurerelative power and since military strength isonly one factor determining duration.Collier, Hoeffler & Sderbom (2004) findonly military interventions on the rebel sideto decrease the duration of war. Largely con-sistent with this, Balch-Lindsay & Enterline(2000) find that interventions on the side ofthe government increase duration. Regan(2002) and Balch-Lindsay & Enterline(2000) also look at the impact of inter-ventions that support both sides of theconflict. Such interventions are likely toincrease the chance of stalemated conflicts,and their results confirm this expectation.DeRouen & Sobek (2004) show that UNinterventions are not associated with shorterwars, but tend to be followed by negotiatedsettlements.

    Counter-insurgency measures do nothave to be military to be effective. Herbst(2004: 362) asks whether Mandelas efficienthandling of the conflict with Inkatha in1994 could serve as a template for other

    African governments.4 However, very fewAfrican governments can afford that type ofagreement. DeRouen & Sobek (2004) show

    journal o fPEA CE RES EA R CH volume 41 / number 3 / may 2004248

    4 The model in Olsson & Fors (2004: 332) also indicates

    how improvement of the rebels institutional environmentmay reduce conflict intensity. The effect of this improve-ment is dependent on the effectiveness of the ruling groupsdefense, however.

    at UNIV OF NOTRE DAME on August 19, 2013jpr.sagepub.comDownloaded from

    http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/
  • 7/21/2019 The Duration and Termination of Civil War

    8/11

    that an efficient bureaucracy increases thechance of reaching negotiated outcomes(although it does not reduce the duration ofcivil wars).

    Policy Implications

    How can the international community con-tribute to ending wars? As all the articlespoint out, civil wars are very heterogeneousin both why and how they start, how theydevelop, and how they end (see Licklider,1993: 303ff.). The policy recommendationsof the articles in the issue are consequently

    diverse.5The first set of policy recommendations is

    to promote equitable development. Severalof the articles note how poor access toalternative employment increases thechances that potential recruits will join andremain in a rebel organization. Increasing theopportunity costs for recruits is likely toreduce both the risk of onset and recurrenceand the duration and intensity of wars.Development assistance is often feasible onlyafter the war has ended, however.

    The empirical articles in this issue showthat the curbing of rebel group finances iseffective to shorten at least some types of con-flicts. One well-known initiative to achievethis is the Kimberley Process, which aims toshut diamonds from conflict zones out of themarket.6 Depending on how successful the

    Kimberley Process turns out to be, thisresearch suggests that correspondinginitiatives could be applied to other com-modities such as other gemstones and timber.However, this type of regulation will be effec-tive only when the goods are legally tradedand where the trading firms are concernedwith their reputation. Hence, rebel groupincome from drugs, extortion, and kidnap-ping have to be curbed by other means.

    Increasing rebel group costs by impeding thetrade in small arms could also be effective.

    Civil wars become long when no partieshave the ability to achieve a decisive victory.

    Shifting the military balance of a conflict istherefore a possible way to end wars. Herbstpoints out how inadequate military mobiliz-ation has served to lengthen many conflictsin Africa, and suggests that assistancetargeted at strengthening governmentsmilitary capabilities may serve both toshorten and deter destructive wars. Suchassistance is always controversial, however,especially in cases where the legitimacy of the

    government is very poor. As Herbst pointsout, governments in Africa have been moreadept at mobilizing against external thaninternal enemies, implying that pure militarypower alone is insufficient to maintaininternal order. DeRouen & Sobeks studysupports this conclusion. Moreover, theempirical studies reviewed here show thatexternal military interventions have at best amixed record of success.

    As Fearon and Herbst point out, it may bedifficult for a government to defeat a rebelgroup decisively even when it is poorlyfunded. Fearon argues that the parties insecessionist conflicts are not able to reach anagreement because of commitment problems:the government, which will retain its militarycapacity after a peace agreement, may have anincentive to renege when it has regained

    strength. This suggests a role for third partiesto serve as guarantors for such agreements.7

    Monitoring military aspects of agreementshas been an explicit objective in UN peace-keeping missions. Such missions are muchless controversial than other military inter-ventions and have had some success (seeDoyle & Sambanis, 2000). Monitoring legalaspects and contributing to fulfilling financialaspects of agreements are other possible ways

    Hvard Hegre INT R OD U CT I ON 249

    5 See Collier et al. (2003) for a more extended discussionof some of the policy recommendations mentioned here.6 See http://www.kimberleyprocess.com.

    7 For studies on third-party mediation in civil war, seePillar (1983) and Pearson (2001).

    at UNIV OF NOTRE DAME on August 19, 2013jpr.sagepub.comDownloaded from

    http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/
  • 7/21/2019 The Duration and Termination of Civil War

    9/11

    for the UN, international financial insti-tutions, and individual governments to helpparties solve the commitment problems.

    Walter shows that partition is a poor

    solution to secessionist conflicts, although itwould terminate the conflict and solve thecommitment problem. Partition dramati-cally increases the risk of recurrent war in thesame country, and may also encourage sepa-ratism in other countries.

