Legal Traditions and Nonbinding Commitments : Evidence From the United Nations ’ Model Commercial Legislation

Non–legally binding agreements provide an important tool for establishing international cooperation. We know little, however, about the variation in the implementation of such agreements. This article identifies a major cause of this variation: legal tradition. Nonbinding agreements, which may be adapted to local needs and circumstances, are consistent with the gradual, organic evolution of common law; by contrast, they are an uneasy fit with the civil-law tradition that neatly distinguishes between “law” and “nonlaw.” Consequently, common-law countries are more likely to implement nonbinding agreements than civil-law countries. Survival analysis of three nonbinding instruments—United Nations model laws aimed at harmonizing commercial legislation— finds strong support for this argument: common-law countries prove significantly more likely to implement these model laws.

While common-law countries are less inclined toward legal commitments, I argue that they hold a more favorable view of non-legally-binding commitments.Such agreements-which states may adapt to local needs and circumstances-are consistent with the legal style of common law.But they uneasily fit the civil-law tradition with its formalistic approach that neatly distinguishes between "law" and "nonlaw."Consequently, common-law countries should prove more likely to implement nonbinding agreements than civil-law countries.I test this argument by examining a rare set of nonbinding instruments for which one may obtain systematic implementation data.These are model laws established by the United Nations Commission on International Trade Law (UNCITRAL) that aim to harmonize states' commercial legislation in three areas: electronic commerce, cross-border insolvency, and international commercial arbitration.Survival analysis of the time to implementation reveals that common-law countries are indeed significantly more likely to pass legislation based on these model laws.
This article provides insight into the link between legal traditions and international cooperation-specifically, how legal traditions shape states' preferences for legal or nonlegal instruments as vehicles for cooperation.I show that while common-law countries may hold a distaste for treaties, they are more favorably disposed toward nonbinding agreements that can fit comfortably in their legal system.This finding matters for the design of international agreements.Furthermore, this article sheds light on additional factors that shape the implementation of nonlegal agreements.Many studies suggest that democracies are more likely to ratify treaties than nondemocracies (Neumayer 2002;Simmons 2009).Yet my analysis reveals that democracy little affects the implementation of nonbinding instruments.Finally, I demonstrate that veto players may obstruct the implementation of nonbinding agreements.This challenges conventional wisdom concerning the relative ease and speed of obtaining international cooperation through such agreements (Lipson 1991).

Legal Traditions and International Commitments
Legal tradition is "a set of deeply rooted, historically conditioned attitudes about the nature of law, about the role of law in the society and the polity, about the proper organization and operation of a legal system, and about the way law is or should be made, applied, studied, perfected, and taught" (Merryman and P erez-Perdomo 2007, 2).Two legal traditions have had a broad geographic reach and long-lasting impact: common law and civil law.The origins of common law lie in the English legal system.From there, it spread to the British colonies, including the United States, Australia, and India.The judge has a central role in the common-law system: law is made by judges who establish binding precedents by resolving case-specific disputes.In the civil-law system, by contrast, statutes and comprehensive codes serve as the primary means of ordering legal material, with a key role for legal scholars who ascertain and formulate rules.Civil law originated in Roman law, which heavily influenced the legal traditions of France and Germany.The significant international influence of these two countries, especially France, spread civil law worldwide (Zweigert and K€ otz 1998;Glendon, Carozza, and Picker 2008; La Porta, Lopez-de-Silanes, and Shleifer 2008).
Research in economics points to important differences in legal arrangements and economic outcomes between common-and civil-law countries (La Porta et al. 1997;Djankov et al. 2002).Yet the influence of legal traditions reaches beyond national boundaries into the realm of international commitments.A series of studies identifies differences between civil-law and common-law countries in terms of their willingness to join treaties or accept the jurisdiction of international courts.Specifically, common-law countries are less likely to show such willingness compared to civil-law countries.Thus far, the distaste of common-law countries for international legal commitments eluded a unifying interpretation, and several mechanisms were offered as possible causes.
One type of explanation focuses on compatibility and similarity of legal characteristics.Mitchell and Powell (2011) argue that civil-law states are more likely to accept the compulsory jurisdiction of the International Court of Justice (ICJ) since the court's rules and procedures resemble those of civil law.Because of this institutional similarity, civil-law states tend to view the ICJ as fair and unbiased.The ICJ creates focal points more easily with civillaw states, and the latter can effectively use the court to signal private information.By contrast, common-law countries expect smaller benefits from an ICJ membership and are less likely to accept the court's jurisdiction.
Common-law countries are also significantly less likely to ratify human-rights treaties (Goodliffe and Hawkins 2006;Simmons 2009).Simmons (2009, 71-7) argues that treaties are the antithesis of an organic, bottom-up law developed by judges.Furthermore, given the power and independence of judges in common-law systems, governments feel uncertain about the consequences of treaty ratification and may find it difficult to escape treaty obligations.
Beyond human-rights treaties, common-law countries are less likely to join BITs, perhaps because these countries already provide strong property-rights protections and have a lesser need for an external commitment mechanism, such as an investment treaty (Elkins, Guzman, and Simmons 2006).2Common-law countries less readily join the International Criminal Court (Simmons and Danner 2010;Mitchell and Powell 2011;Chapman and Chaudoin 2013) and hold reservations toward the international regulation of small arms (Efrat 2010).
Overall, the existing literature suggests that legal traditions strongly influence states' willingness to make legally binding commitments.Do legal traditions also influence the implementation of agreements that are not legally binding?This question has not yet received an answer, and the following section begins to fill this gap.

