The multilateral rules that govern world trade are eroding. The World Trade Organisation (WTO) is struggling to make substantial progress on reform and its Appellate Body, which adjudicates in trade disputes, is disabled. At least twenty unresolved WTO disputes remain in limbo. Meanwhile, the trade war between China and the United States is intensifying. In the US, national security considerations (real or imagined) dominate trade policy.
Preferential trade agreements1 (PTAs), however, continue to proliferate. Capitals react to the undermining of WTO rules by reassessing alternative arrangements to keep trade open and predictable, and to defend themselves against lawlessness.
This is the right response but it is not sufficient. The trend towards preferential trade agreements has dangerous systemic implications. If, in a fraught geopolitical environment, multilateral rules are allowed to fail, world trade will become increasingly regionalised and fragmented. This implies a substantial loss of efficiency and a greater risk of trade conflicts within and between regions2.
After a remarkable advance from the 1950s to 2010, by when they covered about 60 percent of world trade, the expansion of PTAs3 has slowed. Nevertheless, their role has been reinforced, since the composition of agreements has shifted markedly towards PTAs that are reciprocal, deeper, broader and more enforceable.
For example, the export coverage of reciprocal trade agreements (ie. excluding unilateral preference schemes such as the Generalised System of Preferences) increased from 50 percent of world trade in 2010 to 55 percent in 2020.
To be clear, this does not mean that most of world goods trade benefits from preferences since – underscoring the importance of the WTO acquis – about half of all Most Favoured Nation (MFN) applied tariffs are set at zero anyway, and PTAs cover only a part of trade among the parties to the agreement, and even then, the preferences they provide for are not always used. Estimates placed the average trade-weighted preference margin at nearly 1 percent in 2008 (Krishna et al 2012).
However, WTO MFN treatment, the counterfactual in standard calculations of the effect of a bilateral trade agreement, can no longer be considered certain. Accordingly, if countries can no longer assume that WTO treatment will continue, the gain they will expect from a PTA will be far larger.
The most prominent and systemically significant PTAs are mega-regional agreements. The European Union is the oldest and most advanced. No new mega-regional agreement has been concluded since NAFTA in 1994. The last five years have seen the revision of NAFTA – replaced by the US-Mexico-Canada Agreement (USMCA) – and three new agreements: the Comprehensive and Progressive Agreement for Trans-Pacific Partnership (CPTPP), the African Continental Free Trade Area (AfCFTA) and the Regional Comprehensive Economic Partnership (RCEP). Countries accounting for about 78 percent of world GDP are now part of a mega-regional agreement.
Since the global financial crisis, the additional trade covered by PTAs has overwhelmingly concentrated among countries in Asia or has involved an Asian partner. Ten agreements explain 93 percent of the additional export coverage worldwide.
Strikingly, eight of these agreements were among Asian countries or involved an Asian country. EU-Canada and AfCFTA are the only two agreements among the top eight ranked by additional export coverage that do not include an Asian country as a partner.
Meanwhile, the trading powers have followed different paths, with China especially active in negotiating new PTAs and the EU pursuing an active and long-standing programme, while the US has stood back entirely from new agreements beyond the renegotiating of NAFTA into USCMA.
Table 1 shows the share of each country’s exports covered by reciprocal agreements for each of the big trading nations, ie. excluding unilateral preference programmes. The 2020 data does not include RCEP.
Note: * the EU is defined as including its members at that time. The average of EU members includes intra-EU trade; the EU column refers to the share of extra-EU trade covered by the customs union’s trade agreements with other countries or blocs. The decline in export coverage for Brazil from 2010 to 2020 resulted from the rapid growth of its exports to China and other Asian nations, with which it did not have trade agreements at the time.
Source: Bruegel based on UN-COMTRADE and NSF-Kellogg Institute.
Notable in Table 1 is the advance in trade coverage of the EU from 2010 to 2020, accounted for mainly by the entry into force of trade agreements with Japan, Canada and Korea, and the exceptionally high agreement coverage that individual EU countries achieved through EU intra-trade and extra-EU agreements.
