pmjop07

7/28/2019 pmjop07

1/19

The International Court of Justice and the WorldsThree Legal Systems

Emilia Justyna Powell Georgia Southern University

Sara McLaughlin Mitchell University of Iowa

This paper seeks to understand why some countries accept the jurisdiction of the International Court of Justice (ICJ)

more readily than others. The theory focuses on institutional differences between the worlds major legal systems: civillaw, common law, and Islamic law. Important characteristics of these legal systems (stare decisis, bona fides, pacta

sunt servanda) are integrated in an expressive theory of adjudication, which focuses on how adjudication enhancesinterstate cooperation by correlating strategies, constructing focal points, and signaling information. The theory

considers the ability of states to communicate with each other, using acceptance of ICJ jurisdiction as a form of cheaptalk. Empirical analyses show (1) civil law states are more likely to accept the jurisdiction of the ICJ than commonlaw or Islamic law states, (2) common law states place the greatest number of restrictions on their ICJ commitments,and (3) Islamic law states have the most durable commitments.

For nearly 90 years, a World Court (PermanentCourt of International Justice (PCIJ), 192045;International Court of Justice (ICJ), 1946-

present) has been accessible to all countries for thepeaceful settlement of disputes.1 However, initialhopes that states would view the Court as a legitimateand effective conflict manager have not been fully real-ized. Only one-third (63 of 192) of countries in the

world accept the compulsory jurisdiction of theCourt,2 and an overwhelming majority of these states(84%) place reservations on their optional clause dec-larations, which can limit the Courts adjudicationprerogatives. On the other hand, it is much morecommon for states to recognize the ICJs jurisdictionthrough compromissory clauses in bilateral or multi-lateral treaties. Close to 80% of countries in the worldare signatories to one or more treaties that recognize

the ICJs jurisdiction should a dispute arise in thecontext of the treaty.3

Pessimistic views of the ICJ emphasize its under-utilization, flawed internal organization, declininginfluence over time, and external factors, such aspower disparities, that undermine its authority(Elkind 1984; Eyffinger 1996; Goldsmith and Posner2005; Janis 1987; McWhinney 1991; Oduntan 1999;

Posner 2004; Scott and Carr 1987; Scott and Csajko1988). More optimistic views emphasize the ICJs rolein pushing parties towards conflict resolution evenif the disputants never go to court (Bilder 1998;McAdams 2005). The ability for both sides to sue incourt could produce more efficient bargaining out ofcourt, allowing the parties to reach agreements thatthey will comply with more often (Gamble and Fischer1976).

1In our discussion, we often make reference to the International Court of Justice (ICJ), although our theory and analyses focus morebroadly on a World Court, which includes both the PCIJ and ICJ.

2States recognize the compulsory jurisdiction of the ICJ, with or without reservations, through acceptance of the Optional Clause, Article36 (2) of the ICJ Statute, thereby acknowledging the adjudication powers of the ICJ in all legal disputes regarding the interpretation of atreaty, any question of international law, and interpretation of other international obligations (Bederman 2001, 243).

3This figure is based on our own collection of data on compromissory clauses, described in more detail in the research design section. Thebasic data on relevant treaties was obtained from the International Court of Justices website: http://www.icj-cij.org/icjwww/ibasicdocuments/ibasictext/ibasictreatiesandotherdocs.htm.

The Journal of Politics, Vol. 69, No. 2, May 2007, pp. 397415

2007 Southern Political Science Association ISSN 0022-3816

397
http://www.icj-cij.org/icjwwwhttp://www.icj-cij.org/icjwww

7/28/2019 pmjop07

2/19

This research project engages the debate betweenoptimists and pessimists by posing a number ofresearch questions. Why do some states accept the

jurisdiction of the ICJ, while others do not? If twostates accept the jurisdiction of the ICJ, how does thisinfluence their bargaining behavior? Conversely, how

do expectations about international bargaining influ-ence unilateral state decisions to accept or not acceptthe ICJs jurisdiction? To address these questions, wefocus on institutional differences between the worldsmajor legal systems: civil law, common law, andIslamic law.4 We link characteristics of these legalsystems to an expressive theory of adjudication, whichfocuses on how adjudication enhances interstatecooperation by correlating strategies, constructingfocal points, and signaling information (McAdams2005).

In addition to the important role played by the

adjudicator (ICJ), we also consider the ability of statesto communicate with each other, focusing on accep-tance of compulsory jurisdiction as a form of cheaptalk. We argue that civil law states can correlate theirbargaining strategies and generate clear focal pointsfor coordination through the ICJ more easily thancommon law or Islamic law states and hypothesizethat civil law states are most likely to accept compul-sory jurisdiction. Furthermore, among states recog-nizing the ICJs compulsory jurisdiction, we expectcommon law states to place the greatest number ofrestrictions on their ICJ commitments and Islamic law

states to have the most durable commitments. Thusunderstanding the institutional features of domesticlegal systems also gives us purchase for understandingthe design of states international commitments.

Our paper begins with a discussion of the majorinstitutional characteristics of civil, common, andIslamic legal systems, focusing on the use of prece-dents, good faith in contracting, and the conditionsunder which contracts must be fulfilled. This is fol-lowed by a comparison of the legal procedures utilizedin domestic legal systems to those employed by theICJ. We then develop our theoretical argument relat-

ing the institutional features of domestic legal systemsto states unilateral choices to accept or not accept the

jurisdiction of the ICJ, building upon the expressivetheory of adjudication. Finally, we present a series ofempirical analyses of states acceptance of compulsory

jurisdiction, as well as the design of commitments tothe Court. We find that civil law states are more likely

to accept compulsory jurisdiction than common lawor Islamic law states and that they place significantlyfewer reservations on their optional clause declara-tions. However, Islamic and common law states havemore durable commitments in comparison to civil lawstates, which stem from strong norms of contractual

obligation and more precise obligations. Our theoryhighlights the importance of legal systems in worldpolitics and sets the stage for exploring further therelationship between domestic legal institutions, com-mitments to international legal institutions, and inter-state bargaining processes.

Domestic Legal Systems:Civil, Common, and Islamic Law5

Differences among the worlds three legal systemsreflect the great variance in states historical and cul-tural experiences. The origins of the civil legal systemcan be traced to the Roman Empire. Roman law wasestablished by Roman jurists and spread throughoutEurope via the Empires influence (Glenn 2000, 119).It fell out of fashion after the fall of the RomanEmpire, but ius civile was rejuvenated and revised bylegal scholars in European universities in the eleventhto thirteenth centuries. Roman law evolved into aseries of civil codes, including the Civil Code of Napo-leon, the German Civil Code, and the Italian Civil

Code, which influenced not only legal systems withinEurope, but the legal structures of colonies as well(David and Brierley 1985).

On the territories of Great Britain, a distinctsystem of law known as common law developed. Thebirth of this legal tradition is interpreted by some legalscholars as simply the result of a historical accident,the military conquest of England by the Normans(Glenn 2000; Whincup 1992). The Norman invadersestablished the fundamental components of thecommon law system, most notably the absence of thewritten letter of law, and the system was upheld by

English kings resistant to the continental influences ofRoman law. The stare decisis doctrine became wellestablished in common law, where judges were boundprimarily by precedents established by previous judg-ments. Common law practices spread throughout theBritish Empire, influencing many states legal systemson multiple continents.

4We have examined other typologies of legal systems (La Portaet al. 1999), but have found this three-system typology to be bestsuited to our research project.

5The arguments in this section draw upon research by Powell(2006a).

398 emilia justyna powell and sara mclaughlin mitchell

7/28/2019 pmjop07

3/19

The worlds final major legal tradition, Islamiclaw, arose with the birth of Islam in the Arabian Pen-insula and Mesopotamia in the seventh century A.D.

(Mourisi Badr 1978, 187). As the Arab empire ex-panded, Islamic religious and legal traditions becamepredominant in many Central Asian and MiddleEastern states. The Islamic legal tradition is based pri-marily on religious principles of human conduct, andlaw is an integral part of the Islamic religion (Al-Azmeh 1988; Khadduri 1956; Lippman, McConville,and Yerushalmi 1988).

In Figure 1, we plot the percentage of states foreach legal system type (civil, common, Islamic, mixed)from 1920 to 2002. The civil law system is the modalcategory each year, although Islamic and common law

states have become more widespread in recent years.Next, we compare the procedures and legal principlesprevalent in these legal systems, focusing on threeprimary differences: (1) the use of precedents (staredecisis), (2) good faith in contracting (bona fides), and(3) conditions under which contracts must be fulfilled(pacta sunt servanda). Table 1 provides an overview ofthese institutional differences.

Stare Decisis

The use of precedents when making legal judgments is

prevalent in common law systems, but virtually absentin civil law or Islamic law systems (e.g., Opolot 1980).The doctrine of precedent, or stare decisis, states that,when trying a case, a judge is obliged to examine howprevious judges have dealt with similar cases (Darby-shire 2001). In the process of looking back, a judgediscovers principles of law relevant to a case underconsideration and renders judicial decisions consist-ent with existing principles in the law. Stated in ageneral form, stare decisis signifies that when a point oflaw has been previously settled by a judicial decision, it

forms a precedent, which is not to be departed fromafterward (Opolot 1980).6 The stare decisis doctrinedoes not exist in civil law systems based on Roman iuscivile, where law making is a function of the legisla-ture. A judges task is considered to be passive, toimplement legal rules contained mainly in codes, laws,

and statutes.

