Law and Politics in Transnational Judicial Governance: The Case of Forum Non Conveniens

Christopher A. Whytock Associate Professor of Law

University of Utah College of Law whytockc@law.utah.edu Draft: March 19, 2008

Abstract: Domestic courts play a central role in global governance. But international relations scholars have virtually ignored domestic courts. We therefore lack the knowledge to evaluate this role normatively or to assess its implications for positive theories of global governance. This paper helps to fill this gap in our knowledge. I begin by proposing a concept of “transnational judicial governance” that clarifies the ways in which domestic courts contribute to the regulation of transnational activity. Next, I focus on one transnational judicial governance function—the allocation of adjudicative authority among states—and ask: what factors, both legal and political, influence whether U.S. courts dismiss transnational litigation in favor of foreign courts, thus deferring to foreign governance authority, instead of adjudicating transnational disputes themselves, thus asserting U.S. governance authority? I then develop a theory of judicial heuristics, according to which judges make the choice between foreign and domestic authority based largely on decisionmaking shortcuts which allow them to conserve scarce decisionmaking resources while making decisions that achieve acceptable levels of legal quality. I also develop a series of hypotheses about the political determinants of judicial allocation of governance authority. Analyzing an original dataset of 210 allocative decisions based on the common law doctrine of forum non conveniens, I find that U.S. district court judges frequently defer to foreign governance authority. Legal factors have a powerful influence, including two judicial heuristics: deference is less likely when there is a U.S. plaintiff and more likely when the transnational activity giving rise to the dispute occurs mostly or all outside U.S. territory. Traditional political factors, including power and ideology, do not appear to play an important role. But another political factor does have an influence: deference is more likely toward liberal democracies. These findings suggest that U.S. district court judges are not excessively parochial in their outlook, and that transnational judicial governance follows a distinctly legal logic according to which it can help foster the rule of law in world politics—perhaps particularly among democracies. For their helpful criticisms and suggestions on an earlier version of this paper, I thank Tim Büthe, Kevin Clermont, Bruce Jentleson, Judith Kelley, David Klein, and Ralf Michaels.

1

1. Introduction Domestic courts play a central role in global governance. They help allocate governance authority among states, between public and private actors, and between national and international institutions. They allocate risks and resources among transnational actors. And they perform a governance support function, upon which private and interstate governance arrangements largely rely. Their decisions directly affect litigants; and when they publish these decisions, they provide information about how they may make similar decisions in the future, thus influencing the strategic behavior of transnational actors beyond the parties to particular lawsuits. Simply put, domestic courts are global governors. But is this a good thing? Are domestic courts parochial actors that routinely insist on asserting domestic governance authority over transnational activity, or do they have a more cosmopolitan willingness to defer to foreign authority in appropriate circumstances? Are their decisionmaking processes distinctly legal, allowing them to foster transnational rule of law, and making global governance through domestic courts qualitatively different from the more power-oriented mechanisms that are thought to characterize most aspects of world politics? These questions are important not only normatively, but also for positive theory. Political scientists emphasize the political determinants of governmental decisionmaking, with international relations (IR) scholars focusing on the role of power in determining state behavior, and courts scholars on the role of ideology in determining judicial behavior. But when applied to the role of domestic courts in global governance, might these understandings underestimate the importance of legal influences? The problem is that we lack the knowledge to answer questions like these. IR scholars are increasingly relaxing the field’s traditional unitary state actor assumption in an effort to shed light on the impact of domestic institutions on transnational activity, including the role of regulatory agencies and legislatures.1 According to Jeffry Frieden and Lisa Martin, “the impact of domestic institutions and interests on international interaction” is a critical subject of international political economy (IPE) scholarship.2 Yet IR and IPE scholars have virtually ignored domestic courts,3 notwithstanding their growing interest in international courts and transnational arbitration. 4 For their part, those political scientists who do study domestic courts

1 See, for example, Evans, Jacobson, and Putnam 1993; Martin 2000; Milner 1997; Moravcsik 1997; and Putnam 1988. See also the emerging body of interdisciplinary IR/international law scholarship on cross-border cooperation among national regulatory agencies through transgovernmental networks, including Raustiala 2002; Slaughter 2004; Slaughter and Zaring 2006; and Whytock 2005. 2 Frieden and Martin 2002, 119-120. 3 Exceptions include Falk 1964 (on domestic courts and international law); Slaughter 2004 (on transgovernmental networks of judges); Putnam 2006 and Raustiala 2006 (on the extraterritorial application of U.S. law by U.S. courts); Whytock 2006a (on domestic courts, foreign law, and world politics); and the work of various scholars of European integration on the role of domestic courts in the construction of the European legal system, including Alter 2001; Burley and Mattli 1993; and Stone Sweet and Brunell 1998. It also is worth noting that Frieden and Martin include courts in their list of domestic institutions to which IPE scholars should devote more attention: “Future work will need to allow for more nuance and development, incorporating such other domestic institutions as political parties, courts, and central banks and a more sophisticated treatment of domestic interests.” Frieden and Martin 2002, 125. 4 See, for example, Alter 2006; Keohane, Moravcsik, and Slaughter 2000; Voeten 2007; and Volcansek and Stack 2005 (international courts); and Mattli 2001; Stone Sweet 2006; and Whytock 2008 (transnational arbitration).

2

have focused almost exclusively on matters of domestic policy.5 Meanwhile, legal scholarship on transnational litigation in domestic courts is primarily doctrinal. As a result, we know very little about how domestic courts behave as global governors and why they govern the way they do. My goal in this paper is to take some modest steps toward filling this gap in our knowledge, thus providing an empirical basis for critically appraising the role of domestic courts in global governance and for refining positive theories of global governance. I begin by proposing a concept of “transnational judicial governance” that clarifies the ways in which domestic courts contribute to the regulation of transnational activity. Next, I focus on one transnational judicial governance function—the allocation of adjudicative authority among states—and ask: what factors, both legal and political, influence whether U.S. district courts dismiss transnational litigation in favor of foreign courts, thus deferring to foreign adjudicative authority, instead of adjudicating transnational disputes themselves, thus asserting U.S. governance authority? I then develop a theory of judicial heuristics, according to which judges make the choice between foreign and domestic authority based largely on decisionmaking shortcuts which allow them to conserve scarce decisionmaking resources while making decisions that achieve acceptable levels of legal quality. I also develop a series of hypotheses about the political determinants of judicial allocation of governance authority. Analyzing an original dataset of 210 allocative decisions based on the common law doctrine of forum non conveniens, I find that U.S. district court judges frequently defer to foreign adjudicative authority. Legal factors have a powerful influence, including two judicial heuristics: deference is less likely when there is a U.S. plaintiff and more likely when the transnational activity giving rise to the dispute occurs mostly or all outside U.S. territory. Traditional political factors, including power and ideology, do not appear to play an important role. But another political factor does have an influence: deference is more likely toward liberal democracies. These findings suggest that U.S. district court judges are not excessively parochial in their outlook, and that transnational judicial governance follows a distinctly legal logic according to which it can help foster the rule of law in world politics—perhaps particularly among democracies. I conclude by arguing for a broader understanding of the processes of global governance and the relationships between law and world politics than currently prevails in IR and IPE scholarship. 2. Transnational Judicial Governance The concept of “transnational judicial governance” refers to the role of domestic courts in regulating transnational activity.6 By “governance,” I mean the prescription, application, and enforcement of rules.7 The term “judicial” refers to who governs: domestic courts. The term “transnational” refers to what is governed: transnational activity, that is, activity with connections to more than one state, as opposed to purely domestic activity. These connections

5 Exceptions include Davis 2006; Randazzo 2006; and Unah 1998. 6 Whytock 2006b; Whytock 2007. 7 See Kjaer 2004, 10 (defining governance as “the setting of rules, the application of rules, and the enforcement of rules.”). See also Kahler and Lake 2003, 7 (“governance is characterized by decisions issued by one actor that a second actor is expected to obey”); and Keohane and Nye 2000, 12 (“governance” refers to processes and institutions that guide and restrain collective activity”).

3

may be territorial, when the activity or its effects touch the territory of more than one state, or they may be based on legal relationships between a state and the actors engaged in or affected by that activity, such as citizenship. The actors involved in the activity may be either state or nonstate actors.8 The remainder of this section more fully explains the concept of transnational judicial governance. 2.1 Transnational Judicial Governance and Transnational Litigation Domestic court decisionmaking in transnational litigation is the principal mechanism of transnational judicial governance.9 The subject matter of transnational litigation is as diverse as transnational activity itself, ranging from antitrust to securities regulation, from cross-border business transactions to tourism, from human rights to terrorism, from intellectual property to environmental regulation. The parties to transnational litigation are likewise varied. They include not only individuals and private businesses, but also states. For example, between 1995 and 2005, foreign nations were parties to an average of more than 475 U.S. federal district court cases per year.10 Although data is scarce, the Administrative Office of the United States Courts maintains information about one type of transnational litigation: “alienage” cases in the U.S. federal district courts. An alienage case is a case over which a U.S. federal district court has jurisdiction because the dispute is between “citizens of a [U.S.] state and citizens or subjects of a foreign state.”11 Since 1986, there have been well over 120,000 alienage cases in the U.S. district courts.12 In 2005, the most recent year for which there is data, 1,976 alienage jurisdiction cases terminated in the U.S. district courts.13 The total volume of transnational litigation in the United States is almost certainly greater than the alienage jurisdiction figures alone suggest. For one thing, these figures exclude transnational disputes over which U.S. federal courts have jurisdiction based on grounds other than alienage, such as federal question cases, maritime cases, lawsuits against foreign states, and claims by foreign citizens for torts “committed in violation of the law of nations or a treaty of the United States.”14 The alienage jurisdiction figures also exclude litigation not involving foreign parties but nevertheless involving transnational activity. Perhaps most importantly, these figures do not count transnational litigation in U.S. state courts, where by far most litigation in the United States occurs.15

8 I use the term “transnational” rather than “international” to emphasize that the activity being governed may involve not only state actors, but also nonstate actors. 9 I define “transnational litigation” as litigation that arises from transnational activity. 10 Author’s analysis of data collected by the Administrative Office of the United States Courts. Whytock 2006b, 5. 11 U.S. Constitution, Article III; 28 U.S.C. sec. 1332. 12 Clermont and Eisenberg 2007, Table 4. 13 Ibid. 14 U.S. Constitution, Article III; 28 U.S.C. sec. 1330-1333, 1350. 15 Carp, Stidham, and Manning 2007, 67.

4

2.2 The Global Governance Functions of Domestic Courts Domestic court decisionmaking in transnational litigation performs three broad categories of global governance functions: (1) allocation of governance authority, (2) allocation of risks and resources, and (3) governance support. This section explains these functions, which are summarized in Table 1.

[TABLE 1 ABOUT HERE] 2.2.1 Allocation of Governance Authority Who governs transnational activity? This is a fundamental question of global governance.16 But the answer is rarely obvious. Because transnational activity has connections to more than one state, more than one state may have a basis for legitimately exercising the authority to govern it. Insofar as nonstate actors are increasingly claiming authority to govern transnational activity traditionally governed by the state,17 domestic courts are facing new and difficult questions about the allocation of governance authority between state and nonstate actors. And with the spread of international law and organizations, there are equally challenging choices to make about whether they or domestic institutions should govern particular activity. Domestic courts respond to these “who governs” questions by helping to allocate governance authority among states, between state and nonstate actors, and between domestic and international institutions. Two basic types of governance authority are at stake. Dispute resolution authority, which includes adjudicative authority, is the authority to resolve disputes arising from transnational activity.18 Prescriptive authority is the authority to prescribe the rules governing transnational activity. Imagine, for example, a dispute between Party A, a U.S. citizen, and Party B, a State X citizen, arising out of transnational activity with connections to both the United States and State X. Suppose that Party A sues Party B in a U.S. court. Because the choice of dispute resolution forum can strongly influence outcomes,19 Party B may seek to defeat Party A’s choice of forum in various ways. For example, Party B may ask the court to dismiss the case for lack of “personal jurisdiction” over Party B, or for lack of “subject matter jurisdiction” over the parties’ dispute.20 Even if the court finds that it does have jurisdiction, another technique is available: Party B can make a motion asking the U.S. court to dismiss the case in favor of a State X court based on the common law doctrine of forum non conveniens. According to that doctrine, “a federal district court may dismiss an action on the ground that a court abroad is the more appropriate and

16 Kahler 2004, 3; Pollack and Shaffer 2001, 287. 17 Büthe 2004; Coen and Thatcher 2005; Cutler 1999; Cutler 2003; Hall and Biersteker 2002; and Mattli and Büthe 2003. 18 Adjudication refers to dispute resolution in courts. Adjudicative authority is the authority of a state to adjudicate in its courts disputes arising from transnational activity, and is thus one type of dispute resolution authority. 19 Clermont 2005, 183. 20 Federal Rules of Civil Procedure, Rule 12(b)(2) provides for motions to dismiss for lack of jurisdiction over the defendant (that is, for lack of personal jurisdiction), and Rule 12(b)(1) provides for motions to dismiss for lack of subject matter jurisdiction. A defendant may also seek to transfer the case to another forum within the United States under 28 U.S.C. sec. 1404 and 1406.

