1. The Limits of Embedded Liberalism

The logic of embedded liberalism, developed to explain how governments in western Europe in the 1950s and 1960s restored domestic political support for trade liberalization, relies on the coupling of more open trade with generous social protection, effected via government redistribution (Katzenstein, Reference Katzenstein1978; Ruggie, Reference Ruggie1982). Strong labor unions, typically those with close ties to left-leaning political parties and able to effectively mobilize voters, rewarded governments for developing and maintaining such policies. Countries with proportional electoral systems were especially inclined toward redistribution, as those systems tended to favor center-left parties (Rogowski, Reference Rogowski1987; Iversen and Soskice, Reference Iversen and Soskice2006).

The mass public worried less about trade-induced volatility, given the availability of (in various degrees) publicly funded education, unemployment insurance and retraining, health care, old age pensions. and childcare (Burgoon, Reference Burgoon2001, Hays, Reference Hays2009). This claim is echoed in more recent analyses of the anti-globalization backlash, some of which find that access to compensatory programs (such as Trade Adjustment Assistance in the United States) reduces the appeal of populist right-wing political candidates (and softens the effect of the ‘China shock’ on political outcomes (Kim and Pelc, Reference Kim and Pelc2021; Ritchie and You, Reference Ritchie and You2021)).

Since at least the 1990s, however, observers have questioned the viability of the embedded liberalism compromise. First, the compromise emerged at a time when capital account liberalization was quite limited. Capital controls were a central feature of the Bretton Woods monetary and financial system. The International Monetary Fund's Articles of Agreement offered governments the option to restrict capital (financial) account convertibility, obligating members to allow the free exchange of currencies only for current account transactions.Footnote ² In much of Western Europe, even current account convertibility did not fully emerge until the second half of the 1950s.

Limits on capital mobility made it easier for governments to tax financial as well as productive capital. Tax revenues funded many of the public policies that served to embed trade openness in a system of social protection. Firms could not easily threaten to move abroad in response to high taxes (or in response to labor market and other regulations), and bond market investors – an important source of government borrowing in wealthy democracies – could not dramatically raise interest rate premiums in response to government fiscal expansions or imbalances.

With the shift toward capital account liberalization – the result of policy as well as technology, emerging first in wealthy democracies in the late 1970s and 1980s, and then in low- and middle-income countries in the 1990s and after – the balance -between labor and capital shifted (Przeworski and Wallerstein, Reference Przeworski and Wallerstein1988). French president François Mitterrand's ‘U-Turn’ – the abandonment of many of his planned policies between 1981 and 1983, in response to speculative attacks against the franc as well as dramatically higher sovereign interest rates – suggested to many that financial market actors had become less willing to tolerate the interventionist and expansionary policies associated with embedded liberalism. As part of a broader ideational shift toward neoliberalism, many governments in Europe as well as in other regions turned away from state-led intervention and toward capital account liberalization.

Subsequent scholarship called into question the extent to which ‘globalization had gone too far’ (Rodrik, Reference Rodrik1997), in terms of governments being unable to fund generous systems of social protection, lest they provoke capital flight or dramatically increase borrowing costs while also facing greater exposure to trade competition and shocks. Some scholars noted that claims of a ‘race to the bottom’ in, for instance, corporate taxation were overblown; domestic interest groups and political institutions remained important determinants of social and tax policies (Huber and Stephens, Reference Huber and Stephens2001; Basinger and Hallerberg, Reference Basinger and Hallerberg2004). Others pointed out that institutional investors did not necessarily penalize governments for keeping generous social policies in place, provided that governments did not run excessive fiscal deficits or pursue inflationary monetary policies (Mosley, Reference Mosley2003).

At the same time, the question of whether left-leaning parties indeed could continue to ‘act left’ (Garrett and Smith, Reference Garrett and Smith1999) highlights important features of the embedded liberalism compromise. The compromise assumes that governments have an interest in compensating those who suffer negative effects from trade liberalization. Groups that stand to lose from, or to be exposed to, greater volatility as a result of trade liberalization must be important to governments’ political survival. However, this is not always the case. In the United States, for instance, organized labor often has not played a central role in national policy making. The frequent exclusion of labor from national policy making, including in the 1920s and 1930s, left a much more limited – and, for workers, a much less generous – compromise in the United States (Goldstein and Gulotty, Reference Goldstein and Gulotty2021). Meanwhile, with working-class white voters perceived as important to the Biden administration's 2024 electoral prospects, the Trump-era tariffs on many Chinese imports remain in place. Electoral politics have generated a bipartisan unwillingness to rely on reciprocal openness, the hallmark of the post-war international order (Bowen, Broz, and Rosendorff, Reference Bowen, Broz and Rosendorff2022).