7

The stare decisis doctrine is also absent inIslamic law systems, where law is derived from fourprinciple sources: the Quran, the Sunna, judicial con-sensus, and analogical reasoning (Vago 2000).8

Bona Fides

Another major distinction between civil, common,and Islamic law systems stems from the principle, bona

fides, or good faith in contracting. In general, theconcept of good faith requires parties to a contract toabstain from dishonesty and to keep their promises.

The bona fides principle is comprised of three essential

6The main advantage of the doctrine of judicial precedent is that itleads to consistency in the application and creation of principles ineach branch of law. It also enables common law lawyers to forecastwith some degree of certainty what kind of judgment may beexpected in a particular case. Prior to 1966, the highest courts ofEngland and the United States of America took conflicting posi-tions on the question of what a judge should do if he or she isconfronted with an unreasonable or outdated precedent. TheHouse of Lords decided in 1898 that it was bound by its owndecisions. In the United States, on the other hand, the principle ofstare decisis has never been considered an absolute command, andthe duty to follow a precedent is held to be qualified by the right tooverrule prior decisions. The highest courts of the states, as well asthe Supreme Court, have the right to depart from a rule previouslyestablished by them. The British interpretation on precedentsmoved closer to the American view in 1966, when the PracticeStatement of the House of Lords established that previous deci-sions of the House are treated by it as normally binding, but thisis subject to a right to depart from a previous decision when itappears right to do so (Shahabuddeen 1996). Some legal scholarsargue that although formally present in the common law countriesand absent in civil law nation, stare decisis does not constitute themost important disparity between the two systems (Shapiro 1986).

7It is crucial, however, to note that common law countries may alsohave codified bodies of written legal rules, norms, and principles,such as the numerous codes of the State of California. Neverthe-less, as many legal scholars underscore, the codes in common law

nations often recapitulate principles established beforehand injudicial decisions (Merryman 1969).

8The Quran is the sacred book of the Muslims, and it literallymeans the Reading; the Sunna literally means the path taken ortrodden by the Prophet himself, and it contains explanations,deeds, sayings, and conduct of the Prophet (Glenn 2000). Judicialconsensus is constituted by a common religious conviction(Glenn 2000), and it is thus a consensus of traditional Islamicscholars regarding specific points in Islamic law. Analogical rea-soning, the fourth source of Islamic law, is used in circumstancesnot provided for in the Quran or other sources (Vago 2000).According to this method, the provisions of the Quran and Sunnamay be applied to a new problem if there exists a similar operativeor effective cause.

FIGURE 1 Legal System Frequencies, 19202002

1920

1925

1930

1935

1940

1945

1950

1955

1960

1965

1970

1975

1980

1985

1990

1995

2000

Year

0.00

0.20

0.40

0.60

0.80

% Common Law

% Civil Law

% Islamic Law

% Mixed Law

the international court of justice and the worlds three legal systems 399

7/28/2019 pmjop07

4/19

elements: honesty, fairness, and reason (Zimmermannand Whittaker 2000). Originating in Roman law, thebona fides principle is an essential feature of civil lawsystems. A doctrine of good faith establishes principlesfor trustworthy and honorable conduct, permitting

judges to denounce breaches of good faith, whiletaking into consideration the particularities of eachcase (OConnor 1991, 117). Civil law systems treat

bona fides as an overarching legal principle, includingit in general and specific legislation and civil codes.Good faith also constitutes one of the most importantprinciples in the Islamic legal tradition. Both theQuran and Sunna permit trade if it is carried outaccording to the principles of good faith and honesty(Rayner 1991).

On the other hand, common law systems do not,on average, recognize a general duty to negotiate norto perform contracts in good faith. In the UnitedKingdom, for example, the doctrine of good faith isoften perceived by lawyers as threatening and simply

unworkable in the British law system. Some scholarseven state that good faith could well work practicalmischief if ruthlessly implanted into our system oflaw (Bridge 1984, 426, quoted in Zimmermann andWhittaker 2000, 15). While some efforts have beenmade to introduce good faith into the common lawtradition and put limits on the absoluteness of con-tractual rights and obligations (e.g., the doctrine ofeconomic duress), the position of bona fides in thislegal family is much weaker than under civil andIslamic law (Zimmermann and Whittaker 2000).9

Pacta Sunt Servanda

When individuals sign contracts or states sign treaties,such pacts made in good faith are supposed to bebinding. Several scholars have examined whethercommon law nations exhibit higher levels of respectfor the rule of law and commitment to contracts com-pared to their civil law counterparts (Joireman 2001,

2004; Nassar 1995), while others have analyzedwhether contracts are most stable in Islamic legalsystems (Nemeth 2005; Rayner 1991). Common lawsystems recognize that events occurring after thesigning of a contract, or a force majeure, might makethe contract impossible or impracticable to fulfill,which can release all parties from their contractualobligations (Rayner 1991; Whincup 1992). Civil lawsystems view such events as creating only a partialrelease from a contract until the situation conducive tothe fulfillment of contract is restored. On the otherhand, civil law systems do not recognize the parol evi-

dence rule,10

which some have argued strengthens thesanctity of contracts in common law states relative tocivil law states (Nassar 1995).

In the Islamic legal tradition, the principle ofpactasunt servanda is paramount, because it is God Who isthe witness of all contracts (Rayner 1991, 100). Con-tractual obligations governed by Muslim law requireall parties to uphold their commitments: a nationalIslamic state has no vested right to cancel or alter acontract by unilateral action, whether such actiontakes the form of an administrative, judicial or evenlegislative act (Rayner 1991, 87). This obligation of

the faithful to respect their contractual obligations is9Some common law states have gradually introduced the principleof good faith into their legal systems. For example, the UnitedStates Uniform Commercial Code in section 1-304 states: Everycontract or duty within this Act imposes an obligation of goodfaith in its performance. Also, the U.S. Restatement (Second) ofContracts adopted by the American Law Institute in 1979 andpublished in final form in 1981 provides that individuals havenonwaivable duties of good faith. This act stands in a sharp con-trast with the first Restatement of Contracts (1932), which did notinclude a comparable good faith provision (Summers 1982). InGreat Britain, the European Consumer Protection Directive of1994 transplanted good faith directly into the body of Britishcontract law. Despite these developments, we agree with numerous

legal scholars that there are still fundamental differences in thestatus of bona fides in civil and common law traditions. As somescholars state, good faith is a legal irritant in the common lawtradition and the imperatives of a specific Anglo-American eco-nomic culture as against a specific Continental one will bringabout an even more fundamental reconstruction of good faithunder the new conditions (Teubner 1998, 12).

10The parol evidence rule assumes that a written contract embodiesall of the terms of the agreement, and thus that external evidence,such as verbal communication between the parties, could not alterthe parties obligations.

TABLE 1 Characteristics of Legal Systems

TypeThe Use of Precedents

(Stare Decisis)

Good Faith inContracting(Bona Fides)

Keeping Promises(Pacta Sunt Servanda)

Thoroughness ofContracts

Civil Law No Yes Medium Low

Common Law Yes No Medium HighIslamic Law No Yes High Medium


7/28/2019 pmjop07

5/19

binding not only in relation to other Muslims, but alsotowards nonbelievers. According to the Quran, eventhe state of war by itself does not constitute a sufficient

justification for contractual violation (Rayner 1991,87). However, a contract may, under Islamic law, beinvalidated temporarily by subsequent clauses such as

impossibility of performance (rebus sic stantibus) andforce majeure.

The Design of Contracts

We contend that institutional characteristics ofdomestic legal systems will influence both states will-ingness to make commitments in international poli-tics and the design of such commitments. Theweakness of the bona fides and pacta sunt servanda

principles in common law systems should producevery specific and detailed contracts. Unforeseen eventsmay render contracts null and void, thus common lawlawyers will be careful to draft contracts that specifyprecise contractual terms. In addition, because thereare, for the most part, no codes that would spell out allof the general principles applicable to a contract undercommon law, contracting parties must make sure thatall of the principles and rules that are to apply to theiragreement are explicitly addressed in their contract.Contracts in civil legal systems, backed firmly by prin-ciples of good faith will be more frequent although less

precise. Contracts do not spell out all the legal prin-ciples that are to apply to a contract because thewritten codes already enumerate these general over-arching principles. For example, in a civil law state, itwould be unnecessary for contracting parties toinclude good faith as one of the contractual stipula-tions because contractual relations in civil law systemsare automatically governed by this principle.

Islamic law, being rooted firmly in religious prin-ciples, limits parties contracting freedom, whichshould result in a smaller number of contracts.However, strong norms of pacta sunt servanda

produce expectations that contracts negotiated underIslamic law will be upheld, even as circumstanceschange. Islamic states should also be very careful insigning contracts on the international arena. Becausecontracts are sacred, Islamic states should make surethat all of their contractual obligations are clearlyspelled out.

To sum up, there are significant institutional dif-ferences between civil, common, and Islamic lawsystems that influence both the frequency and designof interstate contracts. Next, we describe similarities

and differences between the rules and procedures ofthe ICJ and those employed in civil, common, andIslamic law systems.

