VoxEU Column International trade Politics and economics

Coercive trade policy

In international trade disputes, coercion is often used against governments whose trade practices are deemed unfair. This column presents a theoretical model that offers a new rationale for the greater effectiveness of multilateral compared to unilateral coercion, and hence provides a new argument in favour of commitment to international organisations.

In international trade disputes, coercion is often used against governments whose trade practices are deemed unfair. Trade coercion occurs when a government (the ‘sender’) makes a demand backed by threats to use retaliatory sanctions against a trading partner (the ‘target’) if the latter does not acquiesce to this demand. Trade coercion can be exercised unilaterally or through multilateral institutions. In the case of unilateral coercion, the sender government makes a demand and (if necessary) retaliates one-sidedly, unconstrained by international obligations. A typical example of this access to unilateral coercion was represented by Section 301 of the 1974 US Trade Act, which allowed the US to impose unilateral sanctions on countries whose trade practices were found to be unfair to US interests (Pucket and Reynolds 1996, Schoppa 1999). In the case of multilateral coercion, the sender uses instead an international institution's framework for trade dispute resolution. The WTO Dispute Settlement Mechanism is the leading institution of this kind, and since its inception has handled hundreds of cases. Several preferential trade agreements also include similar institutions, like NAFTA's Dispute Settlement Process or MERCOSUR's Dispute Settlement Mechanism.

How effective is unilateral and multilateral coercion?

While existing evidence is still limited, some interesting results have been uncovered looking at the recent US experience. In particular, examining a large sample of trade threats made by the US, Busch (2000) and Pelc (2010) find that a target of trade coercion from the US is significantly less likely to concede when coercion is unilateral than when it is multilateral. Given that neither GATT nor the WTO enjoys centralised enforcement power, this result represents an interesting puzzle.

Why does unilateral coercion significantly reduce the likelihood of a target conceding?

To answer this question, in a recent paper we develop a theoretical model which analyses the strategic incentives underlying trade coercion under three different institutional settings (Anesi and Facchini 2015). The model depicts a dispute between two states, Home and Foreign, in which the Foreign government is dissatisfied with the trade policy implemented by the Home government. A key feature of trade coercion is that the target government is typically ill informed about the sender government's domestic political constraints (e.g. Bagwell and Staiger 2005). To capture this idea, we assume that the political pressure exerted by the import-competing sector on the government in Foreign is private information, and is only known by the Foreign government. This political pressure plays a key role in shaping its level of resolve – i.e. the severity of its trade sanctions against the Home government – in a potential trade war.

Appraising the actual effectiveness of an international organisation in dispute settlement requires knowing what would happen if that institution did not exist, i.e. if there were no framework of rules governing trade coercion. For this reason, the first setting we examine is one in which unilateral coercion is the only option. The game begins with the Foreign government making a demand. The Home government can concede (ending the game with the implementation of the demanded tariff), or reject it (triggering a retaliatory trade war). In other words, it must decide which concessions are acceptable, that is, which tariff changes it would prefer to make rather than face Foreign's trade sanctions. Since the precise nature of these sanctions is uncertain and crucially depends on the privately observed level of resolve of the Foreign government, the latter has incentives to signal high levels of resolve by making excessive demands about the concessions required from Home to avoid retaliatory measures. Our characterisation of equilibrium outcomes reveals that such incentives lead the Foreign government to make requests that the Home government will not meet, thus causing a retaliatory trade war – even when there exist mutually advantageous policy concessions. This finding provides a possible explanation for the empirically observed lower effectiveness of unilateral coercion in obtaining concessions from target governments.

As we show in the paper, a key factor in determining whether concessions can be obtained with multilateral coercion is the extent to which the sender government can commit to not bypassing the dispute settlement process of the international organisation through which coercion is channelled. To model the different strategic situations that may arise from differences in the sender's ability to commit to the international organisation, we examine two additional distinct variants of the previous model. In the first, the Foreign government is not allowed to bypass the international organisation's dispute settlement process. As a result, multilateral coercion is its only available option. Dispute settlement in this case is modelled by allowing the Foreign government to make a demand to the Home government prior to the international organisation ruling and this assumption is intended to capture the consultations stage of WTO disputes. If the Home government does not concede to the Foreign government's demand, the international organisation issues its ruling, whereas it remains inactive otherwise. As our aim is to investigate the effectiveness of weak international trade institutions – namely, those that have no enforcement power and rely on the sender government itself to implement any retaliatory measures – the Home government is allowed to not comply with the ruling, thus triggering a trade war. Our analysis shows that commitment to the international organisation's ruling makes concessions more likely. Intuitively, the potential IO ruling places a cap on the Foreign government's incentives to signal its resolve with high demands. This results in the latter making more moderate requests, which can be accepted by Home.