In many low- and middle-income countries, workers – who, by virtue of their abundance relative to capital, ought to capture the gains from liberalization – have an even weaker political voice. In Bangladesh, admittedly an extreme case, the government continues to limit the ability of workers to associate freely and bargain collectively. Meanwhile, the owners of apparel factories – which account for the vast majority of the country's exports – are well represented in the country's parliament. It is therefore no surprise that, even in the face of industrial accidents and international pressure to improve labor standards, the country's workers continue to reap only a small share of the gains from trade liberalization.

A first lesson is therefore that compensation for workers exposed to global shocks, as well as distribution of gains to the owners of relatively abundant factors, depends on various features of national political systems. The viability of ‘package treaty’ solutions is also likely to vary with domestic political institutions and interests. We should expect democratic countries (versus non-democratic countries) to be more inclined to implement, and commit to, mitigating measures as well as flanking policies, either unilaterally or via mutual agreement with other like-minded countries. Within democracies, social protections that are more generous are more likely to occur in the presence of higher rates of unionization, in countries with stronger left and center-left political parties, and in countries with proportional representation electoral systems.

A second lesson from the era of embedded liberalism is that providing compensation requires fiscal capacity. Governments must redistribute (via taxation) some of the gains from openness to those who are more vulnerable, or governments must rely on other mechanisms (proceeds from state-owned enterprises, sovereign borrowing, or foreign aid) to fund compensation schemes. Governments generally hesitate to raise taxes; the mobility of international capital and firms increases this reticence.

The regional and global supply chains enabled by earlier efforts at trade liberalization also have made it more difficult for governments to tax firms. Facilitated by communication and transportation technologies, firms often can easily source from abroad or offshore their directly owned production, reducing their tax liability (and their general presence) in home countries. Accordingly, Mansfield and Rudra (Reference Mansfield and Rudra2021) suggest that technological change and the digital revolution have made it difficult for governments to implement embedded liberalism-style compromises.

Multinational firms also engage in tax avoidance strategies, booking corporate profits in lower-tax jurisdictions, making it difficult, thus far, for national governments to take collective action to impose a global corporate minimum tax.Footnote ³ Of course, governments could tax individual (rather than corporate) income, or they could impose consumption taxes. Yet consumption taxes often are regressive, and most governments have been disinclined to tax the wealthiest individuals (Marti, Martinez, and Scheuer, Reference Marti, Martinez and Scheuer2023).

This means that compensatory policies – at least, to the extent that they involve direct government expenditures – often require sovereign borrowing. This is generally cheaper, in terms of interest rates, for governments of high-income countries where investors assume that default risk is low. Governments of low- and middle-income countries, by contrast, face higher borrowing costs on average.Footnote ⁴

Perhaps more importantly, low- and middle-income countries are more exposed to the global capital flow cycle. When global liquidity is high, such countries may have relatively easy access to finance; but when liquidity is low (and risk aversion is high), capital flow booms reverse, and borrowing costs can skyrocket (Wibbels, Reference Wibbels2006; Ballard-Rosa, Mosley, and Wellhausen, Reference Ballard-Rosa, Mosley and Wellhausen2019). Hence, we can expect the capacity for compensation to vary cross nationally with income per capita (and natural resource endowments), as well as over time (with cycles in global commodity prices and liquidity). Hence, the viability of inward-looking flanking mechanisms may vary with the level of economic development, as well as across the global economic cycle.

2. The Use and Abuse of Flexibility

Mechanisms that redistribute directly from winners to losers via taxes and subsidies introduce fewer international distortions and are therefore more efficient from a social welfare point of view. As the above discussion suggests, however, such mechanisms often are not available: states may not have the political or economic capacity to institutionalize such a redistributive system. The post-World War II liberal international order also relied on outward-looking, mutually agreed measures, allowing for the temporary abrogation of liberalization.