Practices and Procedures of the

International Court of JusticeThe creation of the Permanent Court of InternationalJustice was surrounded by significant disagreementabout the legal principles and rules to be utilized bythe newly established court.In addition to differencesgoverning the laws of naval warfare, there was alsobelieved to be a difference between the Anglo-Saxon(i.e., Anglo-American) approach to international lawand the continental (i.e., European) approach tointernational law (Lloyd 1985, 35). The crux of theproblem was a potential clash between opposing legal

orders with judges of divergent legal traditions servingon the Court. Such a concern was expressed by numer-ous English judges and politicians involved in the for-mation of the international court: It was inevitablethat the majority of judges on the Court wouldbe continental lawyers or would follow thatschool . . . By virtue of sitting at the Hague they wouldbe exposed to the pernicious influence of extremeGerman doctrines (Lloyd 1985, 35). The winninginfluence of the continental approach to internationallaw produced many similarities between the Romano-Germanic legal tradition embodied in civil law

systems and the rules and procedures adopted by thePCIJ (and later ICJ).

For example, the doctrine of stare decisis wouldnot be applied in international law. On the contrary,the ICJ in its decision making is bound by Article 59,which states: The decision of the Court has nobinding force except as between the parties and inrespect of that particular case. As most legal scholarsagree, the object of this article is simply to preventlegal principles accepted by the Court in a particularcase from being binding on other states or in otherdisputes (Brownlie 2003). The ICJ is, therefore, forbid-den from formally introducing jurisprudential con-tinuation by invoking its previous judgments.11

11Despite the inability of the ICJ to formally rely on its previousjudgments, the ICJ has retained a form of judicial consistency byinvoking its previous judgments and the judgments of its prede-cessor, the PCIJ, in both arguments and decisions (Jennings andWatts 1992, 41). Some authors, however, accuse the ICJ of defect-ing from this stabilizing practice (Reisman 1989). On a few occa-sions in recent years, the ICJ misstated its own prior holdings byselective quotation (ICJ Reports 1986).


7/28/2019 pmjop07

6/19

Nevertheless, judges often invoke previous decisionsof the Court, in order to support their decision in aparticular case. Invoked previous judgments do not,however, constitute a binding precedent, but aremerely treated as a statement of what the Courtregarded as the correct legal position (Shahabuddeen

1996, 63).

12

The lack of formal judicial precedent inthe activity of the ICJ makes it very similar to civil legalsystems where this doctrine is forbidden for the mostpart (Rosenne 1962).

As far as the remaining two legal principles (bonafides, pacta sunt servanda), international law mostclosely resembles the ius civile tradition. Bona fidesconstitutes one of the general principles of law and isconsidered to be one of the formal sources of interna-tional law (OConnor 1991). The ICJ has recognizedthe doctrine of good faith in several judgments,including the Norwegian Fisheries case (1951), the

North Sea Continental Shelf cases (1969), the NuclearTest cases (1973), and the Arbitral Award made by theKing of Spain on 23 December 1906(1960).13 Addition-ally, the principle of good faith is articulated in theCourts basic documents, including Article 38 of theICJ Statute as well as Article 2 (2) of the UnitedNations Charter. The principle of good faith is some-times viewed as an overarching principle, from whichthe pacta sunt servanda derives, and not surprisingly,

the ICJ also treats contractual compliance as animportant part of international and customary law(OConnor 1991).

Thus far we have described major institutionaldifferences between civil, common, and Islamic lawdomestic systems and argued that the practices of the

ICJ are very similar to those employed in civil lawsystems. What incentives do states have for acceptingthe compulsory jurisdiction of the ICJ and how is thisinfluenced by their domestic legal institutions? In thenext section, we build upon an existing theory of adju-dication and consider how the parties can utilizeoptional clause declarations as information abouttheir willingness to resolve disputes peacefully.

Interstate Bargaining and theInternational Court of Justice

We assume that interstate bargaining experiencesand future bargaining expectations influence stateschoices to accept or not accept the jurisdiction of theICJ. Countries are aware of the multiple contentiousissues that may arise in world politics, including dis-agreements over land borders, maritime zones, andtrade. State leaders have incentives to send strongsignals to other states about their resolve and strength,

yet they also may wish to signal a willingness tobargain peacefully and avoid military contests becausethey realize that conflict is costly (Fearon 1995).

However, signals about peaceful conflict managementare hard to convey, because they are often perceived ascheap talk (Crawford and Sobel 1982; Farrell 1987;Farrell and Gibbons 1989; Farrell and Rabin 1996;Kim 1996; Matthews 1989).14 It is interesting to con-sider what role the ICJ plays in this interstate bargain-ing process.15 Unbiased adjudicators may be effectiveat helping parties strike cooperative agreements bycorrelating strategies, creating focal points, and signal-ing information (Garrett and Weingast 1993; Gins-

12Professor Kisch best describes this feature of the ICJ in theGuardianship Convention case (ICJ Reports 1958, 55), where hestates: Of course I am well aware that the Court is not bound bythe stare decisis principle, British or American style.Of course I usethewordprecedentsin thegeneral and notin the strictly technicalsense, and what I wanted to convey was this: that any lawyer in anycountry, any judge, and any advocate, when confronted with adifficult case, tries to find, if not some support, at least someenlightenment and some inspiration, from what judges, and par-ticularly the best judges, have found in similar cases. I refer toprecedents in that wide sense, including even such situationsaswe have had in this very countrywhere a judge has explicitlydeclared himself inspired by French or English or German deci-sions. Certainly that is not a phenomenon of precedents in theAnglo-American sense (ICJ Pleadings, Application of the Con-

vention of 1902 Governing the Guardianship of Infants, p. 259;Shahabuddeen 1996, 237).

13In the Norwegian Fisheries case, the ICJ upheld Norways right toestablish maritime baselines, arguing that Britain could notcontest them after a long period of inaction (Kolb 2006, 21). In the

North Sea Continental Shelfcases, the ICJ recognized estoppel, animportant doctrine derived from good faith (OConnor 1991). Inthe Nuclear Test cases, the ICJ argued that [o]ne of the basicprinciples governing the creation and performance of legal obli-gations, whatever their source, is the principle of good faith (citedin Virally 1983, 130). In the Arbitral Award case, the ICJ arguedthat Nicaragua could not renege on the territorial award in thearbitration because it had agreed to carry out its terms in goodfaith (Kolb 2006, 2223).

14

Cheap talk can be defined as a statement that may convey infor-mation even though the statement is costless, nonbinding, andnonverifiable (Baird, Gertner, and Picker 1994, 303). One coulddebate whether optional clause declarations are cheap talk orcostly signals, especially if states face significance audience costs(domestic and/or international) for reneging on ICJ judgments.Given the anarchic nature of the international system and theability of states to restrict and/or withdraw optional clause decla-rations at any point in time, we think it is reasonable to view suchdeclarations as cheap talk.

15Like McAdams (2005), we treat interstate bargaining as a coor-dination or mixed-motive game. Future research will explore morecarefully how conclusions are altered if we select different baselinebargaining models (see for example, Goldsmith and Posner 2005).


7/28/2019 pmjop07

7/19

burg and McAdams 2004; McAdams 2005, 1049).Furthermore, states may be able to transmit informa-tion about themselves through the adjudicator by for-mally recognizing its adjudication powers. Thusinterstate bargaining may be influenced both by thepresence and behavior of an adjudicator and by the

parties ability to send information to each otherthrough the adjudicator.Beginning with the adjudicators own behavior,

he/she may convince parties to coordinate theirbehavior by focusing on some random event, or tocorrelate their equilibrium behavior. For example, acoin flip could be used to select between two equilib-rium outcomes, but this solution is problematic. First,if a randomized process for dispute resolution couldbe agreed upon by the disputants, then this wouldpreclude the need for a third-party mediator. Second,it would be tempting for the disputants to renege on

any agreement reached through a randomizeddecision-making mechanism (McAdams 2005, 1057).To overcome this problem, the adjudicator uses cheaptalk to construct a focal point in a coordinationgame(McAdams 2005, 1059), by focusing on particu-lar equilibrium outcomes and conveying this informa-tion to the disputants. Thus adjudicating institutionslike the ICJ and European Court of Justice help tocreate a shared belief system about cooperation anddefection in the context of differential and conflictingsets of individual beliefs that inhibit the decentralizedemergence of cooperation (Garrett and Weingast

1993, 184). The European Court of Justice, forexample, has been successful at creating a focal pointof mutual recognition, which has created a sharedframework about how the common market works(Garrett and Weingast 1993). Adjudicators may alsopromote cooperation by revealing private informationto disputants, such as the players types, which worksbest if the adjudicator is unbiased and has strong repu-tational incentives for being truthful (McAdams2005). All three mechanisms underlie the expressivepower of adjudication and help to explain the emer-gence of a single dominant adjudicator (PCIJ/ICJ), as

well as the parties willingness to comply with theCourts judgments.

