Review of existing literature

To understand the political dynamic of social issues in trade, the extant literature on trade policy-making is the natural departure point. This literature is quite vast and allows for a relatively good understanding of domestic factors responsible for the shape of traditional, at-the-border trade policy issues. Despite some disagreements as to whether the benefits of trade liberalization accrue mainly to societal interests divided along class or industry lines and the role of factor mobility in this process (Frieden 2002; Hiscox 2001; Rogowski 1987), there is broad agreement that international trade creates domestic winners and losers. It is typically argued that the winners from trade liberalization - mostly export-oriented industries -will lobby their governments for more trade openness, while the losers - mostly domestically oriented, import-competing industries - will do the opposite and demand more protection. The respective influence of these competing groups, hinging on their ability to overcome the collective action problem and organize politically, will determine the policy choices that governments will pursue along the continuum between autarchy and free trade (Milner 1988).

Delving deeper into the domestic process of trade policy-making, many studies also emphasize the importance of trade policy coalitions in trade policy formation and agenda-setting (Alt et al. 1996). Their success is thought to depend on the ability to mobilize and use institutional channels provided by the domestic political system, ranging from formal and informal lobbying to electoral mechanisms of interest aggregation (e.g., Bailey et al. 1997; Ehrlich 2008; Lohmann and O'Halloran 1994). Both formal and socio-economic institutions are important for channeling the preferences of various societal groups into policy outcomes because they are the key mechanisms determining different patterns of interest representation (Garrett and Lange 1995; Milner 1999). These institutions influence both the distributional policy demands of the private sector and the macroeconomic constraints on governments. They also determine whose preferences are satisfied in the policy-making process and the overall responsiveness of the government to societal actors. In addition to the material interests of societal actors, constructivist scholarship on trade policy-making also highlights the importance of non-material factors, such as actors' ideas and beliefs, in explaining trade policy outcomes in both the EU and the U.S. (e.g., Goldstein 1993; Siles-Briigge 2011, 2014).

At the same time, very little is known about trade policy-making concerning non-traditional issue areas, including behind-the-border provisions, partly because of the novelty of this phenomenon and partly because of the inadequacy of traditional approaches. Young (2007) argues that policy-making on traditional and social trade issues is characterized by two distinct patterns of domestic politics and the differential role played by society in them. While extant political economy models are well suited to explain the outcomes of traditional trade policy by attributing causal power to the interests of the winners and losers of trade liberalization, they do poorly in trying to explain the outcomes of social trade policy because the distribution of costs and benefits in this case is less clear and significant uncertainty exists regarding the goals of such policy. It has also been noted that societal actors' influence can vary greatly by issue area. For example, Young and Peterson (2006) point out that new trade politics mobilize different kinds of actors, ranging from industries to civil society groups. They also expect that executive actors negotiating trade policy will have more influence on these issues than interest groups, whose stakes might be not as clear-cut as in the case of traditional trade issues.

While the analysis of trade policy-making institutions in the U.S. has achieved great depth, there exist fewer studies of EU trade policy-making institutions that explore its domestic sources as extensively, perhaps owing to the unique supranational nature of the EU and the complexities of interest representation within it (e.g., Conceicao-Heldt 2011; Diir 2010; Meunier 2000, 2005; Siles-Briigge 2014; Woolcock 2012). Some scholars emphasize that the institutional arrangement of

EU trade policy allows the member states to exercise more influence on non-traditional issues when negotiating trade agreements (Meunier and Nicolai'dis 1999). Although the European Commission has full authority to negotiate traditional trade issues on their behalf, its mandate on non-trade issues is more ambiguous, as it shares this authority with governments because of a special clause concerning new trade issues, such as services and intellectual property rights (Dur and Zimmermann 2007; Meunier and Nicolai'dis 1999). Overall, the question about the relative weight of these different actors in the policy-making process pertaining to new trade issues both in the EU and the U.S., as well as the institutional dynamic behind it, are not well explored in the existing literature.