Legal Traditions and Implementation of Nonbinding Commitments
builds on nonbinding arrangements (Skjaerseth, Stokke, and Wettestad 2006) 4 and so do cooperative efforts against criminal activities, such as human trafficking, the illicit arms trade, and money laundering. 5The founding document of the international human rights regime-the Universal Declaration of Human Rights-is a nonbinding instrument, as are some of the other documents constituting this regime. 6espite the common use of nonbinding agreements for establishing international cooperation, they received limited scholarly attention compared to treaties.To date, several accounts sought to explain why states sometimes choose to establish nonbinding agreements (Lipson 1991;Abbott and Snidal 2000;Raustiala 2005;Guzman and Meyer 2010).These studies suggest that nonlegal agreements do not necessarily result from a failure to establish legally binding commitments.Indeed, non-legally binding instruments offer important benefits: they may be easier, faster, and less costly to establish; their flexibility reduces sovereignty costs and allows adaptation to changing circumstances; and they can facilitate compromise between actors with varying interests and degrees of power (Abbott and Snidal 2000;Dreyfus and Patt 2012).
Yet, to influence behavior and achieve compliance, nonbinding agreements typically have to be implemented: states need to incorporate these agreements into their domestic legal system through legislation, executive decree, or some other means.The process of implementation often increases the norm's precision, allows relevant actors to interpret and contest the norm, and results in clear and observable standards that facilitate compliance."[W]ithout implementation compliance may be a fair-weather process since neither states nor other organizations have made strong commitments nor adopted measures to restrict their behavior" (Betts and Orchard 2014, 6).
Using qualitative methods, several studies examined the implementation of nonbinding agreements and the influences that shape it (Shelton 2000;Skjaerseth, Stokke, and Wettestad 2006;Skjaerseth 2010).On the quantitative side, few studies did so.An early study by Ho (2002) assessed the implementation of and compliance with the Basel Accord on capital adequacy.A major limitation of that study, however, is the little variation on the dependent variable.Overall, our systematic understanding of the implementation of nonbinding agreements is limited.
In this study, I aim to shed light on legal traditions as an important influence on the implementation of nonlegally binding agreements.Specifically, I argue that the impact of legal traditions on the implementation of nonbinding commitments is the opposite of their impact on treaty ratification: common-law countries are less willing to enter legally binding treaties relative to civil-law countries; yet when it comes to nonbinding agreements, common-law countries are more prone to implementation than their civil-law counterparts.Why would this be the case?
Compared to treaties, nonlegal agreements better fit the legal culture that characterizes common-law systems.The quintessential quality of common-law systems is their bottom-up evolution: law is made by judges as a means to solve specific social problems.Legal rules evolve gradually; they are sensitive to the social environment in which they operate and correspond with its values (Hutchinson 2005;Gennaioli and Shleifer 2007).Treaties, however, are inconsistent with the notion of local, organic, and socially adaptive law.Foreign documents produced by international political deals, treaties are imposed top-down on the legal system, do not necessarily reflect its values, and may not fit comfortably with its existing rules and practices.An analysis of the Australian approach to international law characterizes that approach as "[a]nxiety . . .fueled by a perception (akin to a form of legal xenophobia) that international law is an intrusion from 'outside' into our self-contained and carefully bounded legal system;" the judiciary worries that "the use of international norms will cause instability in the Australian legal system" (Charlesworth et al. 2003, 424, 446).These concerns lead common-law countries to greatly value the flexibility of international rules and the liberty to modify or disregard provisions that are incompatible with domestic laws and policies.Flexible rules may be implemented in a way that gives due consideration to existing legislative and constitutional obstacles, and they permit adjustment and development in light of experience (McLean 2012, 179).
Treaty reservations may serve as a possible means to introduce flexibility.Indeed, Powell and Pickard (2010) find that common-law countries are more likely to place reservations on their alliance agreements.Yet reservations offer only a partial solution.They allow states to limit their commitments, but do not afford full freedom in implementing the treaty and adjusting it to local conditions.Furthermore, some regard reservations as a problematic tool that might undermine the treaty.A commitment that is heavily qualified by reservations may seem insincere (Simma and Hern andez 2011).In addition, reservations do not allow adaptation of the legal rule in time in light of experience or changing circumstances.
A nonbinding agreement can be integrated into a common-law system more easily than a treaty, even one subject to reservations.Since the agreement lacks legally binding force, it need not be implemented as a whole.Instead, a state may pick and choose only those provisions that meet local needs, address local problems or gaps, and are consistent with existing rules.Moreover, even those provisions of the agreement that are incorporated into the legal system need not be incorporated verbatim: the implementing legislation may adjust and adapt those provisions to local circumstances and to existing law (Abbott and Snidal 2000, 445).Nonbinding agreements thus feel less of a foreign imposition than treaties: they match the core quality of common law as a system that evolves organically and provides tailored solutions to local problems.Nonbinding agreements can offer important guidance and direction as well as ideas and inspiration, while leaving national authorities the ability to shape their own legal measures.
The integration of nonbinding agreements into a common-law system is not only easier as a conceptual matter, but as a practical one as well.When incorporating a treaty into the domestic legal system, local law has to be brought into line with the treaty's legal obligations: an inconsistency might give rise to a treaty violation (Heyns and Viljoen 2001, 497-8).In a common-law system, assessing the conformity of local law with the treaty-and making necessary changes-presents a challenge since the common law is not found in a single major code.Rather, law exists as an amalgam of statutes and legal precedents.Combing through the numerous legal sources and adjusting them to the treaty could be time-consuming and difficult.By contrast, the process of integrating nonbinding agreements into a common-law system is easier: there is no need to ensure a perfect conformity between the agreement and the existing legal framework, and certain inconsistencies may remain.This would not constitute a legal violation, as the agreement is not legally binding.