McAdams (2005) expressive theory focuses on theadjudicators influence as a third-partyactor,16 butdoesnot consider the possibility that some states may havestronger incentives than others to resolve interstate

disputes with an adjudicators assistance. First, thetheory assumes that the adjudicator is unbiased. And

yet as we argued above, the procedures and rules of theICJare extremely similar to those used in domestic civillaw systems, which creates a bias in favor of civil lawstates.17 This institutional similarity between the ICJ

and civil law systems encourages civil law states tocorrelate their equilibrium behaviors naturally becausethe costs of coordination are reduced and because it iseasier for the parties and the adjudicator to agree onwhat each will regard as cooperative and defectivebehavior (McAdams 2005, 1081). Civil law statesaccept similar legal principles domestically, whichmakes it easier for them to correlate their behaviors,and the adjudicator (ICJ) and civil law disputants willconverge naturally on the same outcomes. Civil lawstates are also more likely to view ICJ judgments withlegitimacy due to their recognitionof the principles the

Court applies in reaching its decisions, which produceshigh compliance rates with ICJ rulings.

Second, McAdams theory considers the role ofcheap talk for creating focal points, but does notexamine how the similarities of the disputants pref-erences influence the effectiveness of cheap talk forpromoting coordination.18 Analyses of domesticcourts and their indirect role in resolving disputesprovide insight into the relationship between cheaptalk and bargaining. In the United States, many dis-putes are settled absent of a formal court decision(Bilder 1998). This is most likely to occur when the

dispute lies between parties who have, and expect tocontinue, long-term relationships with each other[for] such relations might be disrupted by resort to thecourts (Bilder 1998, 235). Most disputes never reachthe court, and most of those that do are settled prior toa final decision being made by the court.19

Bilders argument that out-of-court effects arestrongest for parties with similar interests and long-

16McAdams (2005, 1078) argues that a single adjudicator is impor-tant for getting parties to converge on a single outcome and that ex

post compliance with the endorsed outcome is enhanced when asingle actor or institution serves as the adjudicator.

17Posner and Figueiredo (2004) find strong evidence for bias in theICJ at the level of the individual justices, showing that justices tendto vote in favor of their own states and in favor of states that havesimilar wealth, regimes, and culture to their home states.

18Political scientists have examined a variety of sources of statesimilarities including institutions (e.g., regime type), economicinteractions (e.g., trade), and cultural ties (language, history, etc.).We focus on similarities between legal institutions because ourtheory focuses on adjudication, but our theory also explains ourfindings that democratic states are more willing to accept compul-sory jurisdiction. The similarity of preferences between democra-cies and the transparency of their regimes enhances the use ofoptional clause declarations as cheap talk.

19The pattern of judgments for ICJ cases seems to follow a similarpattern, with 61% of cases having no judgment, often due to thedisputing parties reaching a settlement before a judgment by theCourt is rendered (http://www.icj-cij.org/icjwww/idecisions.htm).

http://www.icj-cij.org/icjwww/idecisions.htmhttp://www.icj-cij.org/icjwww/idecisions.htm

7/28/2019 pmjop07

8/19

term relationships meshes well with the equilibriumfindings in cheap talk bargaining models. Crawfordand Sobels (1982) path-breaking model demonstratesthat cheap talk promotes cooperation more readily inbargaining settings if the parties have common inter-ests. Theoretical extensions of Crawford and Sobels

(1982) model confirm Bilders beliefs that the shadowof the future matters as well. Kim (1996), for example,shows that reputation effects in infinitely repeatedinteractions can enhance the credibility of cheap talkand produce more efficient agreements. Long-terminteractions mitigate incentives to lie about ones typebecause bargaining parties seek to avoid future lossesfrom damaged reputation (Sartori 2005).

In the process of bargaining, states can engage incheap talk about their willingness to work with thedesignated adjudicator (ICJ). We can thus extendMcAdams (2005) expressive theory by treating

optional clause declarations as a form of cheap talk.States would like to convince other states that theyprefer to settle interstate disputes peacefully, and rec-ognition of the ICJs jurisdiction sends informationabout a states willingness to view the adjudicator as alegitimate third-party conflict manager. The similari-ties between civil law states and the practices of the ICJproduce great benefits for civil law states use of juris-dictional cheap talk. Civil law systems are the mostfrequent domestic legal systems in the world. From1920 to 2002, civil law states constituted 4878% of allstates in the world (Figure 1).20 The predominance of

civil law states creates high probabilities that any twostates bargaining in international politics will bothhave civil law systems. If we drew two states randomlyfrom the international system in a given year, theprobability of selecting a pair of civil law states wouldat a minimum be .23 (for 48%) and at a maximum be.61 (for 78%). Any dyadic interaction for a civil lawstate has the highest chance of being with another civillaw state. Because civil law states dominate in theinternational arena, and because cheap talk works bestwhen sent to similar states, civil law countries stand tobenefit most from acceptance of the ICJs jurisdiction.

There are more states like themselves in the interna-tional system, so recognition of the Courts jurisdic-tion has very diffuse benefits. In addition to the ICJcreating focal points more easily with civil law states

that share its basic principles, civil law states are betterequipped than common or Islamic law states to useoptional clause declarations as cheap talk. The partiesown communication is just as important as the adju-dicators communication, although talk is easier forcivil law states.

Third, the adjudicators role as a signaler ofprivate information also works more efficiently in civillaw dyads. The similar principles that civil law statesapply to interstate bargaining, such as bona fides, helpto reduce each sides private information. Legal rulesthat govern contracts and their enforcement are clear;hence uncertainties surrounding future complianceare reduced, making it easier to strike an accord. Theseassertions about legal institutional similarity accordwith previous research on regime, economic, orcultural similarity: the likelihood of cooperationincreases as similarity increases. Yet the ICJ as an adju-

dicator plays an important role in interstate bargain-ing between civil law states. The ICJ is more likely to beperceived by civil law states as an unbiased and fairadjudicator, increasing the likelihood that the dispu-tants will believe the signals sent by the institution.The high probability for civil law states to interact withother civil law states opens up more opportunities forthe adjudicator to signal private information effec-tively. In short, the expressive power of adjudication inthe form of correlated strategies, focal points, andinformation signaling are enhanced when the disput-ing parties are civil law states, which leads to our first

hypothesis.21

H1 (Acceptance): States with civil law systems aremore likely to accept the compulsory jurisdiction ofthe International Court of Justice than states withcommon law or Islamic law systems.

In addition to providing leverage for understand-ing why certain states are more likely to recognize theCourts jurisdiction than others, domestic legalsystems and their prominent characteristics can alsogive us insight into the design and success of interna-tional legal commitments. Freedom of contracting

and lack of religious principles will increase interna-tional commitments made by common and civil law

20This information was collected from a variety of sources, includ-ing the CIA Factbook; Glendon, Gordon, and Osakwe (1994);Opolot (1980); and a website created at the Law Faculty of theUniversity of Ottawa: http://www.droitcivil.uottawa.ca/world-legal-systems/eng-generale.html. The remaining 19 states of the192 total not listed here have mixed legal systems, which combinecharacteristics of two or more of the major legal systems.

21Space constraints preclude us from elaborating on the relation-ship between common and Islamic law institutions and the ICJ.Common law and Islamic law states face greater costs when nego-tiating with the assistance of an international court that employsless familiar legal rules and principles. They also experience greateruncertainty with respect to the Courts potential rulings. Further-more, because the International Court of Justice does not base itsdecisions on religious laws (especially Islamic law), Islamic lawstates are least likely to view the Court as a legitimate conflictmanager.

http://www.droitcivil.uottawa.ca/world-legal-systems/eng-generale.htmlhttp://www.droitcivil.uottawa.ca/world-legal-systems/eng-generale.htmlhttp://www.droitcivil.uottawa.ca/world-legal-systems/eng-generale.htmlhttp://www.droitcivil.uottawa.ca/world-legal-systems/eng-generale.htmlhttp://www.droitcivil.uottawa.ca/world-legal-systems/eng-generale.html

7/28/2019 pmjop07

9/19

states relative to Islamic law states. These states aresimply free to sign more contracts. On the other hand,the lack of good faith and compliance principles incommon law systems suggests that common law stateswill be much more cautious and specific about theirinternational obligations. Common law contracts are

elaborate and very detailed, and parties feel obligedto include all of the principles that are to governtheir contractual relations. Thus we anticipate thatcommon law states will be hesitant to accept the ICJscompulsory jurisdiction, but if they do accept theoptional clause, they will place a large number of res-ervations on their commitments. These restrictionsshould enhance the durability of common law statescommitments to the Court, because they will limit theCourts jurisdiction, especially over highly salientmatters.22

The prime position of the bona fides principle in

civil law systems should produce optional clause dec-larations with a small number of reservations. Civillaw states are aware of the fact that the fulfillment oftheir contractual obligation to the ICJ will be governedby the principle of good faith, which should substan-tially decrease the number of reservations on theirdeclarations. Contracts in civil law are not overlydetailed due to the fact that most of the overarchinglegal principles that govern contractual relationshipsare clearly spelled out in codes. The same contractualdesign should carry over from the domestic to theinternational realm. Moderate compliance principles

in civil law states should also produce long-standingcommitments to the Court, although such commit-ments may be shorter than those for Islamic law states.It is simply much easier to break an international com-mitment that is not as clearly specified. Islamic lawstates, while very reluctant to make any optional clausedeclarations, will remain firmly committed to the ICJonce they recognize its jurisdiction due to the preemi-nence of the pacta sunt servanda norm in Islamic law.Islamic states will design their commitments to the ICJcarefully, making them more likely to stay steadfastlycommitted to the Court. It is much easier to keep a

commitment that has been carefully and meticulouslycrafted.