5

convenient forum for adjudicating the controversy.”21 By denying the motion and keeping the case, the court asserts domestic adjudicative authority over the underlying transnational activity; by granting it and dismissing the case in favor of the State X court, the U.S. court defers to foreign adjudicative authority. Personal jurisdiction and subject matter jurisdiction decisions help allocate adjudicative authority by determining whether or not U.S. authority will be asserted, and they are unilateral in the sense that U.S. courts make them regardless of whether the courts of another state may also have jurisdiction.22 In contrast, the forum non conveniens doctrine explicitly calls for a choice between a domestic court and a foreign court.23 In that sense, forum non conveniens decisions are multilateral, providing more complete answers to the “who governs” question than jurisdiction decisions.24 If the U.S. court denies Party B’s motions to dismiss on jurisdictional or forum non conveniens grounds, and proceeds to adjudicate the dispute, a “who prescribes” question arises: should U.S. law or State X law be applied to govern the transnational activity giving rise to the dispute? First, the court might simply decide whether the United States has prescriptive authority or not.25 The most prominent examples of prescriptive jurisdiction decisions in the U.S. federal courts are decisions about whether U.S. federal statutes and regulations should be applied extraterritorially. These decisions help allocate prescriptive authority among states because they determine whether or not the law of one state—the United States—will be applied to transnational activity. But they leave the second half of the “who prescribes” question unanswered: if not U.S. law, then which state’s law?26 A second way that domestic courts help allocate prescriptive authority is by making international choice-of-law decisions; that is, by deciding whether U.S. law or State X law should apply to activity that has connections to both states. International choice-of-law analysis is closely related to prescriptive jurisdiction analysis: unless a state has prescriptive jurisdiction, its law is not eligible for selection as applicable law.27 21 Sinochem International Co. v. Malaysia International Shipping Corp., 549 U.S. _ (2007) (slip opinion available at http://www.supremecourtus.gov/opinions/06pdf/06-102.pdf). 22 [cite: Michaels; confirm] 23 Michaels 2006, 1036-1035. See also Stein 1985, 786 (describing forum non conveniens as “a means of allocating political authority”); Gulf Oil Corp. v. Gilbert, 330 U.S. 501, 507 (1947) (“the doctrine furnishes criteria for choice” between a domestic forum and a foreign forum); and American Dredging Company v. Miller, 510 U.S. 443, 454 (1994) (describing forum non conveniens as a doctrine for “determining which among various competent courts will decide the case”). Some critics argue that the doctrine is used by courts to dismiss lawsuits by foreign individuals against U.S. businesses. However, even if this criticism is valid, it goes only to the court’s motivations, and does not change the allocative implications of these decisions. 24 That said, domestic courts cannot force the parties to actually take their dispute to the foreign court, nor can it force the foreign court to accept the transnational litigation if the parties do so. Indeed, one study suggests that after U.S. courts dismiss cases on forum non conveniens grounds, they frequently are not re-filed in the proposed alternative forum. Robertson 1987. U.S. courts also may also help allocate dispute resolution authority among states when they make decisions based on the act of state, comity, foreign sovereign immunity, and political question doctrines, and when they grant lis pendens stays (stopping litigation of a transnational dispute in a U.S. court pending resolution of that same dispute in a foreign court) or antisuit injunctions (which purport to bar litigation of a transnational dispute in a foreign court when the same dispute is being litigated in a U.S. court). 25 Born and Rutledge 2007, 613. For an important empirical analysis of U.S. federal court decisions to apply (or not apply) U.S. law extraterritorially, see Putnam 2006. 26 In fact, when U.S. courts decide that U.S. federal law does not apply extraterritorially, they typically proceed to dismiss the lawsuit because, given the inapplicability of the U.S. law, either the plaintiff no longer has a legal basis for its claim or the court no longer has the authority to adjudicate. Born and Rutledge 2007, 674; Lowenfeld 1996, 16-17. 27 Born and Rutledge 2007, 561.

6

But whereas prescriptive jurisdiction analysis is unilateral in the sense that it determines whether or not a given state has authority, international choice-of-law analysis is multilateral or “allocative” in the sense that it determines which state’s law should apply. Nevertheless, as Trachtman puts it, “[t]he durable technical legal questions of choice of law and prescriptive jurisdiction resolve into a core normative public policy issue: how should authority be allocated within an interstate or international system.”28 Domestic courts also allocate governance authority between state and nonstate actors and between domestic and international institutions. Suppose that prior to their dispute, Party A and Party B agreed to submit any transnational disputes that may arise between them to binding arbitration, which is a private method of dispute resolution.29 Such a dispute arises, and Party A sues Party B in a U.S. court, based on its ex post assessment that, given the nature of the dispute, litigating instead of arbitrating will give it a strategic advantage over Party B. Party B then asks the U.S. court to enforce the arbitration agreement. If the court does so, it defers to private dispute resolution authority. In contrast, if it declines to enforce the agreement and resolves the dispute itself, it asserts public governance authority.30 Now suppose that Party B, the State X citizen, is suing Party A in a U.S. federal court for its complicity in human rights violations committed against Party B. Should the U.S. court apply domestic law or international law to determine whether Party B’s rights were violated?31 If the answer is international law, then how much deference should the U.S. court give to prior interpretations of that law by an international court?32 In each of these situations, domestic courts help allocate governance authority by deciding whether to assert U.S. dispute resolution or prescriptive authority over transnational activity and, in some cases, whether to defer explicitly to foreign, private, or international governance authority. 2.2.2 Allocation of Risks and Resources among Transnational Actors Domestic courts also perform a substantive allocative function: they allocate risks and resources among transnational actors. When courts resolve disputes, a typical outcome is an order directing one party to transfer economic resources to another as a remedy for harm caused by the first party’s activity, based on a decision that it is that first party that should bear the risks associated with that activity. It is well understood that the allocation of risks and resources among litigants is a general judicial function.33 However, it is only recently that scholars have started to focus on the implications of this function for transnational activity. For example, “transnational public law litigation” is litigation in which “[p]rivate individuals, government

28 Trachtman 2001, 3. 29 Mattli 2001; Whytock 2008. 30 Under Article II of the United Nations Convention on the Recognition and Enforcement of Foreign Arbitral Awards (the “New York Convention”), domestic courts may decline to enforce arbitration agreements if they find them null and void, inoperative, or incapable of being performed. Under Article V, they may decline to enforce arbitral awards on various grounds, including if doing so would be contrary to domestic public policy. 31 See, for example, [ATS case with choice b/w domestic and international law]. 32 See, for example, Sanchez-Llamas v. Oregon, 548 U.S. 331 (2006) (deciding that an International Court of Justice interpretation of the Vienna Convention on Consular Relations was not binding on the U.S. federal courts). 33 See Shapiro 1972, 413 (adopting a definition of politics as the authoritative allocation of values, and arguing that judicial decisionmaking plays a central role in domestic political processes precisely because law is an important instrument of value allocation).

7

officials, and nations sue one another directly, and are sued directly, in a variety of judicial fora, most prominently domestic courts,” based on rights derived from both domestic and international law.34 In “transnational regulatory litigation,”35 domestic courts apply explicitly regulatory domestic legal rules—such as antitrust rules and securities regulations—to transnational activity, thus allocating risks and resources among the participating actors. “Transnational private litigation” involves claims based on private law, including the law of torts, property, and contracts.36 These fields of private law—although generally not explicitly regulatory—are nevertheless policy instruments by which states authoritatively allocate economic risks and resources.37 By applying private law principles in transnational litigation, domestic courts help implement these allocative policies transnationally. As Robert Wai argues, the function of private law in transnational litigation is “not simply facilitation of transactions, but also compensation for harms and social regulation of transnational conduct.”38 2.2.3 Governance Support Function In addition to the allocative functions discussed above, domestic courts perform a “governance support function.”39 That is, their decisions in transnational litigation can either facilitate or frustrate foreign, private, or international efforts to govern transnational activity. Suppose that Party B sues Party A in a State X court, and the State X court enters a judgment against Party A, ordering it to pay compensation to Party B. Party A refuses to pay, but because Party A neither resides nor has assets within State X territory, the State X court is unable to compel payment. Party B may then ask a U.S. court to enforce the State X judgment against Party A, if necessary by attaching assets owned by Party A in the United States. The U.S. court can support foreign State X’s efforts to govern the transnational activity giving rise to the Party A-Party B dispute by deciding to enforce the State X judgment, or it can frustrate that effort by declining to do so.40 By providing transnational dispute resolution services, domestic courts also help support contracting as a form of transnational private governance. Without third-party dispute 34 Koh 1991, 2348-2349. In the United States, these suits are typically based on the Alien Tort Claims Act (ATCA), 28 U.S.C. sec. 1350 (providing that federal district courts have jurisdiction over any civil action brought by an alien for a tort committed in violation of international law), as interpreted by the Second Circuit of the U.S. Court of Appeals in Filartiga v. Pena-Irala, 630 F.2d 876 (2d Cir. 1980). 35 Buxbaum 2006. 36 Wai 2005a. 37 See Shapiro 1972, 413 (arguing that private law is just as much an instrument of authoritative value allocation as “public” or “regulatory”); and Shapiro 1993, 366 (arguing that “[t]he ‘private’ law of property and contract authoritatively allocates most of the values in a capitalist society.”). 38 Wai 2005b, 471. In addition, at the intersection of transnational public law litigation and transnational private litigation are disputes involving the economic activity of sovereign states, including disputes over sovereign debt (Gulati, Buchheit, and Thompson 2007) and, more generally, transnational litigation in which a foreign sovereign does not enjoy immunity from suit in U.S. courts under the commercial activity exception of the Foreign Sovereign Immunities Act. 28 U.S.C. sec. 1602-1611. 39 Whytock 2006b; Whytock 2007. 40 Refusals to enforce foreign judgments typically are based on a determination by the U.S. court that the foreign court that rendered the judgment did not have proper jurisdiction over the dispute, that the defendant did not enjoy certain minimum standards of fairness in the foreign court, that enforcement of the foreign judgment would violate U.S. public policy, or that the foreign court’s judgment is in some other way defective. Born and Rutledge 2007, 1013-1015.

8

resolution, “the costs of exchanges may be prohibitive, since each prospective party may doubt that the other will abide by promises made over the life of a contract.”41 Similarly, domestic courts can support transnational arbitration as a form of private dispute resolution by enforcing arbitration agreements and arbitral awards—or they can undermine it by declining to do so.42 As Slaughter puts it, “the availability of judicial enforcement—the coercive apparatus of the State—undergirds the entire system of international commercial arbitration.”43 Likewise, the effectiveness of international law and international courts depends largely on domestic courts.44 Domestic courts can support these formal interstate governance arrangements by applying and enforcing international legal rules or by following the interpretations of international courts. By declining to do so, they limit the effectiveness of these arrangements.45 Domestic courts also play a leading role in the development of customary international law, deciding whether particular norms are part of customary international law at all and, if so, explaining their meaning. These decisions are, in turn, acknowledged as “subsidiary means” for determining rules of international law.46 Without downplaying the importance of private and international approaches to global governance, the governance support function implies that domestic courts play a more fundamental role in the regulation of transnational activity. It also highlights what may be one of the most promising avenues of global governance research: the complex relationships between different forms of global governance, public and private, national and international. 2.3 Domestic Courts and the Transnational Shadow of the Law Domestic court decisionmaking in transnational litigation is of obvious importance to the parties to that litigation. But an important finding of social science research on domestic legal systems is that the impact of court decisions extends beyond the parties to particular disputes.47 There is, in other words, a “flow of influence outward from courts to the wider world of disputing and regulating.”48 This phenomenon is commonly known as the “shadow of the law.”49 Likewise, domestic court decisionmaking in transnational litigation affects the behavior of 41 Stone Sweet and Brunell 1998, 64. 42 Whytock 2008. This does not mean that arbitration agreements and arbitral awards necessarily go unheeded without judicial recourse. Rather, because of the strong pro-enforcement policy embodied by U.S. Supreme Court precedents and (although arguably to a lesser extent) the New York Convention, transnational actors expect that domestic courts ordinarily will enforce these agreements and awards, and are therefore more likely to comply with them voluntarily. It is more by creating this knowledge than by providing enforcement in particular cases that domestic courts support arbitration as a system of transnational private governance. Whytock 2008. Under some conditions, however, private arbitration may be able to rely instead on private enforcement based on reputational sanctions. Stone Sweet 2006; Whytock 2008. 43 Slaughter 1995, 519. 44 [cites, e.g. Alter/Helfer/Slaughter] 45 See, for example, Sanchez-Llamas v. Oregon, 548 U.S. 331 (2006) (deciding that an International Court of Justice decision applying the Vienna Convention on Consular Relations [VCCR] was not binding on the U.S. federal courts, and that the Supreme Court lacked supervisory authority to impose upon state courts remedy of suppression of evidence for state police officers’ violations of the VCCR). 46 Article 38(1) of the Statute of the International Court of Justice. 47 Cooter, Marks, and Mnookin 1982; Galanter 1983; Mnookin and Kornhauser 1979; Shapiro 1975, 329. 48 Galanter 1983, 118. 49 Mnookin and Kornhauser 1979; see also Shapiro 1975.

9

transnational actors beyond the litigants. In other words, these actors’ activity occurs in the “transnational shadow of the law.”50 By publishing their transnational litigation decisions, domestic courts provide information about how they are likely to make similar decisions under similar circumstances in the future, thus influencing the strategic behavior of transnational actors who acquire this information.51 For example, decisions allocating adjudicative and prescriptive authority affect transnational forum shopping by litigants.52 They can also affect predictability regarding which state’s rules apply to particular transnational activity, thus influencing levels of compliance, and arguably influencing international regulatory competition as well.53 Marc Galanter refers to “the radiating effects of courts,” explaining that “[t]he contribution of courts to resolving disputes cannot be equated with their resolution of those disputes that are fully adjudicated. The principal contribution of courts to dispute resolution is the provision of a background of norms and procedures, against which negotiations and regulation in both private and governmental settings takes place. This contribution includes, but is not exhausted by, communication to prospective litigants of what might transpire if one of them sought a judicial resolution.”54 The important point here is that the radiating effects of domestic courts extend beyond borders. 3. Law and Politics in Judicial Allocation of Adjudicative Authority: Theory and Hypotheses So far I have argued that domestic courts play an important role in global governance—simply put, domestic courts are global governors. But is this a good thing? Are domestic courts parochial actors that routinely insist on asserting domestic governance authority over transnational activity, or do they have a more cosmopolitan willingness to defer to foreign authority in appropriate circumstances? Are their decisionmaking processes distinctly legal, allowing them to foster transnational rule of law, and making global governance through domestic courts qualitatively different from the more power-oriented mechanisms that are thought to characterize most aspects of world politics? In other words, does law or politics predominate in transnational judicial governance? As a step toward answering these questions, the remainder of this paper examines one transnational judicial governance function—allocation of adjudicative authority among states—with a focus on U.S. district court decisions to grant or deny motions to dismiss transnational litigation in favor of foreign courts on forum non conveniens grounds. More precisely, I ask: How often, and under what circumstances—both legal and political—are U.S. district courts likely to grant forum non conveniens motions, thus deferring to foreign governance authority over the underlying transnational activity, instead of denying them, thus asserting U.S. authority? 50 Whytock 2007, 22-31; Whytock 2008. 51 See Stone Sweet and Brunell 1998, 64 (“When a judge decides, the lawmaking effect of the decision is always twofold. First, in settling the dispute at hand, the judge produces a legal act that is particular (it binds the two disputants) and retrospective (it resolves an existing dispute). Second, in justifying the decision, the judge signals that she will settle similar cases similarly in the future; this legal act is a general and prospective one (it affects future and potential [disputants])”). 52 See Whytock 2007, 25-28. 53 Ibid. 54 Galanter 1983, 121.