Looking at the incentives of domestic actors, one would also expect a lesser resistance to nonbinding agreements, compared with treaties.As Simmons (2009) argues, common-law judges and lawyers develop specific legal skills in extant precedent and might resist the insertion of foreignmade treaty rules into the local system (Hofmann 2010).By contrast, nonbinding agreements pose a lesser threat to legal practitioners since they can be implemented in a way that dovetails with existing law.
Whereas common-law systems are more comfortable with nonbinding agreements than with legally binding treaties, the reverse may be true for civil-law systems.Indeed, civil-law countries may also benefit from the flexibility that nonbinding norms afford.Yet the flexibility of the legal rule comes at the expense of its certainty -and it is the latter that the civil law values more.Nonbinding agreements do not easily fit the civil-law's emphasis on certain, clear, and formal legal sources, and its neat distinction between law and nonlaw.
Virtually all legal systems value certainty and clarity.Yet in the civil-law tradition, certainty "has come to be a kind of supreme value, an unquestioned dogma, a fundamental goal."In the interest of certainty, legislation should be clear, complete, and coherent, and importantly, judges are prohibited from making law: certainty requires making the law judge-proof (Merryman and P erez-Perdomo 2007, 48).More broadly, civil law tends to be legalistic and systematic, and it emphasizes respect for existing rules as established in accordance with the formal procedures prescribed by the system.Indeed, legal norms are defined by their pedigree and must meet predefined formal standards to qualify as law (d'Aspremont 2011).Such a formalist and positivist attitude leads civil law to maintain a strict distinction between law and nonlaw to ensure the predictability and security inherent in the legal rule, in contrast with the legal decision or the moral or religious rule.This separation stems from the view of legal reasoning as having its own logic, isolated from any ethical, political, social, or economic considerations (Jouannet 2006).
Applying such conceptions and attitudes to the international legal system, civil law sees formal international rules as the preeminent means for governing interstate relations (Jouannet 2006, 314).Prominent civil-law jurists express such thinking as they emphasize the binary nature of law: an instrument is either law or it is not; the line between law and nonlaw is a bright and clear one and should remain as such (Pauwelin 2012, 127-8).In the words of Jan Klabbers (1996, 181), "law can be more or less specific . . .more or less farreaching; the only thing it cannot be is more or less binding."These proponents of formalism bemoan the current trend of deformalization of international law and express discomfort with the notion of international soft law.In an early expression of such critique, Weil (1983, 415) decried the "blurring of the normativity threshold" that separates legal and nonlegal obligations.More recently, d'Aspremont (2011) suggested that the move away from formalism is counterproductive: while aiming to promote and expand international law, it breeds uncertainty, contributes to the indeterminacy of rules, and diminishes the authority of international law.By contrast, a formal approach to identifying legal norms would preserve the normative character of international law and bolster its legitimacy and efficacy.
Common-law jurists often denounce the formalistic approach of the civil law-the preoccupation with the "purported virtues of clear rules and written texts" (Pildes 2003, 151).They identify a spectrum of legal normativity and a grey zone between law and nonlaw.For example, Michael Reisman (1988, 376) argued that "soft law performs important functions, and, given the structure of the international system, we could barely operate without it." Common-law systems are content with flexible nonbinding instruments that fall along the continuum between law and nonlaw.In fact, as I explained above, common-law systems are more comfortable with nonbinding agreements than with treaties.By contrast, civil law's emphasis on certainty; the preference for formal, established sources; and the separation of law and nonlaw all result in an inclination toward treaties and a distaste toward nonbinding commitments.Treaties fall neatly in the legal domain and are the international equivalent of the civil code that is a main pillar of the civil-law system (Koch 2003, 24-5).Contrarily, nonbinding international instruments are relegated to the nonlaw sphere and are not seen as an established source of international law.Civil-law systems thus feel more comfortable with treaties than with nonbinding agreements.
Overall, given the common law's affinity for nonbinding agreements and the civil law's aversion toward them, we would expect a higher likelihood of implementation of such agreements among common-law countries.This brings us to this study's main hypothesis: H1: Common-law countries are more likely to implement nonbinding international agreements compared with countries whose legal system is not based on common law.
Cursory evidence suggests a common-law preference for nonbinding instruments, consistent with this hypothesis.The Commonwealth, an international organization of fifty-three countries that belonged to the British Empire, consists mostly of common-law countries.This organization works to harmonize members' laws and practices in the areas of human rights, rule of law, good governance, and social and economic development.The harmonization instruments that the Commonwealth produces take the shape of nonbinding documents, such as principles, declarations, best practices, guidelines, schemes, and model laws that "can be appropriately tailored to the common law systems of individual Commonwealth members" (Commonwealth Secretariat 2010, 18).Examples include the 1994 Victoria Falls Declaration of Principles for the Promotion of the Human Rights of Women; Model Law for the Prohibition of Money Laundering; and the 2015 Kigali Declaration to end child marriage.By contrast, the Council of Europe-an international organization of mostly civil-law countries that works on issues similar to the Commonwealth's-usually adopts legally binding conventions, such as those on violence against women and money laundering. 7he case of cybercrime is instructive.In 2001, the Council of Europe adopted a Convention on Cybercrime (the "Budapest Convention").The Commonwealth based its own cybercrime instrument on the Budapest Convention, but chose to establish a nonbinding model law rather than a legally binding convention. 8A preexisting nonbinding instrument-Harare Scheme on mutual assistance in criminal matters-underwent a revision to address cybercrime (Commonwealth Secretariat 2014).This divergence demonstrates common law's affinity for nonbinding instruments versus civil law's treaty inclination.The following analysis offers more systematic evidence of this divergence.