H2 (Durability): Islamic law states will have moredurable commitments to the ICJ than civil or commonlaw states.

H3 (Design): Among states recognizing the ICJscompulsory jurisdiction, common law states will placethe greatest number of reservations on their ICJcommitments.

Research Design

The temporal domain of this study is 19202002,which includes the eras of both the PCIJ (192045)and the ICJ (19462002). We believe that these two

judicial organs can be treated as an equivalent and

functionally unchanged highest court of internationallaw, or as a World Court. Most legal scholars agreethat the ICJ was created in the aftermath of WWII as asuccessorof the PCIJ (Allain 2000; Gamble and Fischer1976; Janis 2003; Jennings and Watts 1992; Shaw2003). The statutes of both courts, the scope of their

jurisdiction, organization, procedures, and their pur-poses are virtually identical. Both courts rely onequivalent sources of international law, both are to becomprised of 15 members that shall be elected for nine

years, and both are to provide states with an alternativeto a forceful resolution of disputes. Declarations

granting jurisdiction to the PCIJ in the optional clauseor treaties/conventions continue to be in force withrespect to the ICJ (Shaw 2003, 980).23 Moreover, nodistinction is made between cases decided by the PCIJand those by the ICJ (Shaw 2003, 960); many PCIJ

judgments have been highly influential for the devel-opment of international adjudication. Given thestrong similarities between the Courts and the conti-nuity that exists across them, we think it is reasonableto combine the PCIJ and ICJ time periods into a singlesample.24

22For example, Canada, a common law state, recognized the com-pulsory jurisdiction of the PCIJ in 1930,placing seven reservationson its initial commitment relating to certain states (ratione perso-nae), certain times of disputes (ratione temporis), and divergentareas of international law (ratione materiae; Alexandrov 1995).Several of these reservations were targeted to deal with very impor-tant issues to Canada, such as their sovereignty rights overresources of the sea off their coasts. Canada added three otherreservations across time, bringing the total to 10, although suchreservations arguably strengthened Canadas long-standing com-mitment to the Court.

23Article 37 of the ICJ Statute declares: Whenever a treaty ofconvention in force provides for reference of a matter to a tribunalto have been instituted by the League of Nations, or to the Perma-nent Court of International Justice, the matter shall, as between theparties to the present Statute, be referred to the InternationalCourt of Justice.

24The most crucial difference between the Statute of the ICJ andthat of its predecessor is that the ICJ was established as the prin-cipal judicial organ of the United Nations. The Statute of the ICJconstitutes an integral part of the Charter, and all members of theUN become ipso facto parties to the ICJ Statute (Gamble andFischer 1976, 4).The PCIJ,although established under theauspicesof the League of Nations, was entirely independent of the Cov-


7/28/2019 pmjop07

10/19

The basic unit of analysis in our empirical modelis the state-year. A state can accept compulsory juris-diction without any reservation; after some time,however, the same state may add a restrictive reserva-tion to its declaration, or else withdraw its declaration.Using state-year as our unit of analysis allows us to

capture the behavior of states over time. One issuewith employing a state-year design is the potential toequate transitions to compulsory jurisdiction accep-tance with continued acceptance year to year. In otherwords, if we simply coded in each year whether a stateaccepts compulsory jurisdiction or not, we would betreating the emergence and survival of commitmentsas equivalent. Such an approach would also assumeimplicitly that the probability of transition is equiva-lent to the probability of survival. We employ theMarkov transition logit model, because this allows usto distinguish between states that transition from not

accepting to accepting compulsory jurisdiction fromthose that continue to recognize the jurisdiction of theICJ year to year. The model can be written as:25

P y y Logit x i,t i,t-1 i,t= =( ) = ( )1 0

P y y Logit x i,t i,t-1 i,t= =( ) = ( )1 0

Our theoretical hypotheses are expressed as relation-ships where the independent variable (e.g., civil legalsystem) increases or decreases the likelihood that astate will accept the compulsory jurisdiction of the

ICJ. The transition model, when the lagged dependentvariable equals zero, is best suited to testing Hypoth-esis 1 because this demonstrates an active decision bya state to deposit an optional clause declaration withthe League of Nations or United Nations. However, wealso examine the influence of our independent vari-ables on the survival of ICJ commitments, in conjunc-tion with analyses on optional clause reservations, toevaluate our second and third hypotheses about thedurability and design of such commitments.

Accepting Compulsory Jurisdiction

The primary dependent variable in this study is accep-tance of compulsory jurisdiction under the PCIJand/or ICJ. We code the dependent variable into twocategories: (1) a state accepts the compulsory jurisdic-

tion with or without reservations or (0) a state doesnot recognize the ICJs jurisdiction.26 The data are col-lected from the annual volumes of the Yearbook of theInternational Court of Justice (http://www.icj-cij.org),noting any declarations by states with respect to theoptional clause and any reservations placed on thesedeclarations. From 1920 to 2002, states accepted com-pulsory jurisdiction in 34.4% of state-years, with amajority placing some reservations on their declara-tions (27.5% of state-years). To evaluate our thirdhypothesis relating to the design of ICJ commitments,we count the total number of reservations states place

on their optional clause declarations in each year.These analyses select out only those countries thathave recognized compulsory jurisdiction. To code res-ervations, we draw upon a typology created by Alex-androv (1995), which identifies restrictions related tocertain states (ratione personae), certain times of dis-putes (ratione temporis), divergent areas of interna-tional law (ratione materiae), general reservations(such as reciprocity), and others.27 The total numberof reservations ranges from 0 to 19, with the averagestate placing five reservations on their optional clausedeclaration in a given year.

Legal Systems

In order to capture the impact of divergent legalsystems on the propensity of states to accept the com-pulsory jurisdiction of the ICJ, we construct fourmutually exclusive dichotomous variables: civil,common, Islamic, and mixed. The first three catego-ries capture our key legal systems, while the mixedcategory captures the legal system of countries where

two or more systems apply interactively or cumula-

enant. Despite the official independence of the PCIJ from the

League of Nations,these two international bodies were inescapablyinterconnected. In spite of the formal autonomy of the two insti-tutions, functionally the Court was a part of the machinery for thesettlement of international disputes envisaged in the very concep-tion of the League of Nations (Gamble and Fischer 1976, 4).

25We are grateful to Matt Golder for suggesting the use of thismodel and providing very useful notes on how to estimate it. Themodel can be estimated in one of two ways, either through esti-mation of the two models separately, conditional on the value ofthe lagged dependent variable (0 or 1) or by creating a series ofinteraction terms which multiply each independent variable by thelagged dependent variable. We employ the first strategy, which wethink facilitates an easier presentation of our results, although theresults are identical in the two procedures.

26We ran some analyses with an ordinal scale (do not accept, acceptwith reservations, accept with no reservations) employing orderedlogit. Ordered logit models make a parallel regression assumptionthat the slope of the line expressing the relationship between inde-pendent and dependent variables is constant across categories(Long 1997). The parallel regression assumption does not hold inthis case, which justifies our use of the simpler logit model. TheMarkov logit model is also superior for evaluating our separateclaims about onset and durability.

27We include the complete typology in Appendix A (online athttp://www.journalofpolitics.org).

http://www.icj-cij.org/http://www.journalofpolitics.org/http://www.journalofpolitics.org/http://www.icj-cij.org/

7/28/2019 pmjop07

11/19

tively.28 Information about domestic legal systems hasbeen gathered using the CIA Fact Book, whichdescribes major characteristics of legal traditions ofeach state in the international system, and severalother subsidiary legal sources.29 Appendix B (online;http://www.journalofpolitics.org) provides a list of

countries for each legal type in 2002.

Control Variables

Power constitutes one of the most frequently men-tioned factors that influence whether states bring theirdisputes to the ICJ (Lloyd 1985). Powerful states preferto bargain bilaterally because their material advan-tages translate into bargaining leverage. In theory,one may expect a particular reluctance to accept com-pulsory jurisdiction by powerful nations, or at least

nations which see themselves as likely to be in a supe-rior bargaining position in the kinds of disputes thatthey think might arise (Bilder 1998, 249). Less pow-erful nations, on the other hand, see impartial adjudi-cation more as a protection than a risk; it allows thesestates to feel legally equal to the worlds powers(Scott and Carr 1987, 57). We therefore expect morepowerful states to be less likely to accept the ICJscompulsory jurisdiction. To measure state power, weemploy the national capabilities index as developed bySinger, Bremer, and Stuckey (1972).30 We obtainedvalues for this variable using the Expected Utility Gen-

eration and Data Management Program (EUGene;Bennett and Stam 2000).31

Some legal scholars compare attitudes towards theICJ of relatively new states with the mind-set of well-established nations (Gamble and Fischer 1976).

Typical of the conventional wisdom in this area is thatnew states view international law as an alien systemthat Western nations have imposed on others (Brierley1963, 43) and the product of European imperialismand colonialism (Rosenne 1962, 173). Newer statescame to view the ICJ as conservative and strongly

determined to preserve the status quo (Gamble andFischer 1976), attitudes that were fueled by severalunpopular ICJ judgments (e.g., the 1966 South West

Africa decision).32 Thus newer states should be morereluctant to make optional clause declarations. Themeasure for state age captures the length of time acountry has been recognized as a state (The CIA FactBook). For countries which trace their origins to timesB.C., as a starting point of a nation, we chose 1200A.D.We calculate the natural logarithm of a states ageto minimize its variance.