10

3.1 Legal Influences Contrary to mainstream political science theories of courts and international relations which emphasize the political determinants of governmental behavior, I argue that legal factors strongly influence domestic court decisionmaking in transnational judicial governance. To explain this influence, I propose a judicial heuristics theory of domestic court decisionmaking.55 3.1.1 A Theory of Judicial Heuristics The theory is based on two assumptions. First, I assume that judges face a combination of heavy caseloads and limited decisionmaking resources that severely limits the amount of time and effort that they can devote to individual decisions. The workload of courts is heavy for all levels of the federal judiciary;56 but unlike the U.S. Supreme Court, district courts are unable to control their caseload. Since 1995 there have been approximately 400 to 500 pending cases per judge in the U.S. district courts each year, each of which may entail many legal decisions.57 Because judicial appointments have not kept pace with growing dockets, “federal judges, at both the trial and circuit court levels, are under severe resource and expertise constraints”; they face “an overwhelming caseload and limited time and resources with which to decide those cases.”58 Simply put, judges have to make a vast number of legal decisions with a limited amount of decisionmaking resources. Second, I assume that a basic operative goal of judges is to make decisions that achieve some minimal acceptable level of legal quality, LQMIN.59 This legal quality assumption has both normative and instrumental foundations. Normatively, as “new institutionalist” scholars of courts argue, judges make decisions in an institutional context that informs the choices they make.60 They “believe that they are required to act in accordance with particular institutional and legal expectations and responsibilities.”61 One of the basic norms constituting the institution of judging is, in Gillman’s words, the obligation “to make the best decisions possible in light of one’s general training” and based on one’s “best understanding of what the law requires,”62 or, as Lindquist and Klein put it, to make “good-faith efforts to find the most persuasive solutions” to legal problems.63

55 For a more detailed statement of the theory, see Whytock 2007, chap. 2. 56 Carp, Stidham, and Manning 2004, 52. 57 The data is from the Federal Court Management Statistics published by the Administrative Office of the U.S. Courts, available at http://www.uscourts.gov/fcmstat/index.html. 58 Bainbridge and Gulati 2002, 86, 102. 59 Baum (1997, 13-14) distinguishes the inherent goals of judges as people from the operative goals of judges that actually affect their judicial behavior. 60 Clayton and Gillman 1999; Gillman and Clayton 1999; Smith 1988. 61 Kahn 1999, 175-176. 62 Gillman 2001, 486. 63 Lindquist and Klein 2006, 136. A central claim of legal realists and their progeny is that legal doctrine is indeterminate: it often is insufficiently precise to justify a unique conclusion and, at an extreme, it may justify any conclusion (Leiter 1996, 265). If legal doctrine were completely indeterminate, then legal quality would be an empty concept: all decisions would be equally legally persuasive. While accepting the limited determinacy of legal doctrine, I assume that it limits the range of decisions that can be legally justified and therefore gives content to the concept of legal quality. As Lindquist and Klein (2006, 138) argue, “[E]ven where various interpretations . . . are possible, some will frequently be more plausible than others—not because the law mandates a particular conclusion,

11

But one risk of these particular formulations of the norm of good faith judging—with their references to the “best decisions possible,” a judge’s “best understanding,” and the “most persuasive solutions”—is that they might be understood as implying that judges seek to maximize the quality of their legal decisions. My first assumption suggests that this cannot be the case. Legal quality is not cost-free: generally speaking, higher legal quality requires greater expenditures of decisionmaking resources. Yet these resources are not unlimited; indeed, they are scarce. Therefore, I assume that judges do not seek to maximize legal quality: they simply do not have the resources to do so.64 Rather, I assume that they seek to achieve some minimal acceptable level of legal quality. Thus, I would modify Gillman’s description of the norm of good faith judging as an “obligation to make the best decisions possible in light of one’s general training and sense of professional obligation”65 by adding “and in light of one’s heavy caseloads and limited decisionmaking resources.” The legal quality assumption also has instrumental foundations. As Baum explains, “Judges spend most of their time interacting with members of their profession and their social groups. . . . Their self-esteem depends heavily on their perception of what these audiences think of them.”66 For this reason, judges seek a good professional reputation in the legal community.67 When judges make decisions that lack a persuasive legal justification, they attract negative attention, and their reputations suffer.68 Their reputations may also suffer if their decisions are reversed by appellate court judges due to legal errors. Based on these assumptions, and building on bounded rationality research, I argue that domestic court decisionmaking in transnational judicial governance is based largely on what I call “judicial heuristics.”69 Bounded rationality research focuses on decisionmaking “within the constraints imposed both by the external situation and by the capacities of the decision maker.”70 The central conclusion of one branch of bounded rationality research is that decisionmakers adapt to the cognitive and environmental constraints of the real world by using decisionmaking shortcuts or rules of thumb called “heuristics.”71 In the language of cognitive psychology, heuristics involve a process of “attribute substitution”: “whenever the aspect of the judgmental but because the methodological and professional conventions of judging make some arguments more persuasive than others.” Thus, as Burton (1992, xii) puts it, the norm of good faith judging does not rely on claims about the determinacy of legal doctrine: “It understands the law as a provider of legal reasons, not necessarily results. It understands the legitimacy of adjudication to depend on respect for the reasons, not agreement with the results, in cases. The good faith thesis claims that judges are bound in law to uphold the conventional law, even when they have discretion, by acting only on reasons warranted by that law as grounds for judicial decision.” 64 I do not mean to suggest that judges will never devote a particularly high level of resources to decisions which they consider particularly important or interesting, in a manner that may appear to approximate maximization of legal quality. I argue only that resource constraints mean that this cannot be done routinely, and that the minimal acceptable legal quality assumption better describes how the average decision is made. 65 Gillman 2001, 486. 66 Baum 2006, 117. 67 Bainbridge and Gulati 2002, 106-108; Baum 1994, 751-753; Posner 1993, 15. 68 Bainbridge and Gulati 2002, 109. As Murphy, Pritchett and Epstein note, “[f]ew jurists have the temerity (or stupidity) to expose themselves to the scorn of fellow judges, scholars, and journalists by announcing decisions for which they cannot give good, even if controversial, reasons” (2002, 312). 69 Whytock 2006b; Whytock 2007. 70 Simon 1985, 294. 71 Gigerenzer and Todd 1999; Kahneman and Frederick 2005.

12

object that one intends to judge (the target attribute) is less readily assessed than a related property that yields a plausible answer (the heuristic attribute),” individuals may substitute the former with the latter.72 In plain English, “difficult questions are often answered by substituting an answer to an easier one.”73 An essential feature of heuristic-based decisionmaking is that it conserves decisionmaking resources: heuristics “often provide an adequate solution cheaply whereas more elaborate approaches would be unduly expensive.”74 Consistent with the premises of bounded rationality, this allows decisionmakers to reach their decisions using realistic amounts of time, information, and computational resources.75 That judges are boundedly rational actors who use heuristics is not a novel proposition.76 But not just any heuristics will do. Judicial heuristics are heuristics that meet two conditions. First, decisions based on the heuristic attributes must entail lower decisionmaking costs than decisions based on the doctrinally-specified target attributes (that is, decisions based on comprehensive analysis of the applicable legal doctrine and the doctrinally relevant facts). This condition follows from the first assumption discussed above and from the very definition of heuristics as decisionmaking processes that conserve decisionmaking resources. For example, say the applicable legal doctrine, Doctrine #1, states that the judge shall defer to foreign governance authority if A=a, B=b, and C=c, and otherwise shall assert domestic governance authority.77 The set of target attributes T is thus defined as T={A, B, C}. Let DCA, DCB, and DCC represent the average decisionmaking costs incurred by the judge to apply target attributes A, B, and C, respectively, such that DCT=DCA+DCB+DCC, representing the total average decisionmaking costs incurred by the judge to make the decision based on Doctrine #1. Let H be the set containing the heuristic attributes, and let DCH represent the total average decisionmaking costs incurred by the judge to make the decision based on the heuristic attributes. The decisionmaking costs condition can then be expressed as DCH<DCT. The second condition follows from the legal quality assumption discussed above: decisions based on the heuristic must achieve some minimal acceptable level of legal quality, LQMIN.78 If LQH is the legal quality of decisions based on the heuristic attribute, then the legal quality condition can then be expressed as LQH>LQMIN.79 To reiterate, the condition is not about maximization of legal quality. The goal of heuristics “is to find a good solution without the

72 Kahneman and Frederick 2005, 269. 73 Ibid. 74 Conlisk 1996, 671. 75 Gigerenzer and Todd 1999, 24. 76 See, for example, Bainbridge and Gulati 2002; Beebe 2006; Gigerenzer 2006, 40; Guthrie, Rachlinski, and Wistrich 2001; and Segal 1986. 77 In this example, the doctrine takes the form of a rule. Alternatively, the legal doctrine might take the form of a standard stating that the judge shall decide whether to defer to foreign governance authority or assert domestic governance authority based on a consideration of factors A, B, and C. 78 Scholars of bounded rationality differ regarding the extent of error costs entailed by the use of heuristics. One line of research focuses on the error costs that heuristics can produce in the form of biases (e.g. Kahneman, Slovic, and Tversky 1982), whereas another has generated models and empirical findings suggesting that heuristics do not necessarily entail significant losses of accuracy (e.g. Gigerenzer and Goldstein 1996; Gigerenzer and Todd 1999). Judicial heuristics, however, must not entail excessive legal error costs, for if they did, the legal quality condition would not be satisfied. 79 This notation is intended to express succinctly the relationship that underlies the legal quality condition, not to imply that either LQH or LQMIN can be easily quantified.

13

fiction of an optimal one,”80 and in this sense judges use judicial heuristics to satisfice, but not maximize, legal quality. There are two basic ways of simultaneously satisfying the decisionmaking costs and legal quality conditions. The first is for the heuristic attributes to be a subset of the doctrinally-specified target attributes, so that H⊂T. I call these “subset heuristics.” Given the set of target attributes T={A, B, C}, as specified by Doctrine #1, plausible sets of heuristic attributes might include H={A} or H={A, B}. In both instances, DCH<DCT because DCA< DCA+DCB< DCA+DCB+DCC. Simply put, decisionmaking costs are saved because a judge using the heuristic analyzes less than all of the doctrinally-specified target attributes. Moreover, since a judge deciding based on the heuristic analyzes some of the same factors that are doctrinally specified, decisions based on the heuristic should be related to decisions based directly on the doctrine; and to the extent of such a relationship, the heuristic-based decision should be legally persuasive. Therefore, in this example it would be plausible to assume that LQH>LQMIN.81 A second basic way of satisfying the two conditions is for the heuristic attributes to be related to one or more target attributes, such that decisions based on the former will be correlated with decisions based on the latter. I call these “correlational heuristics.” For example, H={X} is a plausible correlational heuristic if X∉T; DX< DCT (satisfying the decisionmaking costs condition); and X is correlated with one or more members of T, such that decisions based on the heuristic should be related to decisions based directly on the doctrine, again making it plausible to assume that the legal quality condition is satisfied.82 To be clear, I do not mean to suggest that satisfaction of these conditions demonstrates that judges use a particular heuristic. Rather, these conditions are intended to facilitate ex ante identification of plausible judicial heuristics; whether judges in fact use them can then be investigated empirically.83 Another important point is that judicial heuristics are legal factors. They are heuristics that not only reduce decisionmaking costs relative to comprehensive analysis of the applicable legal doctrine and the doctrinally relevant facts, but also lead to decisions that achieve a minimal acceptable level of legal quality. The resulting decisions must, in other words, be legally persuasive as well as relatively inexpensive. Judicial heuristics accomplish this by specifying heuristic attributes that are a subset of, or correlated with, doctrinally-specified target attributes. This link between legal doctrine and judicial heuristics means that the influences of these heuristics on judicial decisionmaking are legal influences. However, these legal influences are mediated by the realities of judging—including heavy caseloads and limited decisionmaking resources.

80 Gigerenzer 2006, 20. 81 Of course, to assert the plausibility of this assumption falls far short of demonstration. As noted above, the purpose of these conditions is simply to facilitate the ex ante identification of plausible heuristics. 82 In addition to pure subset heuristics and pure correlational heuristics, there may be heuristics that specify a variety of heuristic attributes, some of which are members of T, some of which are correlated with members of T, and some of which are neither members nor correlated, provided the decisionmaking costs and legal error costs conditions are satisfied. 83 In other words, I am not arguing that judicial heuristics can necessarily be deduced (although Gigerenzer and Todd [1999] suggest that this may be possible for some types of heuristics). However, I do suggest that by specifying their conditions, identifying possible judicial heuristics need not be a purely inductive exercise.

14

3.1.2 The Theory Applied: The Case of Forum Non Conveniens Which heuristics might judges use when allocating adjudicative authority among states in the forum non conveniens context? The first step in answering that question is to examine the forum non conveniens doctrine itself. The doctrine specifies three sets of criteria upon which a U.S. district court judge should base his or her choice between assertion of U.S. adjudicative authority and deference to foreign adjudicative authority. First, “[a]t the outset of any forum non conveniens inquiry, the court must determine whether there exists an alternative forum.”84 According to the Supreme Court, “Ordinarily, this requirement will be satisfied when the defendant is ‘amenable to process’ in the other jurisdiction.”85 However, “[i]n rare circumstances . . . where the remedy offered by the other forum is clearly unsatisfactory, the other forum may not be an adequate alternative, and the initial requirement may not be satisfied. Thus, for example, dismissal would not be appropriate where the alternative forum does not permit litigation of the subject matter of the dispute.”86 Second, the district court judge must balance specified private and public interest factors that are intended “[t]o guide trial court discretion.”87 The relevant factors, described by the U.S. Supreme Court in one of its seminal forum non conveniens decisions, are the following:

An interest to be considered, and the one likely to be most pressed, is the private interest of the litigant. Important considerations are the relative ease of access to sources of proof; availability of compulsory process for attendance of unwilling, and the cost of obtaining attendance of willing, witnesses; possibility of view of premises, if view would be appropriate to the action; and all other practical problems that make trial of a case easy, expeditious and inexpensive. There may also be questions as to the enforcibility of a judgment if one is obtained. The court will weigh relative advantages and obstacles to fair trial. It is often said that the plaintiff may not, by choice of an inconvenient forum, “vex,” “harass,” or “oppress” the defendant by inflicting upon him expense or trouble not necessary to his own right to pursue his remedy. But unless the balance is strongly in favor of the defendant, the plaintiff's choice of forum should rarely be disturbed. Factors of public interest also have place in applying the doctrine. Administrative difficulties follow for courts when litigation is piled up in congested centers instead of being handled at its origin. Jury duty is a burden that ought not to be imposed upon the people of a community which has no relation to the litigation. In cases which touch the affairs of many persons, there is reason for holding the trial in their view and reach rather than in remote parts of the country where they can learn of it by report only. There is a local interest in having localized controversies decided at home. There is an appropriateness, too, in having the trial of a diversity case in a forum that is at home with

84 Piper Aircraft Co. v. Reyno, 454 U.S. 235, 255 (1981). 85 Piper, 454 U.S. 235, 255 (1981). 86 Piper, 454 U.S. 235, 254-255 (clarifying that “[w]e do not hold that the possibility of an unfavorable change in law should never be a relevant consideration in a forum non conveniens inquiry. Of course, if the remedy provided by the alternative forum is so clearly inadequate or unsatisfactory that it is no remedy at all, the unfavorable change in law may be given substantial weight; the district court may conclude that dismissal would not be in the interests of justice”). 87 Piper, 454 U.S. 235, 241.

