From President Donald Trump’s perspective, taxing international trade almost certainly makes sense. He has continued to look for ways to tax imports through tariffs, even though the US Supreme Court ruled earlier this year that his sweeping global tariffs were impermissible.
President Trump’s position notwithstanding, the broader question deserves a more careful answer. International economists have long regarded free trade as the optimal policy under standard conditions. But does that prescription continue to hold in a world shaped by environmental externalities, imperfect competition, geopolitical risk and distributional concerns?
New research by Arnaud Costinot and Ivan Werning provides a useful framework for answering this question. It examines when trade taxes or subsidies can improve social welfare, not by reviving mercantilism, but by applying a broadly interpreted Pigouvian principle to trade policy.
The starting point is to view international trade as a form of technology. Access to foreign goods expands the production and consumption possibilities available to an economy, much as technological progress does. Trade policy should therefore be assessed according to whether private decisions to import or export properly reflect their marginal effects on social welfare.
When private agents fail to internalise these effects, an appropriately designed trade tax or subsidy may help correct the resulting distortion.
Free Trade
The authors begin by rejecting mercantilism, the belief that exports are inherently beneficial and imports harmful because countries should seek trade surpluses. Countries are not firms, and international trade is not intrinsically a zero-sum game.
The classical case for free trade follows from this insight. Imports allow consumers and producers to obtain goods at lower costs than domestic production might permit. Access to foreign electric vehicles, for example, can replace a costly domestic production technology with a cheaper alternative.
Trade therefore creates gains by allowing goods to be sourced more efficiently and by enabling the consumption of goods whose marginal benefits exceed their costs. A tariff interferes with this process. By creating a wedge between international and domestic prices, it distorts both production and consumption.
Under the standard assumptions of competitive markets and the absence of other distortions, free trade is therefore optimal.
Real economies, however, rarely satisfy all these assumptions. They contain externalities, imperfect competition, labour-market frictions and distributional concerns. The Theory of the Second Best explains why trade intervention may sometimes improve welfare when such distortions are present.
Under the Costinot-Werning framework, the optimal tariff on an imported good should reflect the social harm caused by additional imports and the extent to which those imports worsen the relevant distortion. The same reasoning can also justify an import subsidy.
If importing a particular good reduces domestic pollution, for instance, subsidising the import may be preferable to taxing it. Where imports affect global carbon emissions, the appropriate policy depends on the relationship between those imports and emissions.
The framework can also accommodate learning-by-doing, geopolitical externalities and domestic misallocation arising from imperfect competition or other market failures.
Distribution matters as well. Trade policy changes domestic prices and wages and, with them, the distribution of real income. If society assigns a higher marginal value to the income of poorer households, a tariff could, in principle, redistribute income towards them.
Trade taxes can also shift income across countries through their effect on world prices. This is the traditional optimal-tariff, or terms-of-trade, argument. A sufficiently large country may restrict imports, lower the prices received by foreign exporters and transfer some income from foreigners to domestic residents.
Targeted Remedies
None of this means that tariffs are usually the best instrument. The Targeting Principle says governments should use the policy tool that addresses the underlying distortion most directly.
If pollution is the problem and a carbon tax is available, a tariff on pollution-intensive imports will generally be inferior because it also distorts consumption and production. Similarly, if redistribution is the objective and effective income or commodity taxes are available, those domestic instruments should ordinarily be preferred to tariffs.
The principle, however, has two important limitations. First, domestic instruments cannot directly address international targets such as foreign carbon emissions or foreign geopolitical behaviour.
A country cannot directly impose a domestic tax on emissions generated abroad. Tariffs, border adjustments or other trade measures may therefore be the most targeted instruments available when imports generate environmental costs in another country.
Second, governments seldom possess the complete set of precise policy instruments assumed in idealised economic models. Domestic taxes and subsidies may be administratively weak, politically constrained or too blunt. In such circumstances, trade policy may remain useful, even when it is only a second-best instrument.
The authors’ examination of the China trade shock demonstrates how narrow the case for tariffs can be. Using estimates of the effect of Chinese import competition on wages across the US wage distribution, together with marginal income tax rates and labour-supply elasticities, they calculate an optimal tariff on Chinese imports of just 0.07%.
The estimate is low because, although Chinese imports had significant labour-market effects, their impact on relative wages was sufficiently modest. Consequently, the efficiency costs of using tariffs for redistribution were large in relation to their redistributive benefits.
The broader conclusion is that free trade is neither an absolute principle nor an obsolete idea. Trade should be taxed or subsidised when private decisions fail to account for their marginal effects on social welfare. But where a more targeted domestic instrument is available, governments should generally use it.
Tariffs retain a role where the target itself is international, as with foreign emissions, geopolitical behaviour or terms-of-trade effects, or where domestic instruments are unavailable or inadequate. Even then, the case is for calibrated intervention tied to a clearly identified distortion, not for tariffs as a general-purpose economic strategy.
.png)