Implementation of Nonbinding Instruments: Statistical Analysis of the United Nations' Model Commercial Legislation
Studying the implementation of international agreements presents an empirical challenge of systematic data collection.Implementation data are difficult to obtain even for treaties: unlike the ratification of a treaty, which is recorded by the body overseeing the treaty, the passage of treaty-implementing legislation is a domestic matter that is often not monitored or recorded internationally.
With nonbinding agreements, one faces an even greater data challenge.Since such agreements do not strictly bind states, they may implement them in whole or in part, or after introducing any modifications.This makes it difficult to ascertain whether a given law indeed qualifies as an incorporation of the nonbinding instrument into the domestic legal system.The current study overcomes this challenge by analyzing a set of United Nations (UN) model laws that aim to harmonize commercial legislation in three areas: electronic commerce, cross-border insolvency, and international commercial arbitration.These laws provide a rare glimpse into the implementation of nonbinding agreements, as they offer systematic data on implementation.
The model laws analyzed here are, in fact, part of broader international efforts to harmonize commercial law.Commercial-law harmonization seeks to facilitate cross-border exchange by reducing uncertainty and transaction costs.Underlying this goal is the assumption that international legal heterogeneity-the diversity of commercial laws across countries-impedes commerce.The absence of legal uniformity makes it harder for private parties to agree on the legal regime that will govern their cross-border transaction and increases the ex-ante costs of crafting contracts.If a deal is ultimately struck, legal diversity might result in misunderstandings and disagreements that could lead to disputes and litigation (Reich 1997;Gillette and Scott 2005).Legal harmonization seeks to solve this problem by providing uniform legal rules applicable to commercial transactions, to be adopted by all countries.These uniform rules supply a standard language that reduces contract-drafting costs and lowers uncertainty about the parties' rights and obligations (Eiselen 1999;Gillette and Scott 2005).By simplifying the legal foundation of trade and removing obstacles resulting from the diversity of legal regimes, harmonization aims to encourage international economic activity and spur trade. 9 A variety of regional and global organizations promote the harmonization of commercial law.Most important among those is UNCITRAL, which works for "the promotion of the progressive harmonization and unification of the law of international trade" (General Assembly Resolution 2205 [XXI], December 17, 1966).UNCITRAL uses legally binding conventions when it seeks to achieve a high degree of harmonization in the participating states and to provide an assurance that the law in each state conforms to the convention.By contrast, UNCITRAL uses a model law as the vehicle for harmonization if flexibility is in order: "when it is expected that States will wish or need to make adjustments to the text of the model to accommodate local requirements that vary from system to system, or where strict uniformity is not necessary or desirable" (UNCITRAL 2013, 13-14).A model law provides a template for legislation: a legislative text that is recommended to states for enactment as part of their national law.As the model law lacks legally binding force, states that choose to implement it may modify and adapt its provisions.Of the various model laws promulgated by UNCITRAL, I focus on three model laws that were established before 2000 and offer an implementation trackrecord that allows an analysis.
The Model Law on International Commercial Arbitration was established in 1985 and amended in 2006.The motivation for harmonization through a model law came from the considerable disparities in national arbitration laws and their various flaws, such as rules drafted with domestic arbitration in mind that poorly suit international arbitration.These problems resulted in unexpected and undesired restrictions, difficulties, and frustration for the parties to international arbitration.To remedy this situation, the UNCITRAL model law presents a special legal regime tailored to international commercial arbitration.This regime covers all stages of the arbitral process, from the arbitration agreement to the recognition and enforcement of the arbitral award (UNCITRAL 2008).Note that the model law addresses international commercial arbitration, which is different from investment arbitration that BITs typically call for (Elkins, Guzman, and Simmons 2006).The latter is a means to resolve a dispute between a foreign investor and the host state.By contrast, international commercial arbitration usually settles a dispute between private parties, and it covers a range of commercial relationships beyond investment, such as trade, financing, and insurance10 (Mattli and Dietz 2014).
The second UNCITRAL model law examined is the 1996 Model Law on Electronic Commerce, aimed at increasing legal predictability and removing legal obstacles for commerce conducted through electronic means.Such obstacles result from existing legislation that imposes restrictions on the use of modern means of communication, for example, by prescribing the use of "written," "signed," or "original" documents.The model law seeks to establish a legal environment conducive to e-commerce through rules based on the principles of nondiscrimination and functional equivalence.The former ensures that a document would not be denied legal effect, validity, or enforceability solely because it is in electronic form.The latter sets out the specific requirements that electronic communications need to meet to fulfill the same functions that certain notions in the traditional paper-8 Commonwealth Model Law on Computer and Computer-related Crime.9 On the trade-facilitating effect of shared legal institutions, see Powell and Pickard (2010).
The third model law analyzed here is the 1997 Model Law on Cross-Border Insolvency, designed to harmonize and modernize the laws for cross-border proceedings concerning debtors experiencing severe financial distress or insolvency.Cross-border proceedings take place when the debtor has assets in more than one state or when some of the creditors are not from the state where the insolvency proceeding is taking place; such cases became more common with the global expansion of international trade and investment.Yet national insolvency laws often fail to deal with cross-border cases, impeding the rescue of financially troubled businesses.Furthermore, the uncertainty over the handling of cross-border insolvency may discourage cross-border investment.Addressing these problems, the model law focuses on certain procedural aspects of crossborder insolvency.Its main features include access to local courts for representatives of foreign insolvency proceedings and for creditors; recognition of foreign insolvency proceedings; relief to assist foreign proceedings; and cooperation with foreign courts or foreign insolvency representatives (UNCITRAL 2014).
UNCITRAL's model laws provide good ground for analyzing the variation in the implementation of nonbinding commitments thanks to the availability of implementation data.As legal harmonization aims to promote cross-border activity, UNCITRAL wishes to publicly identify countries that harmonized their laws and offer a legal environment that welcomes cross-border exchange.On their part, countries that implement the model laws have an incentive to publicize this fact in order to enjoy a boost to trade and investment.When governments or practitioners bring a relevant enactment to the awareness of the UNCITRAL secretariat, the latter verifies that the enactment sufficiently conforms to the model law.The model law's flexibility complicates the evaluation of conformity: states may deviate from the model law by dropping or modifying provisions or by adding new ones.The secretariat thus does not require perfect consistency with the model law.Rather, it verifies that the enactment incorporates the model law's basic principles and that it is a viable piece of legislation, for example, in terms of coherence.The resulting data, available on UNCITRAL's website, give a fairly complete and reliable picture of each model law's implementation worldwide.

