Modeling a US-EU trade war: Tariffs won't improve US global trade balance

America could narrow its merchandise trade deficit with EU countries by imposing tariffs on them, but at a cost of higher inflation and lower GDP than otherwise for the US and nearly all the other EU economies.

The measures, however, would have little impact on the global US trade deficit, according to our new analysis.

President Donald Trump has announced, postponed, and threatened a dizzying variety of trade protection policies since taking office in January. These include, for example, slapping 25 percent tariffs on imported steel, aluminum, and downstream products. The EU has said it will retaliate by imposing tariffs on imports from the US, beginning with iconic products such as bourbon, jeans, and Harley-Davidson motorcycles, and will broaden the list if necessary. Trump has threatened to respond to any such EU measures by imposing 200 percent tariffs on the bloc’s alcoholic beverages.

Trump also threatened on February 26 to put 25 percent tariffs on all EU products, saying that the 27-nation bloc was formed “to screw” the US. He added, “They’ve really taken advantage of us.... They don’t accept our cars, they don’t accept, essentially, our farm products. They use all sorts of reasons why not. And we accept everything of them.”

We find that such tariffs would backfire on the US, pushing up prices and slowing economic growth. And that’s before likely EU retaliation.

To assess the potential global implications of a US-EU trade war, we used the G-Cubed economic model[1] to project the effects of the US imposing a 25 percent tariff on EU goods—in addition to any existing tariffs—and the bloc’s possible retaliation. For tractability, we assume that this retaliation takes the form of mirror image tariffs, though in past disputes, the EU has responded more surgically against US actions.

The model breaks out the three large EU and G20 economies (Germany, France, and Italy), and the smaller, generally more trade open EU members are aggregated as “rest of eurozone.”[2] Model results for trade war effects on real GDP with and without retaliation are presented in figure 1 as deviations from our baseline forecasts for each economy for each year. (A dashboard with greater detail is accessible here.)

Figure 1 shows the US tariffs would damage both the US and EU economies over time, resulting in lower real GDP than if the tariffs were not imposed. In-kind retaliation by the EU would magnify these results.

France and the smaller eurozone countries show as temporary exceptions, seeing higher GDP than otherwise in the first year or so. This is because the depreciation of the euro (driven by the Europe-wide trade shock) would be far greater than the direct impact of tariffs on France relative to other countries in the eurozone. This depreciation would initially offset the GDP effect on these countries, given the difference in trade and production structure of France relative to the rest of Europe.

Germany would see lower GDP than otherwise, as it would suffer a significant hit to durable exports due to the slowdown in investment in Europe and globally. The durable goods sector is an important input into private investment.

After four years, the small, relatively trade-dependent countries would be hurt worst of all. Retaliation would deepen the damage to all the countries analyzed over time.

It is unlikely that these tariffs alone could push the US into recession unless there is a significant increase in uncertainty that impacts financial markets, or the policy is combined with mass deportations from the US. That said, with recession concerns growing, a policy that slows US economic growth would be unhelpful.

Furthermore, the policy would cause inflation to surge temporarily in the US and EU, and would be worse with retaliation, as shown in figure 2, which shows the percentage point deviation from each economy's baseline forecast.

Meantime, the proposed 25 percent US tariff would have virtually no impact on the US global trade deficit—a frequent topic in Trump’s comments.

The measure would increase the Europeans’ global trade surpluses relative to the baseline forecasts, as shown in figure 3. The global competitiveness gain due to the euro depreciation, together with the compression of domestic demand, would outweigh the impact of the tariff. With retaliation, the dollar and the euro would decline against other currencies, and with both areas contracting, the US (temporarily) and the EU (permanently) would experience an increase in their global balances.

The US currently runs a bilateral merchandise trade deficit with the EU of more than $200 billion a year and a services trade surplus of more than $100 billion. Most economists believe that these bilateral balances are not important to the overall health of an economy, but Trump characterizes trade deficits as bad for the US.

If the US imposes the threatened 25 percent tariffs on the EU, the US bilateral trade deficit with the EU would shrink if the EU does not retaliate, as seen in figure 4, which shows the deviations from baseline forecasts in US dollar amounts. But EU retaliation would mute these effects.

The European Commission has pledged to “react firmly and immediately against unjustified barriers to free and fair trade, including when tariffs are used to challenge legal and non-discriminatory policies.”

We have assumed for this analysis that the EU would respond by increasing tariffs on US goods by 25 percent. But European retaliation might not be limited to tariffs or even the merchandise trade sector. Services, digital trade, intellectual property protection, and even government procurement could potentially be subject to retaliation.

Retaliation with the aim of inducing the US to remove the original tariff is a dangerous strategy: desirable if the US backs down but risking additional debilitating protection if the US does not. The US still maintains a 25 percent tariff on light trucks as a legacy of a 1960s trade dispute with Europe over frozen chicken parts.

It is unclear under what law, if any, Trump could claim power to impose tariffs over all imports from the EU. He imposed the 25 percent tariffs on steel, aluminum, and downstream products, under Section 232 of US trade law. This provision gives the president latitude to unilaterally impose tariffs without congressional approval in national security cases. Similar tariffs were introduced in his first term, but allies were given waivers. This time there are to be no exceptions.

The Section 232 tariffs alone will increase the costs of many products and services in the US economy, but particularly those that use steel and aluminum intensively. The costs of automobile production in the US will rise significantly, and consumers will migrate to European and Asian models unless motor vehicles also receive protection.

The Trump administration also has signaled that it would broaden tariffs across products and partners under a misnamed policy of “reciprocity.” The president said he will impose the same tariffs on other countries as they impose on the US. The term carries a connotation of fairness, but the US policy is anything but fair: It appears to only level US tariffs up and contains language allowing America to escape strict reciprocity. Nondiscriminatory value-added taxes in the EU are a particular bête noire. A more accurate characterization of the approach would be arbitrary unilateral tariffs, undertaken without any reference to existing World Trade Organization or free trade agreement obligations.

Looking just at Trump’s planned 25 percent tariffs on the EU, beyond the immediate economic effects, this misguided policy would contribute to the further unravelling of the postwar international trade architecture, which has served the US and the world well. Given the recent turmoil in global financial markets and the series of PIIE studies documenting the downsides of additional trade protection, the US and the world would be better off if Trump forgot about this threat.

Notes

1. The G-Cubed model is a multi-country, multi-sector hybrid dynamic stochastic general equilibrium-computable general equilibrium model (McKibbin and Wilcoxen 1999, 2013); see full documentation here.

4. There is a minor difference between “rest of eurozone” and “rest of EU.” Denmark, Iceland, Liechtenstein, and Sweden are aggregated with “rest of advanced economies.” This difference should not materially affect the results.