15

the state law that must govern the case, rather than having a court in some other forum untangle problems in conflict of laws, and in law foreign to itself.88

Finally, the district court must consider the nationality of the plaintiff: “[T]here is ordinarily a strong presumption in favor of the plaintiff’s choice of forum, which may be overcome only when the private and public interest factors clearly point towards trial in the alternative forum.”89 However, “the presumption applies with less force when the plaintiff . . . is foreign.”90 According to the Supreme Court, “a plaintiff’s choice of forum is entitled to greater deference when the plaintiff has chosen the home forum. When the home forum has been chosen, it is reasonable to assume that this choice is convenient.”91 In contrast, “When the plaintiff is foreign . . . this assumption is much less reasonable. Because the central purpose of any forum non conveniens inquiry is to ensure that the trial is convenient, a foreign plaintiff’s choice deserves less deference.”92 Two judicial heuristics are plausible in the forum non conveniens context, one subset heuristic and one correlational heuristic. The hypothesized subset heuristic is the plaintiff’s nationality: HPNAT={U.S. plaintiff}, with this attribute weighing against dismissal if the plaintiff is a U.S. citizen. The forum non conveniens doctrine specifies the following target attributes: the availability of an alternative forum, private interest factors (balance relative ease of access to sources of proof, relative ease of access to witnesses, enforceability of a judgment if one is obtained), public interest factors (balance extent of administrative difficulties from court congestion, burden on juries, local interest in localized controversies, and familiarity with governing law), and the plaintiff’s nationality: TFNC={alternative forum, private interest factors (proof, witnesses, enforceability, balance), public interest factors (congestion, jury burden, local interest, governing law, balance), plaintiff’s nationality}. Thus, HPNAT⊂T. It is therefore plausible to expect that decisions based on the plaintiff’s nationality heuristic will achieve a minimum level of legal quality—that is, that LQPNAT>LQMIN. Moreover, because decisions based on the plaintiff’s nationality rather than the full set of target attributes allow judges to expend fewer decisionmaking resources, DCPNAT<DCT. Thus, the U.S. plaintiff heuristic appears to satisfy the requirements for subset heuristics.93 This implies the following hypothesis:

H1 (Judicial Heuristic: U.S. Plaintiff): A judge is less likely to grant a motion to dismiss transnational litigation in favor of a foreign court on forum non conveniens grounds, thus deferring to foreign adjudicative authority, when there is a U.S. plaintiff than when there is not.

88 Gulf Oil Corp. v. Gilbert, 330 U.S. 501, 508-509 (1947) (footnotes omitted). See also Piper, 454 U.S. 241, note 6 (reiterating these factors). 89 454 U.S. 235, 255-256. 90 454 U.S. 235, 256. 91 Ibid. 92 Ibid. 93 The other doctrinally-specified factors are less plausible as subset heuristics because they do not exist independently from the judge’s analysis. That is, they are not strictly factual. Rather, they involve subsidiary judgments about the legal implications of various facts. For example, whereas the plaintiff’s nationality usually is a simple yes or no question, the availability of a foreign forum, the adequacy of the remedy it provides, ease of access to evidence, whether a judgment will be enforceable, administrative difficulties, jury burden, and familiarity with the applicable substantive law are not mere facts, but criteria for evaluating facts that are not specified in the doctrine itself. Therefore, these factors are likely to demand substantially greater decisionmaking resources.

16

The hypothesized correlational heuristic is the locus of the transnational activity giving rise to the litigation—HTERR={territoriality}—with this attribute weighing in favor of dismissal if that activity occurred mostly or all outside U.S. territory. The forum non conveniens doctrine does not explicitly specify territoriality as a target attribute (HTERR∉TFNC). However, it is nevertheless plausible to expect that decisions based on the territoriality heuristic generally will be legally persuasive, thus achieving a minimum level of legal quality—that is, that LQTERR>LQMIN. The forum non conveniens doctrine specifies a number of target attributes that are likely to be correlated with the locus of the activity. For example, ease of access to proof and witnesses, which is among the doctrine’s private interest factors, depends on the location of relevant sources of evidence, which in turn will often overlap with the place of the activity giving rise to the litigation. In addition, one of the public interest factors is whether the underlying dispute is a “localized controversy.” The meaning of the phrase is not defined, but it implies that one of the parties or some part of the underlying activity is local.94 Moreover, it would seem that DCTERR<DCFNC—that is, it should be less costly in terms of time and effort for a judge to make decisions based on the territoriality heuristic than on the full range of target attributes specified by the forum non conveniens doctrine. Thus, territoriality is a plausible correlational heuristic. This implies the following hypothesis:

H2 (Judicial Heuristic: Territoriality): A judge is more likely to grant a motion to dismiss a case in favor of a foreign court on forum non conveniens grounds, thus deferring to foreign adjudicative authority, when the activity giving rise to the litigation occurred mostly or all outside U.S. territory than otherwise.

3.2 Political Influences The dominant theory of judicial decisionmaking in the political science literature on courts is known as the “attitudinal model.”95 According to the attitudinal model, the primary operative goal of judges is for their decisions to reflect their personal policy preferences.96 As summarized by its leading defenders, “This model holds that the Supreme Court decides disputes in light of the facts of the case vis-à-vis the ideological attitudes and values of the justices. Simply put, Rehnquist votes the way he does because he is extremely conservative; Marshall voted the way he did because he was extremely liberal.”97 Although more complex methods have been used, a standard way of operationalizing a judge’s ideological position is to use the political party of the president who appointed him or her—Democrat or Republican—as a proxy.98

94 The concept of territoriality also is, to use the language of heuristics research, “accessible” (Kahneman and Frederick 2005, 270-271)—that is, it should easily come to judges’ minds as they make decisions involving the allocation of adjudicative authority. Concepts of territoriality and nationality have played an important role in the historical development of doctrines of adjudicative jurisdiction. See, e.g., Michaels (2004, 105) (“[t]erritory has long shaped our thinking about adjudicative jurisdiction”); Michaels (2006b, 1057) (“[t]erritoriality is still central to jurisdictional thinking”). This accessibility reinforces the plausibility of territoriality as a judicial heuristic. 95 Rohde and Spaeth 1976; Spaeth 1995; Segal and Spaeth 2002. 96 Spaeth 1995, 305. 97 Segal and Spaeth 2002, 86. 98 George and Epstein 1992, 328. See also George 1998, 1650-1655 (defending this approach, and surveying alternatives).

17

The basic implication is that judges’ ideological attitudes will affect the likelihood that they will grant motions to dismiss transnational litigation in favor of foreign courts on forum non conveniens grounds, thus deferring to foreign adjudicative authority. Based on recent survey work conducted by the Pew Research Center for the People and the Press, it is plausible to hypothesize that Republicans generally will be less likely than Democrats to defer to foreign authority. Republicans overall were more likely than Democrats to agree that the United States should follow its own national interests even when its allies strongly disagree and less likely than Democrats to agree that U.S. foreign policy should take into account the interests and views of allies.99 If Republicans are less likely to consider the interests of other states, one might reasonably expect them to be less likely to defer to the governance authority of other states. Moreover, Republicans were less likely than Democrats to agree that “[t]he U.S. should mind its own business internationally and let other countries get along the best they can on their own.”100 Thus, one might reasonably expect Republicans to be more likely to assert U.S. authority over transnational activity. This implies the following hypothesis:

H3 (Attitudinal Model): Republican judges are less likely than Democratic judges to defer to foreign governance authority.

The strategic model of judicial decisionmaking shares the attitudinal model’s basic assumption that “justices, first and foremost, wish to see their policy preferences etched into law. They are . . . ‘single-minded seekers of legal policy.’”101 However, in contrast to the attitudinal model’s emphasis on judges’ freedom to make decisions based on their personal preferences, the strategic model emphasizes the constraints that judges face. To maximize their preferences, judges must act strategically; that is, they must consider not only their own preferences, but also the preferences and actions of other political actors such as Congress, the president, the public, and even other judges, with which the achievement of justices’ policy goals are interdependent.102 Therefore, “the choices of justices can best be explained as strategic behavior, not solely as responses to either personal ideology or apolitical jurisprudence.”103 An important implication of the strategic model is that the political environment has an influence on judicial decisionmaking. Congress and the president have a “vast array of powers over matters important to the Court.”104 For example, Congress can propose constitutional amendments or pass legislation overriding court decisions. It also can use judicial salaries, impeachment, or legislative limits on jurisdiction to “punish” justices for their decisions—and although this rarely occurs, the possibility of such measures gives justices an incentive to consider the policy preferences of the other branches.105 In addition, “government actors can refuse, implicitly or explicitly, to implement particular constitutional decisions, thereby

99 Pew Research Center for the People and the Press 2005a, 21. 100 Pew Research Center for the People and the Press 2005b, 42. 101 Epstein and Knight 1998, 9-10. The strategic model of judicial decisionmaking discussed here was developed specifically to explain the behavior of judges. Therefore, it is distinct from strategic theories of international relations, which generally seek to explain the behavior of states as unitary actors (for example, Powell 1999). 102 Epstein and Knight 1998, 12. 103 Epstein and Knight 1998, xiii. 104 George and Epstein 1992, 325. 105 Epstein and Knight 1998, 141-143.

18

decreasing the Court’s ability to create efficacious policy.”106 Thus, the strategic model suggests that “Congress and the president affect Supreme Court decision making and that the direction of their influence reflects their partisan composition” (George and Epstein 1992, 325). This implies the following hypothesis:

H4 (Strategic Model): Judges are less likely to defer to foreign governance authority in a Republican political environment than otherwise.

Theories of international relations reinforce the attitudinal and strategic models’ emphasis on the political determinants of transnational judicial governance, but with a focus on power rather than ideology. The concept of power occupies a central place in international relations scholarship.107 Because realist theory generally treats states as unitary actors, it is not clear from a realist perspective how national power should affect judicial allocation of governance authority. However, as Alter notes, some scholars of the role of domestic courts in European integration argue that the national interest is the primary determinant of domestic court decisionmaking.108 If this is the case, then insofar as national interests are defined in terms of power, power should influence judges’ decisions. As a general matter, the realist emphasis on power suggests that the likelihood of deference by one political actor to another should be greater when the second is more powerful than the first. Guzman and Simmons explore this basic logic to explain the probability that a state will file complaints with the World Trade Organization’s dispute resolution system.109 They hypothesize that the political costs of filing a complaint are a function of the difference in power between the state filing the complaint and the state against which it is filed.110 The greater the power differential,” they explain, “the greater the ability of the more powerful state to impose costs on the less powerful state . . . . The notion then is that less powerful states are reluctant to challenge more powerful states for fear of retaliation or retribution.”111 The same logic suggests that the more powerful a foreign state is, the more able it is to retaliate against a domestic state that fails to defer to the foreign state’s governance authority.112 This suggests the following hypothesis:

106 Epstein and Knight 1998, 144. 107 Baldwin 2002. Indeed, a focus on power is one of the defining characteristics of realist theories of international relations. Morgenthau (1978, 5) assumes that “statesmen think and act in terms of interest defined as power”; Waltz (1979, 97-98) argues that the distribution of power among states is a fundamental structural variable affecting state behavior; Grieco (1988, 488) emphasizes that “states in anarchy are preoccupied with power and security”; and Mearsheimer (2001, 2) posits that “[t]he overriding goal of each state is to maximize its share of world power.” 108 Alter 2001, 40. 109 Guzman and Simmons 2005. Although neither Guzman nor Simmons identify themselves as “structural realists,” this hypothesis follows from structural realist premises and shares with realism more generally an emphasis on the role of power. 110 Ibid., 569. However, they do not find empirical support for this hypothesis. 111 Ibid. 112 Weaker states can still harm stronger states. Therefore, this logic holds even if the foreign state is weaker than the domestic state: the greater the power of the foreign state, the smaller the power advantage of the domestic state, and the more able the foreign state is to effectively retaliate against the domestic state (even if the domestic state is relatively strong).

19

H5 (Realist IR Theory): The greater the power of a foreign state, the more likely a domestic judge is to defer to its governance authority.

Beyond attitudes and power, Slaughter’s liberal theory of international law suggests that a foreign state’s regime type may influence how domestic courts allocate governance authority.113 Building on liberal international relations theory,114 Slaughter argues that “the distinctive nature of political-military relations among liberal states has an analog in the legal relations among liberal states.”115 More precisely, she argues “that courts of liberal states handle cases involving other liberal states differently from the way they handle cases involving nonliberal states.”116 According to Slaughter’s theory, legal relations among liberal states are based on principles of tolerance and mutual accommodation. Because they share basic core values and institutions, liberal states inhabit what Slaughter refers to as a “zone of legitimate difference,” in which they generally “can disagree with the specific policy choices embedded in each other’s national laws but nevertheless respect those laws as legitimate means to the same ultimate ends.”117 In contrast, disagreements between liberal and nonliberal states are likely to fall outside the zone of legitimate difference because of their fundamentally different values and institutions.118 In addition, because courts in liberal states are unlikely to find similarly professional and independent courts in nonliberal states, there is little possibility of developing the sort of “reciprocal dialogue” that exists, at least figuratively, among the courts of liberal states.119 Liberal international law theory thus suggests the following hypothesis:

H6 (Liberal Theory): Domestic judges in liberal democracies are more likely to defer to the governance authority of a foreign state if the foreign state also is a liberal democracy.

Finally, nationality is a political factor that may influence how domestic courts allocate governance authority. Finding that U.S. parties are substantially more likely to prevail in patent litigation in U.S. courts than foreign parties, Moore concludes that U.S. judges and juries exhibit a “xenophobic bias.”120 Similarly, Bhattacharya, Galpin, and Haslem find that U.S. corporate defendants are less likely to lose than foreign corporate defendants in U.S. federal courts.121 In contrast, Clermont and Eisenberg identify no such pro-domestic party bias. In fact, they provide evidence that foreign parties fare better than U.S. parties when they litigate in the United States: there is, according to them, xenophilia rather than xenophobia in U.S. courts.122 Although the 113 Slaughter 1995, 2000. See also Burley 1992 and Slaughter Burley 1993, 232. 114 Moravcsik 1997. 115 Burley 1992, 1909. 116 Burley 1992, 1917. See also Slaughter 2000, 249 (insisting that “differences in domestic regime type drive differences in positive behavior”). 117 Burley 1992, 1919. On the other hand, liberal states also “recognize each other’s right to reject such laws when they transgress these common values.” Ibid. Slaughter explicitly suggests that the differential treatment by liberal states’ courts of other liberal states and nonliberal states should manifest itself in both international choice of law and forum non conveniens decisionmaking. 118 Burley 1992, 1920-1921. 119 Burley 1992, 1921. 120 Moore 2003. 121 Bhattacharya, Galpin, and Haslem 2006. 122 Clermont and Eisenberg 1996; Clermont and Eisenberg 2007.