Method
To test the impact of legal traditions on the implementation of each of the three UN model laws, I employ eventhistory modeling that estimates the "risk" that an event of interest-the enactment of legislation based on the model law-will occur as time elapses.The primary model used is the Cox proportional hazards model, which is widely applied in the study of treaty ratification (Neumayer 2008;Simmons and Danner 2010;Haftel and Thompson 2013;Schneider and Urpelainen 2013).The results of the Cox model appear as hazard ratios that express the proportionate impact of a given variable on the decision to implement a model law.Values higher than 1 increase and values lower than 1 reduce the likelihood of implementation in any given year for which implementation has not already occurred.Once a country implements a model law, it exits the analysis.The unit of analysis is country-year; the temporal coverage begins in the year in which the model law was established (1985, 1996, and 1997 for the arbitration, e-commerce, and insolvency laws, respectively) and ends in 2012.
In addition to the Cox model, I employ two other event-history models to increase the robustness of the results.The first is a Weibull regression.Whereas in the Cox model, the particular form of the baseline hazard rate is assumed to be unknown and is left unparameterized, in a Weibull model, the baseline hazard rate is assumed to have a particular parametric form.Since I make no assumptions about the nature and shape of the baseline hazard, I employ Cox as the primary model and Weibull as a check.A second check involves discrete event-history analysis, which uses a logistic regression combined with a cubic polynomial to adjust for time dependencies.It is particularly appropriate when data are collected in large increments of time, such as years, as is the case with much of international relations analysis, including the current study (Box-Steffensmeier and Jones 2004).

