Negotiating Military Alliances: Legal Systems and Alliance Formation

Abstract

Do domestic legal systems affect states' propensity to form military alliances? This article, building upon the existing research in international relations, adopts a socio-legal approach to understanding international treaty making. By focusing on the essence of international negotiations—communication between states' representatives—I argue that negotiating parties who share a common legal language have a common a priori understanding concerning concepts under discussion. Domestic laws operating within states impact the process of creation of international law embodied in treaties. Empirical analyses show that states with similar legal systems are more likely to form military alliances with one another. Additionally, domestic legal systems influence the way that states design their alliance commitments. In general, my findings suggest that the influence of domestic laws does not stop at “the water's edge.” It permeates the interstate borders and impacts the relations between states, especially the treaty negotiating and drafting process. International negotiators bring their legal backgrounds to the negotiating table, which influences both their willingness to sign treaties and the design of the resulting agreements.

KEYWORDS:

• international law
• legal systems
• military alliances
• negotiations
• treaties

Acknowledgments

I am grateful to Brett W. Curry, John King Gamble, Brett Ashley Leeds, Sara McLaughlin Mitchell, Richard Pacelle, Charles Wesley Powell, and Krista Wiegand for their comments on this paper.

Notes

¹Memorandum of conversation with Professor Antonio Pereź, conducted by the author on April 02, 2009.

²Memorandum of conversation with Jerzy Szymański, conducted by the author on May 22, 2009.

³Also see Gibler and Wolford (2006) who show that democratic dyads are unlikely to ally. The similarity argument has been extended in the literature to relations between autocracies, as their structural and institutional characteristics may provide explanations for the formation of alliances (Peceny et al. 2002).

⁴ Leeds et al. (2002) replicate Lai and Reiter's (2000) study on a different data set (ATOP—Alliance Treaty Obligations and Provisions). Joint language has also been found to increase interstate cooperation in other areas such as interstate trade (Bliss and Russett 1998; Rose 2004; Souva et al. 2008).

⁵Civil law is the Romano-Germanic legal tradition, largely rooted in the laws of the Roman Empire. In this legal tradition, the written letter of law (codes) constitutes the main source of law. Common law, which originated in the British Isles, is based on the stare decisis doctrine, according to which judges are bound primarily by precedents established by previous judgments (Opolot 1980). Islamic law, the world's third major legal tradition, arose with the birth of Islam in the seventh century a.d., and is based primarily on religious principles of human conduct (al-Azmeh 1988).

⁶An analogical argument can be made in the context of self-executing and non self-executing treaties (see footnote 18).

⁷Of course, treaties are also interpreted by international adjudicative bodies, arbitral tribunals, etc.

⁸The first general method of interpretation used by all domestic legal systems will be literal, but this is only true in cases “when a text expressed in imperative language is clear, precise, unambiguous” (Friesen 1996:10). This is, however, rarely true, as provisions in any international treaty tend to be general and relatively vague. The teleological approach entails detaching the actual text of a treaty from its historical context and giving it a meaning appropriate for present social and economic conditions. Textual, literal approach relies on the ordinary meaning of words used in a text (Friesen 1996:10).

⁹Mu‘āhada, or a treaty according to Islamic law, in order to be valid “must not contradict the legal rulings of the scriptural texts, that is, the Qur'ān and the Sunna” (Bsoul 2008:115).

¹⁰Article 31 of the VCLT spells out the rules of interpretation of treaties, mentioning the direct textual interpretation and the teleological interpretation. However, the International Law Commission does not support the view that in the process of treaty interpretation one must give priority to one particular factor, such as the text (textual interpretation), intention, or the object of the treaty (teleological interpretation) (Aust 2007:231; Tabory 1980:205).

¹¹Perhaps the best description of good faith comes from the medieval jurists. According to them, good faith meant that a party “must keep his word, refrain from deceit and overreaching, and honor obligations that are only implicit in his contract” (Zimmermann and Whittaker 2000:93).

¹²An unconscionable contract is a contract that is so unfair to a party that no informed or reasonable person would agree to it. Usually it results from unequal bargaining positions between the parties. Frustration of a contract occurs when a contract becomes impossible to perform/fulfill or can be performed/fulfilled only in a way significantly different from that originally envisioned.

¹³English common law does not recognize a general duty to negotiate nor to perform contracts in good faith. Despite the fact that some efforts have been made to incorporate bona fides into the common law tradition, the position of bona fides in this legal family is still much weaker than under civil law (Powell and Mitchell 2007; Zimmermann and Whittaker 2000). Some common law states have gradually introduced the principle of good faith into their legal systems (Examples: US: the United States Uniform Commercial Code, the US Restatement (Second) of Contracts, UK: the European Consumer Protection Directive of 1994 (Summerst 1982)).

