By Robert Staiger

In my previous two posts on the value of the World Trade Organization (WTO), I argued that when governments set trade policies independently, they focus on domestic effects but often overlook the costs imposed on trading partners. International trade agreements help governments take account of these cross-border spillovers (terms-of-trade externalities) by encouraging reciprocal negotiations over trade policy. While these negotiations will reduce trade protection, they will not necessarily lead to free trade. But if governments do not generally commit themselves to free trade, why has freedom of transit for goods been seen differently?

Why – from an economic perspective – should ships passing through the Strait of Hormuz or goods crossing a country’s territory enjoy a near-unconditional right of transit while tariffs remain the subject of lengthy negotiations?

This puzzle is clearly visible when comparing the WTO with other areas of international law, such as the regimes of maritime passage under the United Nations Convention on the Law of the Sea (UNCLOS). UNCLOS treats free transit for goods through straits used for international navigation as close to an unconditional entitlement.

The puzzle also appears inside the WTO. Article V of the General Agreement on Tariffs and Trade (GATT) – “Freedom of Transit” – requires that traffic merely passing through a member’s territory – goods that begin and end their journey beyond that member’s frontier – “shall be exempt from customs duties and from all transit duties” apart from charges for transportation or charges commensurate with administrative expenses or the cost of services actually rendered. An imported good is treated differently. A WTO member may set a tariff and remain in full compliance with the GATT; it may not impose a comparable transit duty on goods that simply cross its territory on their way elsewhere.

So, the same puzzle appears twice: once with reference to the WTO and UNCLOS regimes of maritime passage, and once again within the GATT itself. Why would international law allow negotiated tariffs in one situation while imposing an essentially unconditional rule of free transit for goods in another?

The answer lies in the same economic logic developed in my previous posts. This post looks at the puzzle in three steps. First, it revisits the issue of tariffs and market power that underlies the terms-of-trade theory of the GATT/WTO (i.e. when a government raises a tariff, it can affect not only domestic prices and the quantity of imports but also the prices received by foreign exporters). Second, it applies that logic to taxes and other restrictions on goods that merely pass through a territory. Finally, it explains why the same economic reasoning leads to negotiated tariff commitments in one case but to a simple rule of free transit in the other – and why the GATT contains both approaches side by side.

Tariffs, market power and the “political optimum”

The terms-of-trade theory of the GATT/WTO developed in Bagwell and Staiger (1999) starts from an observation already introduced in my previous post: a tariff imposed by an economy with market power does not affect only its own consumers and import-competing firms. It can also depress the price received by foreign exporters, shifting part of the cost of protection onto trading partners. Because governments have an incentive to exploit this effect, unilateral tariff setting can lead to outcomes that are collectively inefficient. This is the cross-border spillover that, as argued in my previous post, the GATT/WTO exists to help governments “negotiate away”.

The important question is what remains once that cost-shifting incentive has been removed. Bagwell and Staiger refer to this as the political optimum: the tariff a government would choose on purely domestic grounds, absent any ability to shift costs onto foreigners. This tariff need not be zero. Even if a government faces all of the costs of its actions, it may still want to set a tariff to raise revenue where other forms of taxation are costly to administer, to protect a sector for domestic political reasons, or to address a domestic market failure when it lacks better policies to do so.

This helps explain why the objective of GATT/WTO negotiations was never free trade as such. The negotiating rounds documented in my previous posts were a process through which governments exchanged reciprocal reductions in the component of their tariffs driven by terms of trade considerations, not a mechanism for reducing every tariff line to zero. Reciprocity – a balanced exchange of market access commitments – helps neutralize the incentive to manipulate terms of trade, while the non-discrimination principle of most-favoured nation (MFN) – where all WTO members are treated equally – helps ensure that the resulting commitments extend across the WTO membership. What remains, in principle, is each government’s own domestic political optimum – and because that optimum differs across economies and products, there is no single tariff level to which the system is expected to converge.

This is also why tariff negotiations are important. The outcome is not predetermined. Governments must use the negotiations to find a set of mutually acceptable commitments, taking account of their own domestic objectives as well as the reciprocal concessions offered by others.

What the same logic implies for goods that are only passing through

Now suppose the same logic is applied to a different policy instrument: a tax or charge imposed on goods that merely pass through a jurisdiction. The goods begin their journey elsewhere and are destined for somewhere else. They are not imported for domestic consumption, and they do not compete with domestic producers.

What is the political optimum for such a charge?