Also of note is the advance of China and Japan since 2000, when neither country was a party to a reciprocal trade agreement. The table does not include RCEP. Once RCEP enters fully into force, 55 percent of Chinese exports and 65 percent of Japanese exports will be covered by PTAs. The US has a low share of exports covered by trade agreements and is falling behind in promoting market access at home and abroad.
Driven by the search for alliances in an increasingly unstable world order, trade policy uncertainty and other powerful motives, the quest for deeper PTAs will soon cover about two-thirds of world trade. PTAs are almost certain to proliferate further in the coming years.
The trend towards preferential trade agreements has dangerous systemic implications. If, in a fraught geopolitical environment, multilateral rules are allowed to fail, world trade will become increasingly regionalised and fragmented
In many ways, the proliferation of PTAs is a good thing. The trade coverage of PTAs is now so extensive that under any assumption about the future of the multilateral trade rules, they will play a key role in keeping large parts of world trade secure.
In many instances PTAs now include more parties, have coverage that is both broader (eg. market access in services, investment) and entail deeper reforms (‘behind the border’) than under the WTO. They are also increasingly enforceable. ‘Deep PTAs’ can benefit.
However, the world’s most important bilateral trade relationships – those between China, the EU and the US – remain uncovered by PTAs or even partial deals, as negotiations between them have failed or stalled. If WTO rules continue to erode, there will be a rules vacuum among the trade giants (and among the 27,000 bilateral trade relationships not covered by PTAs at all).
Although the three big bilateral trade relationships (China-EU, China-US and EU-US) each account for only 3 percent to 4 percent of world exports (11 percent in total), it is in this arena that the possible fragmentation of the world trading system will be decided.
The giants are each part of PTAs/mega-regionals in their respective regions. In these, they play dominant roles, so – as multilateral rules erode – there is a strong temptation for the giants to push ‘friend- shoring’ or ‘near-shoring’. Firms faced with this uncertainty will be inclined to follow and partition their supply chains regionally, implying duplication and inefficiency but not necessarily withdrawal from global markets.
Many politicians, and the increasing number of commentators who pander to them, are comfortable with this course or advocate it, but regional markets will go only so far – economies do not and cannot rely only on their own regions.
In 2021, 70 percent of North America’s exports, 41 percent of Asia’s, and 31 percent of Europe’s went outside their respective regions, while Africa and Latin America depend on other regions for 85 percent of their exports.
Trade is global, as it has been since at least as far back as the great explorers of the fifteenth and sixteenth centuries. Only today, the connections are far deeper.
The prevalence of global value chains (GVCs), which according to the World Bank, now account for more than half of world trade, means that there is heightened dependence on coordinated international production, trade in intermediate inputs and the crossborder sharing of technology.
For these reasons, regionalisation and trade fragmentation is bound to lead to disruption, duplication and vastly reduced efficiency. A WTO study (Bekkers and Goes, 2022), for example, illustrated how “technological decoupling” could be profoundly damaging to the prospects of all nations, especially developing countries.
A World Bank study (Brenton et al 2022) concluded that “steps toward creating a more ‘hostile’ environment for Global Value Chains, with a shift toward global reshoring to high-income countries and China, could drive an additional 52 million people into extreme poverty … In contrast, measures to reduce trade barriers … could lift almost 22 million additional people out of poverty by 2030.”
Furthermore, the regional blocs are not cohesive, reflecting the differing economic and security interests of members, and between the members and the regional hegemon. Smaller members and the middle powers will be forced into asymmetric deals with the hegemon, try to play off the hegemons against each other, or both. This means that the politicisation and weaponisation of trade relations is likely to increase even within regions.
Looking forward, the most damaging scenario is the one to which the world is presently headed, in which the WTO rules become increasingly eroded and the regional hegemons eschew deals with each other. As stressed by those who early on saw PTAs as “stumbling blocks” of multilateralism (Bhagwati, 2008), the dynamics inherent in the proliferation of PTAs are worrisome.