Variables
The dependent variable is the passage of legislation based on each of the three UNCITRAL model laws: electronic commerce, cross-border insolvency, and international commercial arbitration.The key independent variable is a country's legal tradition; more specifically, whether the legal system is based on common law (1) or not (0).Legal-tradition data come from Mitchell and Powell (2007); as a robustness check, I also use the legal-system categorization provided by La Porta, Lopez-de-Silanes, and Shleifer (2008).
While this study focuses on legal traditions, it presents an opportunity to examine additional influences on the implementation of nonbinding agreements.In particular, democracy, veto players, and regional behavior have been identified as important influences on treaty ratification.By including them in the model as controls, we may gain additional insight into the similarity or difference between binding and nonbinding commitments.

Democracy
The tendency of democracies to join treaties is one of the consistent findings in the treaty-ratification literature.For instance, by joining preferential trade agreements, democratic governments can commit to a lower level of protectionism and thereby signal to voters that special interests do not dominate trade policy (Mansfield and Milner 2012).Since the median voter favors environmental protection, election-mindful governments seek to provide such protection by committing to environmental accords (Neumayer 2002;von Stein 2008;Bernauer et al. 2010).
Are democracies more likely to implement nonbinding commitments, similar to their affinity for treaty ratification?Democratic governments that wish to survive politically must deliver policies that benefit a significant number of voters (Bueno de Mesquita et al. 2003).To the extent that international nonbinding instruments can facilitate the provision of public goods and garner voter support, democratic governments may implement them in the domestic legal system.On the other hand, one might expect nonbinding agreements to be less attractive for democratic governments than treaties.The reason is that treaties' appeal as a means to increase popular support hinges on their visibility.Nonbinding agreements have less salience and visibility than treaties and are therefore less likely to come into voters' awareness.They may also be perceived as less credible than treaties since they do not typically include monitoring and enforcement mechanisms (Lipson 1991, 500-1;Raustiala 2005, 597).Nonbinding agreements are therefore less politically expedient for democratic governments: they do not establish a visible, credible commitment that would enhance the government's domestic support.Overall, then, democratic and nondemocratic governments may show a similar tendency of implementing nonbinding agreements. 11

Veto Players
Veto players-domestic actors possessing the ability to block policy change-can hinder the ratification of treaties.These players represent certain domestic groups; through their institutional positions, they can thwart the ratification of agreements that might harm those groups' interests.For the government, ratification of a treaty when many veto players exist is costly: either the treaty will have to be modified to match the preferences of the players and the groups they represent, or the veto players will have to be bribed into approving the agreement.Veto players therefore pose a serious constraint that increases the cost and complexity of securing ratification or might obstruct ratification altogether (Mansfield and Milner 2012, 55-8;Haftel and Thompson 2013).
Yet nothing about the logic of veto players is unique to treaty ratification.As veto players are capable of blocking policy change, they may delay or obstruct the passage of legislation implementing a nonbinding agreement (Shaffer and Pollack 2010, 742-3;Lupu 2015).The agreement itself may not be subject to a formal vote, but a legislative vote is necessary for enacting the implementing law, and it is here that veto players can exercise their influence.As the number of veto players increases, the likelihood of implementation should decline. 12

Regional Implementation
A state's decision to ratify a treaty may be influenced by the ratification choices of countries in its region.Several causal mechanisms may generate such regional influence: socialization, emulation of neighbors' behaviors, and learning from neighbors' experience, as well as competition with them (Dobbin, Simmons, and Garrett 2007;Bernauer et al. 2010, 518-9).Studies found regional influence on the ratification decision in various issue-areas, from human-rights treaties (Neumayer 2008;Simmons 2009, 90-6) to environmental accords (Bernauer et al. 2010).
The logic of regional impact is not unique to treaties: neighbors' behavior may affect a state's decision to implement a nonbinding agreement as well.A state that observes the successful implementation of a nonbinding agreement by its neighbors may learn from that experience or emulate it, or it may view implementation as necessary in competing for capital or export markets.The impact of regional implementation should be particularly strong with respect to the harmonization of commercial law, which is the domain that this article addresses.Since countries often trade heavily with neighboring countries (Disdier and Head 2008), commercial-law compatibility with one's neighbors may hold importance.If legal systems in the region revise their commercial legislation in accordance with the nonbinding instrument, a country may wish to make similar adjustments to facilitate trade with regional partners (Garoupa and Ogus 2006).I therefore control for the rate of legislative implementation in a country's geographic region. 13

Additional Controls
In addition to democracy, veto players, and regional implementation, I control for the size of the population and for gross domestic product (GDP) per capita: 14 the model laws may particularly benefit poor, developing countries that do not already have modern commercial legislation (UNCTAD 2005, 25).Countries that seek to attract foreign direct investment (FDI) may implement the model laws to appeal to investors.I therefore control for the ratio of inward FDI flows to GDP. 15 The level of law-and-order might also influence implementation: countries that rank low on law-and-order may wish to implement the model laws to modernize their legislation and as a signal of a legal environment that is conducive to trade and investment. 16In addition, the model for e-commerce controls for the share of merchandise trade in GDP and for the number of internet users per one hundred people; 17 the model for crossborder insolvency controls for the country's status as a major financial center; 18 and the model for international commercial arbitration controls for the country's membership in the 1958 New York Arbitration Convention: an agreement that lays out an important foundation of international commercial arbitration by providing for the recognition and enforcement of foreign arbitral awards. 19 Table 1 provides descriptive statistics of key variables.Full variable description appears in the online appendix.