The normative explanation for the democratic

peace suggests that when two democratic states dis-agree over an issue, they should be more likely toresolve the dispute peacefully because they realize thattheir adversary is operating under a norm of boundedcompetition, which supports the use of compromise(Dixon 1994). These conciliatory democratic normsshould increase the likelihood of democracies adopt-ing peaceful methods of conflict resolution. Demo-cratic states respect for judicial processes and regardfor constitutional constraints carries over into inter-national relations and democracies are apt to engagethird parties in the resolution of disputes in binding

ways such as adjudication or arbitration due to theirtrust in legal procedures (Raymond 1994).

Our theoretical argument asserts that optionalclause declarations as cheap talk work better betweenparties with similar preferences.While we focus on thesimilarity of legal institutions, we can also evaluate ourargument by looking at the similarity of political insti-tutions. Democratic states should be more likely toaccept the compulsory jurisdiction of the ICJ thannondemocratic states both because of greater prefer-ence similarity and the transparency of democraticregimes, which reduces privately held information

that can impede cooperation. Commitments made bydemocratic countries are more credible, hence demo-cratic states should be more likely to stay committed tothe ICJ once they sign on. To measure each statesdemocracy level in a given year, we use the Polity IVdata set (Jaggers and Gurr 1995), which combinesinformation from four institutional characteristicsinto a single democracy score ranging from 0 (least

28Countries belonging to the mixed category constitute a rathersmall portion of the entire data set. This group includes hybrid, orcomposite legal systems, in which civil, common, and Islamic tra-ditions are mixed with one another, or in which either of these isamalgamated with the customary law of nations. Examples includeBotswana, Brunei, Cameroon, China, Israel, and Japan.

29The following sources have been utilized: Glendon, Gordon, andOsakwe (1994); Opolot (1980); An Analysis of World Legal Tra-ditions, Pilgrimage Press, website created at the Law Faculty of theUniversity of Ottawa: http://www.droitcivil.uottawa.ca/world-legal-systems/eng-generale.html.

30This is an index of a states proportion of total system capabilitiesin six areas: iron/steel production, energy production, urbanpopulation, total population, military expenditures, and militarypersonnel.

31McAdams (2005) points out that power asymmetries can under-mine the expressive power of adjudication. Thus it would be inter-esting for future analyses to look at interactions among our keyvariables and capabilities. 32ICJ Reports (1966, 6).

http://www.journalofpolitics.org/http://www.droitcivil.uottawa.ca/world-legal-systems/eng-generale.htmlhttp://www.droitcivil.uottawa.ca/world-legal-systems/eng-generale.htmlhttp://www.droitcivil.uottawa.ca/world-legal-systems/eng-generale.htmlhttp://www.droitcivil.uottawa.ca/world-legal-systems/eng-generale.htmlhttp://www.droitcivil.uottawa.ca/world-legal-systems/eng-generale.htmlhttp://www.journalofpolitics.org/

7/28/2019 pmjop07

12/19

democratic) to 10 (most democratic). This includesthe competitiveness of political participation, the levelof constraints on the chief executive, and the opennessand competitiveness of chief executive recruitment.We turn now to multivariate analyses to empiricallyevaluate our theoretical hypotheses.

Empirical Analysis

Table 2A reports the percentage of state-years repre-senting each of the legal traditions for the entiresample (19202002), while the annual percentages areplotted in Figure 1. In Table 2B and 2C, we report thepercentages of countries accepting the compulsory

jurisdiction of the PCIJ/ICJ in 1922, 1946, 2002, andall years. Our first hypothesis finds preliminarysupport, as the percentage of civil law states acceptingcompulsory jurisdiction is higher than the percentageof civil law states in the world. Civil law states com-prised 59.8% of all state-years in the data, yet accepted

compulsory jurisdiction in 64.6% of all state-years.Furthermore, civil law states dominate the set of coun-tries accepting the Courts jurisdiction without anyreservations (91%). Islamic countries, making up12.7% of the sample, accept compulsory jurisdictionin only 6.2% of state-years, consistent with Hypothesis1, and they are likely to place reservations on theircommitments when they recognize the Court.Common law states comprise 20.9% of all state yearsand accept jurisdiction in 20.9% of state-years. As

expected (H3), common law states make up a signifi-cant portion of states that place reservations on theirICJ commitments (24.3%).

Turning to multivariate analyses, Table 3 presentsestimates from the Markov transition logit models.Model 1 presents the estimates for transition fromnonacceptance to acceptance, while Model 2 providesestimates for durability, or whether states thataccepted the jurisdiction of the PCIJ/ICJ last year con-tinue to do so this year. Common law states are utilizedas the omitted legal type category. Table 3 also presents

TABLE 2 ICJ/PCIJ Compulsory Jurisdiction Acceptance Rates

2A. Legal Type Frequency

% of all state years

Legal System 1922 1946 2002 All

Civil Law 77.1% 69.6% 54.9% 59.8%

Common Law 14.8% 14.5% 24.0% 20.9%Islamic Law 3.3% 13.0% 11.2% 12.7%Mixed Law 4.9% 2.9% 7.6% 6.5%

2B. Legal Type Frequency for States Accepting Jurisdiction

% of all state years accepting ICJ/PCIJ compulsory

jurisdiction (with or without reservation)

Legal System 1922 1946 2002 All

Civil Law 100% 71% 61.9% 64.6%Common Law 0% 19.4% 20.6% 20.9%Islamic Law 0% 3.2% 6.6% 6.2%

Mixed Law 0% 6.5% 9.5% 8.3%

2C. Legal Type Frequency for States Accepting Jurisdiction, With or Without Reservations

% of all state years accepting ICJ/PCIJ compulsory jurisdiction

Legal System All W/O Reservation With Reservation

Civil Law 64.6% 91.3% 57.9%

Common Law 20.9% 7.2% 24.3%Islamic Law 6.2% .0% 7.9%Mixed Law 8.3% 1.5% 9.9%

Total 100% 100% 100%


7/28/2019 pmjop07

13/19

the predicted probabilities for each model to ascertainthe substantive effect of each variable holding allothers at their mean or mode.

The coefficient for the civil law dummy variable ispositive and statistically significant in the transitionmodel (Table 3, Model 1), providing support for ourfirst hypothesis that civil law countries are more likelythan common or Islamic law states to accept thecompulsory jurisdiction of the International Court ofJustice.33 In Table 3, we see that the transition prob-ability for civil law systems (.019) is nearly twice aslarge as the probability of acceptance for common law

states (.0096), and more than two-and-a-half times aslarge as the probability of acceptance for Islamic lawstates (.0073). States with mixed legal systems have the

highest probability of acceptance overall (.0204),although these states are not very numerous in theinternational system. The low values of these predictedprobabilities reflect the reality that most countries donot make new optional clause declarations in a given

year. However, the variance in domestic legal systemtypes does give us purchase for explaining why somestates are willing to make optional clause declarationsmore readily than others.

The coefficient for the Islamic legal systemdummy is negative as predicted, although not signifi-cantly different from zero. Thus while we cannot

distinguish significantly between common law andIslamic law states in the transition analyses (Model1), we can conclude that countries with an Islamiclegal tradition are less likely to accept compulsory

jurisdiction than civil law states. These results clearlydemonstrate the importance of taking into consider-ation domestic legal systems for understanding theexpressive power of adjudication. Civil law statesbenefit most from acceptance of the Courts jurisdic-tion because they are more likely to view the ICJas an unbiased adjudicator, which facilitates the adju-

33In order to check the robustness of our results, we estimated ourmodels with year as an independent variable. The coefficient forthe year variable is negative and statistically significant indicatinga declining propensity for states to recognize compulsory jurisdic-tion over time. However, the coefficient for the civil law dichoto-mous variable is still positive and statistically significant (p = .09).This shows that our results are not driven by the fact that in theearly years of PCIJ and even ICJ there were far more civil lawcountries than common law or Islamic law states. Clustering by

year produces similar results, although the p-value for civil lawbecomes larger (p = .125). Powell (2006a) finds that even in themost recent years (19842002), the effect of civil law tradition onstates propensity to transition to acceptance is positive andsignificant.

TABLE 3 Markov Transition Logit Model and Substantive Effects ICJ Compulsory JurisdictionAcceptance, 19202002

Markov Transition Logit Model Results Substantive Effects (Probabilities)

Model 1: Transition toCompulsory Jurisdiction

AcceptanceModel 2: Survival of

Existing CommitmentTransition to

AcceptanceContinued ICJ

Acceptance

KeyVariables

Civil Law .687*(.311)

-1.863(1.081)

.0190 .9934

Common Law Reference category .0096 .9990

Islamic Law -.284(.624)

.0073

Mixed Law .760(.490)

-1.465(1.238)

.0204 .9956

ControlVar

iables

Democracy .152*(.029)

.122*(.044)

(Change from 0 to 10).0122 .0532 .9930 .9979

Capabilities 1.986

(3.495)

-14.587*

(4.061)

(Change from 0 to.384)

.0211 .0435 .9956 .4857State Age .009

(.091)-.197(.188)

(Change from .02 to 7.44).0204 .0218 .9964 .9956

Constant -5.208*(.455)

6.751*(1.234)

N = 5,364 N = 2,756

*p < .05.