¹⁴Koran LXXXIII: 1–5.

¹⁵This argument is strengthened by the fact that nowadays professional diplomats “are no longer the only, or even the main, actors in international negotiations” (Cohen 1991:18). Officials from domestic agencies, including justice, trade, etc. also play major roles in modern day diplomatic relations (Cohen 1991; Starkey, Boyer, and Wilkenfeld 2005).

¹⁶Also, words may sound similar in two different languages and two different legal systems, but they still may entail slightly, or even fundamentally different concepts.

¹⁷Examples include the Warsaw Security Pact (1955), the Southeast Asia Collective Defense Treaty (1954), the Mutual Defense Treaty Between the United States and the Republic of Korea (1953), the Treaty of Mutual Cooperation and Security between Japan and the United States of America (1960), the nonaggression pact between Poland and Greece (1996), the Inter-American treaty of Reciprocal Assistance (1947), the nonaggression pact between Angola, Botswana, and twelve other African states (2001), the nonaggression pact between Belarus and Turkmenistan (2002), and many others.

¹⁸Worth mentioning is that international treaties can be either self-executing (such a treaty is effective without the need for supplementary legislation, as it provides to individuals specific rights, which the domestic courts may enforce) or non-self-executing (requires an act of a state's legislative body or of the executive branch, for example the president, to give it a legal effect). In some domestic legal systems, domestic courts decide how to interpret whether a treaty is self-executing or non-self-executing. Importantly, alliance treaties may generate obligations that may ultimately require domestic legal execution (e.g., the funding of an alliance-related war).

¹⁹The courts may alternatively use rules of interpretation established by international law (Gardiner 2008:127).

²⁰For example, domestic legal systems had a substantial impact on the legal set-up of several international courts such as the Permanent Court of International Justice; its successor, the International Court of Justice; and the International Criminal Court (Fernández de Gurmendi 1999; Lloyd 1985:35; Powell and Mitchell 2007).

²¹My arguments dealing with all three counter-explanations concern largely treaty compliance. Following Fearon (1998), I believe that expectations about compliance influence agreement. I expect that the lack of common understanding concerning legal concepts makes the formation of alliances less likely. If, on the other hand, the enforcement and monitoring dilemmas are resolvable, “we should observe serious attempts at international cooperation” (p. 279). It is simply easier to communicate within a legal tradition.

²²My approach to treaty drafting and interpretation is similar to the “contextual approach” to treaty interpretation, which proposes that treaties constitute objective manifestations of the subjective expectations of treaty negotiators. Also see Gardiner (2008:65–68) for the description of the New Haven School/Myres McDougal sociological jurisprudence. Professor McDougal believed that a treaty constitutes “a continuing process of communication” (Gardiner 2008:65).

²³Precise treaty provisions mean fewer (if any) disagreements about interpretation. Vague treaty provisions may lead to a situation where a party may find itself bound by an obligation it had no intent to create. Of course, just as with domestic contracts, treaties are often times ambiguous (Chayes and Chayes 1993). Secondly, ambiguity of treaty provisions may be an indicator of disagreement over more precise provisions.

²⁴Lawyers have also played important roles in states' negotiations in the past. For example, in the 18the century England, the King consulted private legal practitioners or university law professors regarding international negotiations (Fitzmaurice 1965:77). Another more recent example is that of the Association of Soviet Lawyers, which was instrumental in negotiating the Soviet Nuclear Arms Control Agreement and several military alliances (Downs 1997). Nowadays,despite the fact that ministries of foreign affairs play a crucial role in drawing up international treaties, most governments employ their own staff of resident legal advisers, treaty specialists, and legal scholars (both domestic and international) responsible for drafting and interpreting of international agreements.

²⁵Memorandum of conversation with Professor Antonio Pereź, conducted by the author on April 02, 2009.

²⁶Memorandum of conversation with Jerzy Szymański, conducted by the author on May 22, 2009.

²⁷Other examples of contingencies include: establishing an organization (for example, the Inter-American Treaty of Reciprocal Assistance, the Rio Treaty, 1947, establishes an Organ of Consultation); references to the United Nations or other international organizations (the Warsaw Pact, 1955, made references to United Nations (Article 4), the nonaggression pact between Poland and Greece (1996) makes reference to the European Union (Art. 1 and Art. 6), and the United Nations (Art. 2), the nonaggression pact between Belarus and Turkmenistan (2002) refers to the Organization for Security and Co-operation in Europe (Art. 2)); mention of mediation/arbitration or other means of settling disputes among the signatories (the nonaggression pact between Chile and Argentina (1984) in Article III states that “the Parties shall adopt appropriate measures to maintain the best general conditions of coexistence”); specific provisions concerning the length of time the treaty is to last (the nonaggression pact between Belarus and Vietnam in Art. 18 states that this treaty is to last for 120 months).