Governments could provide ‘breathing room’ – time to adjust, retrain and retool – to local industries and workers by suspending their commitments to openness negotiated at the GATT/WTO or in preferential trade agreements. The post-World War II trade regime offered governments a range of flexibility instruments in response to unfair trade practices abroad (dumping; some subsidies), national security considerations, balance of payments problems, and even tariff overhang (the gap between applied and bound tariffs).

These flexibility provisions may sustain international agreements longer, and admit a greater variety of states to participate (Rosendorff, Reference Rosendorff2005; Kucik and Reinhardt, Reference Kucik and Reinhardt2008). More flexibility is also associated with deeper agreements – member states, especially democracies, offer greater concessions at the point of negotiation (Baccini, Dür, and Elsig, Reference Baccini, Dür and Elsig2015). Flexibility, however, comes with a cost. The more frequently governments invoke these exceptions, the lower the gains from trade and integration. The use of flexibility provisions also reduces the overall level of cooperation. The incentives for governments to use these provisions is particularly strong in political systems with higher accountability.

While administered protection was, in theory, a useful alternative to direct redistribution and compensation, it was not very successful. It angered trading partners and delayed adjustment out of declining industries. Governments accused one another of using temporary administered protection as a means to disguise (and unjustified) protectionism; the WTO's dispute settlement body often agreed, instructing governments to remove such measures.Footnote ⁵

3. Issue Linkage to the Rescue?

Many governments recently have found domestic compensation – whether it be taxes and subsidies, unemployment insurance, worker retraining, or publicly funded social security – to be too expensive, economically as well as politically. Similarly, flexibility provisions and administered protection have threatened to undermine international cooperation and sew frustration with international trade institutions, while also failing to adequately address the negative effects of trade liberalization.

The inadequacy of the compensation and flexibility mechanisms is due, in some part, to changes in the international economic environment. As low- and middle-income countries abandoned inward-focused development strategies (sometimes at the behest of international financial institutions and in the wake of severe financial crises) and turned their attention outward, wealthy countries found themselves competing with a broader, and more diverse, set of nations.

For the most part, the compromise of embedded liberalism was largely concentrated in wealthy democracies. Although these countries varied in the precise nature of their political institutions and social policies, they generally were characterized by strong labor and human rights protections. There was, in essence, a consensus about a minimum floor, in terms of (for instance) workplace health and safety protections, the right to organize and bargain collectively, and a prohibition on child as well as forced labor. The limited heterogeneity among GATT's early signatories, in which the conditions under which a good was produced were not legitimate grounds for imposing barriers, rendered the ‘product versus process’ distinction feasible. Factory conditions and labor standards might have varied between Canada and Italy, but the differences were relatively small, and they were deemed irrelevant to the competitiveness of exports.

As low- and middle-income countries increasingly participated in international trade institutions, wealthy countries sometimes crafted trade rules which blunted developing countries' comparative advantages. For instance, barriers to agricultural imports typically remained high in the US and Europe, and unilateral trade preference programs, permitted under GATT's 1979 ‘enabling clause’ (officially, ‘Decision on Differential & More Favourable Treatment, Reciprocity & Fuller Participation of Developing Countries’) allowed developed countries to offer more favorable schemes to developing countries but also to exclude from those schemes certain products. The US Generalized System of Preferences scheme, for instance, did not apply to apparel and footwear.Footnote ⁶ Similarly, between 1974 and 1995, apparel and clothing trade under GATT was governed by the Multifibre Arrangement, which allocated quotas for export to developed countries across a wide range of low- and middle-income producers.

With the expanded participation in global trade networks of countries with political institutions that were more diverse, levels of economic development, degrees of state intervention, and environmental and labor standards, ‘process’ considerations became more salient. While much is made of the tensions generated by China's 2001 entry into the WTO, this dynamic was apparent in the late 1980s and 1990s as well. Governments and interest groups in wealthy democracies more frequently argued that ‘behind the border’ policies generated (unfair) trade advantages. Low standards for worker health and safety, weak environmental regulation, or the lack of minimum wage legislation were often labeled as illegitimate. Similarly, the lack of intellectual property rights protections, or burdensome sanitary or phytosanitary measures, were viewed as putting wealthy country exporting firms at risk. Additionally, developed country labor unions and environmental activists expressed concerns that a ‘race to the bottom’ would ensue, as rich countries found themselves increasingly in competition with lower-standards countries in the Global South.Footnote ⁷

To the extent that rich country governments increasingly viewed their firms, products, and workers as competing with countries with weaker standards and with lower capacity to implement generous social protections, the compromise of embedded liberalism became less appealing. Sustaining the compromise could generate capital flight and offshoring, and it often was insufficient to protect firms and workers from import competition. Governments in the Global North thus faced a domestic political dilemma: many firms, consumers, and workers stood to benefit from trade liberalization. However, other firms and workers were exposed to greater competitive pressures, all whilst governments' willingness and ability to compensate were eroding.