    Conclusion

    This overview has surveyed some of the most

    important topics in the small but growingtheoretical and empirical literature on theduration and termination of civil war. Likecivil wars themselves, the articles are highlydiverse, and they formulate differentpositions in an ongoing debate. The articlesare not in strong disagreement, but the differ-ences in the approaches preclude drawingvery firm conclusions as to where a consen-sus is forming. All the empirical studies haveto confront the fact that theoretical conceptssuch as motivations or perceptions are notdirectly observable, and the studies are forcedto rely on proxies. This may explain why thearticles present fairly divergent results. As inthe study of civil war onset, many differencesin results are due to different civil war datasets(see Ross, 2004). The process of coding whencivil wars start and end is also more difficult

    than just coding that they occurred (seeFearon, 2004: 278279), introducinganother source of variation. There is alsomuch less of a consensus of what are theappropriate explanatory variables to includein studies of civil war duration than in studiesof onset. This is partly due to the hetero-geneity of civil wars and to the variety oftheoretical approaches this inevitably entails.8

    Still, there is convergence on a few issues.Many of the studies note the importance ofrebel group finances and bring evidence thatthe presence of various sources of finance

    prolongs war. This finding seems to be morerobust in studies of civil war duration thanthe claim that sources of finance increase therisk of civil war onset. Many of the studiesalso find that civil wars are longer or morelikely to re-erupt in low-income countries.This is either because of the lack of statecapacity to achieve military victory or toproduce credible negotiated outcomes, orbecause of the low opportunity costs for

    potential rebels. This finding closely reflects asimilar finding in the study of civil war onset.

    The discussion of topics and of possiblepolicy implications above provides only abrief sketch of where the research is headingand what conclusions may be drawn. Thesetopics and several other aspects of civil warare treated in much more detail in the indi-vidual articles in the issue.

    References

    Balch-Lindsay, Dylan & Andrew J. Enterline,2000. Killing Time: The World Politics ofCivil War Duration, 18201992, Inter-national Studies Quarterly44(4): 615642.

    Blainey, Geoffrey, 1973. The Causes of War. NewYork: Free Press.

    Collier, Paul, 2000. Doing Well out of War: An

    Economic Perspective, in Mats Berdal &David Malone, eds, Greed and Grievance:Economic Agendas in Civil Wars. Boulder, CO& London: Lynne Rienner (91111).

    Collier, Paul; Lani Elliot, Hvard Hegre, AnkeHoeffler, Marta Reynal-Querol & NicholasSambanis, 2003. Breaking the Conflict Trap:Civil War and Development Policy. World BankPolicy Research Report. Oxford: Oxford Uni-versity Press (http://econ.worldbank.org/prr/CivilWarPR).

    Collier, Paul & Anke Hoeffler, 1998. On theEconomic Causes of Civil War, OxfordEconomic Papers50(4): 563573.

    journal o fPEA CE RES EA R CH volume 41 / number 3 / may 2004250

    8 The heterogeneity of civil wars is arguably less of aproblem in studies of civil war onset, since it is not possibleto know what type of civil war a country will experiencebefore it starts.

    at UNIV OF NOTRE DAME on August 19, 2013jpr.sagepub.comDownloaded from

    http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/
  • 7/21/2019 The Duration and Termination of Civil War

    10/11

    Collier, Paul & Anke Hoeffler, 2002. Greed andGrievance in Civil Wars, Working PaperSeries 01, Centre for the Study of AfricanEconomies, Oxford University (http://www.csae.ox.ac.uk/workingpapers/pdfs/ 01text.pdf).

    Collier, Paul; Anke Hoeffler & Mns Sderbom,2004. On the Duration of Civil War,Journalof Peace Research 41(3): 253273.

    Collier, Paul & Nicholas Sambanis, eds, 2002.Special Issue on Understanding Civil War,Journal of Conflict Resolution 46(1).

    DeRouen, Karl R., Jr & David Sobek, 2004. TheDynamics of Civil War Duration andOutcome, Journal of Peace Research 41(3):303320.

    Doyle, Michael W. & Nicholas Sambanis, 2000.International Peacebuilding: A Theoreticaland Quantitative Analysis,American PoliticalScience Review94(4): 779802.

    Elbadawi, Ibrahim A. & Nicholas Sambanis,2002. How Much War Will We See? Explain-ing the Prevalence of Civil War, Journal ofConflict Resolution 46(3): 307334.

    Eriksson, Mikael; Peter Wallensteen & MargaretaSollenberg, 2003. Armed Conflict,19892002,Journal of Peace Research 40(5):615637.

    Esty, Daniel C.; Jack A. Goldstone, Ted R. Gurr,Barbara Harff, Marc Levy, Geoffrey D.Dabelko, Pamela T. Surko & Alan N. Unger,1998. State Failure Task Force Report: Phase IIFindings. McLean, VA: Science ApplicationsInternational.

    Fearon, James D., 1995. Rationalist Expla-nations for War, International Organization49(3): 379414.

    Fearon, James D., 2004. Why Do Some CivilWars Last So Much Longer Than Others?,Journal of Peace Research 41(3): 275301.