²⁸According to Lundmark (2001), “a typical American purchase agreement of 3,800 words in length devotes only 550 words to business terms, leaving over 3,000 words for standard definitions and boilerplace [recitations of applicable law]” (p. 121). Interestingly, American common law contracts are longer than their British counterparts.

²⁹Assembly of the necessary data (control variables) has been enhanced by using the Expected Utility Generation and Data Management Program (EUGene) (Bennett and Stam 2000).

³⁰There are several types of military alliance treaties: defense pacts (Barbados and Dominica, 1996; Libya and Niger, 1974), offense pacts (Germany, Austria-Hungary, Russia, 1833; Israel and U S, 1981), neutrality pacts (Paraguay and Argentina, 1856), nonaggression pacts (Germany and Denmark, 1939; U S and Canada, 1975), and consultation pacts (Belarus and Turkmenistan, 2002).

³¹Other scholars propose divergent classifications of legal systems (see for example La Porta Lopez-de-Silanes Schleifer and Vishny (1999) with the following categories: English, Socialist, French, German, and Scandinavian legal origin). Splitting the civil law family into French and German yields an unnecessary subcategorization, since both of these legal families are based on highly similar, and often times identical procedural and substantive concepts. In addition, the LaPorta et al. (1999) measure omits the Islamic legal system, which, I argue, constitutes one of the most important legal traditions of the world, present and past. For robustness checks I run our model with this measure and my results hold up.

³²The mixed legal system/tradition includes hybrid, or composite legal systems, in which civil, common, and Islamic traditions are mixed with one another, or in which either of these is amalgamated with the customary law of nations. Examples include Botswana, Brunei, Cameroon, China, Israel, and Japan. Countries belonging to the mixed category constitute a very small portion of my data set.

³³I decided not to include the socialist legal system in my analyses. The main traits that distinguish socialist law from civil, common, and Islamic legal traditions are inherently bound to the socialist political regime (principles related to the very nature of Man, God, the dominant class, and the special courts, such as comrades' and people's courts), and constitute unique ideas of socialism that permeated from the political system into the domestic legal framework of several socialist states. Because my argument focuses on the substantive characteristics of domestic legal systems, I believe that the link between law and politics does not justify introducing a separate category of domestic legal systems.

³⁴Correlations between same Legal System variable and the cultural similarity variables are relatively low and range from .03 to .08.

³⁵Available at http://www.thecommonwealth.org/ Correlation between British Commonwealth and Common Colonial Heritage is .09.

³⁶The model can be estimated in one of two ways, either through estimation of the two models separately, conditional on the value of the lagged dependent variable (0 or 1) or by creating a series of interaction terms which multiply each independent variable by the lagged dependent variable. I employ the first strategy. The survival of alliance commitments can be estimated via the survival equation, which can be written as follows: P(y_i,t−1 = 1 | y_i,t−1 = 1) = Logit(x_i,t α)

³⁷Since several scholars have shown that a state's reputation as a reliable/unreliable alliance partner can affect alliance formation (Gibler 2008; Leeds 2003), I also estimated my base model with three separate measures of past reliability (past reliability during times of war (Leeds 2003), whether there was a previous violation of an alliance commitment in a dyad in the international system (Gibler 2008), and whether there was a previous violation of an alliance commitment in a dyad in the region (Gibler 2008)). These three measures are negative and statistically insignificant in my models.

³⁸Interestingly, my preliminary analyses (I used Gibler's 2008 measures of alliance violations) show that states with similar domestic legal systems are less likely to violate their alliance commitments. There are more failed promises in military alliances signed between states with different legal systems.

³⁹I have also estimated the model using negative binomial model. Results of this estimation are virtually identical to the OLS model.

⁴⁰In order to make sure that the results are not skewed by the United States, I ran my model excluding the US. Results of this model are virtually identical.

⁴¹Mixed Legal System and Proportion of Mixed Law States constitute the benchmark categories.

⁴² Reiter and Tillman (2000) code an additional category—assembly independent systems—but because Switzerland constitutes the only example of this system, I did not create a separate category for it.

⁴³For example, the Joint Democracy variable became statistically significant in the Domestic Approval and Time Control Models.

⁴⁴This data set includes 1675 peaceful attempts to settle issue claims in the Western Hemisphere, Europe, and the Middle East. A complete list of issue claims is available on the ICOW Web site http://www. paulhensel.org/icow.html

⁴⁵Control variables used include: acceptance of the jurisdiction of the International Court of Justice, democratic dyad, challenger's capabilities, issue salience, recent conflict, and functional/procedural attempts.