The answer turns out to be different from the tariff case. A tariff can retain a domestic component once the motive of cost-shifting to the foreign party has been removed: it may raise revenue from domestic sources, protect a domestic sector, or fulfil other domestic policy objectives. A pure transit charge for goods has no equivalent domestic dimension. The goods are not consumed in the transit economy, do not compete with its industries, and do not generate a domestic policy rationale comparable to the one that may justify imposing a tariff.

The direct burden of a transit charge falls completely on foreign parties: exporters, importers, carriers, or even consumers in third economies. The cost-shifting motive remains, but the domestic component that might otherwise still exist after reciprocal bargaining – the political optimum – disappears. In that sense, the terms-of-trade externality associated with a transit charge appears in unusually pure form.

This is the key difference. For the imposition of a tariff, international bargaining removes the incentive to shift costs abroad while leaving room for legitimate domestic objectives. For a pure transit charge, there is no comparable domestic objective to preserve. The same logic that can justify negotiated tariffs therefore suggests a zero charge on goods traffic that merely passes through a jurisdiction.

Seen in this light, GATT Article V’s flat exemption of transit traffic from customs duties and transit duties reflects the fact that the underlying economic situation is different from the tariff case. UNCLOS reaches a similar result for international straits used for goods navigation. The same economic structure helps explain why both legal regimes converge on a strong rule of freedom of transit for goods.

A tariff, however, is not the only instrument that can have an economic impact on goods traffic. Unnecessary delays at customs, excessive documentation requirements, customs escorts, guarantee deposits and other procedural burdens can reproduce similar effects without being labelled as charges.

This is where the WTO framework becomes especially useful. GATT Article V establishes the basic principle of freedom of transit but Article 11 of the Trade Facilitation Agreement goes further by disciplining the formalities, documentation requirements and customs procedures applied to goods in transit. The economic argument is straightforward: once direct transit duties are prohibited, complex procedural practices should not become an indirect substitute for seeking the same type of economic gain. This can be understood as the transit-law counterpart of the concern raised by James Meade, and discussed in my first post in the value of the WTO series.⁽¹⁾

Charging for specific services

Of course, none of this implies that goods transit need be entirely costless. Moving goods safely and efficiently across a jurisdiction may require infrastructure, border facilities, customs administration, etc.

This point is reflected in GATT Article V. While transit duties are prohibited, the Article allows charges for transportation and charges commensurate with administrative expenses or the cost of services actually rendered. The same point is made in the law of the sea. UNCLOS does not prevent coastal states from charging for specific services they provide, such as pilotage, towage or port facilities. Under separate legal arrangements, tolls charged by the operators of constructed waterways, such as the Suez and Panama canals, are also permitted. All of these share a feature that a pure transit tax lacks: the objective is not to prevent governments from recovering legitimate costs, but to distinguish such charges from measures whose purpose is to seek economic benefits from goods simply passing through the territory.

This is the same distinction made by the GATT in the case of tariffs, and it is worth noting how closely the two texts track each other, despite having been drafted decades apart by different negotiators for different purposes. A tariff that funds a public good or reflects a government’s domestic preferences for income distribution has a legitimate claim to be part of a GATT/WTO trade bargain, because a real domestic cost or benefit underlies it. A transit charge that reflects a real cost of service has an equivalent claim to exist under GATT Article V or under UNCLOS.

What these regimes rule out, each in its own domain, is the component that exists only to seek economic benefit from a foreign party who bears the cost of the policy but has no seat at the bargaining table of the transiting jurisdiction or the tariff-setting government.

Main takeaway

The comparison between tariffs and goods transit illustrates the broader logic of the terms-of-trade theory developed in my previous posts.

The GATT/WTO is not designed simply to promote free trade. Rather, it seeks to remove governments’ incentives to shift the costs of their trade policies onto others while preserving room for legitimate domestic policy objectives. In the case of tariffs, this means negotiated commitments that reflect each government’s own domestic priorities once the terms-of-trade motive has been neutralized.

Transit of goods is different. Because goods merely passing through a jurisdiction have no domestic economic implications, there is no rationale for imposing taxes or unnecessary restrictions on them. The same economic logic that supports negotiated tariffs, therefore, justifies a general rule of free transit for goods.

Viewed from this perspective, the strong protection of freedom of transit found in GATT Article V, reinforced by the disciplines of the Trade Facilitation Agreement and echoed in the international law governing sea navigation, is not an exception to the economic logic of the multilateral trading system. It is one of its clearest applications.

My next post will be the latest instalment in the Value of the WTO series, where I will look at the results from a quantitative trade model to consider the implications of abandoning MFN.