20

empirical evidence is mixed and the theoretical basis for bias underdeveloped, the intuition that domestic judges may favor their fellow citizens is strong, and the normative implications of bias are important.123 Therefore, I take seriously the possibility of pro-domestic-party bias. This suggests the following hypothesis:

H7 (Pro-Domestic-Party Bias): When a U.S. party and a non-U.S. party disagree about the allocation of governance authority, a U.S. judge is more likely to decide in favor of the U.S. party than the non-U.S. party.124

4. Research Design To test these hypotheses, I created a random sample of published forum non conveniens decisions made by U.S. district court judges, and analyzed it using logit analysis and content analysis.125 The sample was generated in three steps. First, I searched the LexisNexis Academic database of U.S. district court decisions for the term “forum non conveniens” between 1990 and 2005.126 Second, I randomly sorted the results. Third, I analyzed each case in the randomly generated order, discarding those that were not actual decisions by U.S. district court judges to either grant or deny a motion to dismiss in favor of a foreign court on forum non conveniens grounds, until I had a sample of 210 decisions.127 To indicate whether each decision in the sample is a decision to defer to foreign adjudicative authority, I created my dependent variable, “Deference to Foreign Adjudicative Authority,” coding it as 1 if the decision was to grant the

123 Clermont and Eisenberg 1996, 1120-1121; Clermont and Eisenberg 2007, 441. 124 For example, a judge would be more likely to defer to foreign governance authority if the U.S. party prefers foreign authority, and more likely to assert domestic governance authority if the U.S. party prefers domestic authority. 125 My reliance on published cases (that is, those published in the Federal Supplement and/or electronically in the LexisNexis database) creates the possibility of selection bias, because unpublished cases may be systematically different from published cases (Clermont and Eisenberg 2002, 125-126). The implications are different for descriptive inferences and causal inferences. Regarding descriptive inferences, the risk is that my estimate of the rate of deference to foreign adjudicative authority may be greater (or less than) the rate in unpublished cases. Regarding causal inferences, selection bias—understood as an omitted variable problem (Heckman 1979, 154-155)—is a threat only if a determinant of judges’ publication decisions both (1) has a causal effect on forum non conveniens decisionmaking and (2) is correlated with one or more of my included independent variables (King, Keohane, and Verba 1994, 169). My analysis of these risks suggests that the threat to descriptive inferences could cut either way, and that the risk that publication decisions introduce selection bias to my causal inferences is low (Whytock 2007). Nevertheless, one must use caution when extending this descriptive finding from the random sample of published decisions upon which it is based to the broader population of published and unpublished opinions. From a governance-oriented perspective, however, such an extension is not essential: published cases are of the greatest importance, for these cases (unlike unpublished cases) provide information that can influence the strategic behavior of transnational actors beyond the parties to particular disputes. Published decisions may only part of the picture, but from this perspective, they are the most important part of the picture. See Olson (1992, 795) (noting that despite the limitations of reliance on only published decisions, such an approach “may be most justifiable for a researcher who is trying to study efficiently the ‘public policy’ output of district courts”). 126 The LexisNexis database includes all decisions published in the Federal Supplement reporter as well as decisions not published in the Federal Supplement. The restriction of the analysis to the post-1990 period helps control for the effect that post-Cold War changes to the international system may have had on judicial allocation of governance authority; and, in the interest of policy relevance, it places an emphasis on contemporary judicial behavior. 127 To obtain this number, I had to screen 848 cases out of a total of 2,791 hits for the search term “forum non conveniens.” Thus, I estimate that there are a total of 691 actual forum non conveniens decisions in the LexisNexis database between 1990 and 2005, with my dataset constituting 30% of those decisions.

21

motion (thus dismissing the case in favor of a foreign court and deferring to foreign adjudicative authority), and 0 if the decision was to deny the motion (thus keeping the case and asserting domestic adjudicative authority). To test the Judicial Heuristic: U.S. Plaintiff hypothesis (H1), I created the variable “U.S. Plaintiff,” indicating whether there is a U.S. plaintiff in the litigation (0 if no, 1 if yes).128 According to this hypothesis, the coefficient for U.S. Plaintiff should be statistically significant and have a negative sign, indicating that a judge is less likely to dismiss a case in favor of a foreign court when there is a U.S. plaintiff. To test the Judicial Heuristic: Territoriality hypotheses (H2), I created the variable “Activity Mostly/All Outside U.S. Territory,” indicating whether the activity giving rise to the litigation occurred mostly or all outside U.S. territory (0 if no, 1 if yes). I coded the variable based on two factors: the place of defendant’s conduct and the place of the plaintiff’s alleged injury, each as stated in the judge’s opinion. I coded it as “yes” if both the place of conduct and the place of injury were entirely foreign, if the place of conduct was entirely foreign and the place of injury mixed, or if the place of injury was entirely foreign and the place of conduct mixed. When the place of conduct and place of injury were both mixed or when either the place of conduct or the place of injury was entirely domestic, I coded Activity Mostly/All Outside U.S. Territory as “no.” According to the territoriality heuristic hypothesis, the coefficient for the variable should be statistically significant and have a positive sign, indicating that a judge is more likely to dismiss a case in favor of a foreign court when the activity giving rise to the litigation occurred mostly or all outside U.S. territory. To test the attitudinal model hypothesis (H3), I created the variable “Republican Judge,” which indicates whether the judge’s party affiliation is Republican (0 if no, 1 if yes). The use of party affiliation as a measure of a judge’s ideology follows George and Epstein.129 According to this hypothesis, the coefficient for Republican Judge should be statistically significant and have a negative sign, indicating that a Republican judge is less likely to dismiss a case in favor of a foreign court than a Democratic judge. To test the strategic model hypothesis (H4), I created the variable “Republican Environment,” which indicates whether, in the year of the court’s decision, the political environment is Republican (0 if no, 1 if yes). I coded the political environment as Republican if the president and at least one house of Congress were Republican.130 According to

128 I coded U.S. Plaintiff based on the citizenship of the plaintiffs indicated in the judge’s opinion or, in the absence of citizenship information, based on residence, domicile, or jurisdiction of incorporation. I counted only parties that joined in or opposed the forum non conveniens motion. 129 George and Epstein 1992, 328; George 1998, 1651-1652. To create the variable, I began by using the political party of the nominating president as a proxy for the judge’s political party, again following George and Epstein (1992, 328) and George (1998, 1651-1652), using data from the Federal Judicial Center’s Biographical Directory of Federal Judges available at http://www.fjc.gov/public/home.nsf/hisj. I then made corrections when Martinek’s (2005) Lower Federal Court Confirmation Database contained data on a judge’s personal partisan identification that was different from that of the nominating president, which was the case for 18 observations in my dataset. For a defense of the use of the nominating president’s party as a proxy for the judge’s partisan affiliation, and a survey of alternatives, see George (1998, 1650-1655). 130 Thus, Republican Environment equals 0 if the president was Democratic or both houses of Congress were Democratic. This approach is based on George and Epstein’s (1992, 328) method of measuring the political environment, whereby their variable equals 0 if both the president and Congress are Democratic, 1 if one branch is Republican and the other Democratic, and 2 if the Senate and the president are Republican. However, I use a dummy variable rather than their three-level measure to help ensure adequate cell size for the logit analysis. For simplification and to account for lame-duck status, I did not account for January when coding presidential transition

22

this hypothesis, the coefficient for Republican Environment should be statistically significant and have a negative sign, indicating that a judge is less likely to dismiss a case in favor of a foreign court in a Republican political environment. To test the realist theory hypothesis (H5), I created the variable, “Foreign State’s GDP (Log),” which equals the natural logarithm of the foreign state’s annual gross domestic product.131 The use of GDP as a measurement of power follows Guzman and Simmons.132 According to this hypothesis, the coefficient for Foreign State’s GDP (Log) should be statistically significant and have a positive effect, indicating that the greater the power of a foreign state, the more likely a judge is to dismiss a case in favor of a court of that state.133 To test the liberal theory hypothesis (H6), I created the variable “Foreign State Liberal Democracy,” which indicates whether the foreign state was rated “Free” by the Freedom House Freedom in the World Survey for the year prior to the court’s decision (0 if no, 1 if yes).134 Finally, to test the pro-domestic-party bias hypothesis (H7), I used the previously described U.S. Plaintiff variable and I created the variable “U.S. Defendant,” indicating whether there is a U.S. defendant in the litigation (0 if no, 1 if yes).135 According to this hypothesis, the coefficient for U.S. Plaintiff should be statistically significant and have a negative sign, indicating that a judge is less likely to dismiss a case in favor of a foreign court if there is a U.S. plaintiff; and the coefficient for U.S. Defendant should be statistically significant and have a positive sign, indicating that a judge is more likely to dismiss a case in favor of a foreign court if

years. For the Senate, I coded 2001 and 2002 as Democratic, because Democrats were in the majority for 8 of 12 months in 2001 and 11 of 12 months in 2002. 131 I use data for the year 2000 and apply a natural logarithmic transformation to improve symmetry. The data source is the World Bank’s World Development Indicators, available at http://devdata.worldbank.org/data-query/. 132 Guzman and Simmons 2005, 574. 133 As an alternative measure, I created the variable “Foreign State’s Military Spending (Log),” which equals the log of the foreign state’s total annual military spending. I again use data for the year 2000 and apply a natural logarithmic transformation to improve symmetry. The data source is again the World Bank’s World Development Indicators. 134 Although these ratings generally are stable over time during the period covered by my dataset, I use a one-year lag on the theory that a boundedly rational judge is unlikely to become immediately aware of even substantial changes in a foreign state’s politics. As an alternative measure, I created the variable “Foreign State Liberal Democracy (Polity IV),” which indicates whether the foreign state had a Polity IV score of 5 or higher for the year prior to the court’s decision (0 if no, 1 if yes). However, for purposes liberal international law theory, Slaughter defines “liberal” states as those “with juridical equality, constitutional protections of individual rights, representative republican governments, and market economies based on private property rights” (Burley 1992, 1909). The Freedom House rankings capture these characteristics by explicitly accounting for equal treatment under the law, protection of individual rights (including freedoms of assembly, open public discussion, and defendants’ rights), representative government (including election of representatives through free and fair elections and other political rights), and private property rights (Freedom House 2006). Polity IV captures the concept of representative government well, but only indirectly captures the other elements of Slaughter’s definition. Therefore, for testing Slaughter’s liberal theory of international law, the Freedom House measure appears more appropriate. 135 I coded U.S. Defendant based on the citizenship of the defendants indicated in the judge’s opinion or, following the same convention used for coding U.S. Plaintiff, in the absence of citizenship information, based on residence, domicile, or jurisdiction of incorporation. I counted only parties that joined in or opposed the forum non conveniens motion.

23

there is a U.S. defendant.136 In contrast, the Judicial Heuristic: U.S. Plaintiff hypothesis (H1) would expect only the U.S. Plaintiff variable to be statistically significant (a bias in favor of U.S. plaintiffs is consistent with the forum non conveniens doctrine, but a general bias in favor of U.S. parties is not). Summary statistics and a correlation table are presented in the Appendix. 5. Results U.S. district court judges defer to foreign adjudicative authority at an estimated rate of 47.1 percent in published forum non conveniens decisions, with 95 percent confidence that the actual rate is between 40.5 and 53.9 percent. Far from indicating that U.S. district courts are parochial actors that routinely insist on asserting U.S. governance authority over transnational activity, this evidence suggests that they may be somewhat cosmopolitan in their outlook, willing to defer to foreign governance authority quite frequently. But under what circumstances are they most likely to do so? What, in other words, are the determinants of judicial allocation of adjudicative authority? And are these determinants predominantly legal or political? Using logit analysis and content analysis to test the hypotheses developed in Section 3, this section proposes some answers. 5.1 Logit Analysis The results of the logit analysis are presented in Table 2. Model 1 is a “legal” model of transnational judicial governance, including the U.S. Plaintiff and Activity Mostly/All Outside U.S. Territory variables to test whether the use of judicial heuristics might be influencing judicial allocation of adjudicative authority. Consistent with the judicial heuristics hypotheses, the coefficients for Plaintiff’s Nationality (H1) and Activity Mostly/All Outside U.S. Territory (H2) have the expected signs and are highly statistically significant (p=.001 and p<.001, respectively): the likelihood of deference is lower when there is a U.S. plaintiff and higher when the activity giving rise to the litigation occurred mostly or all outside U.S. territory.

[TABLE 2 ABOUT HERE] Model 2 is a “political” model of transnational judicial governance, including Republican Judge, Republican Environment, Foreign State GDP, Foreign State Liberal Democracy, U.S. Plaintiff, and U.S. Defendant. The results are mixed. On the one hand, they support the liberal theory hypothesis (H6): the positive sign of the coefficient for Foreign State Liberal Democracy indicates that the probability of deference to foreign adjudicative authority is higher when the foreign state is a liberal democracy, and the effect is highly statistically significant (p=.004).137 136 Recall that in the forum non conveniens context, dismissal in favor of a foreign court necessarily favors the defendant who has moved to dismiss the case on forum non conveniens grounds, and disfavors the plaintiff who initially chose the U.S. forum. 137 It is possible that the statistically significant effect of Foreign State Liberal Democracy may be evidence that judges are engaging in the available and adequate alternative forum analysis called for by the forum non conveniens doctrine—but for the following reasons, this is unlikely. First, the availability prong of the analysis turns strictly on whether the defendant is amenable to process in the foreign state (Piper, 454 U.S. 235, 255), which has no obvious relationship to whether the foreign state is a liberal democracy. Likewise, the adequacy prong of the analysis depends on whether there is a legal remedy for the plaintiff in the proposed foreign forum, and is strictly interpreted to weigh against dismissal only “[i]n rare circumstances . . . where the remedy offered by the other forum is clearly