Sensitive to the tensions between developing and developed countries, the WTO refused to address labor rights issues. Its member countries disagreed over whether lower domestic standards implied a comparative advantage or an unfair trading practice. In 1998, the organization declared that the WTO was a second-best mechanism for the pursuit of core labor rights, and that the ILO, with its 1988 Declaration on Fundamental Principles and Rights at Work, was the most appropriate venue. The WTO's position was consistent with the global trade regime's attention to ‘product’ but not to ‘process.’ The GATT's nondiscrimination principle implied that the conditions under which something was made – labor laws, environmental degradation, domestic regulation – were irrelevant to the treatment of the good by its trading partners.

Developed country governments responded in several ways. While labor-related issues were taken off the table at the WTO level, developed countries did push globally for harmonization in some areas, such as intellectual property rights and health and safety regulations. They also argued, often successfully, that Article XX of GATT allowed for trade-restricting measures protecting ‘human, animal and plant’ life.Footnote ⁸

With respect to labor rights, developed country governments often employed a different strategy. Even before the WTO took labor-related issues off the table, some governments responded to domestic labor groups' demands by including labor rights-related (or civil and political rights-related) clauses in their preferential trade agreements (Hafner-Burton, Reference Hafner-Burton2009). PTAs offered governments wide latitude to link access to trade and investment with labor, environmental, and other conditions (Baccini, Dür, and Elsig, Reference Baccini, Dür and Elsig2015; Lechner, Reference Lechner2016), and developing countries, eager to expand their export networks, were often willing to accept these legally linked flanking measures. For instance, reflecting pressure from labor unions and a few Democratic members of Congress, the 1984 United States Trade Act directed the US Trade Representative to include ‘internationally recognized worker rights’ in all future US preferential trade agreements. The precise way in which these provisions were included varied; they were placed in a side agreement of the North American Free Trade Agreement rather than in the main treaty. There was a separate – and arguably weaker – adjudication procedure for complaints related to compliance with the labor side agreement. In later US PTAs, labor conditions were included in the main treaty text.

The depth of labor relations provisions also grew over time: the US–Jordan agreement (2001) required governments to enforce their domestic labor legislation, whatever its content. In the US–Colombia free trade agreement, which came into force in 2012, the parties committed to ‘adopt and maintain in their laws and practice’ the ILO's 1998 set of core labor rights. The labor rights provisions are subject to the same dispute settlement procedures as other parts of the trade agreement. During the negotiation phase, the US also insisted that Colombia take various steps to address violence against trade unionists and to enforce worker rights provisions.Footnote ⁹ Furthermore, in renegotiating NAFTA, the US government included stronger enforcement for labor provisions. As a result, the 2020 USMCA won the (somewhat unprecedented) support of the AFL-CIO. Given the legal requirement to include labor rights in US trade agreements, as well as the domestic political pressure to do so (Kay, Reference Kay2011), potential partner countries might be inclined to improve their labor protections ex ante, to make themselves more attractive agreement partners (Kim, Reference Kim2012).

Labor-related provisions are now widely used: as of 2022, the ILO identifies approximately 115 trade agreements with labor provisions (ILO 2022). The inclusion of labor-related (and, to an extent, environment-related) provisions in preferential trade agreements often reflects domestic concerns of firms, industries, and labor unions in wealthy countries. Actors who stand to lose from global competition may not have enough political voice to prevent an agreement entirely, but they may have sufficient leverage to insist on the inclusion of non-trade issues. In wealthy democracies, demands for labor-related provisions often were framed as labor internationalism: workers in rich democracies were sincerely concerned about protecting workers elsewhere in the world, but such calls often were interpreted as veiled protectionism: excluding exports from labor-intensive countries protects firms and workers in import-competing sectors. Indeed, in her analysis of the inclusion of ‘non-trade’ issues in PTAs, Lechner (Reference Lechner2016) finds that strong import competition and downward wage pressure motivate domestic ‘losers’ from trade agreements to lobby their governments for design changes, especially when one partner country has high levels of civil and political rights, but other partners do not. This is consistent with Hafner-Burton and Ron's account (Reference Hafner-Burton and Ron2009), which notes that human rights conditionality is explained largely by domestic political pressures in wealthy countries.