As mentioned earlier, social standards have become an integral part of modern PTAs, so International Political Economy (IPE) literature dealing with trade regionalism and bilateralism also needs to be consulted. The first wave of preferential trade liberalization in IPE naturally focused primarily on the causes of PTAformation (Baldwin 1993; Hicks and Kim 2012; Mansfield and Milner 2010; Mansfield and Reinhardt 2003; Ravenhill 2009). Scholars have notably argued that PTAs are fonned because they increase bargaining leverage within the multilateral trading regime (Mansfield and Reinhardt 2003) and that the willingness to form a PTA heavily depends on domestic institutional factors, such as regime type and the number of veto players (Mansfield and Milner 2010). At the same time, it has been noted that domestic economic interests also drive the formation of PTAs, especially “defensive” North-South PTAs (Manger 2009). These interests are wary of possible trade diversion arising from the first mover advantage that occurs as a result of the PTA formation by a competing economic power. It has been further found that such PTAs are also more likely to be formed between countries that are not only economically but politically interdependent (Manger and Shadlen 2014). While this literature has been vast and has illuminated a lot of important political dynamics, it generally tends to ignore new trade issues, including social standards.

Many studies of PTAs have also explored their effects on various economic, political, and social outcomes, including increased volumes of trade (Egger et al. 2011; Gray 2014; Hicks and Kim 2012), foreign direct investment (Blithe and Milner 2014; Manger 2009), exchange rate regimes (Copelovitch and Pevehouse 2013), democratic transition (Pevehouse 2005), domestic economic reform (Bac-cini and Urpelainen 2014), militarized disputes (Haftel 2011), compliance with human rights norms (Hafner-Burton 2009), and even access to potable water (Rudra 2011). Scholars have largely attributed the various positive effects of PTAs to their role as credible commitment devices in inter-state relations and their legally binding effects. While this literature has greatly improved our understanding of PTAs, the causes and consequences of behind-the-border regulatory provisions, including social standards, which now comprise such a significant proportion of PTAs and have proven to be crucial in their negotiation, have not yet received the scholarly attention they deserve.⁷

A more recent wave of PTA studies has started exploring the design of various PTAs, devoting much-needed attention to the variation in institutional features of the “spaghetti bowl” of trade agreements comprising the global economic system, including their coverage of different trade and non-trade issue areas, stemming from human rights to security (Aggarwal and Govella 2013; Diir et al. 2014; Haftel 2011; Johnson and Urpelainen 2014; Kucik 2012). The Design of Trade Agreements (DESTA) project (Baccini and Urpelainen 2014) seeks to provide exhaustive data on PTAs and their various provisions. While this is a very welcome step, the design of social standards is yet to be included in the database. Milewicz et al. (2016), in their comprehensive study, explore the inclusion and replication of non-trade issues, including labor and environmental provisions, across various agreements, focusing on the questions of power and institutional costs. However, they do not explain variation in the recurring patterns of social standards inclusion. Some scholars also tried to explain variation between different legal features of PTAs, drawing on the social constructivist approach (Duina 2006). This literature has examined the economic and political sources of this variation in PTA design but has not yet explored the effects of various design features on different agreement outcomes. While providing a more nuanced understanding of the various institutional forms of PTAs in the global economy, these studies tend to neglect the institutional design of social standards and its effects.

Several existing studies explicitly examined the linkage of social standards in PTAs (Grynberg and Qalo 2006; Horn et al. 2010). Hom et al. (2010) compared labor and environmental standards in EU and U.S. PTAs, pointing at the significant variation between them (e.g., varying legal inflation) that serves as a departure point for this project. However, while providing a thorough description of social standards in PTAs and their comparison across various agreements, these studies leave the domestic determinants of this variation unexplained. Kerremans and Gistelinck (2009) attempt to explain labor standards in EU and U.S. PTAs, highlighting the role of political parties. Yet, little attention is devoted to the role of trade policy executives and societal preferences in this process. Oeliri (2014) compares the governance of labor standards in EU and U.S. PTAs but does not explain where these different modes come from or how they might impact partner states. Morin et al. (2018) provide large-# data on environmental standards in PTAs, pointing at the factors that are most likely to lead to their inclusion. Raess and Sari (2018) do the same for labor standards, exploring their evolution over time. Both studies provide a comprehensive picture of existing labor and environmental commitments in PTAs but do not test their effects, nor explain why there is persistent variation in various countries' approaches to the linkage between trade agreements and social standards. Raess et al. (2018) explore the inclusion of labor standards in PTAs, focusing on the role of labor unions, regime type, and government ideology. The study helps us understand the logic of labor standards' inclusion; however, it does not explore the question of persistent variation in their design between EU and U.S. agreements or the process behind it.