Results
Table 2 presents a series of Cox models that capture influences on the national implementation of the UN's model laws.Model 1 examines the enactment of legislation based on the e-commerce model law; Models 2 and 3 do so for the cross-border insolvency law and the law on international commercial arbitration, respectively.
Consistent with this study's hypothesis, common law is positively and significantly associated with the passage of legislation based on each of the three model laws.Common-law countries are five times as likely to implement the model law on e-commerce as non-common-law countries.They are eight times as likely to incorporate the insolvency model law into the legal system and twice as likely to implement the model law on arbitration.These findings suggest that common-law systems indeed feel 11 Data source: Polity IVd.
12 Source: Henisz's Political Constraint Index.The percentage of countries in the region that have implemented the model law, lagged one year.Classification of countries by region is based on the State Department's categorization.comfortable with nonbinding instruments, such as a model law.Implementation of the model laws allows common-law countries to modernize their legislation, bring it into line with international standards, and remove obstacles to cross-border exchange.At the same time, they can make such changes in a way that ensures smooth integration into and consistency with the existing legal framework.By contrast, the model laws-which are merely a suggested template for legislation-hold less appeal for civil-law systems.Those prefer formal, legally binding texts that enjoy a high degree of certainty, determinacy, and authority.
Figure 1 illustrates the difference that legal tradition makes.This figure plots the cumulative hazard of implementation of the e-commerce model law by common-law countries as well as non-common-law countries.The contrast between the two groups is clear: the cumulative hazard of implementation rises steeply for common-law countries; it rises much more slowly for non-common-law countries.
While common law is positively and significantly associated with the implementation of all three model laws, the control variables are somewhat less consistent, but overall similar, across models.Importantly, democracy does not reach statistical significance in any of the models: democracies are not more (or less) likely to implement the model laws than nondemocracies.The domestic political benefits that a democratic government may reap from nonbinding instruments, such as the model laws, are apparently lower than the political dividends that may come from treaty ratification.Therefore, the democratic inclination for treaty ratification does not manifest itself in the implementation of nonbinding commitments.By contrast, the negative influence of veto players-well familiar in the context of treaty ratification (Mansfield and Milner 2012;Haftel and Thompson 2013)-appears even when the instrument is not legally binding.Veto players are negatively associated with the likelihood of implementation in all three models and are statistically significant in two of them: e-commerce and arbitration.The substantive effect is large: when the veto-players measure increases by one standard deviation, the likelihood of implementing the ecommerce law drops by 38 percent.This shows that the capacity of veto players to block policy change goes beyond the obstruction of treaty ratification: the constraints that they pose could also make it harder to pass implementing legislation pursuant to a nonbinding agreement.
The results also confirm that implementation may be influenced by the behavior of countries in one's own region.The propensity of a country to implement the model laws on e-commerce and arbitration increases with the share of other countries in the region that have implemented these laws.Contrary to expectations, however, GDP per capita is positively associated with the likelihood of implementing the model laws; that is, richer countries are more likely to implement.Perhaps this stems from developing countries' mistrust in international trade initiatives (Wade 2003).Alternatively, poor countries may lack the means, knowledge, or expertise for establishing rules and regulations to supplement the model laws, which provide only a legislative framework (UNCITRAL 1999, 19).Table 3 offers a set of robustness tests.Models 4 and 5 re-estimate Model 1 using alternative estimation methods and alternative measures for two of the variables.As for methods, Model 4 uses a Weibull regression, whereas Model 5 employs discrete-time analysis (logistic regression with a cubic polynomial).In terms of measures, Mitchell and Powell's coding of legal traditions changes to that of La Porta and colleagues.According to Mitchell and Powell's count, there are forty-six common-law countries in the dataset; when using La Porta et al.'s classification, the number of common-law countries rises to sixty-three.The reason is that Mitchell and Powell identify "Islamic law" and "mixed legal systems" as distinct categories, in addition to common law and civil law, while La Porta et al. classify legal systems as having an English origin or as one of the variants of civil law (French, German, or Scandinavian).Thus, some of the systems that Mitchell and Powell consider Islamic (such as Pakistan and Sudan)  The alternative measures and estimation methods do not substantially change the key result: English origin is strongly and positively associated with the likelihood of implementing the e-commerce model law.The same holds true for the other model laws.Models 6 and 7 re-estimate Model 2 (insolvency) and Model 3 (arbitration), respectively, through discrete-time analysis.Once again, common-law countries have a significantly higher likelihood of implementation.The results for the controls hold as well: democracy does not affect the implementation of the model laws, whereas veto players could hinder implementation.
Finally, Powell (2015) suggests that Islamic-law countries generally prefer nonbinding venues of international disputes resolution.Does this preference for nonbindingness lead Islamic-law countries to implement nonlegal instruments?Model 8, which uses Mitchell and Powell's full coding of legal traditions, shows that this is indeed the case.Compared to civil law (the reference category), both common-law countries and Islamic-law countries are more likely to implement the model law on e-commerce.Mixed legal systems, which include some common-law elements, are also more likely to implement (p ¼ .104).