The presence of rights conditions in PTAs offers a mechanism by which domestic and transnational rights advocates can pressure governments to address labor rights violations. Doing so may help boost public support for trade openness in countries exposed to import competition, and to reduce public perceptions that import surges are due to unfair practices abroad. The inclusion of conditions at the negotiation stage, however, offers no guarantee of their implementation.

Many PTAs with labor-related provisions do not provide for regular monitoring of compliance. Rather, just as with commercial policy disputes, these provisions require governments (or, in some cases, other actors with standing) to bring complaints to an administrative agency or dispute settlement body. This means that there must be credible information about violations in partner countries, as well as some expectation of success of the claim. Moreover, even when complaints are brought, national governments and dispute settlement bodies may hesitate to impose penalties, as doing so reduces the material benefits of the trade agreement.Footnote ¹⁰

It therefore is reasonable to question the effectiveness of labor provisions in PTAs. Carrère, Olarreaga, and Raess (Reference Carrère, Olarreaga and Raess2022) consider whether labor provisions in trade agreements are associated with changes in bilateral trade flows, especially between North–South country pairs. Using data for the 1990–2014 period, they find that for Global South countries with weaker labor standards, labor provisions appear to increase exports from Southern countries to Northern countries. This is driven by labor provisions featuring ‘institutionalized cooperation’ (for instance, capacity building), rather than by labor provisions characterized by strong enforcement mechanisms. This finding suggests that not all labor-related PTA provisions are a veil for protectionism – for the subset of governments that are willing to improve their labor standards, such negotiated flanking policies could be a means of increasing exports. Using the same measures of labor rights provisions and a global sample of countries, Raess and Sari (Reference Raess and Sari2020) find that trade agreements with institutionalized cooperation provisions are linked with reductions in labor rights violations in developing countries. This finding stands in contrast to earlier studies of labor provisions, which tend not to differentiate according to their precise type, and which find little effect on labor rights.

Others note, however, that labor provisions typically apply only to workers in the formal sector; high degrees of labor market informality in many countries dilute the overall effectiveness of such conditionalities (LeClerq, Reference LeClercq2022). Moreover, to the extent that the threat to workers in wealthy countries comes more from automation than from workers in countries with lower standards (see Anelli, Colantone, and Stanig, Reference Anelli, Colantone and Stanig2021), labor provisions may lose effectiveness as a possible package treaty component.

Research on PTAs with labor conditions therefore suggests that there is a limited set of conditions under which the elements of “package treaties” might address some of the incidental negative effects of trade liberalization. Indeed, prior research on the determinants of the PTA structure suggests that where governments worry about their ability to monitor the implementation of conditions, they are less likely to sign agreements in the first place. Baccini (Reference Baccini2010) finds evidence for this dynamic in the context of the EU's PTAs. When partner countries are more transparent economically and politically, the EU is more likely to conclude PTAs, and those PTAs are more likely to include flexibility provisions such as escape clauses and anti-dumping protections. Hence, governments of wealthy democracies may view flexibility provisions as appropriate and useful instruments only under certain conditions.

It also is worth investigating more directly whether the inclusion of labor (or other non-trade) provisions in PTAs generates greater public support for those agreements. Do individuals most exposed to trade liberalization – who work in labor-intensive industries, or who are most vulnerable to offshoring – have a more favorable assessment of trade and trade agreements when they are aware of their labor-related clauses?

4. Package Treaties and Issue Linkage

The package treaty solution is a treaty, convention, or agreement ‘that includes legally binding commitments on both trade liberalization and mutually agreed flanking policies’ where these flanking policies (domestic or international) are intended ‘to mitigate negative effects of trade liberalization, or the concerns of domestics shareholders … or both’. The reference to domestic policy suggests a more explicit connection between international liberalization and domestic protection than the embedded liberalism compromise entailed, a more direct (legally or factually) connection between liberalization and ‘behind-the-border’ issues. The package treaty solution also aspires to address trade-related dislocations ex ante and comprehensively (rather than ex post and selectively, as in the case of flexibility provisions).