Already some small and middle-sized countries, including Chile, Mexico and Morocco, and large economies, including Germany and France, are party to numerous PTAs covering 80 percent or more of their trade. In a sense, these nations have already hedged their geopolitical bets, which is a good thing.
But do countries so reliant on PTAs still have a reason to engage in WTO negotiations? And is there a systemic tipping point beyond which reliance on PTAs makes impossible multilateral deals, which require consensus and sacrifice?
This worst-case fragmentation scenario is reminiscent of the trade relations of the eighteenth and nineteenth centuries, when empires and colonial systems competed for markets and natural resources and imposed trade rules and navigation laws in their spheres of influence – frequently resulting in military conflict.
The old idea that trade helps promote peace is often dismissed as simplistic, citing many counterexamples, but it is notable that while wars among the great powers occurred almost continuously over the last six centuries, there have been no wars directly involving the world powers since the 1950-53 Korean War (Roser et al 2016).
The best scenario, which is presently low probability but to which policy must aspire, is a revitalised WTO with enforceable rules and deals among the trade giants that address specifically the key issues that divide them. Indeed, in a globalised economy, it is difficult.
The best-case scenario for the trading system almost certainly begins with an accommodation between China and the United States. In the concluded Comprehensive Agreement on Investment (CAI) negotiations, and in its application to the CPTPP and in other ways, China has signalled it is willing to consider trade reforms that address some of the concerns of partners. Legal experts broadly agree that China is living up to the letter of its WTO obligations (Zhou, 2019).
However, there is no sign that the Chinese leadership is about to change its statist model, and China’s government appears to be reinforcing its reliance on state owned enterprises and to be extending Communist Party control over the private as well as the public sector. The US says that it sees the relationship with China as one of competition, not enmity.
If that is true, eventually, the US leadership will recognise that it will not prevail in that competition by disregarding trade agreements, ignoring WTO rules or decoupling from China, a course that even its closest allies reject.
1. We refer to PTAs as including all forms of preferential trade agreement, including customs unions and free trade agreements under WTO Article 24, which have to cover “substantially all trade”, trade agreements among developing countries that are not subject to the ‘substantially all trade’ provision under the Enabling Clause, and preferences accorded unilaterally under the Generalised System of Preferences (GSP).
2. This analysis is based on Dadush and Dominguez Prost (2023), which is available open access on the World Trade Review portal and includes the calculations and datasets used in this assessment.
3. We refer to PTAs as including all forms of preferential trade agreement, including customs unions and free trade agreements under WTO Article 24, which have to cover “substantially all trade”, trade agreements among developing countries that are not subject to the ‘substantially all trade’ provision under the Enabling Clause, and preferences accorded unilaterally under the Generalised System of Preferences (GSP).
Bekkers E and C Goes (2022) ‘The Impact of Geopolitical Conflicts on Trade, Growth and Innovation’, Staff Working Paper ERSD-2022-09, World Trade Organisation
Bhagwati, J (2008) Termites in the Trading System, Columbia University Press
Brenton, P, M Ferrantino and M Maliszewska (2022) Reshaping Global Value Chains in Light of COVID-19, World Bank Group
Dadush, U and E Dominguez Prost (2023) ‘Preferential Trade Agreements, Geopolitics, and the Fragmentation of World Trade’, World Trade Review 22(2): 278-294.
Mattoo, A, N Rocha and M Ruta (2020) Handbook of Deep Trade Agreements, World Bank Group.
Krishna, P, E Mansfield and J Mathis (2012) ‘World Trade Report 2011. The WTO and Preferential Trade Agreements: From Co-Existence to Coherence’, World Trade Review 11(2): 327-339.
Roser M, J Hasell, B Herre and B Macdonald (2016) ‘War and Peace’, OurWorldInData.org.
Zhou, W (2019) China’s Implementation of the Rules of the World Trade Organization, Bloomsbury Publishing
This analysis is based on Dadush and Dominguez Prost (2023), which is available open access on the World Trade Review portal and includes the calculations and datasets used in this assessment. This article was originally published on Bruegel.