24

On the other hand, the results do not support the attitudinal model (H3) and realist theory (H5) hypotheses: the coefficients for Republican Judge and Foreign State GDP are not statistically significant at traditionally required levels (p=.140 and p=.508, respectively).138 Nor do they support the pro-domestic-party bias hypothesis (H7). If there were such bias, one would expect not only the presence of a U.S. plaintiff to decrease the likelihood of dismissal (as the forum non conveniens doctrine itself provides), but also the presence of a U.S. defendant to increase the likelihood of dismissal. Contrary to this expectation, the coefficient for U.S. Defendant is not statistically significant (p=.295) and in any event has a negative sign. Finally, the implications for the strategic model hypothesis (H4) are unclear: the negative sign of the coefficient for Republican Environment suggests that a Republican political environment may decrease the probability of deference, but this effect is not quite significant at traditionally required levels of confidence (p=.061). Model 3 combines the legal and political factors, and confirms the results of Models 1 and 2. Regarding legal influences, even after controlling for the various political factors, the coefficients for Plaintiff’s Nationality (H1) and Activity Mostly/All Outside U.S. Territory (H2) continue to have the expected signs and are highly statistically significant (p<.001). Regarding political influences, the results are again mixed. The coefficient for Foreign State Liberal Democracy (H5) remains positive and highly statistically significant even after controlling for legal factors (p=.002). On the other hand, the coefficients for Republican Judge, Foreign State GDP, and U.S. Defendant still are not statistically significant at traditionally acceptable levels, and Republican Environment is again on the borderline (p=.092).139 To illustrate the substantive effects of the statistically significant independent variables in Model 3 (plus Republican Environment), Table 3 provides estimates of the impact of a change in each of these variables on the probability of deference to foreign adjudicative authority.140 U.S. district court judges are an estimated 24 percent less likely to defer to foreign adjudicative

unsatisfactory,” that is, “so clearly inadequate . . . that it is no remedy at all” (Piper, 454 U.S. 235, 254-255). Finally, in the 22 cases in the forum non conveniens sample in which a judge concluded that the alternative forum requirement was not satisfied, the foreign state was a liberal democracy in 10 and not a liberal democracy in 12 (as measured by the Foreign State Liberal Democracy Variable), which suggests that judges do not use regime type as a proxy for forum adequacy. 138 Note, however, that the Republican Judge variable can only test whether the allocation of governance authority has an intrinsic left-right (Democratic-Republican) ideological dimension. It may miss significant attitudinal effects that manifest themselves along the left-right dimension of policy issues that are affected by the choice between domestic and foreign governance authority in particular types of cases. For example, conservatives and liberals might have different preferences regarding the assertion of U.S. governance authority depending on the nature of the underlying transnational activity. Similarly, they may have different preferences depending on whether deference to foreign authority in a particular case would benefit a plaintiff or a defendant, or an individual versus a business, or further policies relating to government regulation, consumer protection, or other policy areas with an identifiable left-right dimension. Therefore, my approach can shed light only on the extent to which the attitudes of judges affect the allocation of governance authority in the aggregate, and not on the more subtle attitudinal effects that might be at play in transnational judicial governance. 139 Note, however, that when all independent variables that are not statistically significant at a 95 percent confidence level or higher, except for Republican Environment, are dropped from Model 3, Republican Environment loses statistical significance altogether (p=.173). 140 I used the Clarify software program (Tomz, Wittenberg, and King 2001) to simulate a change in the expected value of the dependent variable caused by increasing each independent variable from 0 to 1, setting each of the other variables at their mode (for dummy variables) or mean (for continuous variables), and based on Model 3.

25

authority when there is a U.S. plaintiff than otherwise; 39 percent more likely to do so when the activity giving rise to the litigation occurred mostly or all outside U.S. territory; and 32 percent more likely to do so when the foreign state is a liberal democracy. Deference is an estimated 15 percent less likely when the domestic political environment is Republican, but the 95 percent confidence interval for Republican Environment includes zero, indicating that we cannot be 95 percent certain that this variable has the hypothesized negative effect on the probability of deference.

[TABLE 3 ABOUT HERE] In summary, the results of the logit analysis suggest that judicial allocation of adjudicative authority in the forum non conveniens context is influenced by both legal factors, including the hypothesized judicial heuristics, and a political factor, whether the foreign state is a liberal democracy. Another political factor—the domestic political environment—may also play a role.141 The measures of fit reported in Table 2—which include the adjusted count R-squared, the percent correctly classified, and the area under the ROC curve142—all suggest that the legal model is better able to predict deference to foreign adjudicative authority than the political model, but the combined model is better yet: it correctly classifies 73 percent of the decisions, which is a 42 percent improvement over the number that would be correctly predicted simply by choosing the modal outcome. 5.2 Content Analysis Do the effects of the U.S. Plaintiff and Activity Mostly/All Outside U.S. Territory variables reflect a transnational judicial governance decisionmaking process based on judicial

141 To verify the results, I re-analyzed Model 3 with a variety of modifications. To check the robustness of the effect of the foreign state’s status as a liberal democracy, I replaced the Foreign State Liberal Democracy variable, which is based on Freedom House measures, with the Foreign State Liberal Democracy (Polity IV) variable. The effect of liberal democracy remained statistically significant (p=.036). To double-check the result that the foreign state’s power does not affect the likelihood of deference, I replaced Foreign State GDP (Log) with Foreign State Military Spending (Log). The new variable was not statistically significant (p=.928). The correlations between democracy on the one hand, and Foreign State GDP (Log) and Foreign State Military Spending (Log) on the other hand, raises the possibility that multicollinearity is causing these variables to lack statistical significance. To check, I repeated the Model 3 logit analysis without the Foreign State Liberal Democracy variable. Neither measure of foreign state power became statistically significant as a result of this modification. Finally, the correlation between U.S. Plaintiff and U.S. Defendant raises the possibility that multicollinearity is causing the U.S. Defendant variable to lack statistical significance. To check, I repeated the Model 3 logit analysis without the U.S. Plaintiff variable. U.S. Defendant did not become statistically significant (p=.708). 142 The correctly classified figure indicates the proportion of outcomes that were correctly classified by the model using a .5 probability cutoff to translate predicted probabilities into dichotomous predictions (Hamilton 2004, 270-271). Thus, it indicates the proportion of outcomes for which the model estimated at least a .5 probability of deference to foreign governance authority, and the court in fact defers. Adjusted count R2 is the proportion of correct predictions beyond the number that would be correctly predicted simply by choosing the outcome with the largest percentage of observed cases, using a .5 probability cutoff (Long and Freese 2006, 111-112). The ROC curve plots 1 minus the specificity (the false positive rate) on the x-axis and sensitivity (the true positive rate) on the y-axis for each possible probability cutoff. The area under the ROC curve is equal to the probability that a random decision granting a forum non conveniens motion has a higher value of the dependent variable than a random decision denying a forum non conveniens motion (Altman and Bland 1994, 188). The larger the area under the curve, the more discriminating the model.

26

heuristics, or do they reflect comprehensive legal analysis of the factors specified in the forum non conveniens doctrine? Judicial heuristics theory would reject the latter interpretation of the evidence on the ground that heavy caseloads and limited decisionmaking resources generally make comprehensive legal analysis unlikely. But the logit analysis cannot discriminate between these two interpretations of the data. Therefore, I conducted a content analysis of each of the 210 judicial opinions in my dataset to determine the extent to which judges actually analyze the factors specified in the forum non conveniens doctrine. I did so with a view toward testing two observable implications of the judicial heuristics theory. First, judges generally should not analyze all of the doctrinally-specified factors in their opinions; doing so would suggest the sort of comprehensive analysis the judicial heuristics theory considers unlikely. Second, those factors that judges most frequently analyze in their opinions should be those that are subset heuristics (in this case, whether there is a U.S. plaintiff).143 To structure the content analysis, I divided the forum non conveniens doctrine, as set forth by the U.S. Supreme Court, into the eleven doctrinally-specified target attributes listed in Table 4. I then analyzed the content of each opinion, giving one point for each target attribute analyzed, so that the sum indicates the total number of attributes analyzed in the judge’s opinion.144

[TABLE 4 ABOUT HERE] Consistent with the first observable implication of judicial heuristics theory, judges do not comprehensively analyze the doctrinally-specified factors in their forum non conveniens opinions. Indeed, they analyze on average only 5 out of 11 of the doctrinally-specified target attributes. The result is consistent across the four main parts of the forum non conveniens doctrine, with only 47 percent of cases analyzing the level of deference owed to the plaintiff’s choice of forum and 55 percent including an analysis of the proposed alternative forum, and with judges analyzing only an average of 2 of 4 private interest factors and 2 of 5 public interest factors. It does indeed appear that judges are using decisionmaking shortcuts rather than engaging in a comprehensive analysis of the forum non conveniens factors. The second observable implication is more difficult to assess. All but 3 of the 210 opinions analyzed indicated whether or not there was a U.S. plaintiff—which is the hypothesized subset heuristic—but only 47 percent of the opinions explicitly connected this fact to the level of deference owed

143 Another observable implication of the judicial heuristics theory might be that the busier judges will rely more heavily on judicial heuristics. This in turn implies an interaction term equal to the product of judicial caseload and Activity Mostly/All Outside U.S. Territory, which should have a positive and statistically significant coefficient. When added to the logit analysis above, the interaction term had the expected positive sign, but it was not statistically significant. Therefore, this test does not provide additional support for the judicial heuristics hypotheses. I thank David Klein for suggesting this additional test. 144 Legal analysis is the application of applicable legal principles to the facts in order to generate a legal conclusion. Therefore, I gave points for combined statements of law and fact, but not for mere recitation of legal principles or summary legal conclusions not explicitly based on some fact. Although mentioned in the Supreme Court’s statement of the doctrine, I did not include view of the premises or “other” private interest factors in my specification of target attributes, because these will not always be factually relevant, which means checking for them as part of the qualitative analysis would have risked creating bias against the legalist theory. The factors that are included in this analysis are relevant regardless of the specific facts of the case and therefore should, in principle, be analyzed in all cases.

27

to the plaintiff’s choice of a U.S. forum.145 And interestingly, the only factors analyzed more than 60 percent of the time are precisely those factors that are likely to be correlated with the territorial heuristic: ease of access to proof and witnesses, and local interest in the dispute. On the one hand, the results of the content analysis suggest that the Activity Mostly/All Outside U.S. Territory variable may reflect the use of additional subset heuristics (proof, witnesses, local interest) rather than the use of a single correlational heuristic (territoriality). On the other hand, these results lend additional support to the basic finding that legal factors strongly influence transnational judicial governance decisionmaking, and that this influence results not so much from comprehensive legal analysis as from the use of judicial heuristics. 6. Conclusion: From “International Law and International Relations” to “Law and World Politics” Domestic courts are global governors. But we know very little about how domestic courts behave as global governors and why they govern the way they do. As a result, it is difficult to appraise the role of domestic courts in global governance, either normatively or in terms of positive theory. As a modest step toward filling this gap in understanding, this paper empirically tested a series of hypotheses about the legal and political determinants of one transnational judicial governance function: the allocation of adjudicative authority among states. The results suggest that U.S. district courts are not parochial actors that insist on U.S. dominance in the regulation of transnational activity. To the contrary, they frequently defer to foreign governance authority. The results also suggest that legal factors strongly influence transnational judicial governance. Consistent with the theory of judicial heuristics developed in this paper, this influence appears to result not so much from comprehensive legal analysis by judges as by the use of decisionmaking shortcuts that allow judges to conserve scarce decisionmaking resources while reaching acceptable levels of legal quality. Contrary to prevailing visions of world politics as primarily power-oriented and judicial decisionmaking as a mostly ideologically-driven process, political factors do not appear to play an important role, with one notable exception: as expected by liberal theory, U.S. district courts are more likely to defer to foreign governance authority when the foreign state is a liberal democracy. These findings suggest that transnational judicial governance follows a distinctly legal logic according to which it can help foster the rule of law in world politics—perhaps particularly among liberal democracies. This paper has focused on the determinants of one transnational judicial governance function (the allocation of governance authority among states), with an emphasis on one example (the allocation of adjudicative authority in the forum non conveniens context), in one set of courts and in one country (the federal district courts in the United States). But this is only the tip of the iceberg. Future research should relax these limitations. Beyond adjudicative authority, how do domestic courts contribute to the allocation of prescriptive authority among states?146 Beyond judicial allocation of governance authority among states, how do domestic courts allocate governance authority between state and nonstate actors, and between national and international governance arrangements? And what is the relationship between transnational judicial governance and other forms of governance—including private and international 145 However, when a district court dismisses the lawsuit on forum non conveniens grounds and there is not a U.S. plaintiff, the opinions analyze the amount of deference owed 67 percent of the time. 146 I have pursued this question in Whytock 2007.

28

governance arrangements—which domestic courts may either support or undermine? Beyond the allocation of governance authority, how do domestic courts allocate risks and resources among transnational actors, including responsibility for the negative externalities of transnational economic activity? Beyond the U.S. district courts, how do U.S. state courts and courts in other countries behave as global governors? And beyond understanding the factors that influence how domestic courts behave as global governors, future research should focus on improving our understanding of the transnational shadow of domestic law—that is, the effect of domestic law and domestic court decisionmaking on the behavior of transnational actors beyond the parties to particular transnational disputes. Notwithstanding the paper’s limited empirical scope, its implications for political science scholarship are broad, particularly for the study of global governance and international law. When people think about global governance, they tend to think about international law, as well as international courts and other international organizations. For some observers, these arrangements are necessary responses to important transnational challenges, while for others they raise fears of an emerging world government. By drawing attention to the role of domestic courts in global governance, this dissertation reinforces an essential point: that global governance does not necessarily require formal international arrangements. To the contrary, many, and possibly most, areas of transnational activity are not governed by such arrangements. But this does not mean that they are ungoverned. These “nonregime”147 areas of activity may be governed by domestic law and domestic courts, as described by the concept of transnational judicial governance. Or they may be governed by transgovernmental networks148 or by primarily private arrangements.149 As these highly decentralized forms of governance suggest, global governance does not necessarily require a high degree of centralization or a hierarchical structure. To be sure, formal interstate arrangements may, under some circumstances—and indeed perhaps many circumstances—be more desirable than these more decentralized approaches. However, transnational judicial governance and other alternative concepts of global governance can help scholars and policymakers avoid a sense of false necessity about the role of international law and international organization in global governance. This in turn can help focus research and policymaking on identifying the most appropriate forms of governance for particular types of transnational activity. Moreover, while pundits and scholars devote considerable attention to understanding and critiquing international arrangements such as the World Trade Organization, the World Bank, and the International Criminal Court, they pay far less attention to alternative forms of governance by which transnational activity is regulated. Yet these alternative arrangements arguably play an even more fundamental role in regulating transnational activity than international arrangements. They therefore deserve far more careful empirical and normative scrutiny than they currently receive.

147 Dimitrov et al. 2007. 148 See, for example, Raustiala 2002; Slaughter 2004; Whytock 2005. As Slaughter (2004, chap. 2) illustrates, domestic courts sometimes do act in the context of transgovernmental networks. However, most transnational judicial governance does not appear to take the form of transgovernmental networks. 149 See, for example, Büthe 2004; Coen and Thatcher 2005; Cutler 1999; Cutler 2003; Hall and Biersteker 2002; and Mattli and Büthe 2003.