7/28/2019 pmjop07

14/19

dicators ability to correlate strategies and create focalpoints. Islamic states prefer not to subject themselvesto an alien legal system because Islamic law putslimits on the types of international commitmentsthat Islamic states can make and because Islamicstates are more likely to view the ICJ as a biased adju-

dicator.However, Hypothesis 2 predicted that Islamic lawand common law states would have more durablecommitments to the ICJ than civil law states. Islamiclaw states have very strong pacta sunt servandanorms, while common law states draft very precisecontracts in international relations, enhancing thedurability of their commitments. The durabilityresults are presented in Table 3, Model 2. The Islamicvariable gets excluded from our survival analysisbecause it perfectly predicts success.34 We can inferfrom this that once they are committed to the ICJ,

Islamic law states do not withdraw optional clausedeclarations. While Islamic states are wary of accept-ing international institutions and rarely do acceptthem, any recognition of the ICJ by Islamic statesremains durable over time due to strong norms ofcontractual compliance. The negative and significantsign for the civil law dummy variable in the survivalmodel shows that civil law states are open to theCourt, but more likely to renege on their commit-ments over time. The substantive effects for the legaltype variables, however, illustrate that all states areextremely likely to remain committed to the ICJ over

time. The predicted probability of continued ICJacceptance ranges from .9934 (civil law states) to 1.00(Islamic law states).

Two of the three control variables exert significantinfluences on commitments to the ICJ: capabilitiesand democracy. The coefficient for capabilities is posi-tive but insignificant in the transition model (Model1), while negative and statistically significant in thesurvival model (Model 2). We conjectured that pow-erful states would be less inclined to recognize the

jurisdiction of the ICJ, due to their bargaining advan-tages in bilateral negotiations. Our results suggest that

both weak and powerful states may be willing to acceptcompulsory jurisdiction, although powerful states aremuch less likely to remain committed to the institu-tion over time. Furthermore, the substantive effect ofcapabilities in the survival model is extremely large.The probability that a country as powerful as theUnited States will continue to recognize the ICJs juris-

diction in a given year is only 49%, while the weakeststates in our sample (e.g., Liechtenstein) almost alwaysremain committed (99% chance of survival). The lackof major power commitment to international legalinstitutions is perhaps best illustrated by the with-drawal of the United States optional clause declara-

tion in 1986. While the creation of new world ordersafter victory in major wars such as World War I andWorld War II may include the creation of internationalcourts for adjudication, the major power victors maybecome less willing to support these institutions whentheir national interests are directly challenged (Posner2004).

On the other hand, our results demonstrate thatdemocratic states are much more willing to acceptcompulsory jurisdiction of the ICJ, and they are alsosignificantly more likely to maintain those commit-ments over time. Perhaps in this regard, the United

States behavior with respect to the Court is anoutlier when compared to the rest of the democraticcommunity. Fully democratic states are four timesmore likely to recognize the ICJs jurisdiction thanfully autocratic states. Furthermore, the commit-ments made by democratic states to these institutionsare durable; there is only a 2-in-1000 chance that ademocratic state will withdraw its optional clausedeclaration. This provides additional support to ourtheoretical argument that optional clause declara-tions work more effectively as cheap talk betweensimilar states.

Finally, we empirically assess our third hypothesisrelating to the design of ICJ commitments. We arguedthat the lack of the bona fides principle in common lawsystems would increase the chances that such statesdraft very detailed and precise international commit-ments. When examining the number of reservationsstates place on their optional clause declarations,common law states should have the highest number ofrestrictions. Table 4A presents the results of a regres-sion analysis, where the dependent variable is the totalnumber of reservations a state places on its PCIJ/ICJcommitment each year. In these analyses, mixed legal

systems are treated as the omitted category. The resultssupport the hypothesis, showing that common lawstates place 2.3 more reservations on their declarationsthan mixed law systems, while Islamic states place 1.7more reservations. Civil law states place the fewestrestrictions, with 1.7 fewer reservations than mixedlaw systems. Democratic, powerful, and older statesplace more reservations on their optional clause dec-larations. The overall model explains quite a bit ofvariance in the dependent variable, with an R-squaredof .36.

34Five Islamic law states have recognized the jurisdiction of the ICJ:Gambia, Sudan, Iran, Egypt, and Pakistan.


7/28/2019 pmjop07

15/19

Discussion

Our three theoretical hypotheses find robust empiricalsupport. However, we focus exclusively on acceptanceof compulsory jurisdiction, which might be problem-atic if common and Islamic law states recognize theCourts jurisdiction frequently through other means,such as the placement of compromissory clauses ininterstate treaties. This form of jurisdiction is grantedmuch more frequently, with 80% of all states belong-ing to one or more compromissory clause treaties.

One might argue theoretically that common law coun-tries prefer recognition of the ICJs jurisdictionthrough compromissory clauses, because such com-mitments are more precise and limited only to thetreaty at hand. The lack of good faith in bargainingcreates specific and limited obligations, which can be

done more easily through a bilateral or multilateraltreaty. Compulsory jurisdiction is a more risky propo-sition, because it can be applied to any legal matter ininternational law. Thus while the placement of reser-vations might help to limit the Courts prerogative,common law states may nonetheless prefer to limit theCourts potential involvement to the more specificsubject matter of interstate treaties. If our theory isapt, on the other hand, then all forms of ICJ jurisdic-tional acceptance should exhibit similar empirical pat-terns. Civil law states have greater incentives to employ

jurisdictional acceptance as cheap talk, and this should

hold whether we are examining compulsory or com-promissory jurisdiction.

To assess these arguments, we have taken an initialcut at identifying the members of all bilateral andmultilateral treaties with compromissory clauses,available on the ICJs website (http://www.icj-cij.org/icjwww/ ibasicdocuments/ ibasictext /ibasictreatiesandotherdocs.htm). Bilateral treaty members are listed onthe website, while multilateral treaties are listed bysimply the name of the treaty and signature year.Using information from the United Nations TreatyDatabase and other sources, we tracked down copies

of each multilateral treaty to determine the memberstates and years of acceptance. We created a countmeasure of the total number of bilateral and multilat-eral treaty memberships for each of the 192 COWsystem members as of 2002. The average state belongsto 28 treaties (26 multilateral, 2 bilateral) with com-promissory clauses, with a standard deviation ofapproximately 15 treaty memberships.

In Table 4B, we present the results of a regressionanalysis, where the dependent variable is the totalnumber of treaty memberships with compromissoryclauses. In addition to our legal system type variables

(omitting civil law), we also include a dummy variablecoded 1 if a state also accepted the compulsory juris-diction of the ICJ, and 0 otherwise. The results providefurther support to our theoretical argument, showingthat civil law states have the highest average numberof treaty memberships. Common law states have sixfewer (p = .01), mixed law states eight fewer (p = .023),and Islamic law states 4.5 fewer treaty memberships(although not significant, p = .136) than civil lawstates. Thus civil law states use both forms of jurisdic-tion as cheap talk. The positive and significant sign for

TABLE 4 Regression Analyses

4A. Number of Optional Clause Reservations

Variable Coefficient (S.E.)

Common Law 2.30

(.21)*Civil Law -1.66

(.19)*Islamic Law 1.65

(.27)*Democracy .12

(.01)*Capabilities 4.45

(1.85)*

State Age .00(.00)*

Year .03(.00)*

Constant -45.89

(4.33)*N = 3,084 (States Accepting Compulsory Jurisdiction)F(7,3076) = 242.84 (p < .0000)

R2 = .3559

4B. Monadic Count of Treaty Memberships withCompromissory Clauses

Variable Coefficient (S.E.)

Common Law -6.15(2.38)*

Mixed Law -7.88

(3.43)*

Islamic Law-

4.52(3.02)Accept ICJ Compulsory Jurisdiction

(in 2002)

11.44

(2.08)*Constant 27.43

(1.55)*N = 192 (Number of States in 2002)

F(4,187) = 11.09 (p < .0000)R2 = .1917

*p < .05.

http://www.icj-cij.org/icjwww/ibasicdocuments/ibasictext/ibasictreatiesandotherdocs.htmhttp://www.icj-cij.org/icjwww/ibasicdocuments/ibasictext/ibasictreatiesandotherdocs.htmhttp://www.icj-cij.org/icjwww/ibasicdocuments/ibasictext/ibasictreatiesandotherdocs.htmhttp://www.icj-cij.org/icjwww/ibasicdocuments/ibasictext/ibasictreatiesandotherdocs.htmhttp://www.icj-cij.org/icjwww/ibasicdocuments/ibasictext/ibasictreatiesandotherdocs.htmhttp://www.icj-cij.org/icjwww/ibasicdocuments/ibasictext/ibasictreatiesandotherdocs.htm

7/28/2019 pmjop07

16/19

the compulsory jurisdiction variable demonstrates thepositive correlation between the two; states recogniz-ing compulsory jurisdiction belong to 11 more treatieswith compromissory clauses.