    Fearon, James D. & David D. Laitin, 2003. Eth-nicity, Insurgency, and Civil War, AmericanPolitical Science Review97(1): 7590.

    Filson, Darren & Suzanne Werner, 2002. A Bar-gaining Model of War and Peace: Anticipatingthe Onset, Duration, and Outcome of War,American Journal of Political Science 46(4):819838.

    Gartzke, Erik, 1999. War Is in the Error Term,International Organization 53(3): 567587.

    Gleditsch, Nils Petter; Peter Wallensteen, Mikael

    Eriksson, Margareta Sollenberg & HvardStrand, 2002. Armed Conflict 19462001: ANew Dataset,Journal of Peace Research 39(5):593607.

    Goemans, Henk E., 2000. War and Punishment:

    The Causes of War Termination and the FirstWorld War. Princeton, NJ: Princeton Uni-versity Press.

    Hegre, Hvard; Tanja Ellingsen, Scott Gates &Nils Petter Gleditsch, 2001. Toward a Demo-cratic Civil Peace? Democracy, PoliticalChange, and Civil War, 18161992,American Political Science Review 95(1):1733.

    Hegre, Hvard & Todd Sandler, eds, 2002.

    Special Issue on Economic Analysis of CivilWars, Defence and Peace Economics13(6).Herbst, Jeffrey, 2004. African Militaries and

    Rebellion: The Political Economy of Threatand Combat Effectiveness, Journal of PeaceResearch 41(3): 357369.

    Lichbach, Mark Irving, 1995. The RebelsDilemma. Ann Arbor, MI: University ofMichigan Press.

    Licklider, Roy, 1993. What Have We Learnedand Where Do We Go From Here?, in RoyLicklider, ed., Stopping the Killing: How CivilWars End. New York & London: New YorkUniversity Press (303322).

    Mason, T. David & Patrick J. Fett, 1996. HowCivil Wars End: A Rational Choice Approach,Journal of Conflict Resolution 40(4): 546568.

    Murshed, S. Mansoob, ed., 2002. Special Issue onConflict, Civil War and Underdevelopment,Journal of Peace Research 39(4).

    Olson, Mancur, 1965. The Logic of Collective

    Action: Public Goods and the Theory of Groups.Cambridge, MA & London: Cambridge Uni-versity Press.

    Olsson, Ola & Heather Congdon Fors, 2004.Congo: The Prize of Predation, Journal ofPeace Research 41(3): 321336.

    Pearson, Frederic S., ed., 2001. Special Issue onDimensions of Conflict Resolution in Eth-nopolitical Disputes,Journal of Peace Research38(3).

    Pillar, Paul R., 1983. Negotiating Peace: WarTermination as a Bargaining Process. Princeton,NJ: Princeton University Press.

    Regan, Patrick M., 2002. Third-Party

    Hvard Hegre INT R OD U CT I ON 251

    at UNIV OF NOTRE DAME on August 19, 2013jpr.sagepub.comDownloaded from

    http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/
  • 7/21/2019 The Duration and Termination of Civil War

    11/11

    Interventions and the Duration of IntrastateConflicts,Journal of Conflict Resolution 46(1):5573.

    Ross, Michael L., 2004. What Do We KnowAbout Natural Resources and Civil War?,

    Journal of Peace Research 41(3): 337356.Sambanis, Nicholas, 2000. Partition as a

    Solution to Ethnic War: An EmpiricalCritique of the Theoretical Literature, WorldPolitics52(4): 437483.

    Sandler, Todd, ed., 2000. Special Issue onEconomic Analysis of Conflict, Journal ofConflict Resolution 44(6).

    Wagner, R. Harrison, 2000. Bargaining andWar, American Journal of Political Science

    44(3): 469484.Walter, Barbara F., 2002. Committing to Peace:The Successful Settlement of Civil Wars. Prince-ton, NJ: Princeton University Press.

    Walter, Barbara F., 2003. Reputation and War:Explaining the Intractability of Territorial

    Conflicts, unpublished paper, University ofCalifornia, San Diego.

    Walter, Barbara F., 2004. Does Conflict BegetConflict? Explaining Recurrent Civil War,Journal of Peace Research 41(3): 371388.

    Wittman, Donald, 1979. How a War Ends: ARational Model Approach,Journal of ConflictResolution 23(4): 743763.

    HVARD HEGRE, b. 1964, cand.polit. inPolitical Science (University of Oslo, 1999).Economist, Development Research Group,the World Bank (200103); Research Fellow,Department of Political Science, University of

    Oslo, and Centre for the Study of Civil War,PRIO (2003 ); Associate Editor, Journal ofPeace Research. Recent publication: co-authorof World Bank Policy Research ReportBreaking the Conflict Trap. Civil War andDevelopment Policy.

    journal o fPEA CE RES EA R CH volume 41 / number 3 / may 2004252

    at UNIV OF NOTRE DAME on August 19, 2013jpr.sagepub.comDownloaded from

    http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/http://jpr.sagepub.com/