29

Another implication of transnational judicial governance is that the ways in which law influences world politics are more varied than prevailing “international relations and international law” approaches suggest. In most IR scholarship on law, the independent variable is international law and the dependent variable is the behavior of the unitary state.150 Thus, the relationship between law and world politics is understood as an “outside-in” process, in which international law is the independent variable, external to the state, that influences state behavior. But these variables are defined too narrowly, and the focus on outside-in processes misses important “inside-out” processes that shape the relationship between law and world politics.151 On the independent variable side, international law is not the only law, and international courts are not the only courts, that influence the behavior of state and nonstate actors in world politics. Indeed, domestic courts and domestic law arguably have a stronger influence on world politics than international law and international courts. These inside-out processes include the transnational judicial governance functions described in this paper: judicial allocation of governance authority among states, between state and nonstate actors, and between national and international institutions; judicial allocation of risks and resources among transnational actors; and support for other forms of global governance (including international law). They also involve the domestic law of foreign relations that helps shape foreign policy processes and outcomes,152 as well as the role of domestic courts in shaping the evolution of international law.153 The implication is that IR and IPE scholars should consider devoting more attention to understanding the inside-out processes that link law and world politics.154 On the dependent variable side, although international law and international relations scholarship focuses on explaining state behavior, legal rules influence the behavior of nonstate actors as well. Of particular importance to IPE scholars, these include transnational economic actors, both individual and corporate. Therefore, a thorough understanding of the role of law in world politics requires an examination not only of the behavior of the unitary state, but also of the behavior of nonstate transnational actors and domestic institutions. The foregoing comments are not intended to downplay the role of international law or the importance of understanding state behavior. Rather, they are meant to suggest that political scientists would benefit from a broader conception of the role of law in world politics, one that encompasses the influence of not only international law and international courts but also domestic law and domestic courts; one that investigates influences on the behavior of not only state but also nonstate actors; and one that examines both inside-out and outside-in legal processes. In short, a conceptual shift is in order, from “international law and international relations” to “law and world politics.”

150 See, for example, Raustiala and Slaughter 2002; Simmons 2000; Von Stein 2005. 151 See Whytock 2006a (using Waltz’s [1959] levels of analysis to distinguish between “second image” and “second image reversed” approaches to the study of law and world politics). 152 Ibid. 153 Indeed, according to Article 38(1) of the Statute of the International Court of Justice, domestic judicial decisions are a “subsidiary means” of determining the rules of international law. 154 A related point is this: the legalization of world politics (Goldstein et al. 2001) does not necessarily require legalized international institutions. Legalization of world politics can also arise through processes of transnational judicial governance, whereby domestic courts help determine the size and shape of the transnational shadow of domestic law within which transnational actors behave.

30

References Abbott, Kenneth W., Robert O. Keohane, Andrew Moravcsik, Anne-Marie Slaughter, and Duncan Snidal. 2000.

“The Concept of Legalization.” International Organization 54: 401-419. Alter, Karen. 2001. Establishing the Supremacy of European Law: The Making of an International Rule of Law in

Europe. Oxford: Oxford University Press. Alter, Karen A. 2006. “Private Litigants and the New International Courts.” Comparative Political Studies 39: 22-

49. Altman, Douglas G., and J. Martin Bland. 1994. “Diagnostic Tests 3: Receiver Operating Characteristic Plots.”

British Medical Journal 309: 188. Alvarez, José E. 2001. “Do Liberal States Behave Better? A Critique of Slaughter’s Liberal Theory.” European

Journal of International Law 12: 183-246. Bainbridge, Stephen M., and G. Mitu Gulati. 2002. “How Do Judges Maximize? (The Same Way Everybody Else

Does—Boundedly): Rules of Thumb in Securities Fraud Opinions.” Emory Law Journal 51: 83-151. Baum, Lawrence. 1994. “What Judges Want: Judges’ Goals and Judicial Behavior.” Political Research Quarterly

47: 749-768. ________. 1997. The Puzzle of Judicial Behavior. Ann Arbor: University of Michigan Press. ________. 2006. Judges and Their Audiences: A Perspective on Judicial Behavior. Princeton: Princeton University

Press. Baldwin, David A. 2002. “Power and International Relations.” In Handbook of International Relations, ed. Walter

Carlsnaes, Thomas Risse, and Beth A. Simmons, 177-191. London: Sage Publications. Beebe, Barton. 2006. “An Empirical Study of the Multifactor Tests for Trademark Infringement.” California Law

Review 94: 1581-1654. Bermann, George A. 2003. Transnational Litigation. St. Paul, MN: West. Bhattacharya , Utpal, Neal E. Galpin, and Bruce Haslem. 2006. “The Home Court Advantage in International

Corporate Litigation.” Unpublished manuscript, available at http://ssrn.com/abstract=509008, forthcoming in Journal of Law and Economics.

Born, Gary, and Peter B. Rutledge. 2007. International Civil Litigation in United States Courts, 4th ed. New York, NY: Aspen.

Buchheit, Lee C., G. Mitu Gulati, and Robert B. Thompson. 2007. “The Dilemma of Odious Debts.” Duke Law Journal 57: 1202-1262.

Burley, Anne-Marie. 1992. “Law among Liberal States: Liberal Internationalism and the Act of State Doctrine.” Columbia Law Review 92: 1907-1996.

Burley, Anne-Marie, and Walter Mattli. 1993. “Europe Before the Court: A Political Theory of Legal Integration.” International Organization 47: 41-76.

Burton, Steven J. 1992. Judging in Good Faith. Cambridge: Cambridge University Press. Büthe, Tim. 2004. “Governance through Private Authority: Non-State Actors in World Politics.” Journal of

International Affairs 58: 281-290. Buxbaum, Hannah L. 2006. “Transnational Regulatory Litigation.” Virginia Journal of International Law 46: 251-

317. Carp, Robert A., Ronald Stidham, and Kenneth L. Manning. 2004. Judicial Process in America, 6th ed. Washington,

DC: CQ Press. Clayton, Cornell W., and Howard Gillman, eds. 1999. Supreme Court Decisionmaking: New Institutional

Approaches. Chicago: University of Chicago Press. Clayton, Cornell W. 1999. “The Supreme Court and Political Jurisprudence: New and Old Institutionalisms.” In

Supreme Court Decision-Making: New Institutionalist Approaches, ed. Cornell W. Clayton and Howard Gillman, 15-41. Chicago: University of Chicago Press. Pp. 15-41.

Clermont, Kevin M. 2005. “The Role of Private International Law in the United States: Beating the Not-Quite-Dead Horse of Jurisdiction.” In Private Law, Private International Law, and Judicial Cooperation in the EU-US Relationship, ed. Ronald A. Brand, 75-114. Eagan, MN: Thomson West.

Clermont, Kevin M., and Theodore Eisenberg. 1995. “Exorcising the Evil of Forum-Shopping.” Cornell Law Review 80: 1507-1535.

Clermont, Kevin M., and Theodore Eisenberg. 1996. “Xenophilia in American Courts.” Harvard Law Review 109: 1120-1143.

31

Clermont, Kevin M., and Theodore Eisenberg. 2007. “Xenophilia or Xenophobia in U.S. Courts? Before and After 9/11.” Journal of Empirical Legal Studies 4: 441-464.

Coen, David, and Mark Thatcher. 2005. “The New Governance of Markets and Non-Majoritarian Regulators.” Governance 18: 329-346.

Conlisk, John. 1996. “Why Bounded Rationality?” Journal of Economic Literature 34: 669-700. Cooter, Robert, and Stephen Marks, with Robert Mnookin. 1982. “Bargaining in the Shadow of the Law: A Testable

Model of Strategic Behavior.” Journal of Legal Studies 11: 225-251. Cutler, A. Claire, Virginia Haufler, and Tony Porter. 1999. Private Authority and International Affairs. Albany, NY:

State University of New York Press. Davies, Martin. 2002. “Time to Change the Federal Forum Non Conveniens Analysis.” Tulane Law Review 77: 309-

386. Davis, Jeffrey. 2006. “Justice Without Borders: Human Rights Cases in U.S. Courts.” Law & Policy 28: 60-82. Dimitrov, Radoslav S., Detlef Sprinz, Gerald M. DiGiusto, and Alexander Kelle. 2007. “International Nonregimes:

A Research Agenda.” International Studies Review 9: 230-258. Easton, David. 1971. The Political System: An Inquiry into the State of Political Science, 2d ed. New York, NY:

Alfred A. Knopf. Epstein, Lee, and Jack Knight. 1998. The Choices Justices Make. Washington, DC: CQ Press. Falk, Richard A. 1964. The Role of Domestic Courts in the International Legal Order. Syracuse, NY: Syracuse

University Press. Federal Judicial Center. 2007. “Biographical Directory of Federal Judges.” Available at

http://www.fjc.gov/public/home.nsf/hisj. Freedom House. 2006. “Freedom in the World 2006.” Database accessible at http://www.freedomhouse.org/. Frieden, Jeffry, and Lisa L. Martin. 2002. “International Political Economy: Global and Domestic Interactions.” In

Political Science: The State of the Discipline, eds. Ira Katznelson and Helen V. Milner, 118-146. New York: W. W. Norton.

Galanter, Marc. “The Radiating Effects of Courts.” In Empirical Theories about Courts, ed. Keith O. Boyum and Lynn Mather, 117-142. New York, NY: Longman.

George, Tracey E. 1998. “Developing a Positive Theory of Decisionmaking on U.S. Courts of Appeals.” Ohio State Law Journal 58: 1635-1696.

George, Tracey E., and Lee Epstein. 1992. “On the Nature of Supreme Court Decision Making.” American Political Science Review 86: 323-337.

Gerken, Joseph L. 2004. “A Librarian’s Guide to Unpublished Judicial Opinions.” Law Library Journal 96: 475-501.

Gigerenzer, Gerd. 2006. “Heuristics.” In Heuristics and the Law, ed. Gerd Gigerenzer and Christoph Engel, 17-44. Cambridge, MA: MIT Press.

Gigerenzer, Gerd, and Daniel G. Goldstein. 1996. “Reasoning the Fast and Frugal Way: Models of Bounded Rationality.” Psychological Review 103: 650-669.

Gigerenzer, Gerd, and Peter M. Todd. 1999. “Fast and Frugal Heuristics: The Adaptive Toolbox.” In Simple Heuristics That Make Us Smart, ed. Gerd Gigerenzer and Peter M. Todd, 3-34. New York: Oxford University Press.

Gillman, Howard. 2001. “What’s Law Got to Do with It? Judicial Behavioralists Test the ‘Legal Model’ of Judicial Decision Making.” Law and Social Inquiry 26: 465-504.

Gillman, Howard, and Cornell W. Clayton. 1999. “Beyond Judicial Attitudes: Institutional Approaches to Supreme Court Decision-Making.” In Supreme Court Decision-Making: New Institutionalist Approaches, ed. Cornell W. Clayton and Howard Gillman, 1-12. Chicago: University of Chicago Press.

Goldstein, Judith L., Miles Kahler, Robert O. Keohane and Anne-Marie Slaughter, eds. 2001. Legalization and World Politics. Cambridge, MA: MIT Press.

Grieco, Joseph M. 1988. “Anarchy and the Limits of Cooperation: A Realist Critique of the Newest Liberal Institutionalism.” International Organization 42: 485-507.

Guthrie, Chris, Jeffrey J. Rachlinksi, and Andrew J. Wistrich. 2001. “Inside the Judicial Mind.” Cornell Law Review 86: 777-830.

Guzman, Andrew T., and Beth A. Simmons. 2005. “Power Plays and Capacity Constraints: The Selection of Defendants in World Trade Organization Disputes.” Journal of Legal Studies 34: 557-598.

Hall, Rodney Bruce, and Thomas J. Biersteker. 2002. The Emergence of Private Authority in Global Governance. Cambridge: Cambridge University Press.

Heckman, James J. 1979. “Sample Selection Bias as a Specification Error.” Econometrica 47: 153-161.

32

Helfer, Laurence R., and Anne-Marie Slaughter. 1997. “Toward a Theory of Effective Supranational Adjudication.” Yale Law Journal 107: 273-391.

Kahler, Miles. 2004. “Global Governance Redefined.” Paper presented at the Conference on Globalization, the State, and Society, Washington University School of Law, available at http://irpshome.ucsd.edu/faculty/mkahler/GlobGov_10.04.doc.

Kahler, Miles, and David A. Lake. 2003. “Globalization and Governance.” In Governance in a Global Economy: Political Authority in Transition, eds. Miles Kahler and David A. Lake, 1-30. Princeton, NJ: Princeton University Press.

Kahn, Ronald. 1999. “Institutional Norms and Supreme Court Decision-Making.” In Supreme Court Decision-Making: New Institutionalist Approaches, eds. Cornell W. Clayton and Howard Gillman, 175-198. Chicago: University of Chicago Press.

Kahneman, Daniel, and Shane Frederick. 2005. “A Model of Heuristic Judgment. In The Cambridge Handbook of Thinking and Reasoning, eds. Keith J. Holyoak & Robert G. Morrison, 267-293. Cambridge: Cambridge University Press.

Kahneman, Daniel, Paul Slovic, and Amos Tversky, eds. 1982. Judgment under Uncertainty: Heuristics and Biases. Cambridge: Cambridge University Press.

Keohane, Robert O. 2002. Power and Governance in a Partially Globalized World. London: Routledge. Keohane, Robert O., and Joseph S. Nye. 2000. “Introduction.” In Governance in a Globalizing World, ed. Joseph S.

Nye and John D. Donahue, 1-41. Washington, DC: Brookings. Keohane, Robert O., Andrew Moravcsik, and Anne-Marie Slaughter. 2000. “Legalized Dispute Resolution:

Interstate and Transnational.” International Organization 54: 457-488. King, Gary, Robert O. Keohane, and Sidney Verba. 1994. Designing Social Inquiry: Scientific Inference in

Qualitative Research. Princeton, NJ: Princeton University Press. King, Gary, Michael Tomz, and Jason Wittenberg. 2000. “Making the Most of Statistical Analyses: Improving

Interpretation and Presentation.” American Journal of Political Science 44: 347-361. Kjaer, Anne Mette. 2004. Governance. Cambridge, U.K.: Polity Press. Koh, Harold Hongju. 1991. “Transnational Public Law Litigation.” Yale Law Journal 100: 2347-2402. Koremenos, Barbara. 2005. “Contracting around International Uncertainty.” American Political Science Review 99:

549-565. Lake, David A., and Robert Powell, ed. 1999. Strategic Choice and International Relations. Princeton: Princeton

University Press. Lindquist, Stefanie A., and David E. Klein. 2006. “The Influence of Jurisprudential Considerations on Supreme

Court Decisionmaking: A Study of Conflict Cases.” Law and Society Review 40: 135-162. Long, J. Scott, and Jeremy Freese. 2006. Regression Models for Categorical Dependent Variables Using Stata, 2d

ed. College Station, TX: Stata Press. Martin, Lisa L. 2000. Democratic Commitments: Legislatures and International Cooperation. Princeton: Princeton

University Press. Mattli, Walter. 2001. “Private Justice in a Global Economy: From Litigation to Arbitration.” International

Organization 55: 919-947. Mattli, Walter, and Tim Büthe. 2003. “Setting International Standards: Technological Rationality or Primacy of

Power?” World Politics 56: 1-42. Mearsheimer, John J. 2001. The Tragedy of Great Power Politics. New York: W. W. Norton. Michaels, Ralf. 2006. “Two Paradigms of Jurisdiction.” Michigan Journal of International Law 27: 1003-1069. Milner, Helen V. 1997. Interests, Institutions, and Information: Domestic Politics and International Relations.