Package treaties therefore increase the dimensionality of agreements. They can broaden the scope of non-trade issues linked to trade agreements and deepen the obligation to address such issues. As such, package treaties can help preserve joint gains from trade, while also allowing special interests to gain protections. Package treaties also may share some of the features of ‘issue linkage’: trade is linked to other dimensions, to address distributional and political concerns.

The standard issue linkage entails a swap of concessions across issue areas – perhaps tariff reductions in exchange for improved respect for worker rights, or reduced barriers to foreign direct investment in exchange for intellectual property rights protections. Linking access to export markets with more stringent regulations abroad may limit the gains from exporters' comparative advantages, while also reducing losses for trade-exposed groups in importing countries. Presumably, the net gains from trade remain positive on both sides, albeit not as large as they might have been absent the upgrading in labor, or environmental standards, or the protection of exposed firms, industries, and workers. Linkage also promises increased domestic political support for liberalization. At the same time, when negotiations include a wider variety of issue areas, the negotiation process is more complex, or longer in duration and less likely to succeed (Koremenos, Lipson, and Snidal, Reference Koremenos, Lipson and Snidal2001). Moreover, linking standards or behaviors with market access involves some degree of coercion: if a participating state does not follow through on its commitments on non-trade issues, or violates treaty obligations related to the linked dimensions, trade sanctions may be applied.

Package treaties also link trade liberalization with non-trade policies, possibly including compensation to affected firms, industries, and workers, as well as upgraded labor, environmental and governance standards, and opportunities for administered protection and temporary safeguards. These linkages, however, may exist within the trade agreement itself (legally linked flanking policies), or across levels and types of governance (factually linked flanking policies). One might even imagine that some governance of non-trade outcomes (labor, climate, development) could be achieved through private market standards, in which firms have (legal and/or material) incentives to commit to improvements throughout their subsidiaries and supplier networks.Footnote ¹¹

It is important to consider how package treaties might overcome the challenges faced in the previous uses of compensation, flexibility provisions, and issue linkages. To what extent will partner governments worry that the compensatory elements of package treaties will be used opportunistically? Under what conditions will governments have the domestic will and capacity to fund compensatory social policies? Additionally, at the negotiation stage, under what circumstances will governments commit to linked sets of policies, and when might they prefer not to be bound by such commitments? It may be the case that when left or center-left governments hold office, but worry about future electoral turnover, they find package treaties an appealing means of tying the hands of future governments, locking in some level of compensation for those likely to bear the negative (necessary or incidental) consequences of trade liberalization.

Partisan shifts also raise the question of compliance with package treaty commitments. Assuming governments are willing to negotiate (and, where necessary, ratify) such deals, under what conditions are governments willing to implement them? Labor provisions in trade agreements and preference schemes often have been weakly implemented. Partner governments may hesitate to bring complaints against one another; civil society groups have limited resources to participate in adjudicatory processes; and adjudicatory bodies may refrain from rendering judgments that reduce the material benefits of trade agreements.

National governments may sometimes face political incentives not to offer flanking and mitigating measures, especially if the beneficiaries of such measures are not electorally important, or if such compensation provokes resentment among some of their constituents. At the international level, and in the case of mutually agreed flanking measures, it is difficult to imagine an enforcement mechanism in which a failure of one government to compensate its domestic constituents was met with the imposition of barriers to that country's exports – the analog to how labor- or environment-related provisions typically operate. Indeed, to be effective ex post, package treaties may need credible mechanisms for compliance and enforcement; but the very existence of such mechanisms makes them difficult to negotiate ex ante.

Supply chain integration adds even more complexity. Evidence suggests that deeper GVC integration makes legal instruments, such as Bilateral Investment Treaties (BITs), less inviting, especially in democracies. When the linkages between domestic and foreign firms are deeper, there are fewer political gains to be had from taxation and expropriation of foreign firms (Johns and Wellhausen, Reference Johns and Wellhausen2016), rendering BITs less important. Ge (Reference Ge2023) explains BIT withdrawals and exit as a consequence of deep supply chain integration, suggesting that deeper integration is not necessarily helpful for treaty survival. However, Fiorini et al. (Reference Fiorini, Hoekman, Altomonte and Colantone2023) find that trade through global value chains (GVCs) increases PTA formation, especially for backward GVC linkages. The findings raise questions for the package treaty approach, suggesting that multidimensionality may affect treaty negotiation and survival in different ways depending on both the direction of flows, and the depth of integration.