The question remains, however, whether simplyaccepting the International Court of Justices jurisdic-

tion in any form has any significant influence onbargaining over contentious issues in internationalrelations. Pessimists would still point to the largenumber of reservations states place on their optionalclause declarations and would argue that even thoughthe number of cases before the Court has increased,the increase in the number of states has far outpacedthis growth (Posner and Figueiredo 2004). Our otherresearch tries to address this question more directly byexamining the influence of joint ICJ acceptance onbargaining (Mitchell et al. 2005) and compliance withICJ judgments (Mitchell and Hensel, 2006). Our initial

look at these important questions provides somereason for optimism.

We find that contending states accepting the ICJscompulsory jurisdiction are 20% more likely to reachagreements to help resolve contentious issue claims.35

States are also 15% more likely to comply with anyagreements reached when both sides jointly accept thecompulsory jurisdiction of the ICJ. Additionally, jointacceptance of the optional clause significantly reducesby 4% the likelihood that conflicting parties will resortto militarized force. Once cases involving territorial,maritime, or cross-border river issues come before the

PCIJ or ICJ, compliance with judgments is almostguaranteed. Contending parties have complied with28 of 29 PCIJ and ICJ decisions over these issues, animpressive record indeed (see Appendix C (online;http://www.journalofpolitics.org) for a list of allcases).36

As noted above, similarity of legal institutionsshould promote cooperation between states ingeneral. However, the logic of our theory suggests thatthe use of optional clause declarations as cheap talkwill work best for civil law states. Peaceful bargainingshould succeed most often when two states have civil

legal systems andwhen they both accept the jurisdic-tion of the ICJ. In these situations, the threat of being

sued in Court is most credible, and thus, out-of-courtnegotiations will be more efficient. In the future,empirical analyses of interstate negotiations need todiscriminate between the general effect of legal systemsimilarity and the specific effect of the ICJ signal, andconsider their interactive effects as well. We also need

to consider the interactive effects of other sources ofstate similarity, such as regime type and sharedculture. Democratic states are more accepting of inter-national legal institutions and make more durablecommitments to these institutions than nondemo-cratic states. The growth in democratic states world-wide37 suggests that a larger percentage of states mayaccept the legitimacy of the International Court ofJustice in the future.

Conclusion

In this paper, we argue that domestic legal systemshave important effects on foreign policy behavior andinterstate bargaining and that civil law countries aremore likely to recognize the jurisdiction of the Inter-national Court of Justice than common law or Islamiclaw countries. On the other hand, common andIslamic law states design more precise commitmentsto the court with more reservations and have moredurable commitments. The lack of attention to legalsystems in the International Relations literature ispuzzling given that most interstate contracts are gov-

erned by a variety of legal principles and are oftennegotiated by lawyers. By focusing on legal systems,our project contributes to the legalization project in IR(Goldstein et al. 1991) and interdisciplinary researchintegrating insights from International Relations andInternational Law (Slaughter, Tulumello, and Wood1998).

Our argument speaks more broadly to the liberalresearch tradition in international relations. Liberalsargue that domestic political variables play an impor-tant role in states behavior on the international arena,and our analyses of domestic legal institutions andregime type support this view. A domestic legalsystem, which constitutes an embodiment of societalpreferences, interests, and ideas, can determine astates behavior towards other states and internationalinstitutions (Moravcsik 1997). We find that the char-acteristics of internal legal systems account for a greatdeal of variance in the acceptance rates and design of

35The issue claims we analyzed were collected by the Issue Corre-lates of War Project (www.icow.org) and include territorial borderdisputes and the usage, access to, or ownership of cross-borderrivers and maritime zones.

36Ginsburg and McAdams (2004, 1229) argue that judicial resolu-tion is more likely for territorial disputes because the underlyingdispute is a coordination problem; high compliance rates areexpected in these cases.

37In our monadic dataset, 55% of states were considered demo-cratic, scoring 6 or higher on the Polity IV scale in 2002.

http://www.journalofpolitics.org/http://www.icow.org/http://www.icow.org/http://www.journalofpolitics.org/

7/28/2019 pmjop07

17/19

international commitments. Our argument also bol-sters democratic peace research, showing that demo-cratic states are more accepting and committed tointernational legal institutions than autocratic states.

Our findings regarding the relationship betweendomestic legal systems and states attitudes towards

the ICJ provide interesting insights into the dynamicsof international institutions. Understanding the char-acteristics of states legal systems may give us leveragefor analyzing a wide variety of international courtsand institutions. Preliminary research by Powell(2006b) suggests that common law states tend to lendhigher levels of support to the European Court ofJustice than civil law states because the ECJ embracesthe stare decisis principle. States behavior towards theICJ may also have important implications for otherinternational courts, such as the Law of the Sea Tribu-nal and the International Criminal Court. Extending

our argument to other international courts willfurther our understanding of the intricate relationshipbetween domestic legal structures and foreign policychoices.

Although this paper has significantly increasedour understanding of the ICJ, numerous questions stillremain unanswered. First, it is plausible that statemembers of other peace-promoting organizations aremore prone to accept the jurisdiction of the ICJ. Inother words, acceptance of the Court may be part of abroader acceptance of international institutions. Ourfinding that democracies are more likely to recognize

the jurisdiction of the ICJ, for example, is consistentwith Russett and Oneal (2001)s argument thatdemocracies are more likely to join international orga-nizations in general. Sorting out the particular effectof international legal institutions on state behaviorwill be an important avenue for future research.

We would also like to explore changes in theCourts practices over time and the implications ofthese changes for our theoretical argument andempirical analyses. Numerous legal scholars havepointed to the fact that the practices of common lawand the ICJ have become increasingly similar over

time (Markesinis 2000; Shapiro 1986; Shahabuddeen1996). The jurisprudence of the World Court hasdeveloped over time in the direction of a powerfulinclination to adhere strongly to its previous decisions.Furthermore, courts of last resort in common lawsystems have come to accept that they are not com-pelled to follow their previous judgments, but withinwell-defined boundaries they might depart from them(Shahabuddeen 1996). The practice of common lawcourts is, therefore, becoming progressively moresimilar to the rules and practices of the International

Court of Justice, which may produce an increasingwillingness on the part of common law states to rec-ognize the Courts jurisdiction. Moreover, statute lawhas recently become a much more important source oflaw in certain common law countries. For example, theoutput of enacted law in the United Kingdom has

noticeably increased mostly as a result of the integra-tion of British law with the European legal culture(Markesinis 2000). These changing processes withinthe ICJ and within common legal systems havebrought the two legal orders closer together (althoughsee Legrand 1996, 1997) and suggest shifting dynamicsof states behavior towards the ICJ. Our theory andempirical analyses show that states may view the Courtnot only as a last-resort forum for resolving disputes,but also as an institution through which countries cancredibly send information to each other about theircommitment to peaceful bargaining practices.

Acknowledgments

Previous versions of this paper were presented at the2005 Midwest Political Science Association meetingand the 2004 Journeys in World Politics Workshop,Iowa City, IA. We are grateful to Fred Boehmke, TerryChapman, Matt Golder, Mike Grieg, Tanya Janulewicz,Kelly Kadera, Eric Posner, Anne Sartori, Ross Schoofs,Chuck Shipan, Jeff Staton, Jana von Stein, ClaytonThyne, and participants in the Political Science Collo-

quium at the University of Iowa for their helpful com-ments, suggestions, and assistance.

Manuscript submitted 6 June 2005Manuscript accepted for publication 3 August 2006

References

Alexandrov, Stanimir A. 1995. Reservations in Unilateral Declara-tions Accepting the Compulsory Jurisdiction of the InternationalCourt of Justice. The Hague, The Netherlands: Martinus Nijhoff

Publishers.Al-Azmeh, Aziz. 1988. Islamic Law: Social and Historical Contexts.

New York: Routledge.

Allain, Jean. 2000. A Century of International Adjudication: TheRule of Law and its Limits. The Hague,The Netherlands: T.M.C.Asser Press.

Bederman, David J. 2001. International Law Frameworks. NewYork: Foundation Press.

Bennett, Scott, and Allan Stam. 2000. A Universal Test of anExpected Utility Theory of War. International Studies Quar-terly44 (3): 45180.

Bilder, Richard B. 1998.International Dispute Settlement and theRole of International Adjudication. In International Law:


7/28/2019 pmjop07

18/19

Classic and Contemporary Readings, eds. Charlotte Ku and PaulF. Diehl. Boulder, CO: Lynne Rienner, pp. 23356.

Bridge, M.G. 1984. Does Anglo-Canadian Contract Law Need aDoctrine of Good Faith? Canadian Business Law Journal9 (4):385434.

Brierley, James L. 1963. The Law of Nations: An Introduction to theInternational Law of Peace. Oxford: Oxford University Press.

Brownlie, Ian. 2003. Principles of Public International Law. 6th ed.Oxford: Oxford University Press.

Case Concerning the Application of the Convention of 1902 Govern-ing the Guardianship of Infants. I.C.J. Reports 1958.

Crawford, Vincent P., and Joel Sobel. 1982. Strategic InformationTransmission. Econometrica 50 (6): 141351.

Darbyshire, Penny. 2001.Eddey and Darbyshire on the English LegalSystem. 7th ed. London: Sweet and Maxwell.

David, Rene, and John E. Brierley. 1985.Major Legal Systems in theWorld Today. London: Stevens.

Di

Documents

pmjop07