Conclusion
This study finds a statistically significant and substantially large impact of a country's legal tradition on its inclination for cooperation.Common-law countries prove much more likely to implement nonbinding international commitments than those outside of the tradition.Further research will need to corroborate this finding for other nonbinding instruments.At this point, however, we can draw some important implications.
First, this study highlights the significance of implementation as an important process that domesticates international norms.Understanding the influences that shape this process allows us to gain a better understanding of international norms' effects and their variation across countries.
Second, this study's result matters to existing work that identifies an aversion of common-law systems toward legally binding commitments.These studies create the impression that common law is generally unfavorable to international agreements and commitments.My analysis corrects that impression.Common-law systems are uncomfortable with legally binding agreements.By contrast, common-law countries hold a more favorable view of nonbinding agreements, as these do not generate the same sense of foreign imposition associated with treaties.For common-law countries, nonbinding agreements achieve the same goal as treaties-international policy coordinationbut in a "gentler" way: one more consistent with the logic and nature of their legal system.This finding comports with Mitchell and Powell (2011) who find that countries are more amenable to making international commitments that reflect the principles governing their domestic legal systems.At the same time, it suggests caution and nuance in characterizing states' attitudes to international commitments: countries' views of legal commitments may differ from their approach to nonlegal commitments.
Third, this study speaks to scholarship that examines how states choose between legal and nonlegal agreements.This literature focuses on functionalist concerns of credibility and flexibility, as well as on power-related considerations (Abbott and Snidal 2000;Guzman and Meyer 2010).Raustiala (2005) weaves domestic political institutions-especially the process of ratification-into his analysis of agreement design.Such work fails to identify legal tradition as a factor that affects the form of international agreements.Yet the finding here suggests a common-law bias toward nonbinding commitments, which may affect the design of agreements.Since common-law systems are more comfortable with nonbinding commitments, they may prefer to establish agreements that are nonlegally binding.This preference could meet the resistance of civil-law countries that have a strong inclination toward treaties.Analysis of institutional design should take these conflicting preferences into account.
Fourth, this study challenges the conventional wisdom regarding the easier domestic acceptance of nonbinding agreements.Lipson (1991, 515, 518), for example, argues that treaty ratification can be "very slow and painful," whereas nonbinding agreements allow governments to act "quickly and quietly."Raustiala (2005, 597-8) similarly suggests that nonbinding agreements are less prominent than treaties and that domestic interest groups are typically less aware of them.Such agreements therefore insulate governments from domestic political pressures.My findings, however, suggest caution: the domestic process of implementing nonbinding agreements may prove faster than that of treaties in the absence of ratification.Yet implementation may still involve the passage of legislation that veto players can derail: the presence of veto players negatively affects the implementation of the model laws on e-commerce and arbitration.We should not treat nonbinding agreements as a panacea for overcoming domestic political constraints; governments should not opt for such agreements with the expectation of smooth domestic acceptance.
The relative ease of domestic acceptance is but one of the benefits of nonbinding agreements-or soft law-that the literature identifies.Abbott and Snidal (2000, 423) argue that "[s]oft law offers many of the advantages of hard law, avoids some of the costs of hard law, and has certain independent advantages of its own."This study suggests, however, that common-law countries may be more likely than civil-law countries to enjoy the advantages of soft law by implementing it in their legal system.Yet the majority of countries in the world-some two-thirds-belong in the civil-law tradition.This means that they might be less enthusiastic about nonbinding agreements and less likely to implement them.We need to keep this in mind when assessing the relative merits of binding and nonbinding agreements. 13

14Source:
World Bank's World Development Indicators.These variables are logged.15 Source: FDI data are from UNCTAD; GDP data are from the World Bank's World Development Indicators.This variable is logged.16 Source: International Country Risk Guide (ICRG). 17 Both are from the World Bank's World Development Indicators. 18 Countries that are members of the Basel Committee on Banking Supervision were coded as financial centers.19 Convention on the Recognition and Enforcement of Foreign Arbitral Awards.Ratification data are from UNCITRAL: http://www.uncitral.org/uncitral/en/uncitral_texts/arbitration/NYConvention_status.html(accessed November 26, 2015).

Figure 1 .
Figure 1.Cumulative hazard of implementation of the ecommerce model law

Table 2 .
(Box- Steffensmeier, Reiter, and Zorn 2003)s model com-The table reports hazard ratios.Standard errors in parentheses.All models are tested for the proportional hazards assumption with the Schoenfeld test.The Law-and-Order variable in Model 2 violates this assumption.The model therefore includes an interaction of that variable with the natural log of time(Box- Steffensmeier, Reiter, and Zorn 2003).*p < .1,**p < .05,***p < .01

Table 3 .
Robustness checks such as Israel and South Africa) are coded by La Porta et al. as having an English origin.In another changing of measures, Models 4 and 5 employ the World Bank's Rule of Law ranking instead of the ICRG's Law and Order.
Model 4 is a Weibull model; Models 5-7 are discrete-time models with a cubic polynomial; Model 8 is a Cox proportional hazards model.The table reports hazard ratios for Models 4 and 9. Standard errors in parentheses.*p < .1,**p < .05,***p < .01 or mixed (