In the second variant of the model, the Foreign government is only partially committed to the international organisation's dispute settlement process, in the sense that it can choose between unilateral and multilateral coercion in an additional stage at the beginning of the game, committing itself to that choice. This setting captures the environment created by Section 301 of the US Trade Act of 1974. We show that the mere availability of the unilateral option prevents the foreign government from obtaining concessions in equilibrium. In fact, using multilateral coercion when unilateral coercion is available is perceived as a sign of the foreign government's weakness. Hence, incentives to signal higher levels of resolve to the Home government will lead the Foreign government to make unilateral demands which, as discussed above, cannot be accepted in equilibrium.

Discussion and policy implications

Our analysis provides several insights on the influence of international trade institutions on coercion outcomes.

  • First, we provide a new rationale for the greater effectiveness of multilateral compared to unilateral coercion.

The received view (Pelc 2010) suggests that it is the perceived illegitimacy of unilateral coercion and the importance of reputation that decrease the likelihood of a target conceding.

While resistance to institutionally constrained demands entails the reputational cost of being branded a violator, resistance to unilateral threats – regarded as illegitimate by the rest of the world – yields a reputational benefit. It decreases the likelihood of being unilaterally targeted again in the future.

Our formal analysis provides an alternative rationale, which focuses on the role played by the sender government's incentives. On the one hand, unilateral coercion creates incentives leading the sender government to make unacceptable demands. On the other, commitment to a multilateral organisation can put a limit on them and allow the sender government to obtain concessions.

  • Second, we offer a novel explanation for why no panel was ever established in the majority of the disputes brought to the GATT/WTO, and a significant share of those for which a panel was established ended prior to the issuance of a formal report.

The existing literature (Busch and Reinhardt 2002) argues that early concessions are due to the normative power of GATT/WTO rulings and to the reputational costs of violating the norm. As a result, if the target government is uncertain about the ruling, then it may prefer to concede beforehand. In our setting, the mechanism at work is different. International rulings do not convey any normative or reputational costs. When the Foreign government anticipates an unfavourable ruling, it expects the Home government to comply with this ruling. This leads the Foreign government to abandon aggressive strategies, and to make more accommodating demands to which the Home government is willing to concede.

  • Third, we develop an alternative rationale for international trade agreements.

Most of the existing literature on this topic suggests that states become members of such institutions to solve the coordination problem created by the terms of trade externality from tariffs (e.g. Bagwell and Staiger 1999).

Our analysis reveals, however, that another driving force may emanate from informational asymmetries in trade coercion. By helping to explain why demands channelled through the multilateral system may be more successful than unilateral demands, our model provides a new rationale for states' commitment to multilateral institutions.

  • Fourth, we provide a new argument in favour of commitment to international organisations.

Our analysis shows that institutions allowing sender governments to choose between unilateralism and multilateralism can reduce the effectiveness of coercion. This idea had already been put forward by Pelc (2010), who suggests that this is due to the perceived illegitimacy of unilateral coercion. Our result stems instead from the Foreign government's strategic incentives created by the presence of a unilateral option.


Anesi, V and G Facchini (2015) “Coercive trade policy”, CEPR Discussion Paper 10687

Bagwell, K and R W Staiger (2005) “Enforcement, private political pressure, and the GATT/WTO escape clause”, Journal of Legal Studies 34. 471-525.

Bagwell, K and R W Staiger (1999) “An economic theory of GATT”, American Economic Review 89. 215-248.

Busch, M L (2000). “Accommodating unilateralism? U.S. section 301 and GATT/WTO dispute settlement”, mimeo Queen's University

Busch, M L and E Reinhardt (2000) “Bargaining in the shadow of the law. Early settlement in GATT/WTO disputes”, Fordham International Law Journal 24. 158-172.

Pelc, K J (2010) “Constraining coercion? Legitimacy and its role in U.S. trade policy 1975-2000” International Organization 64. 65-96.

Puckett, A L and W L Reynolds (1996) “Rules, sanctions and enforcement under section 301. At odds with the WTO?”, The American Journal of International Law 90. 675-689

Schoppa, L J (1999) “The social context in coercive international bargaining”, International Organization 53. 307-342.

1,470 Reads