After Gary Cohn demonstratively resigned from the Trump administration last week in advance of Donald Trump's announcement of tariffs of 25% on steel imports and 10% on aluminum, some analysts panicked that what was coming was nothing short of Smoot-Hawley 2.0. So far, it has turned out to be a tempest in Jamie Dimon's proverbial teapot, with Canada, Mexico and Australia already exempted, and most European allies working hard to obtain footnotes too.
It is also worth noting that the announced tariffs cover only a very small share of US trade. The chart below from Goldman shows that the relevant steel and aluminum imports into the US constitute only 1.8% of total US goods imports and would raise the average US import tariff by 0.3%. The trading partners most affected are (potentially) Canada, Brazil, and South Africa; but even for Canada, the tariffs would only cover 2.9% of total goods exports.
What is more notable, and explains the market's initial panicked reaction, is that the tariffs, coupled with wider talk of greater protection for industries, go against a broad historical trend towards more free trade. The average effective US tariff rate has declined from around 20% in the 1930s to only 1½% in recent times.
It is the fact that Trump is going against this grain that has sparked global anger among trade partners, many of whom have vowed to retaliate to the US should they not be granted exemptions too. And while the threats have so far not been very specific, retaliation seems highly likely according to Goldman, which believes that tariffs on a number of US goods—including jeans, bourbon whiskey, and motorcycles—are likely.
This is summarized in the Goldman table below, which lays out both who is most likely to retaliate and how severe the response is likely to be:
It is what happens next that is most interesting.
According to Goldman's Jan Hatzius, "the pure trade effects of the announced tariffs and the expected retaliation are likely to be very small. For example, using estimates from the literature we find that the announced tariffs will result in a 0.2% fall in US imports when Canada and Mexico are exempt, and 0.4% when they are included."
But this, the Goldman economist notes, focuses narrowly on the trade effects and ignores the broader macro repercussions of protectionism. He adds that the macro costs would rise notably if the trade war escalates.
A severe trade war—in which everyone imposes a tariff on everyone else—leads to higher world inflation, tighter monetary policy and slower growth. This drag is amplified if equity prices drop around the world: financial conditions provide an important channel for negative spillovers. Open economies with trade surpluses—such as the Euro area—are hit hardest in this scenario.
Goldman envisions 4 scenarios of what could happen next, in terms of rising severity:
- US tariffs without retaliation. We round up the announced tariffs to 1% of total imports, partly because the Trump administration has already announced a few specific tariffs (e.g. on Canadian lumber) and partly because some further restrictions are likely even in a mild conflict scenario. We allow interest rate and the exchange rate to respond endogenously to the tariffs, but assume that equity prices remain unchanged.
- A US-focused trade war. We assume that the US tariffs lead to retaliation from trading partners and further US tariff increases. In this scenario we assume that tariffs on all trade to and from the US rise by 5pp. But we assume that trading partners do not put up tariffs between each other; for example, the EU and China both put up a tariff against US imports but do not erect trade barriers between each other.
- A global trade war. We assume that each country imposes a 5% tariff on everyone else. For example, the EU puts up a tariff against China in response to the US steel tariffs in an effort to prevent Chinese steel from flowing to Europe.
- A global trade war with a global equity sell-off. We assume that global equity markets drop by 10%, in addition to the global 5% tariff.
Goldman's simulations of the impact of these 4 scenarios on the global economy are laid out below:
Some further details on each scenario:
- First—so long as its trading partners do not retaliate—US import tariffs have small positive effects on US real GDP (scenario 1). The increase in import prices reduces imports, which boosts domestic production and hence GDP. The growth lift is enough to outweigh the negative effects from higher inflation, which leads to higher interest rates and a stronger dollar. If no retaliation is expected, the simulation highlights the US incentive for erecting tariffs. But it bears emphasizing that the effects are extremely small: US real GDP is around 0.01% stronger after one year and core inflation is 2bp higher.
- Second, everyone loses in a trade war (scenario 2). The effects are now negative for the US as the foreign tariffs slow US exports, inflation remains higher than in the no-tariff baseline, the Fed raises interest rates slightly more quickly in response and the dollar strengthens more. The retaliation of trading partners cushions the negative trade implications of the US tariffs, but growth is still lower due to higher interest rates. That said, the effects remain very small even for a 5% tariff on all trade from and to the US.
- The global cost of protectionism rises more notably if a severe trade war erupts, in which tariffs go up everywhere (scenario 3). Global inflation now rises somewhat more notably, which weighs on world consumer spending and forces central banks to raise interest rates. The dollar no longer appreciates in this scenario. Open economies and countries with trade surpluses are most adversely affected, especially Europe. This is consistent with our European team’s view that the Euro area has much to lose from a trade war.
- Third, the costs of a trade war rise further if risk asset markets fall (scenario 4). A drop in equity prices reinforces the negative effects of the trade war. As one would expect, the adverse market response has more severe consequences in countries with large equity markets, especially the US, and the dollar starts to weaken in this scenario.
The chart below is a summary of the cascading effects of a global trade war. It decomposes the GDP effects in scenario 4 into the four transmission channels. Here's Goldman:
We see that US tariffs (without retaliation) provide a small boost to the US, and are painful for Canada and Mexico given their large trade exposure to the US (dark blue bars). Retaliation hurts the US (making the net effect on output negative) but helps the trading partners, particularly Canada which has a trade deficit and large exposure to the US (light blue bars). An “all out” trade war (grey bars) is particularly painful for the surplus economies, including the Euro area, Japan and China. The 10% equity drop hurts everyone, but the DM economies more so than EM (green bars).
Goldman concedes that for now its model of the world economy under trade war is "highly stylized", and as such it does not capture the microeconomic gains from free trade, including benefits for productive efficiency and productivity.
Moreover, the estimated macro effects are subject to large uncertainties. On the one hand, our simulations might understate the true effects of a trade war because our model does not include adverse confidence, or financial stability effects. On the other hand, our model might overstate the effects of tariffs as we assume that the fiscal revenues from tariffs are not used to stimulate growth. Allowing governments to use the tariff revenues to expand government spending or cut taxes would cushion the tariff effects further.
Nonetheless, Goldman can still conclude that the announced US tariffs and their expected retaliation "pose a modest risk to our optimistic outlook for the world economy. But the risk would rise sharply if a broad trade war erupted and financial conditions tightened in response."
Incidentally, this is precisely what various Fed speakers said last week. Perversely, the adverse scenario may be precisely what Trump ordered: after all, with many forecasters predicting that Trump's fiscal stimulus will be the catalyst that sinks the market as it overheats the economy and forces the Fed to hike more than most expect, the one loophole would be a suddenly cautious Fed which decides that hiking rates is not so prudent if a major potential risk event is hiding just around the corner, keeping rates lower for longer and indefinitely delaying the worst-case scenario for markets.
A risk event such as a global trade war.