Princeton: Princeton University Press. Mnookin, Robert N., and Lewis Kornhauser. 1979. “Bargaining in the Shadow of the Law: The Case of Divorce.”

Yale Law Journal 88: 950-997. Moore, Kimberly A. 2003. “Xenophobia in American Courts.” Northwestern University Law Review 97: 1497-1549. Moravcsik, Andrew. 1997. “Taking Preferences Seriously: A Liberal Theory of International Politics.” International

Organization 51:513-553. Morgenthau, Hans J. 1978. Politics among Nations: The Struggle for Power and Peace, 5th ed., revised. New York:

Alfred A. Knopf. Murphy, Walter F., C. Herman Pritchett, and Lee Epstein. 2002. Courts, Judges, and Politics: An Introduction to the

Judicial Process, 5th ed. Boston: McGraw-Hill. Olson, Susan M. 1992. “Studying Federal District Courts Through Published Cases: A Research Note.” Justice

System Journal 15: 782-800.

33

Pew Research Center for the People and the Press. 2005a. “Beyond Red vs. Blue: Republicans Divided about Role of Government—Democrats by Social and Personal Values,” released May 10, 2005, available at http://people-press.org/reports/display.php3?ReportID=242.

Pew Research Center for the People and the Press. 2005b. “Opinion Leaders Turn Cautious, Public Looks Homeward: America’s Place in the World,” released November 17, 2005, available at http://people-press.org/reports/display.php3?ReportID=263.

Pollack, Mark A., and Gregory C. Shaffer. 2001. “Who Governs?” In Transatlantic Governance in the Global Economy, ed. Mark A. Pollack and Gregory C. Shaffer, 287-305. Lanham, MD: Rowman & Littlefield.

Posner, Richard A. 1993. “What Do Judges and Justices Maximize? (The Same Thing Everybody Else Does).” Supreme Court Economic Review 3: 1-41.

Powell, Robert. 1999. In the Shadow of Power: States and Strategies in International Politics. Princeton: Princeton University Press.

Putnam, Tonya L. 2006. “Courts Without Borders? Extraterritoriality as a Mechanism for Global Governance.” Paper presented at the annual meeting of the International Studies Association.

Randazzo, Kirk A. 2006. “When Liberty and Security Collide: Foreign Policy Litigation and the Federal Judiciary.” Kentucky Law Journal 94: 629-648.

Raustiala, Kal. 2002. “The Architecture of International Cooperation: Transgovernmental Networks and the Future of International Law.” Virginia Journal of International Law 43: 1-92.

________. 2005. “Form and Substance in International Agreements.” American Journal of International Law. 99: 581-614.

________. 2006. “The Evolution of Territoriality: International Relations and International Law.” In Globalization, Territoriality, and Conflict, ed. Miles Kahler and Barbara Walter, 219-250. Cambridge: Cambridge University Press.

Raustiala, Kal, and Anne-Marie Slaughter. 2002. “International Law, International Relations and Compliance.” In Handbook of International Relations, ed. Walter Carlsnaes, Thomas Risse, and Beth A. Simmons, 539-558. London: Sage Publications.

Robertson, David W. 1987. “Forum Non Conveniens in America and England: ‘A Rather Fantastic Fiction.’” Law Quarterly Review 103: 398-_.

Rohde, David W., and Harold J. Spaeth. 1976. Supreme Court Decisionmaking. San Francisco: W. H. Freeman and Company.

Segal, Jeffrey A. 1984. “Predicting Supreme Court Cases Probabilistically: The Search and Seizure Cases, 1962-1981.” American Political Science Review 78: 891-900.

Segal, Jeffrey A. 1986. “Supreme Court Justices as Human Decision Makers: An Individual-Level Analysis of Search and Seizure Cases.” Journal of Politics 48: 938-955.

Segal, Jeffrey A. 1999. “Supreme Court Deference to Congress: An Examination of the Marksist Model.” In Supreme Court Decision-Making: New Institutionalist Approaches, eds. Cornell W. Clayton and Howard Gillman, 237-253. Chicago: University of Chicago Press.

Segal, Jeffrey A., Donald R. Songer, and Charles M. Cameron. 1995. “Decision Making on the U.S. Courts of Appeals.” In Contemplating Courts, ed. Lee Epstein, 296-314. Washington, DC: Congressional Quarterly Press.

Segal, Jeffrey A., and Harold J. Spaeth. 2002. The Supreme Court and the Attitudinal Model Revisited. Cambridge: Cambridge University Press.

Shapiro, Martin. 1972. “From Public Law to Public Policy, or the ‘Public’ in ‘Public Law,’” PS: Political Science and Politics 5: 410-418.

________. 1993. “Public Law and Judicial Politics.” In Political Science: The State of the Discipline II, ed. Ada W. Finifter, 365-_. Washington, DC: American Political Science Association.

Simmons, Beth A. 2000. “International Law and State Behavior.” American Political Science Review 94: 819-835. Simon, Herbert A. 1985. “Human Nature in Politics: The Dialogue of Psychology with Political Science.” American

Political Science Review 79: 293-304. Simon, Steve. 2007. Internet posting on ROC curves, available at http://www.childrensmercy.org/stats/ask/roc.asp. Slaughter Burley, Anne-Marie. 1993. “International Law and International Relations Theory: A Dual Agenda.”

American Journal of International Law 87: 205-239. Slaughter, Anne-Marie. 1994. “A Typology of Transjudicial Communication.” University of Richmond Law Review

29: 99-137. ________. 1995. “International Law in a World of Liberal States.” European Journal of International Law 6: 503-

538.

34

________. 2004. A New World Order. Princeton, NJ: Princeton University Press. Slaughter, Anne-Marie, and David Zaring. 2006. “Networking Goes International: An Update.” Annual Review of

Law and Social Science 2: 211-229. Smith, Rogers M. 1988. “Political Jurisprudence, the ‘New Institutionalism,’ and the Future of Public Law.”

American Political Science Review 82: 89-108. Spaeth, Harold J. 1995. “The Attitudinal Model.” In Contemplating Courts, ed. Lee Epstein, 296-314. Washington,

DC: Congressional Quarterly Press. Spaeth, Harold J., and Jeffrey A. Segal. 1999. Majority Rule or Minority Will: Adherence to Precedent on the U.S.

Supreme Court. Cambridge: Cambridge University Press. Stein, Allan R. 1985. “Forum Non Conveniens and the Redundancy of Court-Access Doctrine.” University of

Pennsylvania Law Review 133: 781-846. Stone Sweet, Alec. 2006. “The New Lex Mercatoria and Transnational Governance.” Journal of European Public

Policy 13: 627-646. Stone Sweet, Alec, and Thomas L. Brunell. 1998. “Constructing a Supranational Constitution: Dispute Resolution

and Governance in the European Community.” American Political Science Review 92: 63-81. Tate, C. Neal, and Torbjörn Vallinder. 1995. The Global Expansion of Judicial Power. New York: New York

University Press. Tomz, Michael, Jason Wittenberg, and Gary King. 2001. CLARIFY: Software for Interpreting and Presenting

Statistical Results. Version 2.0. Cambridge, MA: Harvard University, available at http://gking.harvard.edu. Trachtman, Joel P. 2001. “Economic Analysis of Prescriptive Jurisdiction.” Virginia Journal of International Law

42: 1-39. Unah, Isaac. 1998. The Courts of International Trade: Judicial Specialization, Expertise, and Bureaucratic Policy-

Making. Ann Arbor: University of Michigan Press. Voeten, Erik. 2007. The Politics of International Judicial Appointments: Evidence from the European Court of

Human Rights. International Organization 61: 669-701. Volcansek, Mary L., and John F. Stack, Jr., ed. 2005. Courts Crossing Borders: Blurring the Lines of Sovereignty.

Durham, NC: Carolina Academic Press. Von Stein, Jana. 2005. “Do Treaties Constrain or Screen? Selection Bias and Treaty Compliance.” American

Political Science Review 99: 611-622. Wai, Robert. 2005a. “Transnational Private Law and Private Ordering in a Contested Global Society.” Harvard

International Law Journal 46 (2) (Summer 2005): 471-486. ________. 2005b. “Transnational Private Litigation and Transnational Governance.” In Criticizing Global

Governance, ed. Markus Lederer and Philipp S. Müller, 243-261. New York, NY: Palgrave Macmillan. Waltz, Kenneth N. 1959. Man, the State, and War: A Theoretical Analysis. Reading, MA: Addison-Wesley

Publishing Company. New York, NY: Columbia University Press. ________. 1979. Theory of International Politics. Reading, MA: Addison-Wesley Publishing Company. Waters, Melissa A. 2005. “Mediating Norms and Identity: The Role of Transnational Judicial Dialogue in Creating

and Enforcing International Law.” Georgetown Law Journal 93: 487-574. Whytock, Christopher A. 2005. “A Rational Design Theory of Transgovernmentalism: The Case of E.U.-U.S.

Merger Review Cooperation.” Boston University International Law Journal 23: 1-53. ________. 2006a. “Foreign Law, Domestic Courts, and World Politics.” Paper presented at the annual conference of

the International Studies Association, San Diego, California. ________. 2006b. “Domestic Courts and Global Governance.” Paper presented at the annual meeting of the

American Political Science Association, Philadelphia, Pennsylvania. ________. 2007. Domestic Courts and Global Governance: The Politics of Private International Law. Ph.D.

dissertation, Duke University, available at http://dukespace.lib.duke.edu/dspace/handle/10161/452). ________. 2008. “Litigation, Arbitration, and the Transnational Shadow of the Law.” Duke Journal of Comparative

and International Law 18 (forthcoming).

35

Tables Table 1. Transnational Judicial Governance Functions Allocation of Governance Authority Among States Between Public and Private Actors Between Domestic and International Institutions Allocation of Risks and Resources among Transnational Actors

Among States Among Nonstate Actors Between State and Nonstate Actors

Governance Support Function Support for Foreign Governance

Support for Private Governance Support for International Governance

36

Table 2. Logit Analysis, Dependent Variable: Deference to Foreign Adjudicative Authority (Forum Non Conveniens) Model 1

Legal Model 2 Political

Model 3 Combined

U.S. Plaintiff -1.020** -1.826*** -1.450*** (0.320) (0.376) (0.402) [0.001] [0.000] [0.000] Activity Mostly/All Outside U.S. Territory 1.428*** 1.709*** (0.323) (0.378) [0.000] [0.000] Republican Judge -0.490 -0.529 (0.331) (0.362) [0.140] [0.143] Republican Environment -0.636 -0.629 (0.340) (0.373) [0.061] [0.092] Foreign State’s GDP (Log) -0.061 0.038 (0.093) (0.100) [0.508] [0.704] Foreign State Liberal Democracy 1.234** 1.420** (0.431) (0.456) [0.004] [0.002] U.S. Defendant -0.379 -0.418 (0.362) (0.386) [0.295] [0.279] Constant -0.416 1.369 -1.095 (0.305) (1.200) (1.372) 0.172 [0.254] [0.425] Number of Observations 205 195 193 LR Chi-Squared 43.962 37.419 61.076 Prob.>LR Chi-Squared 0.000 0.000 0.000 Adjusted Count R-Squared .347 .266 .424 % Correctly Classified 69.76 64.62 72.54 Area Under ROC Curve .7490 .7378 .8084 Notes: Standard errors in parentheses, p-values in brackets, * p<0.05, ** p<0.01, *** p<0.001, two-tailed tests.

37

Table 3. Simulated Effects of Independent Variable Changes on Probability of Deference to Foreign Adjudicative Authority Independent Variable Value Change Probability

Change 95%

Confidence Interval

U.S. Plaintiff

no yes -.243 -.402, -.110

Activity Mostly/All Outside U.S. Territory

no yes .386 .213, .538

Republican Environment

no yes -.151 -.316, .013

Foreign State Liberal Democracy

no yes .320 .114, .497

Note: For each change in a variable’s value, all other variables were held constant at their modes.

38

Table 4. Elements of Forum Non Conveniens Analysis Included in Judges’ Opinions Element Points

Possible Mean Std. Error 95% Confidence

Interval Level of Deference

1 .467 .034 .400, .534

Alternative Forum

1 .552 .034 .485, .618

Private Interest Factors • Proof .724 .031 .660, .780 • Witnesses .790 .028 .730, .840 • Judgments .190 .027 .143, .249 • Balance .448 .034 .382, .515 • Subtotal

4 2.152 .081 1.993, 2.311

Public Interest Factors • Congestion .310 .032 .251, .375 • Jury Burden .257 .030 .203, .320 • Local Interest .690 .032 .625, .749 • Governing Law .467 .034 .400, .534 • Balance .386 .034 .322, .453 • Subtotal

5 2.110 .108 1.900, 2.323

Total

11 5.281 .206 4.875, 5.687

39

Appendix Table A1. Summary Statistics Variable Obs. Mean Std.

Dev. Min Max

Decision

210 .471 .500 0 1

U.S. Plaintiff

207 .512 .501 0 1

Activity Mostly/All Outside U.S. Territory

206 .524 .501 0 1

Republican Judge

208 .548 .499 0 1

Republican Environment

210 .381 .487 0 1

Foreign State GDP (Log)

200 12.634 1.799 5.799 15.373

Foreign State Military Spending (Log)

189 17.987 1.675 12.635 19.937

Foreign State Liberal Democracy (Freedom House)

203 .768 .423 0 1

Foreign State Liberal Democracy (Polity IV)

210 .890 .313 0 1

U.S. Defendant

207 .469 .500 0 1

Table A2. Correlation Table Decision U.S. Plaintiff Activity

Mostly/All Outside U.S.

Territory

Republican Judge

Republican Environment

Foreign State GDP (Log)

Foreign State Military

Spending (Log)

Foreign State Liberal

Democracy (Freedom House)

Foreign State Liberal

Democracy (Polity IV)

U.S. Defendant

Decision 1.000 U.S. Plaintiff -0.3515 1.0000 Activity Mostly/All Outside U.S. Territory

0.3965 -0.4068 1.0000

Republican Judge

-0.0513 0.0432 -0.0010 1.0000

Republican Environment

-0.1235 -0.0313 -0.0765 -0.2512 1.0000

Foreign State GDP (Log)

0.0179 0.0689 -0.1989 -0.0758 0.0148 1.0000

Foreign State Military Spending (Log)

-0.0122 0.0805 -0.2012 -0.0974 0.0314 0.9430 1.0000

Foreign State Liberal Democracy (Freedom House)

0.1360 0.1385 -0.1534 0.0202 0.0105 0.3952 0.3595 1.0000

Foreign State Liberal Democracy (Polity IV)

0.1390 0.0212 -0.0476 0.0837 -0.0326 0.2649 0.2474 0.6805 1.0000

U.S. Defendant

0.0947 -0.4189 0.2002 -0.0013 0.1101 0.0387 -0.0290 0.0214 0.0466 1.0000