Some studies in IPE specifically examined the effects of liberalized trade in general, and PTAs in particular, on labor rights and environmental standards. For example, Greenhill et al. (2009) find a positive association between trade and labor rights in developing countries, attributing it to the "California effect.”

Hafner-Burton (2009) examined how various legal features of PTAs might affect compliance with human rights norms and where these norms come from, also looking at the role of labor rights. While the effect is found to be positive, we still lack an understanding of social standards as a rather distinct type of new trade policy, as outlined in the preceding typology. Furthermore, her study points at differences between "hard,” that is, legally binding, and "soft” PTAs in terms of their effectiveness for improving the protection of human rights. However, while being a step in the right direction, it fails to explain the effects of various enforcement provisions in these PTAs, which could matter independently from the degree of legality. Overall, there has been a lack of scholarly attention on the implementation of various PTA provisions, including social standards. For example, a recent study by Gray (2014) based on the survey of PTA bureaucracies, finds that most of the North-South PTAs are never implemented because of the lack of institutional and physical capacity in the developing world. Does this finding travel to the case of social standards? Answering this question begs for a more refined understanding of various implementation mechanisms specific to concrete agreement characteristics, such as the exact design of social standards.

Several studies explore the effects of labor or environmental standards in EU and U.S. PTAs. Kim (2012) looks at the effect of labor standards in U.S. PTAs and finds a significant effect on the improvement of labor rights in partner states, attributing it to their "ex ante due diligence,” whereby U.S. trading partners try to make themselves look more appealing to fair-trade-minded American constituents or risk not attaining an agreement. While this mechanism seems plausible, it is not clear if it would also work in the case of labor standards in EU PTAs, which tend to be less coercive and, therefore, might appear less credible to agreement partners. Furthermore, the enforcement arrangements for labor standards are not theorized. Postnikov and Bastiaens (2014) explore the effects of labor standards in EU PTAs, theorizing that there is a positive ex post effect through the implementation process, contrasting it with Kim's (2012) findings. Bastiaens and Postnikov (2017) apply this logic to the case of environmental standards in EU and U.S. agreements and find support for both the ex ante and ex post hypotheses. Lechner (2016) analyzes variation in the levels of legalization of non-trade issues in North-South PTAs, including labor and environmental standards. While differences in such levels are indeed quite large and vary from one agreement to another, there is a great deal of continuity between EU and U.S. approaches that evolve quite slowly, in stages, and are not greatly influenced by the choice of a partner.

In sum, the literature reviewed here fmds a positive link between trade and social outcomes but fails to examine different institutional mechanisms, such as specific agreement design features, which might lead to this improvement. Furthermore, these studies are often conducted in a large-jV fashion and do not explicitly test the causal mechanisms through which these effects are produced. Case study research can further clarify causal mechanisms at play and how they differ among trade agreements and compare them across various social issues. A case study of the North American Free Trade Agreement (NAFTA) by Aspinwall

(2009) finds that there has been a positive effect on the environmental sector in Mexico as an indirect result of institutional learning by civil society actors. Nolan Garcia (2011) finds a similar effect of NAFTA with regard to labor rights, pointing at the role of transnational advocacy networks. Are such mechanisms at play in bilateral PTAs, and how does the design of social standards matter? Several studies explore how these mechanisms play out in EU PTAs. For example, Orbie et al. (2017) explore the effects of labor standards in the EU-Peru PTA, finding the lack of implementation. Smith et al. (2018) arrive at similar conclusions in the case of the EU-Moldova agreement. While these case studies are a welcome first step needed to understand the impact of social clauses, they do not explore the larger universe of PTAs and do not put EU agreements into a comparative framework, which precludes generalizability. There is also a lack of mutual awareness of the findings of the quantitative studies reviewed here and the emerging qualitative analysis of social standards in trade.

The literature reviewed here also leaves the question of why there is persistent variation in social standards design between EU and U.S. agreements largely unanswered. This might be partly due to the conceptual difficulty of comparing EU and U.S. trade policies within a single analytical framework and partly due to the availability of empirical data. At the same time, the nascent scholarship on social standards' effects primarily consists of disjointed case studies limiting our understanding of causal mechanisms. Arguably, to understand the consequences of social standards, one also needs to know their causes, including the determinants of various institutional characteristics. Such a systematic analysis has been long overdue.